| DF-5833 |
High |
7.8 |
generic_elf_coredump note-size precompute race: LWP insertions between the DRYRUN and WRITE passes overflow the header reservation, the swallowed elf_putallnotes error leaves target->off past off_max, elf_puttextvp computes vnh_max from a size_t-underflowed (off_max - off), and cb_put_fp writes attacker-sized vn_hdr records past the kmalloc'd core-header buffer
Kernel memory corruption (UAF / OOB write) · sys/kern/imgact_elf.c:1012 |
none |
inconclusive |
| DF-5837 |
Low |
6.3 |
cluster_wbuild installs the pbuf KVA window before validating it, with the first component exempt from the page guard and the window outside the bkvareset() protocol
Base system — kernel core (kern) · sys/kern/vfs_cluster.c:1606 |
none |
untested |
| DF-5791 |
Low |
5.2 |
Async-attach thread races driver deregistration/device deletion: device_t has no refcount and device_detach() short-circuits DS_INPROGRESS to success, allowing device_set_driver() to free the softc/kobj while device_attach_thread() is inside DEVICE_ATTACH()
Kernel memory corruption (UAF / OOB write) · sys/kern/subr_bus.c:2005 |
latent UAF write into freed softc/device |
untested |
| DF-5798 |
Low |
4.7 |
unmount_allproc_cb() mutates p->p_textnch without p->p_token, racing exec's drop/copy — namecache reference underflow to UAF potential
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_syscalls.c:777 |
none |
not_reproduced |
| DF-5838 |
Low |
4.7 |
cluster_collectbufs collects buffers matched against last_bp->b_bcount instead of blksize, breaking the documented size contract feeding VOP_REALLOCBLKS
Base system — kernel core (kern) · sys/kern/vfs_cluster.c:1737 |
none |
untested |
| DF-5841 |
Info |
4.7 |
mapped_ioctl: IOC_VOID|IOC_OUT commands copy out size bytes from the 128-byte on-stack ubuf.stkbuf — up to ~8KB kernel-stack overread primitive, gated only on fo_ioctl returning 0
Kernel memory leaks / uninitialized disclosure · sys/kern/sys_generic.c:674 |
none |
not_reproduced |
| DF-5828 |
Info |
4.3 |
Received-message CRCs (hdr_crc, aux_crc) are never validated
Base system — kernel core (kern) · sys/kern/kern_dmsg.c:339 |
undetected message tampering by on-path attacker |
untested |
| DF-5788 |
Low |
3.6 |
kern_getcwd() reads fd_ncdir completely unlocked and walks the mount topology holding no mount reference — latent use-after-free racing chdir() and umount -f
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_cache.c:5076 |
none |
not_reproduced |
| DF-5820 |
Low |
3.6 |
hammer2_vop_nrename masks backend ENOENT as success and commits rename bookkeeping (tip nlink--/DELETING, iparent/name_key rewrite, cache_rename) against an unrenamed media topology
Hammer2 · sys/vfs/hammer2/hammer2_vnops.c:2165 |
none |
not_reproduced |
| DF-5827 |
Low |
3.1 |
freerd_state allocation prologue runs unlocked and is reachable from two threads concurrently — orphaned kdmsg_state leak
Base system — kernel core (kern) · sys/kern/kern_dmsg.c:788 |
slow kernel-memory leak per lost race |
untested |
| DF-5826 |
Info |
3.1 |
Reader never validates DMSGF_SIZE against the base command — undersized extended headers reach command dispatch (latent type confusion)
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_dmsg.c:348 |
latent type confusion |
untested |
| DF-5832 |
Medium |
2.5 |
NCF_WXOK cached-traversal flag resurrection race: chmod revocation of directory search permission can be undone by a concurrent in-flight naccess() and the stale grant is then honored indefinitely
Base system — kernel core (kern) · sys/kern/vfs_nlookup.c:1616 |
none |
not_reproduced |
| DF-5810 |
Low |
2.5 |
vtruncbuf() caches the separately-allocated nc_name pointer across unlocked blocking work, then prints it (narrow UAF read)
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_subr.c:531 |
none |
untested |
| DF-5814 |
Low |
2.5 |
vm_pageout_page(): object->flags use-after-free READ on the failed-vget error path — page hold parks the page on PQ_HOLD but pins neither the vm_object nor (on vget failure) the vnode
Kernel memory corruption (UAF / OOB write) · sys/vm/vm_pageout.c:1254 |
none |
untested |
| DF-5819 |
Low |
2.5 |
hammer2_vop_nsymlink unconditionally discards the symlink-data write error ('/* XXX handle error */ error = 0;')
Hammer2 · sys/vfs/hammer2/hammer2_vnops.c:1847 |
none |
not_reproduced |
| DF-5823 |
Low |
2.5 |
KERN_PROC_PGRP/KERN_PROC_TTY sysctl filters dereference the p_pgrp->pg_session->s_ttyp->t_dev chain with no p_token/pgref — TOCTOU use-after-free read against pgrp teardown (dead p_session NULL guard)
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_proc.c:1736 |
none |
not_reproduced |
| DF-5786 |
Info |
2.5 |
Kernel heap/buffer pointer disclosure to unprivileged dmesg readers via unconditional %p kprintf in hammer2 reptrack debug and check-failure paths
Hammer2 · sys/vfs/hammer2/hammer2_chain.c:2266 |
kernel address disclosure to non-root dmesg readers |
untested |
| DF-5813 |
Info |
2.5 |
Four more raw vm_page_t-pointer kprintfs on attacker-triggerable pagedaemon race-recovery paths (rebusy A/B/C and the scan_cache find/busy warning) disclose kernel heap addresses into the world-readable msgbuf
Base system — virtual memory · sys/vm/vm_pageout.c:1340 |
none |
not_reproduced |
| DF-5821 |
Info |
2.5 |
hammer2_vop_reclaim kprintf discloses vp/ip kernel pointers to the console/msgbuf
Hammer2 · sys/vfs/hammer2/hammer2_vnops.c:183 |
none |
not_reproduced |
| DF-5840 |
Info |
2.3 |
lwkt_rele() never wakes the td_refs drain pollers (tdreap/lwpwait2) — teardown latency and needless scheduler churn only
Base system — kernel core (kern) · sys/kern/lwkt_thread.c:502 |
none |
untested |
| DF-5800 |
Info |
1.1 |
vm_map_delete() no-predecessor branch sets freehint hole_start=vm_map_max, silently disabling the freehint update for bottom-of-map holes (correctness defect, no reachable security impact)
Base system — virtual memory · sys/vm/vm_map.c:3257 |
none |
not_reproduced |
| DF-5805 |
Info |
1.1 |
kern_sigtimedwait() violates the lwp_delsig()/lwp_sigmask locking contract (unlocked mask writes; lwp_delsig fromproc=1 without p->p_token)
Base system — kernel core (kern) · sys/kern/kern_sig.c:1764 |
signal-delivery semantics races (POSIX correctness) |
untested |
| DF-5804 |
Low |
0.8 |
gsignal() leaks the pgfind() reference — permanent pgrp/session leak if ever called
Base system — kernel core (kern) · sys/kern/kern_sig.c:898 |
latent pgrp leak (no in-tree callers) |
untested |
| DF-4036 |
High |
8.8 |
Unclamped VBIOS SCLK/MCLK DPM counts overflow polaris10 SMC GraphicsLevel[8]/MemoryLevel[4] and the 384-byte STACK MC-arb timing table entries[8][4]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:1000 · pass 2 · new |
none |
inconclusive fixed |
| DF-4569 |
High |
8.8 |
drm_ioctl() dispatches on cmd&0xff only — never validates the user ioctl command (size/dir/group) against the canonical table entry; port dropped upstream's cmd=ioctl->cmd + canonical-size guard: kernel stack OOB write via DRM_IOCTL_GET_STATS and kernel-stack disclosure oracle via any 0-returning unauthenticated ioctl
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_ioctl.c:801 · pass 2 · new |
panic |
inconclusive |
| DF-4587 |
High |
8.8 |
ttm_bo_release_mmap drops an unowned OBJT_MGTDEVICE object reference on every BO revocation — premature vm_object termination under live mappings plus early ttm_bo_unref (kernel UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/ttm/ttm_bo_vm.c:821 · pass 2 · new |
none |
inconclusive |
| DF-5320 |
High |
8.8 |
ucom unplug lifecycle UAF: driver softc kfree'd while the never-freed tty keeps dangling t_sc/t_oproc/t_param/t_stop and in-flight cdev dispatch follows si_drv1 into freed memory (multi-port adapters leave orphaned openable device nodes)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/serial/usb_serial.c:543 · pass 2 · new |
none |
not_reproduced |
| DF-3597 |
High |
8.7 |
gfx_v8_0_parse_ind_reg_list: trailing-entry heap OOB read/write past the RLC register-list copy plus post-write BUG_ON panic
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/gfx_v8_0.c:4001 · pass 2 · new |
none |
inconclusive |
| DF-3761 |
Medium |
8.6 |
alc_rxintr() Rx walk has no work bound — the process_limit knob exists but is never read: a remote line-rate flood keeps the return ring perpetually non-empty so the interrupt handler loops forever holding the serializer (NIC + CPU livelock); FreeBSD bounds this loop, DFly msk(4) bounds its own
Base system — drivers · sys/dev/netif/alc/if_alc.c:3236 · pass 2 · new |
none |
inconclusive |
| DF-3904 |
Medium |
8.6 |
sk_rxeof() Rx walk has no work bound — for(;;) loop's only exit is the OWN bit, so a line-rate remote flood keeps the walk running forever under the card-global serializer (DF-3761 family)
Base system — drivers · sys/dev/netif/sk/if_sk.c:1820 · pass 2 · new |
none |
inconclusive |
| DF-3907 |
Medium |
8.6 |
xl_rxeof() interrupt-mode Rx walk has no work bound — the count>=0 budget exists but every interrupt-path caller passes -1, and non-905B chips goto-again restart the walk on uplist stall (DF-3761 family)
Base system — drivers · sys/dev/netif/xl/if_xl.c:1960 · pass 2 · new |
none |
inconclusive |
| DF-3748 |
High |
8.3 |
STA-mode AID→WCID truncation: rt2860_newassoc writes node pointer past wcid2ni[254] into txq[0] (remote OOB write via rogue AP)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ral/rt2860.c:885 · pass 2 · new |
none |
not_reproduced |
| DF-4654 |
Medium |
8.2 |
twa: unchecked CAM cdb_len overflows the 16-byte firmware CDB field (heap/DMA-pool overflow)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/twa/tw_osl_cam.c:276 · pass 2 · new |
none |
not_reproduced |
| DF-3264 |
High |
8.1 |
pf_state use-after-free: pf_send_tcp() drops pf_token mid-state-operation, letting the purge thread free the in-flight state
Kernel memory corruption (UAF / OOB write) · sys/net/pf/pf.c:2254 · pass 2 · new |
none |
not_reproduced |
| DF-5737 |
Medium |
8.0 |
Event-context link teardown races the syscall transmit path with no shared locking -> use-after-free write in hci_acl_send (forged RESET completion / CREATE_CON status / discon_compl frees links a send(2) is still touching)
Kernel memory corruption (UAF / OOB write) · sys/netbt/hci_event.c:271 · pass 2 · new |
none |
untested |
| DF-3672 |
Medium |
7.9 |
MPT_TAG_2_REQ production macro indexes tgt_cmd_ptrs[tag_id>>18] with no bounds check — user-controlled XPT_CONT_TARGET_IO tag_id gives heap OOB read + wild request_t* dereference
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_cam.c:4355 · pass 2 · new |
none |
inconclusive |
| DF-3619 |
High |
7.8 |
Heap OOB write in ci_get_svi2_voltage_table: VBIOS dependency-table count (u8, up to 255) copied into fixed entries[32] before any trim
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/ci_dpm.c:2264 · pass 2 · new |
none |
inconclusive |
| DF-3683 |
High |
7.8 |
Heap OOB write in phm_get_svi2_voltage_table_v0: VBIOS V0 dependency count (u8, up to 255) copied into fixed entries[32] before any trim
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:219 · pass 2 · new |
none |
inconclusive |
| DF-3684 |
High |
7.8 |
Kernel stack overflow: smu7_get_sclks/smu7_get_mclks copy up to 255 VBIOS entries into amd_pp_clocks.clock[16]/latency[16] held in the caller's stack frame
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:4613 · pass 2 · new |
none |
inconclusive |
| DF-3701 |
High |
7.8 |
vega10 clock getters overflow the caller's 16-entry stack array: unbounded VBIOS dep-table counts (DF-3684 twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:4115 · pass 2 · new |
none |
inconclusive |
| DF-3703 |
High |
7.8 |
vega10 ODN tables are fixed entries[8] but filled/iterated with unbounded VBIOS dep-table counts — heap overflow of struct vega10_hwmgr
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:310 · pass 2 · new |
none |
inconclusive |
| DF-3705 |
High |
7.8 |
vega10 SMC PPTable populate loops write 8-entry SMC arrays (CksEnable/CksVidOffset, StaticVoltageOffsetVid, UvdDpmVoltageIndex, SocVid) with unbounded dep-table counts
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:1700 · pass 2 · new |
none |
inconclusive |
| DF-3787 |
High |
7.8 |
METEORSSIGNAL stores td->td_proc without a reference and video_close() never clears it → frame-completion interrupt ksignal()s a freed struct proc (UAF from interrupt context)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/bktr/bktr_core.c:1485 · pass 2 · new |
none |
inconclusive |
| DF-3790 |
High |
7.8 |
bktr_filter_detach() calls knote_insert() instead of knote_remove() → destroyed knote left on (or self-looped into) the VBI klist → KNOTE() from the frame interrupt dereferences freed memory / hard hang (root cause bktr_os.c, sink bktr_core.c:705)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/bktr/bktr_core.c:705 · pass 2 · new |
none |
inconclusive |
| DF-3796 |
High |
7.8 |
ena_down tears down IO queues and frees RX buffers without draining the deferred RX cleanup task (mbuf UAF / double free)
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/amazon/ena/ena.c:2611 · pass 2 · new |
none |
not_reproduced |
| DF-3838 |
High |
7.8 |
32-bit wraparound family in evergreen CS surface-size validation ((slice+1)*64 wrap → nby=0/layer_size=0; layer_size*mslice u32 product wrap; <<8 base-offset truncation vs 40-bit GPU addresses; htile nbx*nby*4 wrap; streamout <<8 and *4 wrapped before the u64 check) → CS validator bypass → GPU OOB read/write past BO, cross-process GPU memory corruption
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/evergreen_cs.c:404 · pass 2 · new |
none |
inconclusive |
| DF-3914 |
High |
7.8 |
Unbounded firmware NumSlots heap OOB write in the direct-attached enclosure reservation path (_mapping_get_dev_info/_mapping_get_pcie_dev_info)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_mapping.c:1243 · pass 2 · new |
none |
not_reproduced |
| DF-4009 |
High |
7.8 |
NVMe-encapsulated pass-thru: unvalidated u32 DataSize drives PRP walk past the 8MiB arena (deterministic OOB write); kmalloc result unchecked
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_user.c:904 · pass 2 · new |
none |
inconclusive |
| DF-4013 |
High |
7.8 |
mps FW-diag state machine: shared DMA trio + stale per-type size yields OOB kernel disclosure, cross-type free, UAF read, and double free (DF-4008 mpr twin, worse)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps_user.c:1376 · pass 2 · new |
panic |
inconclusive fixed |
| DF-4028 |
High |
7.8 |
CS validator rewrites ib[0] after validation: cb_color_size_idx defaults to 0 (full r600 CS validator bypass)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_cs.c:309 · pass 2 · new |
none |
not_reproduced |
| DF-4031 |
High |
7.8 |
Tracked offsets wrap in u32 while the GPU addresses 40 bits: streamout size raw*4 and base<<8 truncation validate the WRONG offset
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_cs.c:721 · pass 2 · new |
none |
not_reproduced |
| DF-4085 |
High |
7.8 |
Unvalidated IOC-pg8 MaxNumPhysicalMappedIDs used as num_slots: heap OOB write in _mapping_process_dpm_pg0 reservation loop
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps_mapping.c:1512 · pass 2 · new |
none |
not_reproduced |
| DF-4086 |
High |
7.8 |
_mapping_get_ir_maprange off-by-one: end_idx == max_devices -> OOB reads and conditional 32-byte OOB write of mapping_table[] on IR events
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps_mapping.c:221 · pass 2 · new |
none |
not_reproduced |
| DF-4092 |
High |
7.8 |
Unvalidated IOC-pg8 MaxNumPhysicalMappedIDs used as num_slots: heap OOB write in mpr _mapping_process_dpm_pg0 reservation loop (twin of DF-4085)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_mapping.c:2216 · pass 2 · new |
none |
not_reproduced |
| DF-4093 |
High |
7.8 |
mpr _mapping_get_ir_maprange off-by-one: end_idx == max_devices -> OOB reads and two 32-byte OOB write sinks on IR paths (twin of DF-4086)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_mapping.c:254 · pass 2 · new |
none |
not_reproduced |
| DF-4231 |
High |
7.8 |
txp(4) trusts NIC-fed rxd->rx_len: unbounded m_copydata heap OOB read+write past mbuf clusters
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/txp/if_txp.c:676 · pass 2 · new |
none |
inconclusive |
| DF-4232 |
High |
7.8 |
txp(4) walks device-fed ring cursors unvalidated: OOB descriptor walk via kernel pointer in DMA memory, wild rx_sd deref, non-terminating loops
Base system — drivers · sys/dev/netif/txp/if_txp.c:658 · pass 2 · new |
none |
inconclusive |
| DF-4233 |
High |
7.8 |
txp_init() ignores txp_rxring_fill() failure and enables RX over a stale buffer ring: firmware DMA into freed clusters + NULL deref
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/txp/if_txp.c:1127 · pass 2 · new |
none |
inconclusive |
| DF-4412 |
High |
7.8 |
radeon_vm_get_bos() TOCTOU on vm->max_pde_used — kernel heap out-of-bounds write via racing GEM_VA/CS
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_vm.c:135 · pass 2 · new |
none |
not_reproduced |
| DF-4510 |
High |
7.8 |
drm_syncobj: lockless drm_syncobj_fence_get() races drm_syncobj_replace_fence()'s outside-the-lock dma_fence_put() of the old fence — rcu_read_lock() is only preempt_disable() on DragonFly and the last put kfree()s synchronously, so a cross-CPU waiter kref_get()s freed memory and then dispatches through the freed fence's ops: unprivileged kernel heap UAF (read+write) via render-node syncobj ioctls on i915/amdgpu
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_syncobj.c:182 · pass 2 · new |
none |
inconclusive |
| DF-4531 |
High |
7.8 |
drm_clients_info() walks dev->filelist under the wrong lock with a sleeping kmalloc between a count pass and an unbounded copy pass: concurrent open()/close() yields a kernel heap OOB write (or UAF read)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_file.c:268 · pass 2 · new |
none |
inconclusive |
| DF-4604 |
High |
7.8 |
radeon_gem_va_ioctl: bo_va use-after-free — set_addr() drops the BO reservation internally, then radeon_gem_va_update_vm() dereferences bo_va unprotected while a racing GEM_CLOSE on the same DRM fd frees it
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_gem.c:693 · pass 2 · new |
none |
inconclusive |
| DF-4635 |
High |
7.8 |
radeon_audio: enc_priv type confusion — radeon_encoder_lvds* consumed as radeon_encoder_atom_dig* on LCD-carrying digital encoders (eDP) → dig->pin/dig->afmt OOB read/write into the ADJACENT 288-byte slab chunk + wild MMIO deref
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_audio.c:262 · pass 2 · new |
none |
inconclusive |
| DF-4853 |
High |
7.8 |
evergreen_hdmi_enable / evergreen_dp_enable consume lvds-typed enc_priv as radeon_encoder_atom_dig on the eDP dpms path: dig->afmt OOB read at +288 into the adjacent slab chunk → guard falls through → wild deref feeding six WREG32 MMIO sites + dig->afmt->enabled write-through (DF-4635/DF-4828 family, new sinks; corrects DF-4828's dpms-route dismissal)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/evergreen_hdmi.c:429 · pass 2 · new |
none |
inconclusive |
| DF-4864 |
High |
7.8 |
amdgpu_vm_clear_bo() 64-byte IB overrun: the ATS root-PD clear emits two 10-dword set_pte_pde packets + SDMA pad (24 dwords) into a 16-dword (64-byte) SA sub-allocation — deterministic 32-byte kernel heap OOB write per GPU-VM create on every AMD Raven APU
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_ring.c:107 · pass 2 · new |
none |
not_reproduced |
| DF-5292 |
High |
7.8 |
Beacon mbuf write-after-free across disjoint lock domains: rum_update_beacon's unlocked ieee80211_beacon_update window vs m_freem in rum_newstate/rum_vap_delete
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_rum.c:2825 · pass 2 · new |
none |
inconclusive |
| DF-2949 |
Medium |
7.8 |
cttyclose() takes its tty-vnode reference with vref() racing the terminal vrele() of concurrent ctty teardowns — vref: bad refcnt panic (INVARIANTS) / freelist vnode resurrection (stock)
Kernel memory corruption (UAF / OOB write) · sys/kern/tty_tty.c:153 · pass 2 · new |
none |
not_reproduced |
| DF-3481 |
Medium |
7.8 |
ng_ppp mutates its entire private state (frag TAILQ, qlen, mseq/xseq, activeLinks, conf) with zero synchronization across concurrent v1 execution contexts — double TAILQ_REMOVE/kfree/m_freem (UAF-class) if raced; empirically contained on stock socket-fed topologies but one feeder refactor away
Kernel memory corruption (UAF / OOB write) · sys/netgraph/ppp/ng_ppp.c:1144 · pass 2 · new |
none |
not_reproduced |
| DF-3845 |
Medium |
7.8 |
OCR return-to-CAM flush races the lock-free ISR completion path: only a 1-second DELAY heuristic separates two unsynchronized mrsas_cmd_done callers (double xpt_done / duplicated free-list entry); timeout callout re-arms 600 s on recycled commands
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mrsas/mrsas.c:2324 · pass 2 · new |
none |
inconclusive |
| DF-4054 |
Medium |
7.8 |
SMU-firmware-fed DPM level count stored unclamped: vega12_setup_single_dpm_table fill loop writes past dpm_levels[16] — firmware-controlled linear heap overflow of the kzalloc'd vega12_hwmgr backend
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/vega12_hwmgr.c:481 · pass 2 · new |
none |
inconclusive fixed |
| DF-4148 |
Medium |
7.8 |
UAF/double-free race: in-flight cuse_client_open vs final cuse_server_unref (async destroy_dev, refs++ after the final decrement)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/cuse/cuse.c:1509 · pass 2 · new |
none |
not_reproduced |
| DF-4255 |
Medium |
7.8 |
agp_i915_write_gtt programs GTT PTEs into the MMADR register BAR (sc_res[0]) instead of GTTADR (sc_res[1]): GPU-register corruption incl. PGTBL_CTL hijack at PTE index 0x808, and OOB writes past the mapped MMADR resource for PTE indices >= 32768
Kernel memory corruption (UAF / OOB write) · sys/dev/agp/intel-gtt.c:998 · pass 2 · new |
none |
inconclusive |
| DF-4320 |
Medium |
7.8 |
SIOCGATHDIAG/SIOCGATHPHYERR dispatched with no privilege check: unprivileged users get arbitrary device MMIO read/write (incl. plaintext WEP/TKIP keycache dump), EEPROM reads, and NULL-deref panics on ATH_DIAGAPI kernels
Privilege escalation (general use) · sys/dev/netif/ath/ath/if_ath_ioctl.c:299 · pass 2 · new |
none |
inconclusive |
| DF-4439 |
Medium |
7.8 |
drm_atomic_state holds no drm_device reference while nonblocking commits outlive the ioctl on system_unbound_wq — missing backport of Linux 4e076c73e4f6 (stable-fix use-after-free)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_atomic.c:77 · pass 2 · new |
none |
inconclusive |
| DF-4469 |
Medium |
7.8 |
drm_crtc_helper_set_config() sizes save_connector_encoders with the racy num_connector while iterating the live list — concurrent DP-MST connector registration (spinlock-only) writes one pointer past the kcalloc'd array, and the fail path reads the OOB slot back into connector->encoder (heap OOB write + forged-pointer injection)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_crtc_helper.c:571 · pass 2 · new |
none |
inconclusive |
| DF-4497 |
Medium |
7.8 |
sbsh: deactivate() drops xq/rq mbufs but leaves stale pointers - TXS/RXS/UFL processing after carrier loss (or after ifconfig down) double-frees / uses freed mbufs
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/sbsh/if_sbsh.c:726 · pass 2 · new |
none |
inconclusive |
| DF-5033 |
Medium |
7.8 |
UAF/OOB-read race in amdgpu_bo_list_get(): DragonFly's idr_find() is lockless while idr_grow() kfree()s the node array synchronously (Linux RCU-safe-idr assumption does not hold)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_bo_list.c:167 · pass 2 · new |
none |
inconclusive |
| DF-5186 |
Medium |
7.8 |
DRM_UNLOCKED dumb-buffer MAP_DUMB races the unsynchronized legacy GEM offset-hash publication: live-key free_unr double-free, same-node dual-bucket hlist link, and wrong-object resolution in the mmap path (cross-object / UAF-class mapping ceiling)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dumb_buffers.c:115 · pass 2 · new |
none |
inconclusive |
| DF-5371 |
Medium |
7.8 |
Unlocked shared dmat->segments array across all load paths — concurrent loads on one tag corrupt each other's scatter/gather lists (vkernel never got pc64's bus_dma_tag_lock/BUS_DMA_CACHE_SEGMENTS fix)
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/busdma_machdep.c:314 · pass 2 · new |
device-visible corrupted S/G lists (DF-5371) / 72B-per-failure leak (DF-5373) |
untested |
| DF-5420 |
Medium |
7.8 |
pmap_inval_smp()/pmap_inval_smp_cmpset() shortcut samples pmap->pm_active outside the CPULOCK_EXCL interlock (and takes none for non-PMAP_MULTI pmaps); originator also returns before target done-bits drain
Kernel memory corruption (UAF / OOB write) · sys/platform/pc64/x86_64/pmap_inval.c:282 · pass 2 · new |
none |
not_reproduced |
| DF-2900 |
Low |
7.8 |
procfs_rwmem ignores kmem_alloc_pageable failure — kva=0 reaches pmap_kenter_quick(0,pa), overwriting pml4[0] of the current user pmap (page-table pivot) plus uiomove on VA 0
Privilege escalation (general use) · sys/vfs/procfs/procfs_mem.c:101 · pass 2 · new |
arbitrary_write |
not_reproduced |
| DF-4629 |
Low |
7.8 |
mpt_user: DMA page buffer freed on ioctl timeout while the deliberately-abandoned request still points at it — late IOC completion DMAs into freed kernel heap
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_user.c:259 · pass 2 · new |
Kernel heap corruption by device DMA: late IOC completion writes up to len bytes into freed (reallocatable) heap |
untested |
| DF-5143 |
Low |
7.8 |
GFP_DMA32 <4GiB constraint for I965G/GM internal objects is dead code on DragonFly — shim alloc_pages ignores gfp_mask, so pre-gen6 32-bit GTT PTEs truncate >4GiB physical addresses and the GPU DMAs to aliased physical pages
Base system — drivers · sys/dev/drm/i915/i915_gem_internal.c:72 · pass 2 · new |
latent 8B OOB write on sub-word call / GPU DMA to aliased physical pages |
untested |
| DF-3177 |
Medium |
7.7 |
hpfs NFS file handles are a bare 32-bit LSN — generation field commented out of vptofh, no hpfid_len/ino validation in fhtovp, and vget accepts any sector whose first u32 is FN_MAGIC: forged handles turn attacker-written file content into live vnodes, composing into remote arbitrary-device R/W and heap corruption on exported mounts
Privilege escalation (general use) · sys/vfs/hpfs/hpfs_vfsops.c:424 · pass 2 · new |
none |
untested |
| DF-3844 |
Medium |
7.7 |
mrsas_issue_blocked_cmd times out and lets callers release mfi commands still owned by the firmware — late completions land on recycled commands, double-inserting them into the free pool (shared sc->chan ident wakes the wrong waiter)
Base system — drivers · sys/dev/raid/mrsas/mrsas.c:2865 · pass 2 · new |
none |
inconclusive |
| DF-4671 |
Medium |
7.7 |
bktr_mmap is re-executed at page-fault time and reads bktr->bigbuf/alloc_pages with no lock while METEORSETGEO's grow path kmem_free()s and reassigns the buffer — fault in the window maps pages via a dangling KVA pointer (stale mappings; ceiling: RW kernel-memory alias)
Base system — drivers · sys/dev/video/bktr/bktr_os.c:677 · pass 2 · new |
none |
not_reproduced |
| DF-3881 |
High |
7.6 |
Unbounded MST topology depth (LCT) -> OOB stack/heap writes and unbounded kernel recursion from malicious hub
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dp_mst_topology.c:1052 · pass 2 · new |
none |
not_reproduced |
| DF-3820 |
Low |
7.6 |
aac(4) dereferences adapter-controlled completion/AIF indices with no bounds validation at five sites — malicious or faulty controller gets wild kernel pointers and unlink writes
Base system — drivers · sys/dev/raid/aac/aac.c:853 · pass 2 · new |
none |
inconclusive |
| DF-3823 |
Low |
7.6 |
asr_intr maps the adapter-posted FromFIFO reply offset into a kernel pointer with no validation — forged reply frame yields a forged ccb with LIST_REMOVE/callout_stop writes; FAIL path uses device-controlled MMIO offsets unbounded
Base system — drivers · sys/dev/raid/asr/asr.c:2890 · pass 2 · new |
none |
inconclusive |
| DF-3388 |
High |
7.5 |
wg peer teardown races in-flight noise consumers: wg_peer freed via uncounted r_arg and r_index re-linked after handshake_clear (use-after-free)
Kernel memory corruption (UAF / OOB write) · sys/net/wg/wg_noise.c:596 · pass 2 · new |
none |
inconclusive |
| DF-3843 |
High |
7.5 |
mrsas_sync_map_info writes firmware-controlled ldCount × 4-byte entries into a fixed 172,200-byte RAID-map DMA buffer — unbounded linear kernel heap overflow (computed bound size_sync_info is dead code)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mrsas/mrsas.c:3207 · pass 2 · new |
none |
inconclusive |
| DF-3886 |
High |
7.5 |
drm_dp_sideband_msg_build msg[256] accumulation overrun: up to 61 hub-controlled bytes past msg_rx->msg (distinct sink from DF-1264)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dp_mst_topology.c:371 · pass 2 · new |
none |
not_reproduced |
| DF-3908 |
High |
7.5 |
Kernel stack OOB write in tonga_program_memory_timing_parameters: VBIOS-controlled SCLK/MCLK DPM counts index fixed stack table arb_regs.entries[8][4]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/tonga_smumgr.c:1492 · pass 2 · new |
none |
inconclusive fixed |
| DF-3909 |
High |
7.5 |
Heap OOB writes into smc_state_table.GraphicsLevel[8]/MemoryLevel[4] via unclamped VBIOS SCLK/MCLK DPM counts (incl. unguarded MemoryLevel[count-1] store)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/tonga_smumgr.c:709 · pass 2 · new |
none |
inconclusive fixed |
| DF-3939 |
High |
7.5 |
Kernel stack OOB write in ci_program_memory_timing_parameters: VBIOS SCLK/MCLK DPM counts index fixed stack table arb_regs.entries[8][6] (DF-3908 twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/ci_smumgr.c:1653 · pass 2 · new |
none |
inconclusive |
| DF-3940 |
High |
7.5 |
Heap OOB writes into kzalloc'd struct ci_smumgr via unclamped VBIOS MCLK DPM count vs SMU7_MAX_LEVELS_MEMORY=6: MemoryLevel[6..7], unguarded MemoryLevel[count-1], mc_regs->data[6..7], mclk_levels[MemoryDpmLevelCount] (DF-3909 twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/ci_smumgr.c:1314 · pass 2 · new |
none |
inconclusive |
| DF-4000 |
High |
7.5 |
Unclamped VBIOS SCLK/MCLK DPM counts overflow the iceland SMC state tables and the 384-byte stack MC-arb table: GraphicsLevel[8], MemoryLevel[4], arb_regs.entries[8][4] (kernel STACK), mc_regs.data[4] — DF-3908/3909/3939/3940 family with iceland's tighter mclk bound (count 5 suffices)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/iceland_smumgr.c:980 · pass 2 · new |
none |
inconclusive fixed |
| DF-5246 |
High |
7.5 |
Write-after-free race on run_vap::beacon_mbuf: run_update_beacon (ic lock) vs run_newstate/run_vap_delete (sc lock) - disjoint lock domains
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_run.c:4887 · pass 2 · new |
none |
inconclusive |
| DF-5254 |
High |
7.5 |
Beacon mbuf use-after-free write: urtwn_update_beacon's unlocked ieee80211_beacon_update window races setup_beacon/vap_delete across disjoint lock domains
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_urtwn.c:2206 · pass 2 · new |
none |
inconclusive |
| DF-3547 |
Medium |
7.5 |
Use-after-free of struct ngd_private in all cdevsw methods: disconnect frees priv immediately after DragonFly's asynchronous destroy_dev()
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ng_device.c:299 · pass 2 · new |
none |
build_failed |
| DF-3679 |
Medium |
7.5 |
igb_rxeof trusts NIC-DMA'd descriptor length with no posted-buffer bound: unbounded m_len/m_pkthdr.len → heap OOB read up to 63,487 bytes past the 2046/2048-byte RX cluster (DF-3604/3639/3663/3667 family)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/igb/if_igb.c:2744 · pass 2 · new |
none |
inconclusive |
| DF-3713 |
Medium |
7.5 |
em_rxeof trusts the 16-bit NIC descriptor length with no posted-buffer bound: unbounded m_len/m_pkthdr.len -> heap OOB read up to ~63.5KB past the 2048-byte RX cluster (positive direction of DF-1188)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/em/if_em.c:3692 · pass 2 · new |
none |
inconclusive |
| DF-3759 |
Medium |
7.5 |
alc_rxeof trusts the controller-DMA'd 14-bit RRS frame length with no bound vs nsegs*alc_buf_size/MCLBYTES: heap OOB read up to ~14.3KB past a 2KB RX cluster (kernel heap info leak) and negative m_len (panic)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/alc/if_alc.c:3307 · pass 2 · new |
none |
inconclusive |
| DF-3782 |
Medium |
7.5 |
vmxnet3_rxq_eof() RX completion walk has no work bound: interrupt-thread livelock holding the RXQ lock under a remote line-rate flood (DF-3761/3778 family, vmx(4) instance)
Base system — drivers · sys/dev/virtual/vmware/vmxnet3/if_vmx.c:2201 · pass 2 · new |
dos |
inconclusive |
| DF-3990 |
Medium |
7.5 |
RX walk has no work bound (ALE_PROC_* knobs are dead code) — serializer-held livelock on line-rate flood; ale_txeof takes the 16-bit TPD_CONS_IDX unmasked vs the 256-ring — TX over-reclaim of still-owned mbufs (DF-3761/3778/3934 family)
Base system — drivers · sys/dev/netif/ale/if_ale.c:2026 · pass 2 · new |
none |
inconclusive |
| DF-4045 |
Medium |
7.5 |
stge_rxeof trusts the NIC-DMA'd 16-bit RxDMAFrameLen with no per-buffer bound: jumbo-MTU FrameEnd claim yields m_len up to 9022 on the 2048-byte cluster (up to 6974-byte heap OOB read into delivered frames); framelen < rxlen yields negative tail m_len -> m_copydata unsigned count -> ~4GB bcopy panic
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/stge/if_stge.c:1606 · pass 2 · new |
none |
inconclusive |
| DF-5165 |
Medium |
7.5 |
request_irq admits duplicate (irq,dev_id) registration and its failure path releases/NULLs pdev-global IRQ state owned by live registrations — armed handler left on released resource/MSI vectors (interrupt-time UAF), orphaned handler with dangling dev_id, NULL-resource crash on retry, double-release at teardown
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/linux_irq.c:75 · pass 2 · new |
interrupt-context UAF / orphaned armed handler / cross-device teardown |
untested |
| DF-5581 |
Medium |
7.5 |
Intel graphics stolen-memory reservation is compiled out; drm/i915 trusts the unreserved firmware values and its only conflict guard is a no-op rman stub
Base system — bus / platform / cpu · sys/bus/pci/x86_64/pci_early_quirks.c:264 · pass 2 · new |
none |
not_reproduced |
| DF-2740 |
Low |
7.5 |
Inconsistent locking of linker shared state: sys_kldload userrefs++/id read outside kld_lock (UAF-write window), unlocked refs mutations, and unlocked linker_files/found_modules traversals racing KPI loaders
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_linker.c:814 · pass 2 · new |
none |
untested |
| DF-4259 |
Low |
7.5 |
my_stop frees all RX/TX mbufs and zeroes descriptor rings without waiting for the TX/RX engines to go idle — trailing DMA into freed clusters (packet-controlled kernel heap corruption)
Base system — drivers · sys/dev/netif/my/if_my.c:1651 · pass 2 · new |
none |
inconclusive |
| DF-5162 |
Low |
7.5 |
bktr_mem stored-address registry is never invalidated when the METEORSETGEO grow path frees the frame buffer — driver reload reinstalls the freed KVA, bzeros ~3.5 MB and RISC-DMAs video into reallocated kernel memory (write-after-free)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/bktr/bktr_mem.c:103 · pass 2 · new |
none |
not_reproduced |
| DF-3938 |
Info |
7.5 |
Latent device-controlled-length memcpy in ena_com_get_hash_function(): keys_num (u32 in device-DMA buffer) << 2 as copy size
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/amazon/ena/ena-com/ena_com.c:2354 · pass 2 · new |
none |
inconclusive |
| DF-3520 |
High |
7.4 |
Negative-length bcopy in ng_pptpgre ack processing: one in-window ack converts the DF-0596 window violation into an unbounded kernel OOB read+write
Kernel memory corruption (UAF / OOB write) · sys/netgraph/pptpgre/ng_pptpgre.c:641 · pass 2 · new |
none |
not_reproduced |
| DF-3962 |
High |
7.4 |
mvfrey CCB-timeout callout runs iop_intr without the adapter lock, racing the locked hard-IRQ drain: double srb completion (double free to freelist + double xpt_done → CAM UAF); plus 60s DELAY busy-loop in softclock and the timed-out CCB is never completed
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptiop/hptiop.c:2647 · pass 2 · new |
none |
not_reproduced |
| DF-4066 |
High |
7.4 |
ti_rxeof indexes the mbuf chain arrays with the NIC-DMA'd uint16_t return-descriptor ti_idx and no bounds check: OOB pointer read used as struct mbuf (wild-pointer stores at :1690 + if_input type confusion) and OOB 8-byte NULL write up to +512KB past the chain arrays inside struct ti_softc
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ti/if_ti.c:1638 · pass 2 · new |
none observed (hardware-gated); source-level: kernel OOB write (8-byte NULL store up to +512KB past chain array) + wild-pointer mbuf writes + type confusion into ether_input via NIC-DMA'd uint16_t descriptor index |
inconclusive |
| DF-5258 |
High |
7.4 |
sc_epq[MTW_BULK_RX] is a zero-length array: the 6 TX endpoint queues alias 14 KB of live mtw_softc fields; every ifconfig-up stamps over them and every channel change (sc_chan_group) forges the upper dword of epq[0].tx_data[3].ni, a live ieee80211_node pointer
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_mtw.c:4466 · pass 2 · new |
none |
inconclusive |
| DF-3518 |
Medium |
7.4 |
ng_pptpgre priv UAF: rmnode kfree(priv) races in-flight rcvdata/rcvmsg dispatched inline without a node reference
Kernel memory corruption (UAF / OOB write) · sys/netgraph/pptpgre/ng_pptpgre.c:437 · pass 2 · new |
none |
not_reproduced |
| DF-3519 |
Medium |
7.4 |
Non-atomic node->refs++ / priv->timers++ in ng_pptpgre timer start vs ng_unref — lost update frees live ng_node (UAF)
Kernel memory corruption (UAF / OOB write) · sys/netgraph/pptpgre/ng_pptpgre.c:768 · pass 2 · new |
none |
not_reproduced |
| DF-3934 |
Medium |
7.4 |
RX walk has no work bound (AGE_PROC_* knobs are dead code) and 16-bit CMB producer indices are not masked to ring sizes — serializer-held livelock DoS and TX over-reclaim
Base system — drivers · sys/dev/netif/age/if_age.c:2149 · pass 2 · new |
none |
inconclusive |
| DF-3961 |
Medium |
7.4 |
SG-capacity heap overflow: firmware-fed max_sg_count writes scatter-gather entries past the 34-entry capacity of srb->req[600] into srb->hba/ccb/next/dma_map (itl non-highmem, mv, mvfrey paths)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptiop/hptiop.c:2471 · pass 2 · new |
none |
not_reproduced |
| DF-5242 |
Medium |
7.4 |
xpt_release_device() uses non-joining callout_stop before kfree(device) — xpt_release_devq_timeout can already be blocked on CAM_SIM_LOCK(device->sim) and then operate on the freed cam_ed (UAF; upstream defers the free)
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/cam_xpt.c:4790 · pass 2 · new |
heap OOB comparison oracle / race-gated UAF / accounting drift |
untested |
| DF-4421 |
Low |
7.4 |
ich_pci_attach() publishes the 0666 /dev/mixerN (mixer_init) before five failure paths whose bad: unwind frees the AC97 codec and softc — dangling mixer cdev UAF on open/ioctl (DF-4156/4193/4299/4334/4344 family, 6th instance, widest window in family)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/ich.c:1031 · pass 2 · new |
none |
inconclusive |
| DF-4726 |
Low |
7.4 |
aureal: BAR loop stores tags/handles for up to 6 BARs into au->st[3]/au->sh[3] — OOB kernel write over sh[], parent_dmat and lock at attach (crafted-PCI-device threat model)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/aureal.c:571 · pass 2 · new |
none |
untested |
| DF-4568 |
Info |
7.4 |
CT teardown frees the shared CT page while GuC interrupts are still enabled and guc->handler still points at CT code — interrupt-context use-after-free read+write of the released descriptor mapping
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_guc_ct.c:898 · pass 2 · new |
none |
untested |
| DF-4321 |
Medium |
7.3 |
ath_hal_getregdump(): signed/unsigned comparison break + unchecked do-while first write => unbounded attacker-length kernel-heap overflow with attacker-known register values
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_hal/ah.c:848 · pass 2 · new |
none |
inconclusive |
| DF-4836 |
Medium |
7.3 |
Firmware-facing frame->hdr.sge_count never reconciled with the validated user_ioc->sge_count: FW consumes stale SGEs (device DMA to freed ioctl buffers) and can read its SGL past the 1024-byte frame allocation
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mrsas/mrsas_ioctl.c:205 · pass 2 · new |
corruption |
untested |
| DF-4854 |
Low |
7.3 |
vega10_ih: unvalidated device-DMA'd IH wptr (DF-4843 twin) — permanent amdgpu interrupt-thread spin on misaligned wptr, plus 20-byte OOB ring read via RB_OVERFLOW recovery
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/vega10_ih.c:194 · pass 2 · new |
dos |
untested |
| DF-3911 |
Medium |
7.2 |
mpt_download_fw consumes the IOC-uploaded firmware image with zero validation: unchecked ImageSize and unbounded ext-image walk give wild kernel reads streamed to the device / page-fault panic
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt.c:2467 · pass 2 · new |
none |
inconclusive fixed |
| DF-3018 |
Low |
7.2 |
spec VOP dispatch (devfs_spec_read/write/ioctl/kqfilter, D_QUICK close) calls dev_d* on vp->v_rdev with no cdev reference and the vnode unlocked — vfs_timestamp(&node->atime) after teardown is a UAF write if the device dies during the dispatch
Kernel memory corruption (UAF / OOB write) · sys/vfs/devfs/devfs_vnops.c:1099 · pass 2 · new |
none |
not_reproduced |
| DF-4759 |
Low |
7.2 |
Module teardown is unsynchronized: vinum_inactive() TOCTOU vs vinumopen(), the state > volume_down gap skips open-but-down volumes, and VF_OPEN/VF_DAEMONOPEN are refcount-free booleans (UAF-class teardown window)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinum.c:171 · pass 2 · new |
none |
untested |
| DF-4567 |
Info |
7.2 |
CTB ring math trusts GuC-firmware-written descriptor fields with only compiled-out GEM_BUG_ON validation — desc->size==0 gives divide-by-zero panic, inflated size/tail gives FW-offset wild write past the 1-page CT object
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_guc_ct.c:324 · pass 2 · new |
none |
untested |
| DF-3749 |
High |
7.1 |
wcid2ni[] never cleared on node free → TX-status drain use-after-free (hostap remote)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ral/rt2860.c:887 · pass 2 · new |
none |
not_reproduced |
| DF-3882 |
High |
7.1 |
drm_dp_mst_handle_down_rep txmsg use-after-free race vs drm_dp_mst_wait_tx_reply timeout kfree
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dp_mst_topology.c:2337 · pass 2 · new |
none |
not_reproduced |
| DF-4444 |
High |
7.1 |
DragonFly-only drm_vma_node_unmap() cannot revoke a single user mapping: user CPU PTEs into the physical GGTT aperture survive VMA unbind/evict/fence-steal/object-free and read+write whatever object later reuses the aperture range (unprivileged cross-object/cross-user GPU memory access)
Base system — drivers · sys/dev/drm/i915/i915_vma.c:857 · pass 2 · new |
none |
inconclusive |
| DF-4791 |
High |
7.1 |
ttm_shrink ↔ ttm_bo_swapout unbounded recursion: the TTM memory-accounting shrinker re-enters itself through the lazy page-populate of the very BO it is swapping out — attacker-controlled nesting depth exhausts the 16KiB kernel stack
Base system — drivers · sys/dev/drm/ttm/ttm_memory.c:502 · pass 2 · new |
none |
not_reproduced |
| DF-3686 |
Medium |
7.1 |
Zero-count DPM tables (ignored setup failures) drive unsigned index 0xFFFFFFFF into dpm_levels[]: wild read/write ~48 GB out plus OOB reads and divide-by-zero across OD paths
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:3588 · pass 2 · new |
none |
inconclusive |
| DF-3712 |
Medium |
7.1 |
2-byte DMA heap overflow in em(4) RX when MTU in (1500, 2028]: ETHER_ALIGN front-trim coexists with RCTL_LPE long-frame reception on 2048-byte hardware buffers
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/em/if_em.c:3378 · pass 2 · new |
none |
inconclusive |
| DF-3930 |
Medium |
7.1 |
mly_cam_complete() INQUIRY rewrite writes vendor/product/revision at data_ptr[8..36) regardless of dxfer_len — heap OOB write via short pass(4) INQUIRY CCB on a logical device
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mly/mly.c:2341 · pass 2 · new |
none |
not_reproduced |
| DF-4008 |
Medium |
7.1 |
mpr diag-buffer registration: shared DMA trio + stale per-type size yields OOB kernel disclosure, cross-type free, and NULL-deref/UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_user.c:1508 · pass 2 · new |
none |
inconclusive |
| DF-4429 |
Medium |
7.1 |
radeon_verify_access returns 0 unconditionally: GEM ownership check dropped — any video-group user can mmap every other user's GPU buffer objects
Privilege escalation (general use) · sys/dev/drm/radeon/radeon_ttm.c:234 · pass 2 · new |
Any local user in group video can mmap and read/write every other user's radeon GEM buffer object (Xorg scanout contents, other users' textures/uploads) by sweeping mmap offsets on /dev/dri/cardN: radeon_verify_access (radeon_ttm.c:234) is an unconditional 'return 0'; the drm_vma_node_verify_access ownership check that upstream radeon and the in-tree sibling amdgpu_ttm.c:311-325 perform was dropped in this port. Kernel-internal BOs (ttm_bo_type_kernel) carry no vma node and are not exposed; ceiling is user GEM objects. |
untested |
| DF-2990 |
High |
7.0 |
newblk hash chains mutated without the softdep lock: lock-free insert/lookup races locked remove+kfree on shared chains (list corruption, UAF write into freed M_NEWBLK, 'lost block' panics)
Kernel memory corruption (UAF / OOB write) · sys/vfs/ufs/ffs_softdep.c:1003 · pass 2 · new |
none |
not_reproduced |
| DF-3188 |
High |
7.0 |
Unlocked pfsm RB-tree lookup-vs-final-remove race: hammer_load_pseudofs/hammer_rel_pseudofs lack any serialization — UAF + double free + RB tree corruption, sprayable unprivileged via the ungated HAMMERIOC_GET_PSEUDOFS on a nonexistent PFS
Kernel memory corruption (UAF / OOB write) · sys/vfs/hammer/hammer_pfs.c:68 · pass 2 · new |
none |
not_reproduced |
| DF-3629 |
High |
7.0 |
iwm_rx_tx_cmd: firmware-supplied qid indexes sc->txq[IWM_MAX_QUEUES=31] with no bounds check - OOB ring access, NULL writes at OOB offsets, wild node deref (iwm twin of DF-1122)
Base system — drivers · sys/dev/netif/iwm/if_iwm.c:3598 · pass 2 · new |
none |
inconclusive |
| DF-3699 |
High |
7.0 |
lkpi queue_work() per-CPU double-enqueue: unprivileged irq-vs-waitboost race on rps->work corrupts the system workqueue STAILQs
Privilege escalation (general use) · sys/dev/drm/i915/i915_irq.c:1838 · pass 2 · new |
none |
inconclusive |
| DF-3813 |
High |
7.0 |
mpr_wait_command()/mpr_request_polled() miss the realloc check on the concurrent-reinit wakeup/poll path — use-after-free on struct mpr_command
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr.c:3578 · pass 2 · new |
none |
inconclusive |
| DF-3880 |
High |
7.0 |
dsp_oss_syncstart use-after-free: lksleep drops the syncgroup lock mid-walk; timeout wake resumes SLIST_FOREACH over sm/sg freed by a concurrent SYNCSTART
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pcm/dsp.c:3027 · pass 2 · new |
none |
not_reproduced |
| DF-4014 |
High |
7.0 |
mps_user_pass_thru frees the in-flight command and its DMA buffer the instant mps_wait_command() times out (use-after-free with device DMA writeback)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps_user.c:975 · pass 2 · new |
none |
inconclusive |
| DF-4927 |
High |
7.0 |
radeon_sa_bo_next_hole leaves dangling borrowed fence pointers across inner-loop iterations → UAF write, UAF read, and double kfree in radeon_sa_bo_new (missing upstream fix 0381ac3ca2e7)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_sa.c:267 · pass 2 · new |
none |
inconclusive fixed |
| DF-5013 |
High |
7.0 |
mfi LD lifetime UAF: device_delete_child frees the mfi_disk softc under open descriptors and in-flight bios (AEN offline path has no gate at all)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mfi/mfi_disk.c:178 · pass 2 · new |
none |
not_reproduced |
| DF-5049 |
High |
7.0 |
mfi syspd lifetime UAF: AEN PD removal (MR_EVT_PD_REMOVED) deletes the syspd locklessly with no gate and mfi_syspd_detach frees the softc under open descriptors and bios parked on the controller bioq (DF-5013's syspd twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mfi/mfi_syspd.c:155 · pass 2 · new |
none |
not_reproduced |
| DF-5259 |
High |
7.0 |
Beacon mbuf use-after-free write: mtw_update_beacon manipulates rvp->beacon_mbuf with no MTW_LOCK (the port dropped FreeBSD's RUN_LOCK), racing m_freem in mtw_newstate(S_RUN)/mtw_vap_delete under the disjoint sc-lock domain
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_mtw.c:3855 · pass 2 · new |
none |
inconclusive |
| DF-2692 |
Medium |
7.0 |
lwpsignal() writes p->p_sigacts->ps_frominfo[] on the token-swapped delivery path holding only the lwp token, racing kern_execve()'s sigacts replacement/free (UAF write)
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_sig.c:1425 · pass 2 · new |
none |
untested |
| DF-2842 |
Medium |
7.0 |
vnode_pager_reference() lacks the vp↔object interlock — TOCTOU use-after-free on vm_object ref_count
Kernel memory corruption (UAF / OOB write) · sys/vm/vnode_pager.c:203 · pass 2 · new |
none |
not_reproduced |
| DF-3394 |
Medium |
7.0 |
vlan SIOCGIFMEDIA uses an unpinned parent ifnet pointer across a serializer drop: UAF (indirect if_ioctl call through freed ifnet) racing privileged parent destroy
Kernel memory corruption (UAF / OOB write) · sys/net/vlan/if_vlan.c:951 · pass 2 · new |
none |
not_reproduced |
| DF-3800 |
Medium |
7.0 |
mprsas_realloc_targets resizes the target array to stale pre-reset IOC facts and never updates sassc->maxtargets while the mapping layer is rebuilt from new facts: heap OOB write past sassc->targets[] (and past mapping_table[] in the opposite direction)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_sas.c:3911 · pass 2 · new |
none |
not_reproduced |
| DF-3928 |
Medium |
7.0 |
Unvalidated firmware channel counts overflow mly_cam_sim[6] and mly_btl[6][16] at attach (~4KB pointer writes + ~290KB strided flag writes past the softc)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mly/mly.c:1947 · pass 2 · new |
none |
not_reproduced |
| DF-3929 |
Medium |
7.0 |
mly_done() completion slot has no lower-bound check: command_id 0 indexes mly_command[-1] — OOB field writes + TAILQ pointer writes + indirect call through firmware-fed memory
Base system — drivers · sys/dev/raid/mly/mly.c:1535 · pass 2 · new |
none |
not_reproduced |
| DF-4007 |
Medium |
7.0 |
eb_move_to_gpu error path double-unreserves vmas: pin/fence/GEM-object refcount underflow → UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/i915_gem_execbuffer.c:1864 · pass 2 · new |
none |
inconclusive fixed |
| DF-4069 |
Medium |
7.0 |
Unclamped VBIOS SCLK DPM count overflows GraphicsLevel[8] in vegam_populate_all_graphic_levels (heap OOB write past vegam_smumgr)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:886 · pass 2 · new |
none |
inconclusive |
| DF-4070 |
Medium |
7.0 |
Unclamped VBIOS MCLK DPM count overflows MemoryLevel[4] in vegam_populate_all_memory_levels — count=5 already OOB
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:1047 · pass 2 · new |
none |
inconclusive |
| DF-4071 |
Medium |
7.0 |
Kernel STACK overflow in vegam_program_memory_timing_parameters: unclamped sclk x mclk counts fill a 640B stack SMU75_Discrete_MCArbDramTimingTable (entries[8][4])
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:1291 · pass 2 · new |
none |
inconclusive |
| DF-4157 |
Medium |
7.0 |
ADC DMA-ring completion feed uses the raw APU CURPTR register as its only bound — desync/power-state race yields unbounded kernel OOB write
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/maestro.c:1066 · pass 2 · new |
none |
untested |
| DF-4191 |
Medium |
7.0 |
sdhci_start_data treats bus_dmamap_load()==EINPROGRESS as a plain failure: the deferred sdhci_adma2_getaddr callback stays armed against the shared ADMA2 descriptor table (and the map against a freed buffer) — potential mid-transfer descriptor rewrite -> controller DMA to stale addresses
Base system — drivers · sys/dev/disk/sdhci/sdhci.c:1326 · pass 2 · new |
none |
untested |
| DF-4356 |
Medium |
7.0 |
postclose teardown order reversed: amdgpu_vm_fini before amdgpu_ctx_mgr_fini (DF-port divergence) - VM page tables freed under still-live context jobs (device-side UAF / GPU loss)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_kms.c:1073 · pass 2 · new |
none |
inconclusive |
| DF-4566 |
Medium |
7.0 |
VCN rings ship with zero CS ib validation — no .parse_cs hook, userspace decode messages (session handle, fb buffer addresses/sizes) reach firmware unvalidated
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_vcn.c:465 · pass 2 · new |
none |
untested |
| DF-5064 |
Medium |
7.0 |
drm_device_detach never unregisters or releases the drm_device: /dev/dri minors stay live and dispatch into unloaded driver-module memory after kldunload (UAF-class) plus unbounded drm_device leak; i915 never runs any driver teardown
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dragonfly.c:251 · pass 2 · new |
none |
not_reproduced |
| DF-5191 |
Medium |
7.0 |
sio_pccard_detach mutates com->gone/com->tp/com->ibuf and calls ttyclose without tty_token, racing siopen/sioclose/siosetwater into kernel heap UAF and double-free of com->ibuf
Kernel memory corruption (UAF / OOB write) · sys/dev/serial/sio/sio_pccard.c:117 · pass 2 · new |
none |
not_reproduced |
| DF-5260 |
Medium |
7.0 |
Deferred mtw_cmdq entries hold raw vap/ni pointers; mtw_vap_delete frees the mtw_vap without draining sc->cmdq_task - mtw_update_beacon_cb/mtw_newassoc_cb/mtw_key_set_cb execute on freed objects
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_mtw.c:909 · pass 2 · new |
none |
inconclusive |
| DF-5262 |
Medium |
7.0 |
Lockless mtw_cmdq[] ring: producers store .func before the args with no barrier while mtw_cmdq_cb consumes concurrently under MTW_LOCK (which producers never take) - torn publish executes callbacks with stale pointers
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_mtw.c:1937 · pass 2 · new |
none |
inconclusive |
| DF-2959 |
Low |
7.0 |
ucom_get_data TX path passes unclamped request length to clist_qtob — latent heap OOB write past USB page-cache segment (consumer-side defect found in tty_subr.c pass-2 API sweep)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/serial/usb_serial.c:1898 · pass 2 · new |
none |
untested |
| DF-2961 |
Low |
7.0 |
kobj_class_free() reclaims the compiled ops table with no grace period while the lock-free dispatch hot path reads it (force multiplier turning any consumer refcount slip into a freed function-pointer table call)
Kernel memory corruption (UAF / OOB write) · sys/kern/subr_kobj.c:179 · pass 2 · new |
corruption |
not_reproduced |
| DF-3316 |
Low |
7.0 |
tcp_pcblist marker-dance walks per-cpu inpcb lists from a preemptible user thread with no synchronization — DF-3315's non-atomic advance race, here with fully unprivileged churn (UAF read/write on struct inpcb)
Kernel memory corruption (UAF / OOB write) · sys/netinet/tcp_subr.c:1271 · pass 2 · new |
none |
not_reproduced |
| DF-3588 |
Low |
7.0 |
Double free of adev->pm.dpm.ps on allocation-failure path in si_parse_power_table (plus leak of already-parsed si_ps states)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/si_dpm.c:7256 · pass 2 · new |
OOM-gated double free |
untested |
| DF-3811 |
Low |
7.0 |
intel_ddi_init error path kfrees intel_dig_port while a fully-registered DP connector still references it (OOM-gated UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_ddi.c:3863 · pass 2 · new |
none |
untested |
| DF-4156 |
Low |
7.0 |
agg_attach() error path frees the AC97 codec while the 0666 /dev/mixerN mixer_init() created survives — post-failure mixer ioctl UAF through freed kobj
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/maestro.c:1894 · pass 2 · new |
none |
untested |
| DF-4334 |
Low |
7.0 |
via_attach() pcm_register()-failure path destroys the AC97 codec while the 0666 /dev/mixerN mixer_init() created survives — dangling mixer devinfo UAF on open/ioctl
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/via8233.c:1335 · pass 2 · new |
none |
untested |
| DF-4344 |
Low |
7.0 |
atiixp_chip_post_init mixer_init/pcm_register error-path asymmetry leaves a live 0666 /dev/mixerN over freed codec+softc — UAF on open/ioctl and NULL-deref suspend (DF-4156 family, 5th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/atiixp.c:1085 · pass 2 · new |
none |
inconclusive |
| DF-4457 |
Low |
7.0 |
cmi_attach error path after mixer_init leaves a live 0666 /dev/mixerN pointing at freed sc — UAF on open/ioctl (DF-4156 family, 7th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/cmi.c:986 · pass 2 · new |
none |
inconclusive |
| DF-4504 |
Low |
7.0 |
t4dwave: attach-failure paths after mixer_init() leave a live 0666 /dev/mixerN over freed codec/tr_info — mixer-UAF family (9th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/t4dwave.c:882 · pass 2 · new |
none |
untested |
| DF-4537 |
Low |
7.0 |
cs4281: attach-failure path after mixer_init() leaves a live 0666 /dev/mixerN over freed codec/softc - mixer-UAF family (10th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/cs4281.c:845 · pass 2 · new |
none |
untested |
| DF-4550 |
Low |
7.0 |
als_pci_attach error path after mixer_init leaves a live 0666 /dev/mixerN pointing at freed sc — dangling mixer devinfo UAF on open/ioctl (DF-4156 family, 11th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/als4000.c:842 · pass 2 · new |
none |
not_reproduced |
| DF-4610 |
Low |
7.0 |
nm_pci_attach failure paths after mixer_init() leave a live 0666 /dev/mixerN over freed codec+softc — mixer-UAF family, 13th instance (DF-4156/4193/4299/4334/4344/4421/4457/4468/4504/4537/4550)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/neomagic.c:695 · pass 2 · new |
none |
untested |
| DF-4724 |
Low |
7.0 |
aureal: attach failure after mixer_init() leaves the live 0666 /dev/mixerN over the freed au softc — DF-4156 mixer-UAF family, 15th instance (premise live-verified on the guest)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/aureal.c:634 · pass 2 · new |
none |
untested |
| DF-4743 |
Low |
7.0 |
via_attach(): five failure paths after mixer_init() free the AC97 codec while the 0666 /dev/mixerN survives — dangling mixer devinfo UAF on unprivileged open/ioctl (DF-4156 mixer-UAF family, 16th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/via82c686.c:548 · pass 2 · new |
none |
untested |
| DF-3585 |
High |
6.8 |
Heap OOB write in si_get_svi2_voltage_table (and lowSMIO sink in si_populate_smc_voltage_table): BIOS dependency-table count (u8, up to 255) copied into fixed entries[32]/lowSMIO[32] arrays
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/si_dpm.c:4410 · pass 2 · new |
none |
inconclusive |
| DF-3981 |
High |
6.8 |
login_callout never cancelled when sbp_dev is freed (stale-LUN invalidation, !VALID_LUN loop, post_explore GC) — softclock fires sbp_login_callout on the kfree'd sdev and runs sbp_mgm_orb writes into freed slab + fwdma_free'd ORB pages
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/sbp/sbp.c:625 · pass 2 · new |
none |
inconclusive |
| DF-3736 |
Medium |
6.8 |
emx_rxeof trusts the 16-bit NIC descriptor length with no posted-buffer bound: unbounded m_len/m_pkthdr.len -> heap OOB read up to ~63.5KB past the 2046/2048-byte RX cluster (DF-3713 em(4) twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/emx/if_emx.c:3369 · pass 2 · new |
none |
inconclusive |
| DF-3887 |
Medium |
6.8 |
VBIOS usVoltageIndex (USHORT 0..65535) OOB heap write in sumo_construct_vid_mapping_table — up to 5 chosen-offset 16-bit-pair writes into/after the kv_power_info heap object
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/kv_dpm.c:167 · pass 2 · new |
none |
inconclusive |
| DF-4107 |
Medium |
6.8 |
vge_rxeof trusts the NIC-DMA'd 14-bit RX descriptor length with no MCLBYTES clamp: single-fragment path sets m_len/pkthdr.len up to 16379 on a 2048-byte cluster (~14KB heap OOB read into if_input/BPF) and signed-underflows to -4 for BUFSIZ<4 (DF-3604/4055/4083 family)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/vge/if_vge.c:1268 · pass 2 · new |
none |
untested |
| DF-2904 |
Low |
6.8 |
mbr_extended()'s signed int slice counter can wrap negative after 2^31 accepted EBR entries, defeating the `slice >= MAX_SLICES` guard → OOB write of attacker-controlled ds_offset/ds_size before dss_slices[]
Kernel memory corruption (UAF / OOB write) · sys/kern/subr_diskmbr.c:423 · pass 2 · new |
none |
untested |
| DF-3402 |
Low |
6.8 |
slstart drains sc_fastq/ifsq under tty locks while sloutput fills them under the ifsq serializer — no common lock, concurrent ifqueue corruption
Kernel memory corruption (UAF / OOB write) · sys/net/sl/if_sl.c:494 · pass 2 · new |
none |
not_reproduced |
| DF-3644 |
High |
6.7 |
Heap OOB write in ci_get_svi2_voltage_table: VBIOS dependency-table count (u8, up to 255) copied into fixed entries[32] before any trim
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/ci_dpm.c:2159 · pass 2 · new |
none |
inconclusive |
| DF-3739 |
High |
6.7 |
BIOS-count-driven OOB pointer write past rdev->i2c_bus[16] in radeon_atombios_i2c_init
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_atombios.c:172 · pass 2 · new |
none |
inconclusive |
| DF-3740 |
High |
6.7 |
Stack buffer overflow in radeon_atom_get_mclk_range_table: uncapped (u8) num_entries fill of mclk[16] into a caller stack local
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_atombios.c:3938 · pass 2 · new |
none |
inconclusive |
| DF-3698 |
Medium |
6.7 |
cu_info walkers index bitmap[4][4]/ao_cu_bitmap[4][4] with unclamped max_shader_engines/max_sh_per_se taken from gpu_info firmware or VBIOS → heap OOB write + read
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/gfx_v9_0.c:937 · pass 2 · new |
none |
not_reproduced |
| DF-3704 |
Medium |
6.7 |
vega10_odn_edit_dpm_table signed-compare lets negative user 'long' pass the bounds check -> wild kernel write (DF-3687 twin)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:4884 · pass 2 · new |
none |
inconclusive |
| DF-3966 |
Medium |
6.7 |
rxeof trusts the device-fed 12-bit RX frame length against a 2048-byte cluster: m_len up to 4095 → up to 2047-byte OOB read past the mbuf cluster into ieee80211_input (kernel heap disclosure / parser crash)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ral/rt2661.c:1101 · pass 2 · new |
none |
not_reproduced |
| DF-4067 |
Medium |
6.7 |
ti_rxeof trusts the NIC-DMA'd 16-bit return-descriptor ti_len with no bound vs the backing buffer: std-ring m_len up to 65535 on a 2048-byte cluster (63489-byte heap OOB read into delivered frames), jumbo up to 56513 bytes past a 9024-byte slot — DF-3604/3989/4024/4045/4055 family, ti(4) instance
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ti/if_ti.c:1636 · pass 2 · new |
none observed (hardware-gated); source-level: kernel heap OOB read up to 63489 bytes past a 2048-byte std cluster (65535 on u16 ti_len) delivered into received frames via BPF/sockets; jumbo path up to 56513 bytes past a 9024-byte slot incl. past the contigmalloc block; page-fault panic variant |
inconclusive |
| DF-4967 |
Medium |
6.7 |
radeon_gart_bind/unbind take a 32-bit `unsigned offset`: GTT aperture offsets >= 4 GiB truncate and alias GART entries across buffer objects (cross-process GPU data redirection)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_gart.c:241 · pass 2 · new |
none |
untested |
| DF-5293 |
Medium |
6.7 |
Deferred rum_cmdq entry holds a raw struct ieee80211vap *; rum_vap_delete() frees the rum_vap without draining sc->cmdq_task - rum_update_beacon_cb/rum_set_beacon execute on freed memory
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_rum.c:2830 · pass 2 · new |
none |
inconclusive |
| DF-5294 |
Medium |
6.7 |
rum_ratectl_task re-arms rvp->ratectl_ch unconditionally; the re-armed callout survives rum_vap_delete's drain order and fires on the freed rum_vap (post-free callout/task cycle)
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_rum.c:3140 · pass 2 · new |
none |
inconclusive |
| DF-3805 |
Low |
6.7 |
Unbounded csio->cdb_len bcopy into the 32-byte MPI2 CDB field overflows the request frame in the req_frames DMA pool
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_sas.c:2102 · pass 2 · new |
none |
not_reproduced |
| DF-3920 |
Low |
6.7 |
OsSendCommand bcopy's user-controlled cdb_len (<=255) into UCHAR Cdb[16] - kernel stack overflow
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptmv/entry.c:2771 · pass 2 · new |
none |
not_reproduced |
| DF-4019 |
Low |
6.7 |
debugfs amdgpu_iomem: upstream page-ownership check is #if 0'd, turning the debug file into arbitrary physical-memory read/write for debugfs holders
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_ttm.c:2424 · pass 2 · new |
debugfs amdgpu_iomem: the upstream page-ownership check is #if 0'd ('XXX DRAGONFLY - dev_mapping NULL atm'), so the file gives its holder arbitrary physical-memory read/write by llseek+read/write (only pfn_valid is checked). |
untested |
| DF-4001 |
High |
6.6 |
Unclamped VBIOS SCLK/MCLK DPM counts overflow fiji SMC GraphicsLevel[8]/MemoryLevel[4] and the 384-byte STACK MC-arb timing table
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/fiji_smumgr.c:1026 · pass 2 · new |
none |
inconclusive fixed |
| DF-4002 |
High |
6.6 |
Unclamped VBIOS MM-dependency count overflows fiji UvdLevel[8]/VceLevel[8]/AcpLevel[8] SMC arrays
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/fiji_smumgr.c:1433 · pass 2 · new |
none |
inconclusive fixed |
| DF-4307 |
Medium |
6.6 |
adw_intr walks device-fed 32-bit completion tokens (areq_ba/next_ba) through unchecked acbbotov/carrierbotov pointer arithmetic — OOB kernel read/write and interrupt livelock
Base system — drivers · sys/dev/disk/advansys/adwcam.c:1280 · pass 2 · new |
none |
untested |
| DF-3688 |
Low |
6.6 |
Unbounded sprintf loops in smu7_print_clock_levels overflow the 4096-byte sysfs page buffer once the VBIOS DPM count exceeds ~233
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:4428 · pass 2 · new |
none |
inconclusive |
| DF-4027 |
High |
6.5 |
sili deferred timeout hands a port-owned ccb (state ATA_S_ONCHIP) back to poll callers which sili_ata_put_xfer() it — ccb/slot double lifecycle: reissue on an active slot, stale sili_unload_prb of the new command's live dmamap, and double CAM completion
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/sili/sili.c:1359 · pass 2 · new |
none |
not_reproduced |
| DF-4404 |
High |
6.5 |
u32 integer overflow in amdgpu_uvd_cs_msg_decode defeats DPB/ctx buffer size validation — UVD runs physical addressing, so undersized decode sessions GPU-DMA past validated buffers (cross-process VRAM/GTT corruption)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_uvd.c:524 · pass 2 · new |
none |
inconclusive |
| DF-3158 |
Medium |
6.5 |
deget() publishes the denode in the global hash before taking its reference (refcnt store outside the token): SMP TOCTOU permits duplicate denodes for one directory entry
Base system — filesystems · sys/vfs/msdosfs/msdosfs_denode.c:308 · pass 2 · new |
none |
not_reproduced |
| DF-3360 |
Medium |
6.5 |
pf scrub max-mss: 2-byte stack OOB read/write past opts[40] in pf_normalize_tcpopt() via malformed MSS option
Kernel memory corruption (UAF / OOB write) · sys/net/pf/pf_norm.c:1915 · pass 2 · new |
none |
not_reproduced |
| DF-3576 |
Medium |
6.5 |
Remote kernel panic: division by zero in iwn_set_timing() via AP beacon with bintval=0 (net80211 stores the invalid value anyway)
Base system — drivers · sys/dev/netif/iwn/if_iwn.c:5596 · pass 2 · new |
panic |
inconclusive |
| DF-4221 |
Medium |
6.5 |
pvscsi_process_completion() writes through a NULL ccb on duplicate/spurious device completion (missing guard its own timeout handler has)
Base system — drivers · sys/dev/virtual/vmware/pvscsi/pvscsi.c:903 · pass 2 · new |
none |
inconclusive |
| DF-4362 |
Medium |
6.5 |
Atomic-ioctl out-fence path leaks a struct file per commit against the port's stubbed fd/sync_file shims: get_unused_fd_flags() really allocates while fd_install/fput/put_unused_fd are no-op stubs and sync_file_create() returns NULL — permanent nfiles exhaustion (system-wide ENFILE DoS until reboot)
Privilege escalation (general use) · sys/dev/drm/drm_atomic_uapi.c:1048 · pass 2 · new |
dos |
untested |
| DF-4413 |
Medium |
6.5 |
radeon_vm_bo_set_addr() commits the mapping before page-table allocation — allocation failure leaves NULL page tables and the next update panics in radeon_vm_update_ptes()
Base system — drivers · sys/dev/drm/radeon/radeon_vm.c:519 · pass 2 · new |
none |
not_reproduced |
| DF-4458 |
Medium |
6.5 |
drm_gem_mmap_single passes the user-controlled mmap length and in-object offset (bits 0..39 of the fake offset) to the GEM device pager with no bounds check — the only upstream clamp is dead in #if 0
Base system — drivers · sys/dev/drm/drm_gem.c:1082 · pass 2 · new |
none |
inconclusive |
| DF-4564 |
Medium |
6.5 |
radeon: dropped radeon_bo_unreserve() in radeon_driver_open_kms permanently holds the IB-pool BO reservation — first DRM close (and every later open) hangs uninterruptible, wedging the GPU for all users
Base system — drivers · sys/dev/drm/radeon/radeon_kms.c:662 · pass 2 · new |
dos |
not_reproduced |
| DF-4607 |
Medium |
6.5 |
4x-timeout DANGER1 direct completion races the real completion: concurrent ata_completed -> UAF + double objcache_put
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-queue.c:140 · pass 2 · new |
none |
not_reproduced |
| DF-4882 |
Medium |
6.5 |
Reachable KASSERT panic in ath_rate_ctl_start() from a zero-length Supported Rates IE beacon — remote DoS on default (INVARIANTS) DragonFly kernels with ath+onoe+IBSS
Base system — drivers · sys/dev/netif/ath/ath_rate/onoe/onoe.c:285 · pass 2 · new |
none |
inconclusive |
| DF-5108 |
Medium |
6.5 |
lockrange() commits pos one past plex->lock[PLEX_LOCKS] when the table fills while a conflicted requester sleeps (16-byte heap OOB write + usedlocks overrun defeating the == fullness checks)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumlock.c:139 · pass 2 · new |
memcorrupt |
not_reproduced |
| DF-5223 |
Medium |
6.5 |
rv770_copy_dma under-reserves the DMA ring for radeon_sync_rings() semaphore waits: with two pending cross-ring fences, 16 dwords are written past the reserved window, overwriting un-executed GPU commands (user-reachable rv770 twin of DF-1943)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/rv770_dma.c:59 · pass 2 · new |
none |
not_reproduced |
| DF-5266 |
Medium |
6.5 |
MTW_AID2WCID hashes 64 station AIDs into 8 WCIDs: in hostap with >8 stations the 9th+ pairwise key overwrites a colliding station's key slot, and downlink frames are then encrypted with ANOTHER station's key (cross-station confidentiality break on WPA2-Enterprise)
Privilege escalation (general use) · sys/bus/u4b/wlan/if_mtw.c:37 · pass 2 · new |
none |
untested |
| DF-3863 |
Low |
6.5 |
u32 truncation of nptes in amdgpu_vm_bo_update_mapping undersizes the update IB; amdgpu_vm_update_ptes then emits an unbounded command stream past its end (kernel heap overflow)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_vm.c:1807 · pass 2 · new |
none |
not_reproduced |
| DF-4235 |
Low |
6.5 |
rl(4): any RX-error frame triggers a full adapter re-init in interrupt context under the ifnet serializer (remote per-frame DoS amplifier)
Base system — drivers · sys/dev/netif/rl/if_rl.c:1028 · pass 2 · new |
none |
inconclusive |
| DF-4488 |
Low |
6.5 |
vo destruction runs under the 'video' serial only while all vo sysctl handlers run under 'video_output' only — use-after-free race on vo/vo_levels with an unprivileged 'levels' reader
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_video/acpi_video.c:300 · pass 2 · new |
none |
untested |
| DF-4513 |
Low |
6.5 |
vmbus: vmbus_msghc_wait_result blocks uninterruptibly with no timeout - a silent or malicious Hyper-V host permanently wedges device attach at boot (and later wedges vmbus_msghc_ctx_destroy/kldunload too)
Base system — drivers · sys/dev/virtual/hyperv/vmbus/vmbus.c:517 · pass 2 · new |
Permanent boot-hang / sticky msghc wedge with no panic or recovery (availability twin of DF-1668's memory corruption) |
untested |
| DF-3607 |
Medium |
6.4 |
ncr(4): timeout daemon completes an nccb still armed in the chip's DSA — later phase-mismatch resumes DMA from stale/reused scatter descriptors (memory corruption) or wedges the HBA
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/ncr/ncr.c:5148 · pass 2 · new |
none |
inconclusive |
| DF-3660 |
Medium |
6.4 |
isp_detach() never terminates or reaps the per-channel FC kthreads nor stops ldt/gdt callouts; isp_pci_detach() then frees the lock, isp_param and bus resources they still use - teardown use-after-free
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/isp/isp_freebsd.c:159 · pass 2 · new |
none |
not_reproduced |
| DF-3661 |
Medium |
6.4 |
isp_attach_chan() cam_simq_release()s the shared devq once per channel while isp_attach held only one reference - premature cam_devq free + write-after-free on multi-channel adapters
Base system — drivers · sys/dev/disk/isp/isp_freebsd.c:95 · pass 2 · new |
none |
not_reproduced |
| DF-3812 |
Medium |
6.4 |
isp_add_wwn_entry / isp_del_wwn_entry index FCPARAM(isp, chan) with an unchecked firmware-controlled channel (raw in_vpidx) — kernel heap OOB read/write up to ~17.8 MB past isp_param; isp_del_all_wwn_entries off-by-one admits chan == nchan
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/isp/isp_library.c:2398 · pass 2 · new |
none |
inconclusive |
| DF-4029 |
Medium |
6.4 |
SET_RESOURCE VALID_BUFFER relocates with 8-dword stride into 7-dword resource slots (validated/executed desync)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_cs.c:1953 · pass 2 · new |
none |
not_reproduced |
| DF-4083 |
Medium |
6.4 |
sis_rxeof trusts the 12-bit DMA-written RX descriptor length (SIS_RXBYTES = (sis_ctl & 0xFFF) - 4, up to 4091) into m_len/m_pkthdr.len of a 2048-byte cluster with no clamp: up to ~2 KB kernel heap OOB read delivered to the network stack/BPF (DF-3604 length-trust family)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/sis/if_sis.c:1374 · pass 2 · new |
none |
untested |
| DF-4178 |
Medium |
6.4 |
twe_immediate_request timeout releases a request still posted to the controller and linked on twe_busy — late completion double-processes the reused request (TAILQ corruption / biodone of in-flight bio)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/twe/twe.c:1003 · pass 2 · new |
none |
inconclusive |
| DF-4341 |
Medium |
6.4 |
Unserialized ISR/netgraph concurrency over per-channel descriptor rings: UAF writes, hook write-after-free, descriptor double-free, and device DMA into freed mbufs (ignored channel-zap status)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mn/if_mn.c:1329 · pass 2 · new |
none |
inconclusive |
| DF-4414 |
Medium |
6.4 |
radeon_vm_fini() DragonFly rewrite spins forever when radeon_bo_reserve() fails — unkillable kernel livelock (upstream single-pass merely leaks)
Base system — drivers · sys/dev/drm/radeon/radeon_vm.c:1250 · pass 2 · new |
none |
not_reproduced |
| DF-4614 |
Medium |
6.4 |
ips command-queue lifecycle double-free/UAF: unchecked ips_cmdqueue_init at ips.c:414 plus unconditional, non-NULLing kfree in ips_cmdqueue_free (also frees under live commands on detach)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ips/ips.c:116 · pass 2 · new |
none |
not_reproduced |
| DF-4762 |
Medium |
6.4 |
fwe_stop() frees RX bulkxfer chunks and their mbufs while fwohci_rbuf_update()/fwe_as_input() still manipulate or walk them — teardown race causes UAF write/read and mbuf double-free
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/fwe/if_fwe.c:258 · pass 2 · new |
none |
inconclusive |
| DF-4884 |
Medium |
6.4 |
Double-free of adev->bios when amdgpu_get_bios() fails: every VBIOS read method kfrees without NULLing and amdgpu_device_fini() frees the dangling pointer again on the probe-failure teardown path
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_bios.c:116 · pass 2 · new |
none |
not_reproduced |
| DF-4910 |
Medium |
6.4 |
Unsynchronized aconnector->dc_sink teardown: lockless dm_dp_destroy_mst_connector (destroy work) races mode_config.mutex-held dm_dp_mst_get_modes check-then-set — dc_sink leaked with dangling dc_sink->priv, remote-sink slot exhaustion makes the known DF-1966 NULL-deref panic deterministic, plus UAF-read windows
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c:232 · pass 2 · new |
none |
untested |
| DF-5247 |
Medium |
6.4 |
Deferred run_cmdq entries outlive the vap/node they point at: run_vap_delete frees run_vap without draining sc->cmdq_task
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_run.c:1030 · pass 2 · new |
none |
inconclusive |
| DF-3278 |
Low |
6.4 |
bridge_fragment() re-header loop double-frees the fragment chain and UAF-reads freed mbufs when any M_PREPEND(M_NOWAIT) fails
Kernel memory corruption (UAF / OOB write) · sys/net/bridge/if_bridge.c:4386 · pass 2 · new |
none |
not_reproduced |
| DF-3732 |
Low |
6.4 |
ciss_init_pci: unvalidated firmware transport_offset maps ciss_perf outside the BAR window; ciss_init_perf writes through it
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ciss/ciss.c:694 · pass 2 · new |
none |
inconclusive |
| DF-3733 |
Low |
6.4 |
ciss_periodic re-arms after unlock vs callout_stop in ciss_free -> post-free callout execution (UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ciss/ciss.c:3431 · pass 2 · new |
none |
inconclusive |
| DF-3756 |
Low |
6.4 |
tulip_srom_decode() unbounded media-info pointer walk: OOB WRITE past tulip_softc.tulip_mediainfo[10] into the softc tail and kernel heap
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/de/if_de.c:2080 · pass 2 · new |
none |
inconclusive |
| DF-3923 |
Low |
6.4 |
hpt_attach bus_setup_intr failure kfree()s the newbus-owned IAL_ADAPTER_T softc while it stays linked on gIal_Adapter (use-after-free)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptmv/entry.c:2026 · pass 2 · new |
none |
not_reproduced |
| DF-4076 |
Low |
6.4 |
Unvalidated chip-reported IOCRequestFrameSize sizes every request slot; all request builders write fixed sizeof(struct) with no minimum check
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps.c:750 · pass 2 · new |
none |
inconclusive |
| DF-4193 |
Low |
6.4 |
es_pci_attach error path after mixer_init leaves a live /dev/mixerN pointing at freed codec/es (UAF on open/ioctl)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/es137x.c:1810 · pass 2 · new |
es_pci_attach error path after mixer_init leaves a live 0666 /dev/mixerN pointing at freed codec/es (UAF on open/ioctl). Attach-failure precondition only (IRQ/dma-tag/pcm_register failure after mixer_init) - not user-triggerable on demand. Same family as DF-4156 (maestro). |
untested |
| DF-4274 |
Low |
6.4 |
twa(4) deferred-busdma passthru error path returns the request to the free queue twice — queue corruption and cross-thread request reuse (latent UAF/double-free; sibling of twe DF-4178)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/twa/tw_osl_freebsd.c:1282 · pass 2 · new |
none |
untested |
| DF-4611 |
Low |
6.4 |
nm_init() programs ring/coeff buffers from unvalidated card-provided buftop — cbuf/rbuf/pbuf/acbuf wrap to ~4G offsets used by nm_wrbuf() and sndbuf_setup()
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/neomagic.c:576 · pass 2 · new |
none |
untested |
| DF-4628 |
Low |
6.4 |
bxt_exec_gpio: function-static bxt_gpio_table caches devm-allocated gpio_descs across device unbind/rebind — use-after-free on gpiod_set_value
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_dsi_vbt.c:304 · pass 2 · new |
UAF write (gpiod_set_value mutates desc->flags/value) on first panel-enable after rebind; potentially exploitable if the freed chunk is reclaimed |
untested |
| DF-4963 |
Low |
6.4 |
snd_hdspe attach has no error unwind: failed attach leaks snd lock/IRQ/DMA resources and leaves hdspe_intr live on a dead device — after kldunload the IRQ vector dispatches into unmapped module text
Base system — drivers · sys/dev/sound/pci/hdspe.c:293 · pass 2 · new |
corruption |
untested |
| DF-5326 |
Low |
6.4 |
mp_bootaddress() integer underflow on degenerate base-memory size corrupts the kernel physical memory map
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/mp_machdep.c:212 · pass 2 · new |
none |
untested |
| DF-2645 |
Medium |
6.3 |
Flush-driven indirect collapse contradicts its own skip guard: concurrent child COW-modify during hammer2_chain_indirect_maintenance leaves a live child, tripping the unconditional repchange emptiness KKASSERT (panic) or orphaning the subtree (silent data loss)
Hammer2 · sys/vfs/hammer2/hammer2_flush.c:1039 · pass 2 · new |
none |
not_reproduced |
| DF-3000 |
Medium |
6.3 |
hammer_vop_write UIO_NOCOPY (putpages/msync) path: bqrelse(bp) followed by bread(&bp) makes breadnx reuse the RELEASED buffer; write-out then runs on an unlocked free-queued buffer (guaranteed panic('buffer is not busy') + buffer-recycle race window); identical pattern in hammer2
Base system — filesystems · sys/vfs/hammer/hammer_vnops.c:690 · pass 2 · new |
none |
not_reproduced |
| DF-3093 |
Medium |
6.3 |
MSDOSFS_LOCK_MP expands to nothing: the free-cluster allocator (clusteralloc/chainalloc/usemap_alloc + pm_inusemap scan) is unsynchronized on SMP — double-allocation/cross-linked files (release) or KASSERT panic (INVARIANTS)
Base system — filesystems · sys/vfs/msdosfs/msdosfs_fat.c:340 · pass 2 · new |
none |
not_reproduced |
| DF-3116 |
Medium |
6.3 |
msdosfs_rename() restores stale de_fndoffset/de_fndcnt after dropping the target-directory lock — createde() blindly overwrites a concurrently-created dirent (TOCTOU)
Base system — filesystems · sys/vfs/msdosfs/msdosfs_lookup.c:314 · pass 2 · new |
none |
not_reproduced |
| DF-3118 |
Medium |
6.3 |
Signed 32-bit flusher sequence ring misorders all completion/mis-sequencing comparisons past 2^31; wrapped state causes permanently stuck flush group and self-sustaining synchronous-finalize storm (system wedge)
Kernel memory corruption (UAF / OOB write) · sys/vfs/hammer/hammer_flusher.c:114 · pass 2 · new |
dos |
inconclusive |
| DF-3345 |
Medium |
6.3 |
lagg_lacp_detach drops the lagg WLock mid-detach while sc_* protocol callbacks remain installed: ioctl re-entry into half-torn-down LACP state (incl. unprivileged SIOCGLAGG)
Kernel memory corruption (UAF / OOB write) · sys/net/lagg/if_lagg.c:1085 · pass 2 · new |
none |
not_reproduced |
| DF-3604 |
Medium |
6.3 |
bce_rx_intr trusts NIC-DMA'd l2_fhdr_pkt_len with no MCLBYTES bound: heap OOB read up to ~63.5KB past the 2048-byte RX cluster, and unsigned underflow to ~4GB m_len (guaranteed panic)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/bce/if_bce.c:4396 · pass 2 · new |
none |
inconclusive |
| DF-3630 |
Medium |
6.3 |
iwm_handle_rxb: PAGE_SIZE sc_cmd_resp copy from possibly-offset response packet reads up to ~4KB past the 4096-byte RX cluster; iwm_alive_fn v3 fallback reads fixed struct without payload-length check
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwm/if_iwm.c:5511 · pass 2 · new |
none |
inconclusive |
| DF-3639 |
Medium |
6.3 |
bnx_rxeof trusts NIC-DMA'd return-ring bge_len with no MCLBYTES bound: heap OOB read / kernel info leak up to ~63.5KB past the 2048-byte RX cluster
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/bnx/if_bnx.c:2848 · pass 2 · new |
none |
inconclusive |
| DF-3663 |
Medium |
6.3 |
bge_rxeof trusts NIC-DMA'd return-ring bge_len with no MCLBYTES/jumbo bound: heap OOB read / kernel info leak up to ~63.5KB past the RX cluster (OOB write on i386 ALIGNBUG path)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/bge/if_bge.c:3138 · pass 2 · new |
none |
inconclusive |
| DF-3765 |
Medium |
6.3 |
32-bit multiply overflow in r100_cs_track_check() PW=1 vertex-array bound (esize * max_indx * 4) -> CS validator bypass -> GPU OOB index/vertex fetch
Base system — drivers · sys/dev/drm/radeon/r100.c:2326 · pass 2 · new |
leak |
inconclusive |
| DF-3766 |
Medium |
6.3 |
(value + 1) 32-bit wrap in r100_cs_track_check_pkt3_indx_buffer() bypasses the PACKET3_INDX_BUFFER offset check -> GPU OOB index fetch up to ~4GB past the BO
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r100.c:1909 · pass 2 · new |
leak |
inconclusive |
| DF-3849 |
Medium |
6.3 |
dc_rxeof trusts the NIC-DMA'd 14-bit RX frame length with no bound against the 2040-byte posted cluster: heap OOB read up to ~14 KB handed to the stack, and negative m_len on FRAMESZ<4 (DF-3604 family, dc(4) instance)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/dc/if_dc.c:2523 · pass 2 · new |
none |
inconclusive |
| DF-3933 |
Medium |
6.3 |
Single-segment RX frame length taken unclamped from the return descriptor: m_len/m_pkthdr.len up to 65535 on a 2046-byte cluster — heap OOB read (kernleak)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/age/if_age.c:2000 · pass 2 · new |
none |
not_reproduced |
| DF-3973 |
Medium |
6.3 |
Ignored btc_dpm_init failure leaves dangling rdev->pm.dpm.ps and un-NULLed dyn_state pointers: double free in btc_dpm_fini (two paths) plus eg_pi/ps_priv leaks — btc twin of DF-3645/3755/3955
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/btc_dpm.c:2574 · pass 2 · new |
none |
inconclusive |
| DF-4024 |
Medium |
6.3 |
re_rxeof trusts the NIC-DMA'd 14-bit GFRAGLEN (RE_RXBYTES) with no bound vs MCLBYTES: single-fragment m_len up to 16379 on a 2048-byte cluster -> 14,331-byte heap OOB read into delivered frames; multi-fragment m_pkthdr.len past the fragment chain -> m_copydata NULL-walk panic — DF-3604/3759/3933/3989 family, re(4) instance
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/re/if_re.c:1142 · pass 2 · new |
none |
inconclusive |
| DF-4211 |
Medium |
6.3 |
mpt_refresh_raid_vol frees a request on ETIMEDOUT that is still queued to the IOC and on request_pending_list → request-pool double-link / cross-command use-after-free
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_raid.c:1327 · pass 2 · new |
none |
untested |
| DF-4253 |
Medium |
6.3 |
SATA-ID timeout recovery frees in-flight commands (SMID reuse confusion), can Target-Reset the wrong target, and overwrites targ->tm without NULL check
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_sas_lsi.c:992 · pass 2 · new |
none |
inconclusive |
| DF-4371 |
Medium |
6.3 |
Dangling rdev->pm.dpm.dyn_state entries pointers after error-path kfree in r600_parse_extended_power_table — r600_free_extended_power_table at teardown kfrees them again (heap double free; root site shared by all six DPM family callers)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_dpm.c:939 · pass 2 · new |
none |
not_reproduced |
| DF-4398 |
Medium |
6.3 |
tws_scsi_err_complete touches the ccb after xpt_done (callout_stop on a surrendered ccb) — CAM ccb use-after-free race on every error completion
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws_cam.c:593 · pass 2 · new |
panic |
inconclusive |
| DF-4400 |
Medium |
6.3 |
tws_intr_resp invokes req->cb with no busy-state/completer validation — firmware replay of an in-range req_id double-completes surrendered ccbs or calls a NULL cb (distinct from DF-1590's index OOB)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws_cam.c:1161 · pass 2 · new |
panic |
inconclusive |
| DF-5021 |
Medium |
6.3 |
dp_retrain_link_dp_test walks dc->current_state with no lock and no state reference while concurrent atomic commits release it — use-after-free driven by unauthenticated HPD-RX automated-test input from a malicious DP sink
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/core/dc_link_hwss.c:274 · pass 2 · new |
kernel UAF/NULL-deref per finding |
untested |
| DF-2725 |
Low |
6.3 |
cluster_rbuild() under-counts component pages vs the pbuf's embedded xio_pages[]/KVA window — latent OOB pointer write + PTE install past MAXPHYS window for non-page-tiling VFS geometry
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_cluster.c:954 · pass 2 · new |
none |
not_reproduced |
| DF-3719 |
Low |
6.3 |
Firmware load: registration-metadata size drives an unbounded SRAM PIO copy (*limit initialized to sram_size then blindly overwritten) and short-inflate copies uninitialized kernel heap to the NIC
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mxge/if_mxge.c:685 · pass 2 · new |
none |
inconclusive |
| DF-3891 |
Low |
6.3 |
Double free of adev->pm.dpm.ps (and leak of parsed ps_priv states) on the kzalloc-failure path of kv_parse_power_table
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/kv_dpm.c:2750 · pass 2 · new |
none |
inconclusive |
| DF-4317 |
Low |
6.3 |
lge_jref() check-then-increment runs without the jslot serializer: racing lge_jfree() recycles the slot onto the free list with inuse==1 — double ownership of a 9KB RX jumbo cluster (DF-1384 nge twin)
Base system — drivers · sys/dev/netif/lge/if_lge.c:810 · pass 2 · new |
none |
untested |
| DF-4725 |
Low |
6.3 |
aureal: bad: unwind frees au BEFORE bus_teardown_intr (live IRQ handler over freed softc — UAF write in au_intr) and releases NULL reg[] entries after a mid-list BAR failure (NULL-deref panic)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/aureal.c:599 · pass 2 · new |
none |
untested |
| DF-4977 |
Low |
6.3 |
cardbus_detach_card ignores device_detach/device_delete_child failures and pci_freecfg's the child devinfo anyway: surviving child keeps freed ivars (UAF) while its bus resources are ripped out from under the running driver
Kernel memory corruption (UAF / OOB write) · sys/dev/pccard/cardbus/cardbus.c:236 · pass 2 · new |
none |
untested |
| DF-5274 |
Low |
6.3 |
cdgetccb() stores its result in a function-static 'union ccb *' — concurrent callers on different SIM locks can receive the same ccb (double xpt_release_ccb) or leak one
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_cd.c:1384 · pass 2 · new |
CCB double-release UAF family (DF-5274) / attach-time leak (DF-5275) |
untested |
| DF-4828 |
High |
6.2 |
r600_hdmi_buffer_status_changed / r600_hdmi_update_audio_settings consume lvds-typed enc_priv as radeon_encoder_atom_dig on the audio-work walk: dig->afmt OOB read at +288 into the adjacent slab chunk → wild deref + last_buffer_filled_status write-through (DF-4635 family, new sinks, interrupt-driven trigger)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_hdmi.c:283 · pass 2 · new |
none |
inconclusive |
| DF-3902 |
Low |
6.2 |
sioclose() gone-path double-free of com->ibuf: kfree guarded by '!= NULL' but pointer never NULLed — every further close on the detached unit re-frees the same M_DEVBUF block
Kernel memory corruption (UAF / OOB write) · sys/dev/serial/sio/sio.c:1475 · pass 2 · new |
none |
inconclusive |
| DF-3967 |
Low |
6.2 |
tx_intr STA_CSR4 completion desync: qid 5-15 garbage mapped to the mgt ring, results matched to the driver's stat cursor rather than a device index, unguarded txq->queued-- underflow defeating the ring-full gate, and m leaked on the ni==NULL path
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ral/rt2661.c:896 · pass 2 · new |
none |
not_reproduced |
| DF-3706 |
Medium |
6.1 |
vega10_setup_default_pcie_table reads bios entries[1] past the allocation when the VBIOS PCIE table has a single state (NUM_LINK_LEVELS=2)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:1232 · pass 2 · new |
none |
inconclusive |
| DF-3872 |
Medium |
6.1 |
mfi_check_for_sscd() dereferences NULL cm_data on MFI_CMD DCMD CFG_ADD with zero data_len (and heap-OOB-reads short buffers, bypassing pre/post checks)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mfi/mfi.c:2679 · pass 2 · new |
dos |
inconclusive |
| DF-3915 |
Medium |
6.1 |
Wild OOB read (and NVRAM garbage write) in _mapping_commit_enc_entry via unvalidated enclosure start_index ingested from DPM pg0 (num_slots==0 bypasses DF-1282)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_mapping.c:129 · pass 2 · new |
none |
not_reproduced |
| DF-4075 |
Medium |
6.1 |
mps_request_polled has no timeout cancellation: mps_update_events prints a NULL reply (NULL-deref panic) and callers free still-posted commands that late-complete against freed/reallocated cm state
Base system — drivers · sys/dev/raid/mps/mps.c:1724 · pass 2 · new |
none |
inconclusive |
| DF-4424 |
Medium |
6.1 |
Negotiated SEG_MAX smaller than a MAXPHYS transfer overflows the per-queue sglist: KASSERT panic on INVARIANTS kernels, truncated/mislabeled descriptor chain in production
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/virtio/block/virtio_blk.c:582 · pass 2 · new |
none |
not_reproduced |
| DF-4447 |
Medium |
6.1 |
VCE two-pass parse desynchronization: amdgpu_vce_cs_reloc patches the NEXT command's len/cmd header via in-IB dword aliasing — attacker-synthesized command stream skips loop-1 validation and len=0 causes an infinite in-kernel loop in the CS ioctl
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_vce.c:656 · pass 2 · new |
dos |
untested |
| DF-4452 |
Medium |
6.1 |
Wild out-of-bounds read in acpi_pci_link_choose_irq(): unbounded _PRS extended-IRQ values and FADT SciInterrupt used as index into pci_link_interrupt_weights[256]
Kernel memory leaks / uninitialized disclosure · sys/dev/acpica/acpi_pci_link.c:1010 · pass 2 · new |
none |
not_reproduced |
| DF-4500 |
Medium |
6.1 |
radeon_uvd CS parser/hardware desync: stale-or-unset DATA0/DATA1 dword indices patch the validated IB and make the UVD engine consume unvalidated GPU addresses
Base system — drivers · sys/dev/drm/radeon/radeon_uvd.c:583 · pass 2 · new |
none |
untested |
| DF-5479 |
Medium |
6.1 |
efi_enter()/efi_leave(): per-CPU CR4.SMEP disable with no critical section or CPU pinning (the comment claims one) — preempt+migrate inside the runtime window leaves the source CPU with SMEP permanently off and the migrated firmware call SMEP-fault-looping with efi_lock held forever
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/efirt.c:327 · pass 2 · new |
none |
inconclusive |
| DF-3689 |
Low |
6.1 |
Heap OOB read via unvalidated u8 vddInd/vddcInd from VBIOS indexing the voltage-lookup tables
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:1858 · pass 2 · new |
none |
inconclusive |
| DF-3821 |
Low |
6.1 |
Unvalidated GETCOMMPREF geometry: aac_max_fib_size of zero divides by zero at attach (boot panic); sizes below the 28-byte FIB header make the unconditional header copyin overflow the undersized FIB slots
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/aac/aac.c:1802 · pass 2 · new |
none |
inconclusive |
| DF-4172 |
Low |
6.1 |
iir_intr: unvalidated controller completion index reads/writes up to 2 gdt_ccb entries past sc_gccbs[124] (OOB dispatch, SLIST corruption, wild gc_ccb/gc_ucmd derefs)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/iir/iir.c:1580 · pass 2 · new |
none |
untested |
| DF-4889 |
Low |
6.1 |
iir_pci_attach drives the i960 MPR register layout at fixed offsets 0x2c-0x3fc0 with no BAR0 size validation — short BAR0 => MMIO reads/writes past the resource mapping (panic or up to ~12KB wild zero-write) before the magic read-back can reject the card
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/iir/iir_pci.c:186 · pass 2 · new |
Out-of-bounds device-memory (MMIO) access from a short BAR0: kernel page-fault panic or up to ~12KB bulk zero-write past the resource mapping, before the magic read-back check; requires crafted/broken PCIe hardware with the probed Intel/ICP IDs (CWE-125/CWE-787 device-memory variant) |
untested |
| DF-5781 |
Low |
6.1 |
hci_route_lookup walks hci_unit_list and unit->hci_links with no synchronization against attach/detach (crit_enter-only) and privileged hci_disable teardown — UAF reads and 6-byte freed-heap disclosure via getsockname
Kernel memory corruption (UAF / OOB write) · sys/netbt/hci_misc.c:65 · pass 2 · new |
none |
untested |
| DF-3383 |
Medium |
6.0 |
bpfdetach() frees only the FIRST bpf_if of an interface: leftover entries (wlan radiotap) keep bif_ifp/bif_driverp pointing at the freed ifnet -> UAF writes/reads on later close or rebind; bp==NULL branch dereferences NULL instead of reporting
Kernel memory corruption (UAF / OOB write) · sys/net/bpf.c:1566 · pass 2 · new |
none |
not_reproduced |
| DF-3598 |
Medium |
6.0 |
gfx_v8_0_init_microcode ignores amdgpu_ucode_validate() result for the RLC firmware (asymmetric size gate)
Base system — drivers · sys/dev/drm/amd/amdgpu/gfx_v8_0.c:1082 · pass 2 · new |
none |
inconclusive |
| DF-3645 |
Medium |
6.0 |
UAF + double free of rdev->pm.dpm.{ps,priv}: ci_parse_power_table/ci_dpm_init free and leave dangling pointers, and radeon_pm_init_dpm ignores the dpm.init failure
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/ci_dpm.c:5620 · pass 2 · new |
none |
inconclusive |
| DF-3697 |
Medium |
6.0 |
RLC v2_1 save-restore-list offset/size fields parsed without any bounds check → wild-pointer memcpy (OOB kernel heap read / panic)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/gfx_v9_0.c:516 · pass 2 · new |
none |
not_reproduced |
| DF-3852 |
Medium |
6.0 |
Voltage-GPIO MMIO register index taken raw from VBIOS and executed by r100_pm_misc() — arbitrary in-BAR register RMW and out-of-BAR MM_INDEX writes (up to 0x3FFFC)
Base system — drivers · sys/dev/drm/radeon/radeon_combios.c:2748 · pass 2 · new |
none |
inconclusive |
| DF-3859 |
Medium |
6.0 |
vega20_odn_edit_dpm_table VDDC_CURVE: one-sided signed bounds check (input_index > 2) lets negative user index through -> OOB read of od8_settings_array with values disclosed via pr_info (DF-3704 twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega20_hwmgr.c:2678 · pass 2 · new |
none |
not_reproduced |
| DF-3983 |
Medium |
6.0 |
sbp_orb_pointer fw_asyreq-failure branch xpt_done()s the CCB while leaving the OCB queued on sdev->ocbs, timeout-armed and dmamap-loaded — later timeout/device completion fires a second xpt_done on the stale ccb and/or duplicate-frees the OCB
Base system — drivers · sys/dev/disk/sbp/sbp.c:1175 · pass 2 · new |
none |
inconclusive |
| DF-3631 |
Low |
6.0 |
iwm_read_firmware TLV walk: size_t underflow on len -= roundup2(tlv_len,4) (and datasize < header) walks the parser past the firmware image
Base system — drivers · sys/dev/netif/iwm/if_iwm.c:607 · pass 2 · new |
none |
inconclusive fixed |
| DF-3687 |
Low |
6.0 |
Signed-compare bypass in smu7_odn_edit_dpm_table: negative sysfs level passes the bounds check and wraps to a u32 wild write index
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:4854 · pass 2 · new |
none |
inconclusive |
| DF-3695 |
Low |
6.0 |
ata_raid_delete frees the ar_softc without draining in-flight I/O — use-after-free in ata_raid_done/strategy
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-raid.c:1281 · pass 2 · new |
none |
untested |
| DF-3720 |
Low |
6.0 |
Unvalidated firmware-returned SRAM offsets used as kernel write pointers (lanai TX/RX rings, intr_coal_delay_ptr, irq_claim/irq_deassert, RSS table and key): OOB MMIO writes past the BAR mapping on a hostile device; int16_t ring-mask truncation on device ring sizes
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mxge/if_mxge.c:1207 · pass 2 · new |
none |
inconclusive |
| DF-3824 |
Low |
6.0 |
ASR_init_message writes the mandatory sense SGL element 8 bytes past asr_action's 512-byte stack buffer when one CCB spans >=58 physical runs; the adapter-derived ha_SgSize limit is computed and never used
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/asr/asr.c:1812 · pass 2 · new |
none |
inconclusive |
| DF-4174 |
Low |
6.0 |
gdtexecuteccb: unbounded physical scatter/gather copy — CAM_SCATTER_VALID|CAM_SG_LIST_PHYS CCBs with sglist_cnt > GDT_MAXSG overflow the 324-byte gccb->gc_cmd (pass(4)-reachable heap overflow)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/iir/iir.c:876 · pass 2 · new |
none |
untested |
| DF-4399 |
Low |
6.0 |
Passthrough SGL written at user-controlled offset: gcmd->size (u8, verbatim from TWS_IOCTL_FIRMWARE_PASS_THROUGH) used as SGL destination offset — lands on neighboring request's header_desc
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws_cam.c:1034 · pass 2 · new |
panic |
inconclusive |
| DF-4704 |
Low |
6.0 |
get_pixel_size() integer-divides depth by 8 and the blit/fill dispatch falls through to 32-bit ops — 8bpp KMS console gets 4-byte/pixel writes on a 1-byte/pixel stride (OOB writes past the framebuffer mapping)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/sckmsrndr.c:138 · pass 2 · new |
Kernel page faults / adjacent-memory corruption on the first console clear at 8bpp KMS console (4-byte writes on a 1-byte stride); root/boot-config gated |
untested |
| DF-5025 |
Low |
6.0 |
ipsd open/detach TOCTOU: detach can free the ipsdisk softc under an open descriptor — ipsd_close/strategy then write freed memory (DF-5013 pattern in ips)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ips/ips_disk.c:92 · pass 2 · new |
none |
not_reproduced |
| DF-3480 |
Medium |
5.9 |
so_rcv mutation from the SWI task without the sockbuf token: ng_btsocket_l2cap_data_input's sbappendrecord() races soreceive's ssb_token-protected unlink surgery (lost wakeups / torn record chains / potential freed-mbuf write)
Base system — network stack · sys/netgraph7/bluetooth/socket/ng_btsocket_l2cap.c:1455 · pass 2 · new |
none |
not_reproduced |
| DF-3482 |
Medium |
5.9 |
ng_ppp_mp_strategy uses the file-scope global compareLatencies as kqsort comparator state and clears it to NULL on exit — the DF-0330 twin is LIVE in v1 wherever unfuneled transmit contexts exist (ng_iface output): NULL-index kernel panic
Base system — network stack · sys/netgraph/ppp/ng_ppp.c:363 · pass 2 · new |
none |
inconclusive |
| DF-3778 |
Medium |
5.9 |
jme_rxeof() RX walk has no work bound on the interrupt path (count=-1 at all three call sites): a remote line-rate small-packet flood keeps the ring perpetually non-empty so the interrupt handler loops forever holding the serializer — NIC + CPU livelock, whole-interface wedge on legacy/MSI (DF-3761/alc family, jme(4) instance)
Base system — drivers · sys/dev/netif/jme/if_jme.c:2494 · pass 2 · new |
none |
inconclusive |
| DF-3074 |
Low |
5.9 |
nfs_mountdiskless() double-frees the dup_sockaddr() on mountnfs() failure (mountnfs's bad: path frees nm_nam via nfs_free_mount, then haderror kfrees it again)
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_vfsops.c:767 · pass 2 · new |
none |
not_reproduced |
| DF-4722 |
Low |
5.9 |
ath_tdma_update passes the 0xff rateCodeToIndex sentinel into ath_hal_pkt_txtime — OOB read of rates->info[255] (64-entry table) and possible divide-by-zero panic from a received TDMA beacon
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath/if_ath_tdma.c:407 · pass 2 · new |
none |
untested |
| DF-2786 |
Medium |
5.8 |
mtx_abort_link stale-state double-unlink: mtx_delete_link leaves MTX_LINK_LINKED_* set on an unlinked link until an unlocked IDLE reset, letting a racing abort write through dangling next/prev (UAF-write / wait-queue poisoning)
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_mutex.c:914 · pass 2 · new |
none |
not_reproduced |
| DF-2896 |
Medium |
5.8 |
cnwrite() constty use-after-free: unsynchronized constty->t_dev captured without a reference and dispatched through dev_doperate() after a sleepable log_console() walk, while an unprivileged user (default UCONSOLE) controls the cdev's lifetime
Kernel memory corruption (UAF / OOB write) · sys/kern/tty_cons.c:465 · pass 2 · new |
none |
not_reproduced |
| DF-2923 |
Medium |
5.8 |
vq_vptomp check-then-use race on vp->v_pfsmp → mount use-after-free with indirect call through freed memory
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_quota.c:420 · pass 2 · new |
none |
not_reproduced |
| DF-3862 |
Medium |
5.8 |
amdgpu_device_gpu_recover dereferences the timed-out job across the multi-second GPU reset although the fence-completion path can already have kfree'd it (UAF, concurrent-reinit family)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_device.c:3414 · pass 2 · new |
none |
not_reproduced |
| DF-4335 |
Medium |
5.8 |
twa CL: tw_cli_get_param/tw_cli_set_param re-queue a request after ETIMEDOUT — FREE_Q double-insert (unwalkable queue, spin under gen_lock) or BUSY_Q+FREE_Q dual membership (request reuse under live firmware DMA)
Base system — drivers · sys/dev/raid/twa/tw_cl_io.c:804 · pass 2 · new |
none |
untested |
| DF-4584 |
Medium |
5.8 |
radeon_bo_get_surface_reg steals surface registers from other BOs under no lock: NULL-deref panic and 4-byte UAF write into freed radeon_bo slab
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_object.c:609 · pass 2 · new |
none |
inconclusive |
| DF-4783 |
Medium |
5.8 |
mmcsd_detach() does not exclude or drain the data path: strategy/open callers race lockuninit() and the newbus softc free (DF-4646 family)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mmcsd/mmcsd.c:209 · pass 2 · new |
none |
not_reproduced |
| DF-3315 |
Low |
5.8 |
in_pcblist_range walks non-shared pcbinfo (ripcbinfo) with no synchronization; non-atomic marker advance vs same-cpu netisr pcb teardown yields UAF read/write and list corruption
Kernel memory corruption (UAF / OOB write) · sys/netinet/in_pcb.c:2452 · pass 2 · new |
none |
not_reproduced |
| DF-3656 |
Low |
5.8 |
Error-path UAF: amdgpu_dm init failure paths kfree registered i2c adapters / connectors / encoders / planes that remain in DRM core lists
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm.c:4080 · pass 2 · new |
none |
inconclusive |
| DF-4216 |
Low |
5.8 |
dc_submit_i2c indexes dc->links[] with an unvalidated, wrong-domain index (DDC line number from VBIOS): NULL/wild deref panic and silent wrong-connector DDC routing
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc.c:1742 · pass 2 · new |
none |
untested |
| DF-4468 |
Low |
5.8 |
ess_attach() pcm_register-failure unwind kfrees sc while the 0666 /dev/mixerN that mixer_init() published survives — dangling mixer devinfo UAF on open/ioctl (DF-4156 family, 8th instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/solo.c:1045 · pass 2 · new |
none |
inconclusive |
| DF-4486 |
Low |
5.8 |
REPROBE destroy path leaves the per-output ACPI notify handler registered against freed memory (AcpiRemoveNotifyHandler(NULL) is a no-op)
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_video/acpi_video.c:386 · pass 2 · new |
none |
untested |
| DF-3310 |
Medium |
5.7 |
Port change on a running wg interface re-runs lockinit() on the live so_lock and closes/swaps the UDP sockets without holding it
Base system — network stack · sys/net/wg/if_wg.c:974 · pass 2 · new |
none |
not_reproduced |
| DF-3311 |
Medium |
5.7 |
Interface down/destroy tears down tasks/queues/softc while the receive upcall can still re-enqueue them (sockets closed last): late handshake/decrypt task fires on freed softc
Privilege escalation (general use) · sys/net/wg/if_wg.c:2166 · pass 2 · new |
dos |
inconclusive |
| DF-3429 |
Medium |
5.7 |
DWDS discovery frames staged on ic_stageq are leaked forever on age-out: ieee80211_node_timeout discards the mbuf list returned by ieee80211_ageq_age (remote kernel memory-exhaustion DoS)
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_wds.c:321 · pass 2 · new |
none |
inconclusive |
| DF-3593 |
Medium |
5.7 |
Duplicate firmware TX-done re-completes a freed TX slot: NULL-deref/KASSERT panic and TX-ring accounting underflow
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn.c:5996 · pass 2 · new |
panic |
inconclusive |
| DF-3594 |
Medium |
5.7 |
bwn_intr_noise missing re-entry guard: 9th noise-sample interrupt writes 4 bytes past noi_samples into mac_phy
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/bwn/bwn/if_bwn.c:5198 · pass 2 · new |
panic |
inconclusive |
| DF-3716 |
Medium |
5.7 |
mxge_rx_done_big trusts the NIC-DMA'd 16-bit completion length with no posted-cluster bound: heap OOB read up to ~63.5KB past the 2/4KiB RX cluster (DF-3604/3639/3663/3679/3713 family)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/mxge/if_mxge.c:2345 · pass 2 · new |
none |
inconclusive |
| DF-3931 |
Medium |
5.7 |
Detach/shutdown lifetime: self-rescheduling mly_periodic can survive callout_stop and in-flight commands are never drained — use-after-free after mly_free()
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mly/mly.c:1397 · pass 2 · new |
none |
not_reproduced |
| DF-4023 |
Medium |
5.7 |
Dangling rdev->pm.dpm.ps after kzalloc-failure kfree in rv7xx_parse_power_table (:2308): ignored rv770_dpm_init error -> double free in rv770_dpm_fini at unload/reboot + leak of parsed ps_priv states — shared root site of DF-3645/3755/3955/3973, rv770-dispatch instance
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/rv770_dpm.c:2306 · pass 2 · new |
none |
inconclusive |
| DF-4110 |
Medium |
5.7 |
tl_intvec_rxeof trusts device-written 16-bit tlist_frsize: unbounded mbuf length on an MCLBYTES cluster (OOB read propagated up the network stack)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/tl/if_tl.c:1405 · pass 2 · new |
none |
untested |
| DF-4113 |
Medium |
5.7 |
nge_rxeof trusts device-written 16-bit descriptor BUFLEN: unbounded mbuf length on the jumbo minicluster (OOB read propagated up the network stack)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/nge/if_nge.c:1225 · pass 2 · new |
none |
untested |
| DF-4117 |
Medium |
5.7 |
Dangling rdev->pm.dpm.ps after swallowed cypress_dpm_init failure: double free in cypress_dpm_fini at unload/reboot (cypress dispatch of the DF-4023/DF-3973 family)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/cypress_dpm.c:2050 · pass 2 · new |
none |
inconclusive |
| DF-4162 |
Medium |
5.7 |
radeon_pm.c:1448 discards radeon_dpm_init() return (and the asic *_init files discard radeon_pm_init's): the root enabler of the six-file NULL-boot_ps / dangling-ps family — central hub fix consolidating DF-4022/3972/4117/4118/4124/4125/4143/4144
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_pm.c:1448 · pass 2 · new |
none |
inconclusive |
| DF-4503 |
Medium |
5.7 |
ahci_pm_read/write recycle the ata_xfer while the command is still programmed on the chip (deferred timeout returns ATA_S_ONCHIP): CCB freed and re-issuable while chip owns the slot -> ahci_ata_cmd_done KKASSERT panic or cross-command completion corruption
Base system — drivers · sys/dev/disk/ahci/ahci_pm.c:976 · pass 2 · new |
none |
inconclusive |
| DF-5218 |
Medium |
5.7 |
dc_sink retain/release contract violated by DM forced-connector paths: emulated_link_detect strands references (permanent ~1KB dc_sink leak per forced-detect cycle) and update_connector_after_detect branch A releases a borrowed dc_sink (premature free with dangling link->local_sink, or kref_put UAF write + dc_sink_free double-free on freed slab after EDID-change detect)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/core/dc_sink.c:68 · pass 2 · new |
none |
inconclusive |
| DF-2709 |
Low |
5.7 |
hammer2_volconf_update() uses iocom->conn_state unlocked across a peer-driven drop (TOCTOU): kdmsg_lnk_conn_reply mutates conn_state without msglk
Base system — kernel core (kern) · sys/kern/kern_dmsg.c:231 · pass 2 · new |
none |
not_reproduced |
| DF-3153 |
Low |
5.7 |
volume-del on a volume hosting UNDO-zone big-blocks: bare KKASSERT(0) panics INVARIANTS kernels; on production kernels the case falls through and bzero()s live undo FIFO entries (silent crash-recovery corruption)
Base system — filesystems · sys/vfs/hammer/hammer_volume.c:554 · pass 2 · new |
none |
untested |
| DF-3171 |
Low |
5.7 |
mfsclose() deactivates the MFS I/O server on ANY device close — no finality check (speculative: would destroy_dev/kfree the serving mfsnode under a live mount if a non-final close ever slips through)
Kernel memory corruption (UAF / OOB write) · sys/vfs/mfs/mfs_vfsops.c:136 · pass 2 · new |
none |
not_reproduced |
| DF-3628 |
Low |
5.7 |
Unvalidated VBT backlight.controller used as power-sequencer register index (out-of-range MMIO writes on BXT)
Base system — drivers · sys/dev/drm/i915/intel_dp.c:864 · pass 2 · new |
none |
inconclusive |
| DF-3875 |
Low |
5.7 |
mfi_intr bounds reply-queue contexts by mfi_max_fw_cmds but mfi_commands[] is allocated MIN(128, mfi_max_fw_cmds) -> OOB command object under firmware misbehavior
Base system — drivers · sys/dev/raid/mfi/mfi.c:1228 · pass 2 · new |
none |
inconclusive |
| DF-3963 |
Low |
5.7 |
mv family queue-pointer registers used unmasked: device-fed outbound_tail/inbound_head index inbound_q/outbound_q[512] MMIO — OOB bus-space read and 8-byte OOB MMIO write up to ~32GB past BAR2
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptiop/hptiop.c:229 · pass 2 · new |
none |
not_reproduced |
| DF-4163 |
Low |
5.7 |
radeon_pm_fini_dpm never cancels pm.dpm.thermal.work nor clears dpm_enabled while the thermal IRQ outlives pm_fini: post-fini UAF in radeon_dpm_thermal_work_handler
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_pm.c:1690 · pass 2 · new |
none |
inconclusive |
| DF-4173 |
Low |
5.7 |
iir_shutdown: shared flush ucmd freed (wrong malloc type) while still queued/in flight, and re-queued for the next drive while a gccb owns it — use-after-free at shutdown
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/iir/iir.c:1493 · pass 2 · new |
none |
untested |
| DF-4223 |
Low |
5.7 |
Attach error-path asymmetry: double bus_dmamap_destroy (double kfree of M_DEVBUF maps) and callout_drain/bus_dmamap_destroy walked through a NULL hcb array
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/vmware/pvscsi/pvscsi.c:615 · pass 2 · new |
none |
inconclusive |
| DF-4286 |
Low |
5.7 |
acpi_thinkpad LED async task (AcpiOsExecute) not drained on detach — use-after-free of the softc
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_thinkpad/acpi_thinkpad.c:391 · pass 2 · new |
use-after-free of acpi_thinkpad softc (read/write + AcpiEvaluateObject with handles loaded from freed memory) racing kldunload/devctl-detach against /dev/led/thinklight toggles; root-only, hardware-gated |
inconclusive |
| DF-4328 |
Low |
5.7 |
sf_attach() bus_setup_intr() failure path double-invokes ether_ifdetach() and passes a NULL intr cookie to bus_teardown_intr()
Base system — drivers · sys/dev/netif/sf/if_sf.c:777 · pass 2 · new |
none |
untested |
| DF-4352 |
Low |
5.7 |
dce110_link_encoder_create indexes 7/6/6-entry const register tables with unvalidated VBIOS-derived transmitter/channel/hpd_source — OOB const reads become the encoder's MMIO register bases (DF-4312/4319 twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dce110/dce110_resource.c:588 · pass 2 · new |
none |
untested |
| DF-4405 |
Low |
5.7 |
amdgpu_uvd sw_init/resume trust unvalidated firmware ucode extent fields (ucode_size_bytes/ucode_array_offset_bytes) — sw_init bo_size u32 wrap, resume ~4GiB OOB read past the firmware blob and underflowed OOB memset_io (DF-1946 family instance)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_uvd.c:250 · pass 2 · new |
none |
inconclusive |
| DF-4438 |
Low |
5.7 |
mrsas_cam_attach error paths free sim_0/sim_1 without clearing sc pointers — attach_fail_cam unwind re-frees via mrsas_cam_detach: cam_sim_path() UAF read, second xpt_bus_deregister, cam_sim_free() double-free into M_CAMSIM (plus devq refcount underflow)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mrsas/mrsas_cam.c:150 · pass 2 · new |
none |
inconclusive |
| DF-4495 |
Low |
5.7 |
dce100_link_encoder_create indexes 7/6/6-entry const register tables with unvalidated VBIOS-derived transmitter/channel/hpd_source — OOB const reads become the encoder's MMIO register bases (DF-4312/4319/4352/4375 family twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dce100/dce100_resource.c:573 · pass 2 · new |
none |
untested |
| DF-4512 |
Low |
5.7 |
vmbus: selected/registered vmbus cputimer never stopped or deregistered on detach - per-tick reload runs on the freed vmbus_softc and through unloaded module text (UAF write + dangling call target) after kldunload
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/hyperv/vmbus/vmbus.c:280 · pass 2 · new |
Per-tick UAF write (8-byte host-influenced value into freed softc) + call through unloaded module text after kldunload |
untested |
| DF-4543 |
Low |
5.7 |
agp: /dev/agpgart cdev outlives chipset detach - dangling si_drv1 UAF on open fds, plus free_cdev unit-vs-minor mismatch leaves a stale node
Kernel memory corruption (UAF / OOB write) · sys/dev/agp/agp.c:240 · pass 2 · new |
none |
not_reproduced |
| DF-4719 |
Low |
5.7 |
Pending AAC_EVENT_CMFREE fires after aac_cam_detach freed the sim -> xpt_release_simq/xpt_done through freed objects (UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/aac/aac_cam.c:165 · pass 2 · new |
none |
untested |
| DF-4908 |
Low |
5.7 |
ath_descdma_setup() returns error==0 (success) after tearing everything down when the ath_buf array kmalloc fails — and the unvalidated RWTUN hw.ath.*buf sysctls feeding it make the failure deterministic (and can truncate bsize into a small kmalloc walked OOB by the setup loop)
Base system — drivers · sys/dev/netif/ath/ath/if_ath_descdma.c:272 · pass 2 · new |
panic |
untested |
| DF-4909 |
Low |
5.7 |
__drm_legacy_pci_free consumes never-initialized tag/map: the DF port broke the _DRM_CONSISTENT map's alloc/free contract — drm_pci_alloc's descriptor is kfree'd at addmap, the rmmap path forges it on the stack with only vaddr/busaddr/size, and the DF bus_dma teardown then runs bus_dmamap_unload/bus_dmamem_free/bus_dma_tag_destroy on uninitialized stack pointers while the real tag/map/DMA memory leak permanently
Base system — drivers · sys/dev/drm/drm_pci.c:152 · pass 2 · new |
corruption |
untested |
| DF-4916 |
Low |
5.7 |
amdgpu_virt_init_data_exchange: host-controlled pf2vf_size drives a 1 KB memset (and the header reads) past the kmap'ed fw VRAM reservation — no validation that the region contains the 64 KB-offset exchange area
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_virt.c:402 · pass 2 · new |
none |
not_reproduced |
| DF-5057 |
Low |
5.7 |
mlxd lifetime TOCTOU: MLX_DETACH_DRIVE checks MLXD_OPEN then sleeps in mlx_flush before device_delete_child; mlxd_detach has no recheck and no in-flight drain — softc freed under racing open fds / in-flight bios (DF-5025/DF-5046 family, last LD sibling)
Base system — drivers · sys/dev/raid/mlx/mlx_disk.c:98 · pass 2 · new |
none |
untested |
| DF-5301 |
Low |
5.7 |
TOCTOU: uhub_child_location_string()/uhub_child_pnpinfo_string() dereference sc->sc_udev->hub after a device_is_attached() guard that stays true while uhub_detach() frees and NULLs it (NULL-deref panic / UAF read)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_hub.c:1668 · pass 2 · new |
none |
not_reproduced |
| DF-5442 |
Low |
5.7 |
Cross-CPU MTRR reprogram race: amd64_mrstore() fire-and-forget IPI vs amd64_mrfetch() immediate rewrite of the shared mr_desc[] array leaves a persistent per-CPU memory-type desync (torn descriptor generation)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/amd64_mem.c:289 · pass 2 · new |
none |
not_reproduced |
| DF-4829 |
Info |
5.7 |
md(4) mdstrategy_malloc sector-array growth uses 32-bit sector number and 32-bit size arithmetic: wrapped growth kmalloc(0)/(undersized chunk) -> ZERO_LENGTH_PTR bcopy / heap overflow + wild write (LATENT — unreachable behind dscheck)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/md/md.c:253 · pass 2 · new |
none |
not_reproduced |
| DF-4893 |
Info |
5.7 |
Dead AH_DEBUG_ALQ block: unlocked global `ath_hal_alq` TOCTOU — use-after-free of `struct alq` if ever compiled (latent, triple-unreachable on stock DF)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_hal/ah_osdep.c:230 · pass 2 · new |
none |
untested |
| DF-3029 |
Medium |
5.6 |
fuse_io_thread lost-wakeup window (no tsleep interlock between empty-queue check and sleep) can strand a bio indefinitely
Base system — filesystems · sys/vfs/fuse/fuse_vnops.c:1682 · pass 2 · new |
none |
not_reproduced |
| DF-3667 |
Medium |
5.6 |
ix_rxeof trusts NIC-DMA'd descriptor length with no rx_mbuf_sz bound: unbounded m_len/m_pkthdr.len → heap OOB read up to 63,487 bytes past the RX cluster (DF-3604/3639/3663 family)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ix/if_ix.c:2974 · pass 2 · new |
none |
inconclusive |
| DF-4030 |
Medium |
5.6 |
u32 (size + offset) wraparound defeats VALID_BUFFER BO bound check (+ forced-size underflow)
Base system — drivers · sys/dev/drm/radeon/r600_cs.c:2001 · pass 2 · new |
none |
not_reproduced |
| DF-4032 |
Medium |
5.6 |
Texture mipmap BO bounds check is commented out (GPU OOB mip walk past mipmap BO)
Base system — drivers · sys/dev/drm/radeon/r600_cs.c:1399 · pass 2 · new |
none |
not_reproduced |
| DF-4046 |
Medium |
5.6 |
stge_stop() deadlocks the callout subsystem against stge_tick (AB-BA): synchronous callout_stop() issued while holding the driver serializer vs a tick callback parked in lwkt_serialize_enter — mechanism PROVEN on the guest via protocol replica; stge_tick re-arms UNCONDITIONALLY under the lock
Base system — drivers · sys/dev/netif/stge/if_stge.c:1758 · pass 2 · new |
none |
inconclusive |
| DF-4415 |
Low |
5.6 |
atkbdc drain/poll loops reset their budget on every received byte with no overall cap — unbounded kernel busy-loop under i8042/AUX data flood (empty_both_buffers got the cap, these four did not)
Base system — drivers · sys/dev/misc/kbd/atkbdc.c:865 · pass 2 · new |
none |
not_reproduced |
| DF-4511 |
Low |
5.6 |
acpi_wmi: _WDG result consumed unvalidated — NULL-deref boot panic on missing Return; ACPI_OBJECT union type-confusion memcpy on non-Buffer _WDG return
Base system — drivers · sys/dev/acpica/acpi_wmi/acpi_wmi.c:725 · pass 2 · new |
Boot-time NULL-deref panic on missing Return; type-confused wild memcpy read with contents rendered via world-readable /dev/wmistat |
untested |
| DF-4617 |
Low |
5.6 |
hdacc_stream_intr/hdacc_stream_free index codec->streams[dir][stream] with an unvalidated controller-derived stream tag — negative DF-4115 BSS tags dispatch through an OOB-read device_t (wild indirect call)
Base system — drivers · sys/dev/sound/pci/hda/hdacc.c:665 · pass 2 · new |
none |
not_reproduced |
| DF-4630 |
Low |
5.6 |
mpt_user: shared static user_handler_id over a global reply-handler table — detaching one controller unregisters mpt_user_reply_handler for all others, leaking requests until pool exhaustion
Base system — drivers · sys/dev/disk/mpt/mpt_user.c:90 · pass 2 · new |
Storage DoS: each RAID_ACTION on the surviving controller leaks one of 512 request slots until ENOMEM takes down all mpt I/O including disk traffic |
untested |
| DF-3589 |
Medium |
5.5 |
Unprivileged kernel panic: CTLFLAG_ANYBODY psm sysctls allow the smoother divisor product div_x*multiplicator to wrap to zero (divide-by-zero in psmsmoother)
Base system — drivers · sys/dev/misc/psm/psm.c:4337 · pass 2 · new |
none |
not_reproduced |
| DF-3653 |
Medium |
5.5 |
NULL pointer dereference in amdgpu_dm_atomic_commit_tail() via update_stream_scaling_settings() with NULL dc_stream_state (underscan path)
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm.c:4882 · pass 2 · new |
none |
inconclusive |
| DF-3665 |
Medium |
5.5 |
Divide-by-zero (#DE panic) in evergreen_latency_watermark(): zero available_bandwidth from tiny user mode clocks, u32 fixed-point wrap of mclk/sclk, or zero VBIOS clocks
Base system — drivers · sys/dev/drm/radeon/evergreen.c:2048 · pass 2 · new |
dos |
inconclusive |
| DF-3685 |
Medium |
5.5 |
NULL-subtable dereference family at backend init: V1 parser treats mm/lookup/PowerTune subtables as optional, smu7 derefs them unconditionally with ignored error returns
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:1842 · pass 2 · new |
none |
inconclusive |
| DF-3702 |
Medium |
5.5 |
vega10 backend-init NULL-derefs optional pptable subtables (tdp/mm/lookup/dep tables) — missing-offset VBIOS panics the kernel (DF-3685 twin)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:269 · pass 2 · new |
none |
inconclusive |
| DF-3763 |
Medium |
5.5 |
r100_cs_track_cube() NULL-pointer dereference via unbound cube-face BOs -> kernel panic from unprivileged CS ioctl (pre-R300)
Base system — drivers · sys/dev/drm/radeon/r100.c:2143 · pass 2 · new |
panic |
inconclusive |
| DF-3828 |
Medium |
5.5 |
Divide-by-zero family in DCE11 display watermark math: dce_v11_0_latency_watermark divides by available_bandwidth and disp_clk before any guard (DF-3665 twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/dce_v11_0.c:908 · pass 2 · new |
none |
not_reproduced |
| DF-3839 |
Medium |
5.5 |
SET_RESOURCE SQ_TEX_VTX_VALID_BUFFER bound-check bypass: (size + offset) added in u32 (+ forced-size underflow) → GPU vertex fetch up to ~4 GB past the BO
Base system — drivers · sys/dev/drm/radeon/evergreen_cs.c:2416 · pass 2 · new |
none |
inconclusive |
| DF-3857 |
Medium |
5.5 |
Divide-by-zero in dce_v10_0_latency_watermark: divides by available_bandwidth/disp_clk before any guard (DF-3828 DCE11 twin)
Privilege escalation (general use) · sys/dev/drm/amd/amdgpu/dce_v10_0.c:882 · pass 2 · new |
none |
not_reproduced |
| DF-3858 |
Medium |
5.5 |
vega20_get_sclk_od/get_mclk_od: count-1 underflow wild read (+34 GB) and DIV_ROUND_UP divide-by-zero on zero golden clock via world-readable pp_sclk_od/pp_mclk_od (DF-3686/DF-3707 twins)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega20_hwmgr.c:1326 · pass 2 · new |
none |
not_reproduced |
| DF-3869 |
Medium |
5.5 |
Zero-divisor (#DE kernel panic) in calculate_bandwidth: exact-zero divisor at dce_calcs.c:839 from UABI-legal degenerate modes, plus u32 division-by-zero at :1945
Privilege escalation (general use) · sys/dev/drm/amd/display/dc/calcs/dce_calcs.c:839 · pass 2 · new |
dos |
inconclusive |
| DF-3871 |
Medium |
5.5 |
MFI_CMD ioctl: devclass_get_softc() result used without NULL check -> one-ioctl kernel NULL-deref panic
Base system — drivers · sys/dev/raid/mfi/mfi.c:2953 · pass 2 · new |
dos |
inconclusive |
| DF-3873 |
Medium |
5.5 |
mfi ioctl handlers dereference user ioctl buffers without copyin/copyout -> trap_fatal panic on unmapped args (and kernel writes through raw user pointers)
Base system — drivers · sys/dev/raid/mfi/mfi.c:2909 · pass 2 · new |
dos |
inconclusive |
| DF-3878 |
Medium |
5.5 |
Divide-by-zero family in DCE8 display watermark math: dce_v8_0_latency_watermark divides by available_bandwidth and disp_clk before any guard (DF-3828 DCE11 / DF-3857 DCE10 twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/dce_v8_0.c:817 · pass 2 · new |
none |
not_reproduced |
| DF-3946 |
Medium |
5.5 |
dce110_apply_ctx_for_surface programs a front end for a head pipe whose plane was detached (plane_state NULL) -> kernel NULL-deref panic
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:2743 · pass 2 · new |
none |
inconclusive |
| DF-3993 |
Medium |
5.5 |
User-selected UVD/VCE CS ring dereferences NULL asic->ring[] slot on asic families whose dispatch table lacks the ring — kernel panic from unprivileged CS ioctl
Base system — drivers · sys/dev/drm/radeon/radeon_asic.c:214 · pass 2 · new |
panic |
inconclusive |
| DF-4042 |
Medium |
5.5 |
Hostile SCSI target panics the kernel via MESSAGE REJECT of the driver's own MSG_ABORT/MSG_NOOP: amdsentmsg panic("Inconsistent msg buffer")
Base system — drivers · sys/dev/disk/amd/amd.c:1435 · pass 2 · new |
none |
inconclusive fixed |
| DF-4043 |
Medium |
5.5 |
amdsettags panic("Implement me!") reached when a target rejects a tagged command (MESSAGE REJECT handling)
Base system — drivers · sys/dev/disk/amd/amd.c:846 · pass 2 · new |
none |
inconclusive fixed |
| DF-4044 |
Medium |
5.5 |
amdparsemsg MSG_DISCONNECT writes through NULL active_srb after an unexpected reselection
Base system — drivers · sys/dev/disk/amd/amd.c:1218 · pass 2 · new |
none |
inconclusive fixed |
| DF-4183 |
Medium |
5.5 |
amdgpu fb creation never validates framebuffer size against the backing GEM object (twin of DF-4181; covers BOTH legacy DCE and DC scanout stacks) — display engine scans out megabytes past a tiny BO, disclosing adjacent VRAM on the physical display
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_display.c:506 · pass 2 · new |
none |
inconclusive |
| DF-4217 |
Medium |
5.5 |
ae_stop_rxmac()/ae_stop_txmac() decrement their idle-wait loop counter (i--): a non-idle device busy-wedges the kernel thread forever, timeout branch unreachable
Base system — drivers · sys/dev/netif/ae/if_ae.c:288 · pass 2 · new |
none |
inconclusive |
| DF-4219 |
Medium |
5.5 |
ae_rx_intr() for(;;) descriptor walk has no work bound: a device that keeps AE_RXD_UPDATE set livelocks the interrupt handler holding ifp->if_serializer (DF-3990 family)
Base system — drivers · sys/dev/netif/ae/if_ae.c:643 · pass 2 · new |
none |
inconclusive |
| DF-4225 |
Medium |
5.5 |
Divide-by-zero (kernel panic) in drm_mode_vrefresh: user-supplied umode with htotal*vtotal == 0 (mod 2^32) passes drm_mode_validate_basic and wraps den to zero
Base system — drivers · sys/dev/drm/drm_modes.c:783 · pass 2 · new |
dos |
untested |
| DF-4262 |
Medium |
5.5 |
CS ioctl with num_chunks==0 returns early success, amdgpu_cs_ib_fill dereferences NULL parser->entity — unprivileged kernel panic
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_cs.c:114 · pass 2 · new |
none |
untested |
| DF-4263 |
Medium |
5.5 |
Duplicate BO_HANDLES/FENCE/SYNCOBJ_OUT chunks overwrite parser state, permanently leaking bo_list + all BO refs, user-fence BO refs, and syncobj refs
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_cs.c:88 · pass 2 · new |
none |
untested |
| DF-4370 |
Medium |
5.5 |
amdgpu_bo_placement_from_domain walks up to 6 domain entries into placements[AMDGPU_BO_MAX_PLACEMENTS=3] and BUG_ONs: unprivileged GEM_CREATE with domains=VRAM|GTT|CPU panics the kernel; domains=0x3f writes places[3..5] OOB (over bo->placement/bo->tbo) before the panic
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_object.c:139 · pass 2 · new |
none |
not_reproduced |
| DF-4373 |
Medium |
5.5 |
Divide-by-zero in dce_transform_get_max_num_of_supported_lines: prescaler_on + fully-off-CRTC plane (or underscan-inset stream dst) yields pixel_width==0 — ASSERT is WARN_ON-only, integer #DE kernel panic from an unprivileged atomic TEST_ONLY check
Base system — drivers · sys/dev/drm/amd/display/dc/dce/dce_transform.c:675 · pass 2 · new |
none |
not_reproduced |
| DF-4390 |
Medium |
5.5 |
NULL request-descriptor dereference in mfi_tbolt_send_frame for PD_SCSI_IO with READ_10/WRITE_10 CDBs - any I/O to a Thunderbolt system PD panics the kernel
Base system — drivers · sys/dev/raid/mfi/mfi_tbolt.c:1156 · pass 2 · new |
none |
not_reproduced |
| DF-4526 |
Medium |
5.5 |
ar5416AniControl NULL-pointer dereference when driven before the first channel reset (ah_curani == NULL): kernel panic via HAL_DIAG_ANI_CMD
Privilege escalation (general use) · sys/dev/netif/ath/ath_hal/ar5416/ar5416_ani.c:184 · pass 2 · new |
none |
not_reproduced |
| DF-4576 |
Medium |
5.5 |
amdgpu_gem: GEM_VA MAP/REPLACE commits page tables for the entire attacker-chosen range before amdgpu_vm_bo_map() validates the mapping — every failed map strands ~(map_size>>18) bytes of page-table BOs until fd close; unprivileged render-node GTT/kernel exhaustion (cross-process GPU wedge)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_gem.c:638 · pass 2 · new |
none |
untested |
| DF-4580 |
Medium |
5.5 |
radeon_cs_parser_init: fully attacker-controlled u32 num_chunks/length_dw drive unbounded wired GFP_KERNEL allocations — on DragonFly GFP_KERNEL is M_WAITOK without M_NULLOK, so failure is panic('malloc limit exceeded')/panic('kernel_map ran out of space') or an eternal unkillable vm_wait() loop (local kernel DoS); latent NULL-return path would additionally crash radeon_cs_parser_fini via parser->chunks[i]
Base system — drivers · sys/dev/drm/radeon/radeon_cs.c:287 · pass 2 · new |
none |
not_reproduced |
| DF-4582 |
Medium |
5.5 |
rs690 watermark fixed-point divide-by-zero via pixel-clock 2^20 wrap: dfixed_const(1048576 kHz) == 0 → do_div by zero → kernel panic (trap 18), reachable from an unprivileged DRM master via DRM_IOCTL_MODE_SETCRTC on RS690/RS740/RS780/RS880
Base system — drivers · sys/dev/drm/radeon/rs690.c:326 · pass 2 · new |
panic |
inconclusive fixed |
| DF-4608 |
Medium |
5.5 |
NULL-pointer dereference panic in radeon_vce_cs_reloc when a VCE command stream is submitted without a relocation chunk
Base system — drivers · sys/dev/drm/radeon/radeon_vce.c:478 · pass 2 · new |
dos |
untested |
| DF-4696 |
Medium |
5.5 |
Divide-by-zero (#DE kernel panic) in dpp1_dscl_calc_lb_num_partitions: UABI-legal off-CRTC plane yields recout.width==0 (line_size==0 → 816/0 at :439, also :440/:441) — DCN10 commit-path twin of DF-4373
Base system — drivers · sys/dev/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c:411 · pass 2 · new |
none |
not_reproduced |
| DF-4819 |
Medium |
5.5 |
i915 GEM shrinker is dead code: never registered with any memory-reclaim path — purgeable/unbound GPU buffer pages are never reclaimed under system memory pressure (DF-4507 family)
Base system — drivers · sys/dev/drm/i915/i915_gem_shrinker.c:313 · pass 2 · new |
none |
untested |
| DF-2843 |
Low |
5.5 |
vnode_pager_generic_putpages() undirties pages on short writes — silent stale-data substitution (write-side DF-2663 family)
Base system — virtual memory · sys/vm/vnode_pager.c:771 · pass 2 · new |
none |
untested |
| DF-3242 |
Low |
5.5 |
ufs_check_export hands out &np->netc_anon after dropping NE_LOCK — concurrent export update (mount -u MNT_DELEXPORT) frees the netcred while nfsrv_fhtovp still reads the embedded credential across a window that includes VFS_FHTOVP disk I/O: UAF read feeding garbage uid/groups into the NFS root-squash mapping
Kernel memory corruption (UAF / OOB write) · sys/vfs/ufs/ufs_vfsops.c:246 · pass 2 · new |
leak |
not_reproduced |
| DF-3396 |
Low |
5.5 |
netmap_mem_rings_create cleanup path leaks every allocated packet buffer: repeated failing ring creation permanently drains the global buffer pool (latent — module does not build on DFly)
Base system — network stack · sys/net/netmap/netmap_mem2.c:987 · pass 2 · new |
none |
build_failed |
| DF-3633 |
Low |
5.5 |
iwm_notif_intr: 12-bit closed_rb_num compared against 256/512-entry ring positions - non-terminating interrupt-context loop, hard system hang (iwm twin of DF-3581)
Base system — drivers · sys/dev/netif/iwm/if_iwm.c:5710 · pass 2 · new |
none |
inconclusive |
| DF-3819 |
Low |
5.5 |
aac_ioctl_send_raw_srb sleeps on channel NULL while its CMFREE callback wakes channel &cm — permanent unkillable process when the FIB pool is exhausted
Privilege escalation (general use) · sys/dev/raid/aac/aac.c:3094 · pass 2 · new |
none |
not_reproduced |
| DF-4015 |
Low |
5.5 |
mps_user_command leaks one mps_command per timed-out MPSIO_MPS_COMMAND (shared pool exhaustion starves the storage stack)
Base system — drivers · sys/dev/raid/mps/mps_user.c:705 · pass 2 · new |
mps_user_command leaks one mps_command per timed-out MPSIO_MPS_COMMAND (goto Ret skips RetFreeUnlocked) - operator drains the finite command pool shared with SCSI I/O -> ENOMEM for management ioctls and eventual I/O starvation of the boot disk. |
untested |
| DF-4245 |
Low |
5.5 |
dpt_get_conf ready-wait is an unbounded, interrupts-disabled spin driven only by HBA register values — hostile status shape hangs the kernel at attach
Base system — drivers · sys/dev/raid/dpt/dpt_scsi.c:498 · pass 2 · new |
none demonstrated on guest (no DPT hardware); source-level impact: unbounded crit-section MMIO spin at attach → kernel hang (boot-time DoS) from hostile/faulty HBA status |
untested |
| DF-4355 |
Low |
5.5 |
Four unbounded MMUCR_BUSY busy-waits (two in hard interrupt context) plus no-exit device-driven loops: unrecoverable system hang on SMC91xx glitch or PCMCIA ejection
Base system — drivers · sys/dev/netif/sn/if_sn.c:275 · pass 2 · new |
none |
untested |
| DF-4425 |
Low |
5.5 |
VIRTIO_BLK_F_MQ with device-reported num_queues=0: modulo-by-zero panic in vtblk_attach before any validation
Base system — drivers · sys/dev/virtual/virtio/block/virtio_blk.c:292 · pass 2 · new |
none |
not_reproduced |
| DF-4670 |
Low |
5.5 |
ar5416GetDiagState: HAL_DIAG_CHECK_HANGS dereferences args==NULL when the ioctl omits ATH_DIAG_IN — NULL-deref panic (DF-1521 sibling, in-file)
Privilege escalation (general use) · sys/dev/netif/ath/ath_hal/ar5416/ar5416_misc.c:543 · pass 2 · new |
Unprivileged local kernel panic on ath(4)+ATH_DIAGAPI machines while DF-4320's missing priv check stands; NULL read at fixed address 0 - pure DoS |
untested |
| DF-4843 |
Low |
5.5 |
Unvalidated device-DMA'd IH wptr (alignment never checked, only masked) + unbounded amdgpu_ih_process ring walk — a misaligned wptr can never equal the 16-byte-stepping rptr, permanently hard-spinning the amdgpu interrupt thread (all GPU interrupt delivery dead)
Base system — drivers · sys/dev/drm/amd/amdgpu/tonga_ih.c:196 · pass 2 · new |
dos |
untested |
| DF-3092 |
Info |
5.5 |
fillinusemap() tail-marking loop compares a cluster number against a word-index bound (never executes) and chainlength()'s word-relative clamp misses — phantom "free" clusters beyond pm_maxcluster drive usemap_alloc KASSERT panic on INVARIANTS kernels
Kernel memory corruption (UAF / OOB write) · sys/vfs/msdosfs/msdosfs_fat.c:320 · pass 2 · new |
none |
not_reproduced |
| DF-4018 |
Info |
5.5 |
amdgpu userptr machinery is dead-but-armed: usertask is never set, so re-enabling GEM_USERPTR instantly NULL-derefs in amdgpu_ttm_tt_get_user_pages; gup shim has an inverted KKASSERT; ttm->sg is freed without NULLing (double-free shape)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_ttm.c:834 · pass 2 · new |
amdgpu userptr machinery is dead-but-armed: usertask never set (#if 0) so re-enabling GEM_USERPTR instantly NULL-derefs at amdgpu_ttm.c:834 (guard after deref); gup shim has an inverted KKASSERT; pin failure path kfrees ttm->sg without NULLing (double-free shape); bind-error paths leak pinned pages. |
untested |
| DF-4003 |
Medium |
5.4 |
fiji_populate_cac_table overflows BapmVddcVid{Lo,Hi}Sidd[16] and OOB-reads vddc_voltage_table.entries[32]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/fiji_smumgr.c:772 · pass 2 · new |
none |
inconclusive fixed |
| DF-4004 |
Medium |
5.4 |
fiji clock-stretcher walk overflows Sclk_voltageOffset[8] using the raw VBIOS SCLK-dependency count
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/fiji_smumgr.c:1706 · pass 2 · new |
none |
inconclusive fixed |
| DF-3404 |
Low |
5.4 |
parse_beacon VENDOR branch stores pre-draft HT-info IE into scan->htcap instead of scan->htinfo (copy-paste typo)
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_input.c:652 · pass 2 · new |
none |
not_reproduced |
| DF-3830 |
Low |
5.4 |
trm_detach frees I/O port, SRB array and DMA tags before bus_teardown_intr and SIM deregistration, with no drain of in-flight CCBs — use-after-free / permanent CAM wedge on kldunload
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:3696 · pass 2 · new |
none |
inconclusive |
| DF-3832 |
Low |
5.4 |
trm_eepromBuf[unit] indexed by unbounded adapter unit — 128-byte OOB read+write of adjacent kernel globals at attach with a 5th Tekram card
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:212 · pass 2 · new |
none |
inconclusive |
| DF-4850 |
Low |
5.4 |
amrr ath_rate_getxtxrates() writes hardware rate codes into ath_rc_series.rix (rate-table INDEX) fields and clobbers the correct series-0 index — wrong-entry rt->info[] derefs on every multi-rate-retry frame, HT mis-programming on 11n HALs, and a second OOB read sink when compounded with DF-1931
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_rate/amrr/amrr.c:135 · pass 2 · new |
none |
untested |
| DF-3184 |
High |
5.3 |
procfs_exit() writes pfs->pfs_pid through a pfsnode a concurrent vnode reclaim may already have kfree()'d (use-after-free write, 4-byte OR at offset 20 in the M_PROCFS zone)
Kernel memory corruption (UAF / OOB write) · sys/vfs/procfs/procfs_subr.c:542 · pass 2 · new |
none |
not_reproduced |
| DF-3582 |
High |
5.3 |
Type-confused FCPARAM() write in RQCS_RESET_OCCURRED handler: on dual-bus SCSI cards a routine bus reset performs a 32-bit RMW ~21 KB past the 408-byte isp_param heap allocation
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/isp/isp.c:6105 · pass 2 · new |
none |
inconclusive |
| DF-2719 |
Medium |
5.3 |
Coredump each_segment traverses the vm_map RB-tree unlocked while a pre-counted-'stopped' sibling LWP can still be mutating it (UAF read race)
Kernel memory corruption (UAF / OOB write) · sys/kern/imgact_elf.c:1193 · pass 2 · new |
none |
not_reproduced |
| DF-2924 |
Medium |
5.3 |
cmd_get_usage_all walks accounting RB trees without ac_spin → RB_NEXT traversal cycle (unkillable kernel spin) and torn 64-bit reads
Base system — kernel core (kern) · sys/kern/vfs_quota.c:183 · pass 2 · new |
none |
not_reproduced |
| DF-3673 |
Medium |
5.3 |
mpt_timeout() uses the CCB after xpt_done() hands it back — callout_stop() does not interlock against the in-progress timeout callback (freed/recycled CCB -> wild deref or wrong-request abort)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_cam.c:1240 · pass 2 · new |
none |
inconclusive |
| DF-3735 |
Medium |
5.3 |
r600_pm_init_profile hardcodes clock-mode indices without num_clock_modes validation — default-config OOB heap read of per-state clock_info
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r600.c:685 · pass 2 · new |
none |
inconclusive |
| DF-3750 |
Medium |
5.3 |
Raw-xmit path lacks TXWI pool-empty guard → NULL-pointer write panic under Tx saturation
Base system — drivers · sys/dev/netif/ral/rt2860.c:1536 · pass 2 · new |
none |
not_reproduced |
| DF-3757 |
Medium |
5.3 |
tulip_txput_setup() ring-full guard off-by-one (`ri_free == 1`): ri_free underflows to -1 and the setup frame is programmed over a chip-owned TX descriptor
Privilege escalation (general use) · sys/dev/netif/de/if_de.c:3631 · pass 2 · new |
none |
not_reproduced |
| DF-3760 |
Medium |
5.3 |
alc_stop()/alc_tick() AB-BA deadlock: self-re-arming tick callout blocks on the driver serializer while alc_stop calls the synchronous callout_stop() holding it — permanent kernel wedge on ifconfig down/MTU change/detach/suspend (DF-3715 family, independent instance)
Base system — drivers · sys/dev/netif/alc/if_alc.c:3414 · pass 2 · new |
none |
inconclusive |
| DF-3894 |
Medium |
5.3 |
XPT_RESET_DEV allocates from the normal command pool but frees via the high-priority list, and underflows tm_count: both pools' accounting corrupted, TM simq-freeze discipline permanently disabled (DF-1223 twin)
Base system — drivers · sys/dev/raid/mps/mps_sas.c:2765 · pass 2 · new |
none |
not_reproduced |
| DF-3945 |
Medium |
5.3 |
CVE-2019-14615 carried: Gen9 indirect-context WA batch lacks the EU residual-state cleardown PIPE_CONTROL (cross-context GPU info leak on SKL/KBL/BXT/GLK/CFL)
Kernel memory leaks / uninitialized disclosure · CVE-2019-14615 · sys/dev/drm/i915/intel_lrc.c:1581 · pass 2 · new |
none |
not_reproduced |
| DF-4037 |
Medium |
5.3 |
polaris10_populate_cac_table overflows BapmVddcVid{Lo,Hi,Hi2}Sidd[16] and OOB-reads vddc_voltage_table.entries[32]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:712 · pass 2 · new |
none |
inconclusive fixed |
| DF-4038 |
Medium |
5.3 |
polaris10 clock-stretcher walk overflows Sclk_voltageOffset[8] using the raw VBIOS SCLK-dependency count; AVFS walk OOB-reads entries[i<NUM_VFT_COLUMNS] when the count is lower
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:1555 · pass 2 · new |
none |
inconclusive fixed |
| DF-4152 |
Medium |
5.3 |
Object-table walks dereference unvalidated VBIOS u16 offsets and stride unbounded/unchecked sizes — OOB kernel heap reads at attach (amdgpu twin of DF-3745)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_atombios.c:318 · pass 2 · new |
none |
inconclusive |
| DF-4154 |
Medium |
5.3 |
amdgpu_atombios_init_mc_reg_table: zero-stride reg_data walk never terminates; unbounded block walk and unvalidated offsets read past the VBIOS buffer (amdgpu twin of DF-3741)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_atombios.c:1607 · pass 2 · new |
none |
inconclusive |
| DF-4155 |
Medium |
5.3 |
amdgpu_atombios_lookup_voltage_object_v3: zero usSize strides forever; walk bounded only by an unvalidated usStructureSize (hang / OOB read at DPM init)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_atombios.c:1415 · pass 2 · new |
none |
inconclusive |
| DF-4693 |
Medium |
5.3 |
sili_pm: pm_read/pm_write recycle the ata_xfer while the chip still owns the slot — deferred poll timeout returns ATA_S_ONCHIP and the accessors put the ccb back on the free list with ap_expired processing pending
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/sili/sili_pm.c:657 · pass 2 · new |
none |
inconclusive |
| DF-4694 |
Medium |
5.3 |
sili_pm_softreset polls with sili_quick_timeout whose sole-active KKASSERT is violated by co-active PM-target commands — XPT_RESET_DEV enters softreset without draining the port
Base system — drivers · sys/dev/disk/sili/sili_pm.c:450 · pass 2 · new |
none |
inconclusive |
| DF-4844 |
Medium |
5.3 |
drm_sched_entity_fini dependency cleanup check-then-uses entity->dependency while drm_sched_entity_clear_dep NULLs it unsynchronized from fence-signaling context — dma_fence_remove_callback(NULL/stale) NULL-deref lockmgr panic or UAF (tight-window TOCTOU, beyond DF-1927's scheduler-thread race)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/scheduler/sched_entity.c:279 · pass 2 · new |
none |
untested |
| DF-4971 |
Medium |
5.3 |
mfip_detach frees the CAM sim without draining mfi_cam_ccbq or clearing mfi_cam_start: queued passthrough ccbs leak permanently (unkillable hang) and are later dispatched through the stale callback into deregistered/freed CAM state
Base system — drivers · sys/dev/raid/mfi/mfi_cam.c:142 · pass 2 · new |
none |
inconclusive |
| DF-5249 |
Medium |
5.3 |
Lockless sc->cmdq[] ring: producers store .func before args with no barrier and consumer reads are unordered; torn publish yields func(NULL), 16-slot wrap yields type-confused callbacks and silently lost commands
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_run.c:1041 · pass 2 · new |
none |
inconclusive |
| DF-5309 |
Medium |
5.3 |
usb_fifo_open() curr_cpd single-owner check is unlocked — concurrent opens of the same ugen endpoint bypass EBUSY and yield two live owners on one fifo
Base system — bus / platform / cpu · sys/bus/u4b/usb_dev.c:701 · pass 2 · new |
single-owner invariant break -> concurrent-setup entry / latent dangling curr_cpd |
untested |
| DF-2906 |
Low |
5.3 |
Unchecked mbuf allocations in mb_init/mb_reserve/mb_put_mem/md_init — NULL-pointer kernel panic under allocation failure
Base system — kernel core (kern) · sys/kern/libmchain/subr_mchain.c:75 · pass 2 · new |
panic |
not_reproduced |
| DF-3013 |
Low |
5.3 |
_hammer_ip_reseek() KKASSERTs on an EDEADLK-capable seek: INVARIANTS panic (or silent early iteration-EOF on release builds) when the btree re-seek deadlocks
Base system — filesystems · sys/vfs/hammer/hammer_object.c:1443 · pass 2 · new |
none |
not_reproduced |
| DF-3014 |
Low |
5.3 |
hammer_ip_sync_record_cursor() error path frees the bulk record's data reservation on retryable EDEADLK and ignores hammer_blockmap_finalize() failure — double-finalize / blockmap accounting corruption on the flush retry
Base system — filesystems · sys/vfs/hammer/hammer_object.c:1256 · pass 2 · new |
none |
not_reproduced |
| DF-3072 |
Low |
5.3 |
tmpfs_alloc_vp KKASSERT(!TMPFS_VNODE_DOOMED) at subr.c:402 is an unguarded race window: NFS fhtovp lookups on an unlinked file during inactive's DOOMED..reclaim interval panic INVARIANTS kernels
Kernel memory corruption (UAF / OOB write) · sys/vfs/tmpfs/tmpfs_subr.c:395 · pass 2 · new |
none |
not_reproduced |
| DF-3265 |
Low |
5.3 |
Recursive exclusive lockmgr panic ('locking against myself') in pf_state_key_attach under pf_state_item allocation failure
Base system — network stack · sys/net/pf/pf.c:915 · pass 2 · new |
panic |
not_reproduced |
| DF-3275 |
Low |
5.3 |
Non-constant-time authentication compares: PAP password and CHAP response checked with early-exit bcmp
Base system — network stack · sys/net/sppp/if_spppsubr.c:4077 · pass 2 · new |
none |
not_reproduced |
| DF-3348 |
Low |
5.3 |
Unlocked/wrong-lock readers of the port list vs lagg_port_destroy: unlocked lagg_sysctl_active walk, and port if_ioctl/if_output/if_lagg torn down without the port serializer
Privilege escalation (general use) · sys/net/lagg/if_lagg.c:763 · pass 2 · new |
none |
not_reproduced |
| DF-3600 |
Low |
5.3 |
sq_edc_source_names[] indexed by unbounded 3-bit SQ_EDC_INFO.SOURCE field: OOB rodata pointer handed to %s
Base system — drivers · sys/dev/drm/amd/amdgpu/gfx_v8_0.c:6811 · pass 2 · new |
none |
inconclusive |
| DF-3608 |
Low |
5.3 |
ncr(4) on >4GB amd64 systems: all chip DMA descriptors are 32-bit but filled from unrestricted 64-bit vtophys() — silent address truncation drives transfers to wrong physical memory
Base system — drivers · sys/dev/disk/ncr/ncr.c:4149 · pass 2 · new |
none |
inconclusive |
| DF-3718 |
Low |
5.3 |
mxge_tx_done walks the TX ring toward the unvalidated NIC send_done_count: premature TX-buffer free (device DMA-after-free) and permanent interrupt-context spin under tx/main serializers
Base system — drivers · sys/dev/netif/mxge/if_mxge.c:2369 · pass 2 · new |
none |
inconclusive |
| DF-3737 |
Low |
5.3 |
emx_tx_collect walks the TX ring toward an unvalidated TDH register value: out-of-range write-back (0xFFFFFFFF on surprise removal) -> all TX buffers freed (DMA-after-free window) then KKASSERT panic / permanent serializer-held spin (DF-3714 em(4) twin)
Base system — drivers · sys/dev/netif/emx/if_emx.c:2817 · pass 2 · new |
none |
inconclusive |
| DF-4095 |
Low |
5.3 |
emu_intr2 tests and calls sc->mpu_intr/sc->mpu without the lock, racing emu_muninit + mpu401_uninit's kfree at detach — heap UAF on struct mpu401
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/emu10k1.c:1196 · pass 2 · new |
none |
inconclusive |
| DF-4300 |
Low |
5.3 |
vkbd_dev_close check-then-sleep TASK handshake can miss the drain task's wakeup and pin the kernel mplock indefinitely (system-wide device-I/O stall until signalled)
Base system — drivers · sys/dev/misc/vkbd/vkbd.c:307 · pass 2 · new |
none |
not_reproduced |
| DF-4323 |
Low |
5.3 |
pcm_unregister() can complete while dsp_clone() sleeps in PCM_WAIT (which holds no SD_F_BUSY), leaving the woken handler to run snd_clone_alloc(NULL)/snd_clone_disabled on freed state and PCM_RELEASE on the freed d->lock — NULL-deref / lock use-after-free panic
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pcm/sound.c:1165 · pass 2 · new |
dsp_clone validates under PCM_LOCK and only then does PCM_WAIT/ACQUIRE |
untested |
| DF-4419 |
Low |
5.3 |
Use-after-free race: unprivileged bktr ioctls race privileged msp_detach and write into the freed msp3400c struct (no locking, admitted by 'XXX mutex lock required')
Kernel memory corruption (UAF / OOB write) · sys/dev/video/bktr/msp34xx.c:1206 · pass 2 · new |
UAF write into freed M_DEVBUF block via msp_wake_thread racing msp_detach during kldunload; heap corruption ceiling on dedicated allocation, realistically panic |
untested |
| DF-4562 |
Low |
5.3 |
mpssas_add_device/mpssas_volume_add re-initialize a target slot without checking the asynchronous removal is still in flight (INREMOVAL flag cleared under an outstanding TM; TAILQ_INIT over possibly-linked command lists)
Base system — drivers · sys/dev/raid/mps/mps_sas_lsi.c:653 · pass 2 · new |
none |
untested |
| DF-4789 |
Low |
5.3 |
mpr_print_sgl walks the request-frame SGL unbounded, dereferences a possibly-NULL chain, and hexdumps a fixed 128 bytes — unconditional console path on firmware MPI2_IOCSTATUS_INVALID_SGL
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_table.c:547 · pass 2 · new |
none |
untested |
| DF-4868 |
Low |
5.3 |
mps_print_sgl walks the request-frame SGL unbounded, dereferences a possibly-NULL chain, and hexdumps fixed 128 bytes — unconditional console path on firmware MPI2_IOCSTATUS_INVALID_SGL (chain deref sits in the only walk, unlike the mpr twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps_table.c:449 · pass 2 · new |
none |
untested |
| DF-4901 |
Low |
5.3 |
cz_ih (Carrizo/Stoney APU): unvalidated device-DMA'd IH wptr — 4th instance of the DF-4843/4854/4887 family: a misaligned wptr permanently hard-spins the amdgpu interrupt thread, and the RB_OVERFLOW +16 recovery imports the bad phase, parking rptr within 12 bytes of ring end for an up-to-12-byte OOB read past the zero-slack 64 KiB GTT ring BO
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/cz_ih.c:192 · pass 2 · new |
dos |
untested |
| DF-4902 |
Low |
5.3 |
iceland_ih (Topaz): 5th and last unfiled instance of the unvalidated-wptr IH family — byte-identical to cz_ih.c in all vulnerable logic (permanent ithread spin on misaligned wptr + RB_OVERFLOW phase-import ring-end OOB read)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/iceland_ih.c:192 · pass 2 · new |
dos |
untested |
| DF-5170 |
Low |
5.3 |
evergreen_copy_dma discards radeon_sync_resv()/radeon_sync_rings() returns — Evergreen/Cayman twin of DF-4871/DF-5089: DMA copy committed without cross-ring ordering on sync failure, silent bo data corruption
Base system — drivers · sys/dev/drm/radeon/evergreen_dma.c:130 · pass 2 · new |
none |
untested |
| DF-5338 |
Low |
5.3 |
kern_trap() broken fall-through for unhandled kernel-mode traps (T_PROTFLT/T_SEGNPFLT/T_BPTFLT/T_TRCTRAP): NULL p/lp deref for kernel threads, trapsignal(lp, i=0) sig-0 OOB read/call, and handler-returns-to-faulting-instruction livelock — pc64 ends these cases in trap_fatal
Kernel memory leaks / uninitialized disclosure · sys/platform/vkernel64/x86_64/trap.c:636 · pass 2 · new |
NULL-deref recursion / sig-0 OOB indirect call / silent livelock on any vkernel kernel fault |
untested |
| DF-5681 |
Low |
5.3 |
Kernel _PROP_ONCE_RUN cmpset-once has no completion barrier: first concurrent prop_number use races the interned-tree lazy init (lockmgr on mid-lockinit lock, rbt_ops still NULL)
Base system — libkern / libprop / ddb · sys/libprop/prop_number.c:167 · pass 2 · new |
none |
not_reproduced fixed |
| DF-5693 |
Low |
5.3 |
prop_bool singleton lazy-init shares the kernel cmpset-once with no completion barrier: concurrent first-ever <true/> internalize can hand out a NULL-typed, stale-valued, or refcount-corrupted TRUE/FALSE singleton (DF-2225/DF-5681 class, third instance)
Base system — libkern / libprop / ddb · sys/libprop/prop_bool.c:111 · pass 2 · new |
none |
not_reproduced fixed |
| DF-3987 |
High |
5.2 |
Poisoned cur_link_settings.lane_count outlives link training: dp_test_send_phy_test_pattern OOB stack WRITE past lane_settings[4] (~3KB span) and hpd_rx_irq_check_link_loss_status OOB stack READ past the 6-byte union hpd_irq_data
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/core/dc_link_dp.c:1772 · pass 2 · new |
none |
inconclusive |
| DF-4337 |
High |
5.2 |
atom_context has no bios_size: every VBIOS offset dereference in the ATOM interpreter and master-table parsers is unbounded (OOB heap read up to 128KB; interpreter executes past the image) — central hub fix for the DF-4104/4152/4289 family
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/atom.c:1194 · pass 2 · new |
none |
inconclusive |
| DF-4345 |
High |
5.2 |
amdgpu atom_context has no bios_size: every VBIOS offset dereference in the ATOM interpreter and master-table parsers is unbounded (OOB heap read up to 128KB; interpreter executes past the image) — central hub fix port of DF-4337 to the amdgpu twin
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/atom.c:268 · pass 2 · new |
none |
inconclusive |
| DF-4288 |
Medium |
5.2 |
VRAM_Info MC-register-table path: unvalidated VBIOS NULL/offset/stride — NULL-deref panic, OOB read walk, and stride-0 infinite loop at attach
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:148 · pass 2 · new |
none |
inconclusive |
| DF-4289 |
Medium |
5.2 |
SS-assignment walk bounded only by the table's own usStructureSize — up to 64KB OOB read past the ATOM table
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:1231 · pass 2 · new |
none |
inconclusive |
| DF-4773 |
Medium |
5.2 |
apple_smc: detach frees sensor-key state while its per-device sysctl oids stay registered until device_delete — post-detach UAF read / NULL-ioport panic window (DF-4647 family)
Kernel memory corruption (UAF / OOB write) · sys/dev/apple/smc/smc_sysctl.c:169 · pass 2 · new |
leak |
inconclusive |
| DF-3618 |
Low |
5.2 |
ath_vap_delete second TX drain races live taskqueue workers before kfree(avp) — author-acknowledged use-after-free window on vap teardown
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath/if_ath.c:1785 · pass 2 · new |
none |
inconclusive |
| DF-3669 |
Low |
5.2 |
ix_detach never terminates the self-re-arming per-ring tx_gc_timer: pending callout fires on freed tx_rings → use-after-free (write into freed slab + callout-queue corruption)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ix/if_ix.c:373 · pass 2 · new |
none |
inconclusive |
| DF-3681 |
Low |
5.2 |
igb_detach never terminates the self-re-arming sc->timer and per-ring tx_gc_timer (callout_stop only, which does not interlock against an in-progress callback): re-arm after detach → use-after-free on freed softc/rings and, under kldunload, execution of unmapped module text
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/igb/if_igb.c:327 · pass 2 · new |
none |
inconclusive |
| DF-3822 |
Low |
5.2 |
aac_detach frees the container list with no aac_container_lock while the still-live interrupt path aac_handle_aif and the aac_query_disk ioctl traverse it — use-after-free race
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/aac/aac.c:679 · pass 2 · new |
none |
inconclusive |
| DF-3955 |
Low |
5.2 |
Double free of rdev->pm.dpm.ps (and leak of parsed ps_priv states) on the kzalloc-failure path of kv_parse_power_table
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/kv_dpm.c:2675 · pass 2 · new |
none |
inconclusive |
| DF-4082 |
Low |
5.2 |
oce_tx returns 0 (success) when oce_insert_vlan_tag fails on !M_WRITABLE mbufs: the dequeued mbuf is leaked and the packet silently dropped while oce_start_locked believes it was sent
Base system — drivers · sys/dev/netif/oce/oce_if.c:833 · pass 2 · new |
none |
untested |
| DF-4276 |
Low |
5.2 |
cbb_detach() deletes the bridge children before stopping the slot event thread — cbb_insert/cbb_removal operate on freed child devices (UAF race), and the RUNNING-flag startup window lets detach free the softc under a live thread
Kernel memory corruption (UAF / OOB write) · sys/dev/pccard/pccbb/pccbb.c:285 · pass 2 · new |
none |
not_reproduced |
| DF-4312 |
Low |
5.2 |
dcn10_link_encoder_create indexes 4-entry const register tables with unvalidated VBIOS-derived transmitter/channel/hpd_source values (−1 and 4..9 / −1 and 4..7 / 4..6) — OOB kernel reads that become wild-MMIO register bases
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dcn10/dcn10_resource.c:744 · pass 2 · new |
none |
inconclusive |
| DF-4319 |
Low |
5.2 |
dce80_link_encoder_create indexes 7/6/6-entry const register tables with unvalidated VBIOS-derived transmitter/channel/hpd_source (DF-4312 family)
Base system — drivers · sys/dev/drm/amd/display/dc/dce80/dce80_resource.c:661 · pass 2 · new |
none |
untested |
| DF-4332 |
Low |
5.2 |
Read-only GUI ioctls walk live array state with no driver lock while DELETE_ARRAY frees it - use-after-free race (panic or kernel-pointer copyout)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptmv/gui_lib.c:163 · pass 2 · new |
none |
inconclusive |
| DF-4375 |
Low |
5.2 |
dce112_link_encoder_create indexes 7/6/6-entry const register tables with unvalidated VBIOS-derived transmitter/channel/hpd_source — OOB const reads become the encoder's MMIO register bases (DF-4312/4319/4352 twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dce112/dce112_resource.c:577 · pass 2 · new |
none |
inconclusive |
| DF-4615 |
Low |
5.2 |
ips_copperhead_reinit ignores ips_copperhead_queue_init's return: freed status queue left dangling → interrupt-context UAF read/write; every forced reinit also leaks the previous queue mapping
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ips/ips.c:630 · pass 2 · new |
none |
not_reproduced |
| DF-4677 |
Low |
5.2 |
cmx_detach lifetime holes: dying is checked unlocked before CMX_LOCK (TOCTOU across lockuninit) and registered knotes are never detached from sc->kq.ki_note — knote cleanup after softc free is a use-after-free
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/cmx/cmx.c:265 · pass 2 · new |
cmx_detach lifetime holes: unlocked dying TOCTOU across lockuninit, and registered knotes never detached from ki_note - post-free knote cleanup is a UAF write |
untested |
| DF-4707 |
Low |
5.2 |
amr_pci_detach's EBUSY guard is defeated: AMR_STATE_OPEN is cleared on ANY close (not last) and is TOCTOU-open — kldunload frees the softc while a /dev/amr0 descriptor still references it via dev->si_drv1 (UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/amr/amr_pci.c:365 · pass 2 · new |
none |
inconclusive |
| DF-4875 |
Low |
5.2 |
afd_describe(): ksprintf of device-controlled u64 mediasize into char[16] — deterministic 1-byte NUL stack overflow at attach
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-fd.c:446 · pass 2 · new |
none |
untested |
| DF-5116 |
Low |
5.2 |
dcons_crom_detach NULLs post_busreset with no synchronization against fw_busreset's unlocked check-then-call dispatch running in the fwohci interrupt handler — a physical bus reset racing kldunload executes dcons_crom_post_busreset on freed softc: interrupt-context UAF write (bzero of sc->unit + crom_add_* stores)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/dcons/dcons_crom.c:196 · pass 2 · new |
none |
untested |
| DF-5250 |
Low |
5.2 |
run_key_set defers a raw ieee80211_key pointer that can be node-embedded (pairwise) and freed before run_key_set_cb executes - UAF read/write including k->wk_pad write
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_run.c:2368 · pass 2 · new |
UAF read/write on freed node (DF-5250) / repeatable mbuf+node leak DoS (DF-5251) / device-bound 4B OOB (DF-5252) |
untested |
| DF-5656 |
Low |
5.2 |
USB dynamic dispatch unload hooks use a fixed 1-second tsleep as the only grace between live consumers and freed module text
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_dynamic.c:113 · pass 2 · new |
none |
not_reproduced |
| DF-3649 |
Medium |
5.1 |
Wild OOB kernel read in ci_get_std_voltage_value_sidd when the CAC leakage table count is 0: idx=(u16)(count-1)=65535 past a ZERO_LENGTH_PTR that passes the entries!=NULL guard
Base system — drivers · sys/dev/drm/radeon/ci_dpm.c:2374 · pass 2 · new |
none |
inconclusive |
| DF-3650 |
Medium |
5.1 |
OOB read in ci_parse_power_table's VCE state fill: clk_idx (ucClockInfoIndex & 0x3f) indexes clockInfo[] without the ucNumEntries bound used everywhere else
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/ci_dpm.c:5654 · pass 2 · new |
none |
inconclusive |
| DF-3835 |
Medium |
5.1 |
bwn_phy_lp_calc_rx_iq_comp divides by hardware-controlled ie_ipwr under a sum-only guard (ie_ipwr==0 with ie_qpwr>=2 passes) — kernel #DE panic from the init/60s calibration path
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn_phy_lp.c:2340 · pass 2 · new |
none |
inconclusive |
| DF-4068 |
Medium |
5.1 |
ti_rxeof/ti_txeof/ti_handle_events terminate on raw equality with NIC-DMA'd 32-bit producer/consumer words: one out-of-range word makes the loop condition unsatisfiable — interrupt-context livelock with host interrupts masked and the ifnet serializer held (DF-3761 unbounded-walk family, ti(4) instance)
Base system — drivers · sys/dev/netif/ti/if_ti.c:511 · pass 2 · new |
none observed (hardware-gated); source-level: hard CPU livelock with host interrupts masked and ifnet serializer held (permanent DoS of the system) from a single out-of-range DMA'd 32-bit producer/consumer word; txeof variant also underflows sc->ti_txcnt per spin |
inconclusive |
| DF-4179 |
Medium |
5.1 |
twe_add_unit off-by-one: TWEIO_ADD_UNIT with unit==16 passes `unit > TWE_MAX_UNITS` and writes one struct past twe_drive[16] (OOB write into softc + ghost disk with unit truncated to 0)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/twe/twe.c:212 · pass 2 · new |
none |
inconclusive |
| DF-4502 |
Medium |
5.1 |
ahci_pm_softreset leaks the error-CCB reservation on final-retry second-FIS failure: stuck AP_F_ERR_CCB_RESERVED + permanently zeroed ap_active/ap_sactive -> port I/O wedge and KKASSERT panic on next error recovery/probe
Base system — drivers · sys/dev/disk/ahci/ahci_pm.c:805 · pass 2 · new |
none |
inconclusive |
| DF-4613 |
Medium |
5.1 |
ips_timeout performs the forced adapter reinit while holding queue_lock and then tsleeps forever on a completion only the lock-blocked interrupt handler can deliver — permanent watchdog-path deadlock (timer twin of DF-2163)
Base system — drivers · sys/dev/raid/ips/ips.c:337 · pass 2 · new |
none |
not_reproduced |
| DF-4814 |
Medium |
5.1 |
parse_csr_fw bounds its stepping-search loop with attacker-controlled package_header->num_entries while the walked array is fixed fw_info[20]: unbounded kernel heap OOB read (panic / OOB dmc_offset)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_csr.c:134 · pass 2 · new |
code-level: kernel heap OOB read -> page-fault panic (DoS) plus OOB-read dmc_offset steering further fw->data indexing; kernel path not executable on the audit guest (only display device is QEMU std-VGA 0x1234:1111, parse_csr_fw gated on gen9+ HAS_CSR hardware); the exact parse loop + packed structs were replicated and executed on the guest, faulting past the buffer in both variants (run.log) |
untested |
| DF-4987 |
Medium |
5.1 |
tws_passthru trusts driver_pkt.buffer_length with no bound: arbitrary-size M_ZERO kernel alloc, EFBIG ignored → command submitted with truncated SGL, protected-dmamap EINPROGRESS panic on >4GB-RAM hosts, unkillable 60s sleep + controller-wide reset
Base system — drivers · sys/dev/raid/tws/tws_user.c:117 · pass 2 · new |
none |
not_reproduced |
| DF-3606 |
Low |
5.1 |
MSI-X allocation failure silently falls back to one MSI/legacy vector while rx_ring_cnt stays >1: RSS rings 1..N are configured but never serviced — traffic blackhole on BCM5709/5716
Base system — drivers · sys/dev/netif/bce/if_bce.c:6748 · pass 2 · new |
RSS rings un-serviced traffic blackhole |
untested |
| DF-3834 |
Low |
5.1 |
Inverted KASSERT in bwn_phy_lp_tblinit_r2 asserts rev < 2 although it is only ever called for rev >= 2 — deterministic panic at PHY init on INVARIANTS kernels (upstream-inherited)
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn_phy_lp.c:2742 · pass 2 · new |
none |
inconclusive |
| DF-4120 |
Low |
5.1 |
Second divide-by-zero window in cypress_populate_mclk_value: clk_s truncating to 0 makes (clk_s * 625) a zero divisor at :567 (sibling of DF-1387, different expression and condition)
Base system — drivers · sys/dev/drm/radeon/cypress_dpm.c:565 · pass 2 · new |
none |
inconclusive |
| DF-4227 |
Low |
5.1 |
cyapa detach/module-unload leaves open fds, blocked readers and kqueue hooks on a vanished softc — NULL-deref panic / dispatch through freed module ops
Kernel memory corruption (UAF / OOB write) · sys/dev/smbus/cyapa/cyapa.c:545 · pass 2 · new |
none |
untested |
| DF-4272 |
Low |
5.1 |
dcn_bw_notify_pplib_of_wm_ranges dereferences unchecked res_pool->pp_smu (kzalloc-failure NULL) — probe-time panic on DCN1
Base system — drivers · sys/dev/drm/amd/display/dc/calcs/dcn_calcs.c:1401 · pass 2 · new |
none |
untested |
| DF-4273 |
Low |
5.1 |
twa(4) firmware-passthrough buffer_length completely unvalidated: unbounded kernel kmalloc/copyin (memory-exhaustion DoS) and silent no-payload execution for lengths >= 0xFFFFFE01
Base system — drivers · sys/dev/raid/twa/tw_osl_freebsd.c:927 · pass 2 · new |
none |
inconclusive |
| DF-4287 |
Low |
5.1 |
acpi_thinkpad attach: kfreeenv(NULL) on the nosmbios path → panic("trying to free NULL pointer") when exactly one SMBIOS env var is unset
Base system — drivers · sys/dev/acpica/acpi_thinkpad/acpi_thinkpad.c:622 · pass 2 · new |
boot-time / attach-time kernel panic (kfree of NULL pointer) on ThinkPad firmware with a partial SMBIOS Type-1 table; no user interaction |
inconclusive |
| DF-4555 |
Low |
5.1 |
dce_aux: ignored acquire() failure + unconditional release_engine() closes NULL ddc — near-NULL deref panic under AUX/GPIO contention
Base system — drivers · sys/dev/drm/amd/display/dc/dce/dce_aux.c:53 · pass 2 · new |
none |
not_reproduced |
| DF-4813 |
Low |
5.1 |
ata_marvell_edma_allocate() dereferences ch->dma unconditionally (work_bus/work/work_tag): NULL-pointer panic when ata_dmaalloc() failed at channel attach — DF-4809 class, but in ->allocate (marvell is the only nata chipset whose allocate touches ch->dma)
Base system — drivers · sys/dev/disk/nata/chipsets/ata-marvell.c:209 · pass 2 · new |
none |
untested |
| DF-4930 |
Low |
5.1 |
atphy_attach 'no media present' branch skips mii_phy_add_media, so miibus_mediainit calls ifmedia_set on an EMPTY media list → DragonFly panic("ifmedia_set") at attach
Base system — drivers · sys/dev/netif/mii_layer/atphy.c:135 · pass 2 · new |
none |
not_reproduced |
| DF-5003 |
Low |
5.1 |
truephy_attach 'no media present' branch skips mii_phy_add_media, so miibus_mediainit calls ifmedia_set on an EMPTY media list → panic("ifmedia_set") at attach (4th leaf of the DF-4930/DF-4957/DF-4981 family)
Base system — drivers · sys/dev/netif/mii_layer/truephy.c:171 · pass 2 · new |
none |
not_reproduced |
| DF-5077 |
Low |
5.1 |
acphy_attach 'no media present' branch skips mii_phy_add_media, so miibus_mediainit calls ifmedia_set on an EMPTY media list → panic("ifmedia_set") at attach (6th leaf of the DF-4930/4957/4981/5003/5053 family; hub DF-5001)
Base system — drivers · sys/dev/netif/mii_layer/acphy.c:169 · pass 2 · new |
none |
not_reproduced fixed |
| DF-5105 |
Low |
5.1 |
amdgpu_do_test_moves success path leaks all n pinned+reserved+referenced GTT test BOs (~entire GTT) with reservation locks held forever
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_test.c:92 · pass 2 · new |
none |
not_reproduced |
| DF-5120 |
Low |
5.1 |
vtrnd_harvest runs unbounded virtqueue_poll() busy-wait inside the periodic softclock callout — a stalled host entropy backend permanently hangs the guest kernel
Base system — drivers · sys/dev/virtual/virtio/random/virtio_random.c:215 · pass 2 · new |
none |
untested |
| DF-3463 |
Low |
5.0 |
gre decapsulation leaves m->m_pkthdr.rcvif pointing into gre_softc while ip_reass can hold the fragment chain far past if_detach's netmsg_service_sync barrier — post-destroy reassembly processes the packet through freed gre_softc memory (delayed UAF reads in ip_input)
Kernel memory corruption (UAF / OOB write) · sys/netinet/ip_gre.c:184 · pass 2 · new |
none |
not_reproduced |
| DF-4130 |
Low |
5.0 |
Firmware-controlled NIC_CREATE_WQ response db_offset used unvalidated as MMIO write offset on every TX doorbell
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/oce/oce_mbox.c:1710 · pass 2 · new |
none |
not_reproduced |
| DF-3476 |
Medium |
4.8 |
DLC teardown deadlock: untimeout() called under pcb->pcb_lock races the DLC timeout callback blocking on pcb_lock - DragonFly's synchronous callout_stop() turns the overlap into a permanent non-interruptible AB-BA wedge (softclock thread + task/user thread halted)
Base system — network stack · sys/netgraph7/bluetooth/socket/ng_btsocket_rfcomm.c:3554 · pass 2 · new |
dos |
untested |
| DF-3365 |
Low |
4.8 |
sta_entry lifetime: three additional unlocked-use windows beyond DF-0395/0396 — PICK1ST match_bss after TABLE_UNLOCK, select_bss result used/joined after unlock, roam/assoc sta_lookup uses
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_scan_sta.c:397 · pass 2 · new |
none |
not_reproduced |
| DF-3772 |
Low |
4.8 |
Hardware-fed CI/SACT consistency KKASSERT (and CI==0 asserts) panic on chipset misbehavior the driver documents and recovers from elsewhere
Base system — drivers · sys/dev/disk/ahci/ahci.c:2622 · pass 2 · new |
none |
not_reproduced |
| DF-3773 |
Low |
4.8 |
ahci_beg_exclusive_access() drain loop never processes deferred timeouts -> permanent port wedge with the port lock held
Base system — drivers · sys/dev/disk/ahci/ahci.c:2266 · pass 2 · new |
none |
not_reproduced |
| DF-4723 |
Low |
4.8 |
ath_tdma_settimers computes DBA/SWBA alerts with 32-bit (nexttbtt<<3) while nexttbtt is the full TSF-derived TU value — inconsistent timer wrap corrupts beacon timers after ~6.2 days of continuous TDMA operation (the in-file unaudited wrap TODO)
Base system — drivers · sys/dev/netif/ath/ath/if_ath_tdma.c:144 · pass 2 · new |
none |
untested |
| DF-5110 |
Low |
4.8 |
SATA hotplug connect tasks are unsynchronized and never drained at channel teardown: racing ATTACH/DETACH tasks yield NULL-softc panic or duplicate disk children
Base system — drivers · sys/dev/disk/nata/ata-sata.c:45 · pass 2 · new |
NULL-deref panic / duplicate disk children |
untested |
| DF-2646 |
Medium |
4.7 |
hammer2_flush_core LOST CHILD3 panics on the NULL->non-NULL parent transition (unconditional KKASSERT / hammer2_chain_unlock(NULL))
Hammer2 · sys/vfs/hammer2/hammer2_flush.c:681 · pass 2 · new |
none |
inconclusive |
| DF-2649 |
Medium |
4.7 |
NULL parent->data dereference in hammer2_bulkfree_scan's PFSROOT kprintfs when a PFS-root chain fails to load with a non-CHECK error
Hammer2 · sys/vfs/hammer2/hammer2_bulkfree.c:143 · pass 2 · new |
none |
not_reproduced |
| DF-2660 |
Medium |
4.7 |
hammer2_sync_insert() TOCTOU re-lookup races concurrent same-key creates on mounted rw clusters: KKASSERT(chain == NULL) panics INVARIANTS kernels (duplicate-key chain_create on production)
Hammer2 · sys/vfs/hammer2/hammer2_synchro.c:741 · pass 2 · new |
none |
not_reproduced |
| DF-2712 |
Medium |
4.7 |
bsd4_kick_helper issues IPIs/wakeups while holding the global bsd4_spin: FIFO-full liveloop enables same-cpu schedclock recursion into the spinlock (whole-system wedge), unprivileged-triggerable when usched_bsd4 is active
Base system — kernel core (kern) · sys/kern/usched_bsd4.c:1590 · pass 2 · new |
none |
not_reproduced |
| DF-3012 |
Medium |
4.7 |
hammer(1) in-memory record permanently leaked when hammer_blockmap_reserve() fails in hammer_ip_add_bulk() (and on the namekey-exhaustion ENOSPC path of hammer_ip_add_direntry)
Kernel memory leaks / uninitialized disclosure · sys/vfs/hammer/hammer_object.c:711 · pass 2 · new |
none |
not_reproduced |
| DF-3096 |
Medium |
4.7 |
mount -u ronly→rw transition: RW internal state, FWRITE device opens, and recovered-volume-header media writes are not rolled back when stage2 recovery fails
Base system — filesystems · sys/vfs/hammer/hammer_vfsops.c:478 · pass 2 · new |
none |
untested |
| DF-3193 |
Medium |
4.7 |
fuse_unmount() frees fmp under in-flight statfs/statvfs: FUSE is MPSAFE (VFS_MPLOCK no-op) and fuse_statfs/statvfs take no fmp reference, so a forced umount plus daemon fd-close kfree's fmp while a statfs thread sleeps on its IPC — the wake/timeout path locks and walks freed (recycled) kernel memory
Kernel memory corruption (UAF / OOB write) · sys/vfs/fuse/fuse_vfsops.c:190 · pass 2 · new |
none |
not_reproduced |
| DF-3194 |
Medium |
4.7 |
nfs_nget/nfs_nget_nonblock race-loser paths vx_put() a never-connected VNON vnode — vx_put forces VREF_FINALIZE for VNON, vnode_terminate runs VOP_INACTIVE, and nfs_inactive dereferences v_data == NULL at nfs_node.c:393 → guaranteed kernel panic whenever the duplicate-validation branch executes (unprivileged local race on any NFS mount)
Base system — filesystems · sys/vfs/nfs/nfs_node.c:194 · pass 2 · new |
none |
not_reproduced |
| DF-3208 |
Medium |
4.7 |
cd9660 ihash publish-before-validate: vget_internal leaves the half-initialized iso_node in the hash on every post-publish error path; a concurrent cd9660_ihashget blocked in vget wakes with the orphaned VNON vnode (i_mnt == NULL, i_devvp == NULL) and returns it as SUCCESS — NULL-deref panic in cd9660_getattr/read on first use
Base system — filesystems · sys/vfs/isofs/cd9660/cd9660_node.c:118 · pass 2 · new |
none |
not_reproduced |
| DF-3239 |
Medium |
4.7 |
p_fd NULL-deref/UAF in vn_fullpath()->cache_fullpath() against an exiting target read via /proc/pid/map
Kernel memory corruption (UAF / OOB write) · sys/vfs/procfs/procfs_map.c:181 · pass 2 · new |
none |
not_reproduced |
| DF-3303 |
Medium |
4.7 |
free_bw_list() runs unschedule_bw_meter() on bw_meter_timers[] without mroute_token (del_mfc / del_bw_upcall DELETE_ALL) — race vs bw_meter_process panic/UAF
Kernel memory corruption (UAF / OOB write) · sys/net/ip_mroute/ip_mroute.c:1115 · pass 2 · new |
none |
inconclusive |
| DF-3584 |
Medium |
4.7 |
isp_control() ISPCTL_RESET_DEV/ABORT_CMD index fcp->isp_dev_map[tgt] without validating tgt; the CAM glue packs bus<<16 into tgt so every channel>0 device reset reads ~108 KB out of bounds
Base system — drivers · sys/dev/disk/isp/isp.c:4522 · pass 2 · new |
none |
inconclusive |
| DF-3620 |
Medium |
4.7 |
Wild OOB kernel read in ci_get_std_voltage_value_sidd when CAC leakage table count==0 (ZERO_LENGTH_PTR passes the entries!=NULL guard)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/ci_dpm.c:2487 · pass 2 · new |
none |
inconclusive |
| DF-3621 |
Medium |
4.7 |
Integer divide-by-zero in spread-spectrum programming when VBIOS SS rate == 0 (or truncates to 0 via /=100)
Base system — drivers · sys/dev/drm/amd/amdgpu/ci_dpm.c:3337 · pass 2 · new |
none |
inconclusive |
| DF-3622 |
Medium |
4.7 |
Integer divide-by-zero on zero MPLL reference clock in ci_calculate_mclk_params (usMemoryReferenceClock stored unchecked)
Base system — drivers · sys/dev/drm/amd/amdgpu/ci_dpm.c:2977 · pass 2 · new |
none |
inconclusive |
| DF-3623 |
Medium |
4.7 |
NULL-pointer dereference of adev->pm.dpm.boot_ps when the VBIOS power table has no ATOM_PPLIB_CLASSIFICATION_BOOT state
Base system — drivers · sys/dev/drm/amd/amdgpu/ci_dpm.c:5297 · pass 2 · new |
none |
inconclusive |
| DF-3657 |
Medium |
4.7 |
XPT_ABORT passes the abort CCB itself to ISPCTL_ABORT_CMD - every SCSI command abort fails (upstream passes accb)
Base system — drivers · sys/dev/disk/isp/isp_freebsd.c:4541 · pass 2 · new |
none |
not_reproduced |
| DF-3700 |
Medium |
4.7 |
synchronize_irq() is an empty lkpi stub: interrupt-drain barriers in i915_irq.c are void (MMIO after power-well off, work resurrected after cancel_work_sync)
Base system — drivers · sys/dev/drm/i915/i915_irq.c:3710 · pass 2 · new |
none |
inconclusive |
| DF-3738 |
Medium |
4.7 |
callout_stop(&sc->timer) under main_serialize deadlocks against emx_timer's blocking lwkt_serialize_enter (AB-BA): emx_stop/emx_intr_body/emx_npoll_status vs the 1 Hz emx_timer — mechanism PROVEN on the guest kernel via protocol replica; fix validated at protocol level (DF-3715 em(4) twin)
Base system — drivers · sys/dev/netif/emx/if_emx.c:1914 · pass 2 · new |
none |
inconclusive fixed |
| DF-3744 |
Medium |
4.7 |
Unbounded LVDS mode-patch record walk and fake-EDID copy read past the VBIOS image buffer (conditional userspace EDID leak)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_atombios.c:1706 · pass 2 · new |
none |
inconclusive |
| DF-3758 |
Medium |
4.7 |
tulip_rx_intr() RX-replenish path panics on bus_dmamap_load_mbuf() failure instead of backing off (local DoS)
Base system — drivers · sys/dev/netif/de/if_de.c:3134 · pass 2 · new |
none |
not_reproduced |
| DF-3801 |
Medium |
4.7 |
mprsas_resetdev_complete dereferences a NULL task-management reply: XPT_RESET_DEV outstanding during a Diag Reset panics the kernel
Base system — drivers · sys/dev/raid/mpr/mpr_sas.c:3392 · pass 2 · new |
none |
not_reproduced |
| DF-3802 |
Medium |
4.7 |
mprsas_send_abort/mprsas_send_reset early -1 returns are ignored by every caller: targ->tm stays set forever, the timed-out CCB is never completed (unkillable I/O hang) and the high-priority TM leaks
Base system — drivers · sys/dev/raid/mpr/mpr_sas.c:1609 · pass 2 · new |
none |
not_reproduced |
| DF-3807 |
Medium |
4.7 |
mskc_handle_events(): 'continue' on a status LE naming an unattached port skips the loop's only consumer-index advance -> permanent mskc_intr() drain spin with the serializer held and the IRQ never re-enabled
Base system — drivers · sys/dev/netif/msk/if_msk.c:3174 · pass 2 · new |
none |
inconclusive |
| DF-3814 |
Medium |
4.7 |
Reallocating diag reset tears down all interrupt handlers and never re-establishes them — controller permanently dead (hang/panic DoS)
Base system — drivers · sys/dev/raid/mpr/mpr.c:801 · pass 2 · new |
none |
inconclusive |
| DF-3818 |
Medium |
4.7 |
NVMe PRP page pool exhaustion: NULL dereference / KASSERT panic above 2048 concurrent native-SGL NVMe I/Os (unprivileged local DoS)
Base system — drivers · sys/dev/raid/mpr/mpr.c:2716 · pass 2 · new |
none |
inconclusive |
| DF-3860 |
Medium |
4.7 |
gpu_info firmware payload pointer built from unvalidated ucode_array_offset_bytes — kernel heap OOB read feeding AMDGPU_INFO disclosure and the DF-3698 walkers
Privilege escalation (general use) · sys/dev/drm/amd/amdgpu/amdgpu_device.c:1354 · pass 2 · new |
none |
not_reproduced |
| DF-3864 |
Medium |
4.7 |
amdgpu_vm_init error paths dereference vm->root.base.bo before amdgpu_vm_bo_base_init assigns it — NULL-pointer panic on a render-node-reachable error path (unprivileged DoS)
Privilege escalation (general use) · sys/dev/drm/amd/amdgpu/amdgpu_vm.c:3097 · pass 2 · new |
none |
not_reproduced |
| DF-3892 |
Medium |
4.7 |
cm_callout is callout_init_mp (lock-free) but mpssas_scsiio_timeout asserts and manipulates mps_lock-protected state: deterministic KKASSERT panic on the first SCSI IO timeout (stock INVARIANTS/GENERIC kernels), lock-free race with interrupt completion otherwise
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps_sas.c:1490 · pass 2 · new |
none |
not_reproduced |
| DF-3893 |
Medium |
4.7 |
mpssas_resetdev_complete dereferences a NULL task-management reply: XPT_RESET_DEV outstanding during a Diag Reset panics the kernel (DF-3801 twin)
Base system — drivers · sys/dev/raid/mps/mps_sas.c:2799 · pass 2 · new |
none |
not_reproduced |
| DF-3895 |
Medium |
4.7 |
mpssas_send_abort/mpssas_send_reset early -1 returns are ignored by every caller: targ->tm wedged forever, TM leaked, timed-out CCBs never returned to CAM, simq frozen for the life of the adapter (DF-3802 twin)
Base system — drivers · sys/dev/raid/mps/mps_sas.c:1303 · pass 2 · new |
none |
not_reproduced |
| DF-3922 |
Medium |
4.7 |
hpt_timeout (takes driver_lock) vs callout_stop under driver_lock: DragonFly's synchronous callout_stop creates an AB-BA sleep deadlock - permanent storage wedge
Base system — drivers · sys/dev/raid/hptmv/entry.c:2629 · pass 2 · new |
none |
not_reproduced |
| DF-3935 |
Medium |
4.7 |
age_stop() (serializer held) vs age_tick (blocking lwkt_serialize_enter): AB-BA callout deadlock — DF-3715 family instance, mechanism PROVEN on the guest via replica (stopper stuck D5 wchan 'costp')
Base system — drivers · sys/dev/netif/age/if_age.c:2194 · pass 2 · new |
none |
inconclusive |
| DF-3941 |
Medium |
4.7 |
cxm(4): cxm_read() consumes the DMA ring with no serialization — a stale ring offset paired with a different buffer underflows `size - offset` (size_t) and uiomove()s past the 61440-byte DMA buffer; read/write desync also discloses never-DMA'd (uninitialized) kernel heap via read(2)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/cxm/cxm.c:2262 · pass 2 · new |
leak |
inconclusive |
| DF-3989 |
Medium |
4.7 |
ale_rxeof trusts the controller-DMA'd 14-bit RRS frame length with no bound vs page/max_frame_size: m_devget heap OOB read up to ~14.8KB past the DMA Rx page into delivered mbufs (kernel heap info leak) — DF-3604/3759/3933 family, ale instance
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ale/if_ale.c:2204 · pass 2 · new |
none |
inconclusive |
| DF-3991 |
Medium |
4.7 |
ale_stop() (serializer held) vs ale_tick() (blocking lwkt_serialize_enter in a self-re-arming NON-MP callout): AB-BA deadlock against synchronous callout_stop — permanent kernel wedge + softclock starvation; mechanism PROVEN on the guest via protocol replica (DF-3715/3760/3935 family)
Base system — drivers · sys/dev/netif/ale/if_ale.c:2263 · pass 2 · new |
none |
inconclusive |
| DF-4051 |
Medium |
4.7 |
nfe_stop() deadlocks the callout subsystem against nfe_tick (AB-BA): synchronous callout_stop() issued while holding the driver serializer vs a tick callback parked in lwkt_serialize_enter — mechanism PROVEN on the guest via nfe-shaped replica (DF-3715/4025/4046 family instance)
Base system — drivers · sys/dev/netif/nfe/if_nfe.c:636 · pass 2 · new |
dos |
inconclusive |
| DF-4153 |
Medium |
4.7 |
amdgpu_atombios_get_clock_info stores VBIOS reference clocks unchecked — zero reference_freq reaches PLL/SS divisions (#DE panic) (amdgpu twin of DF-3742)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_atombios.c:592 · pass 2 · new |
none |
inconclusive |
| DF-4159 |
Medium |
4.7 |
xe_intr() RX livelock: packet-drop paths 'continue' without releasing the packet to the card — permanent hard hang
Base system — drivers · sys/dev/netif/xe/if_xe.c:742 · pass 2 · new |
dos |
untested |
| DF-4161 |
Medium |
4.7 |
Missing !conn_state->crtc guard in intel_panel_set_backlight: unprivileged hw.backlight_level sysctl write NULL-derefs to_intel_crtc(NULL)->pipe in vlv_set_backlight (VLV/CHV)
Base system — drivers · sys/dev/drm/i915/intel_panel.c:1148 · pass 2 · new |
none |
inconclusive |
| DF-4164 |
Medium |
4.7 |
radeon_unmap_vram_bos walks rdev->gem.objects without gem.mutex while unprivileged GEM create/close list_add/kfree the same list: UAF read / list corruption race in the PM clock-change path
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_pm.c:146 · pass 2 · new |
none |
inconclusive |
| DF-4194 |
Medium |
4.7 |
drm_mode_create_tile_group() ERR_PTR vs NULL contract mismatch poisons connector->tile_group — wild-pointer kref_put/deref at four sites (kernel panic)
Base system — drivers · sys/dev/drm/drm_connector.c:1917 · pass 2 · new |
panic |
untested |
| DF-4196 |
Medium |
4.7 |
Lock-free vblank->hwmode.crtc_clock = 0 store in drm_crtc_vblank_off() races the reloaded divisor in drm_calc_vbltimestamp_from_scanoutpos() — TOCTOU integer divide-by-zero (#DE kernel panic)
Base system — drivers · sys/dev/drm/drm_vblank.c:618 · pass 2 · new |
dos |
inconclusive |
| DF-4218 |
Medium |
4.7 |
ae_tick() (non-MP, self-re-arming, blocking lwkt_serialize_enter) vs ae_stop() (sync callout_stop under the serializer): permanent AB-BA deadlock (DF-3715 family)
Base system — drivers · sys/dev/netif/ae/if_ae.c:232 · pass 2 · new |
none |
inconclusive |
| DF-4237 |
Medium |
4.7 |
sili(4): sili_cam_probe() dereferences unchecked sili_ata_get_xfer() NULL — kernel panic when a port reprobe runs against a saturated ccb pool (IDENTIFY write at :290, NCQ-renegotiation read at :372)
Base system — drivers · sys/dev/disk/sili/sili_cam.c:289 · pass 2 · new |
none |
not_reproduced |
| DF-4303 |
Medium |
4.7 |
bfe_stop() can deadlock the callout subsystem against bfe_tick (AB-BA): sync callout_stop under the serializer waits forever for a callback blocked on that serializer — interface + one softclock CPU wedged (DF-3715 family)
Base system — drivers · sys/dev/netif/bfe/if_bfe.c:1494 · pass 2 · new |
none |
inconclusive |
| DF-4304 |
Medium |
4.7 |
bfe_rxeof/bfe_txeof completion walks trust the 12-bit DMA current-descriptor field with no range validation and no work bound: encodable off-ring index 511 (byte offsets 4088..4095) makes the interrupt loop non-terminating — serializer-held livelock / KKASSERT panic
Base system — drivers · sys/dev/netif/bfe/if_bfe.c:1064 · pass 2 · new |
none |
inconclusive |
| DF-4310 |
Medium |
4.7 |
ste_stop()/ste_intr STATS_OFLOW issue the synchronous callout_stop() while holding if_serializer while the 1 Hz non-MP ste_stats_update callback must first block in lwkt_serialize_enter on that serializer — AB-BA deadlock (DF-3715 family), plus STOP-override timer resurrection at detach (UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ste/if_ste.c:790 · pass 2 · new |
dos |
inconclusive |
| DF-4314 |
Medium |
4.7 |
GPU-reset vs overlay-off lifetime race: intel_overlay_reset() clears overlay->crtc while an in-flight overlay-off retire still dereferences it — NULL+offset kernel write or GEM_BUG_ON panic
Base system — drivers · sys/dev/drm/i915/intel_overlay.c:373 · pass 2 · new |
none |
not_reproduced |
| DF-4316 |
Medium |
4.7 |
lge_stop() callout_stop-under-serializer AB-BA deadlock: lge_tick() parks on if_serializer every stop-path caller holds (DF-3715 family) — permanent kernel hang + CPU-wide callout starvation
Base system — drivers · sys/dev/netif/lge/if_lge.c:984 · pass 2 · new |
none |
inconclusive |
| DF-4365 |
Medium |
4.7 |
pcn_stop()/pcn_tick() AB-BA callout deadlock: synchronous callout_stop() under the ifnet serializer vs the 1 Hz self-re-arming non-MPSAFE tick parked in blocking lwkt_serialize_enter — DF-3715 family, pcn(4) instance (mechanism PROVEN on guest via protocol replica)
Base system — drivers · sys/dev/netif/pcn/if_pcn.c:879 · pass 2 · new |
none |
inconclusive |
| DF-4369 |
Medium |
4.7 |
hpt_timeout()/os_cmddone() completion race: non-synchronizing callout_stop() frees the PCOMMAND while the timeout handler dereferences it (UAF -> wild ldm_reset_vbus -> panic)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hpt27xx/hpt27xx_osm_bsd.c:411 · pass 2 · new |
none |
not_reproduced |
| DF-4374 |
Medium |
4.7 |
Sysctl read handlers unlock before touching the HAL and never test sc_invalid — racing device_detach frees the HAL object and unmaps the BARs before the oids are removed, giving an unprivileged reader a use-after-free panic
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath/if_ath_sysctl.c:256 · pass 2 · new |
none |
not_reproduced |
| DF-4433 |
Medium |
4.7 |
sln_mii_cmd() busy-waits on SL_MIISTAT_BUSY with no timeout while holding the interface serializer — unprivileged SIOCGIFMEDIA turns a wedged PHY/PCI readback into a permanent serializer-held kernel busy loop
Base system — drivers · sys/dev/netif/sln/if_sln.c:389 · pass 2 · new |
Unkillable kernel busy-loop holding the interface serializer (SIOCGIFMEDIA path) / wedged interrupt thread (media_intr path); permanent network DoS |
untested |
| DF-4461 |
Medium |
4.7 |
amdtemp_gettemp17to19h KASSERTs on mutable CCD-valid hardware state — unprivileged sysctl read or the periodic sensor task panics stock INVARIANTS kernels
Base system — drivers · sys/dev/powermng/amdtemp/amdtemp.c:960 · pass 2 · new |
none |
not_reproduced |
| DF-4507 |
Medium |
4.7 |
ttm_page_alloc: TTM page pools have no memory-pressure reclamation — the vm_lowmem eventhandler is never invoked anywhere in the kernel and register_shrinker does not exist
Base system — drivers · sys/dev/drm/ttm/ttm_page_alloc.c:442 · pass 2 · new |
none |
untested |
| DF-4621 |
Medium |
4.7 |
amdgpu_dm debugfs: OR-combined valid_input accepts invalid DP lane_count; stored unvalidated and consumed by a lane_settings[LANE_COUNT_DP_MAX=4] write loop -> stack OOB write (LATENT: file not compiled)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c:195 · pass 2 · new |
none |
untested |
| DF-4643 |
Medium |
4.7 |
dce_virtual's crtc_irq .set callback invokes the DragonFly hrtimer_cancel emulation, which may sleep (taskqueue_drain) and migrate CPUs, while amdgpu_irq_update holds adev->irq.lock — atomic-context contract violation, display-irq stall/lock-protocol hazard
Base system — drivers · sys/dev/drm/amd/amdgpu/dce_virtual.c:715 · pass 2 · new |
none |
untested |
| DF-4692 |
Medium |
4.7 |
sili_pm: sili_pm_read/sili_pm_write/sili_pm_set_feature dereference unchecked sili_ata_get_xfer() NULL — kernel panic when any PM register access runs against an exhausted ccb pool
Base system — drivers · sys/dev/disk/sili/sili_pm.c:549 · pass 2 · new |
none |
inconclusive |
| DF-4763 |
Medium |
4.7 |
Unchecked M_PREPEND(M_NOWAIT) failure in fwe_as_output() NULL-derefs the already-freed mbuf — guaranteed kernel panic under mbuf pressure
Base system — drivers · sys/dev/netif/fwe/if_fwe.c:506 · pass 2 · new |
none |
inconclusive |
| DF-4874 |
Medium |
4.7 |
afd_detach() in-flight teardown race: UAF of afd_softc/struct disk + shared dscheck/dsgone disk-layer race — twin of reproduced DF-4797 (ad)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-fd.c:134 · pass 2 · new |
none |
untested |
| DF-4915 |
Medium |
4.7 |
DragonFly port removed upstream's in_interrupt() guard from amdgpu_virt_kiq_rreg/wreg: unconditional console kprintf per SR-IOV MMIO access and msleep() from hardirq context on KIQ fence timeout
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_virt.c:162 · pass 2 · new |
none |
not_reproduced |
| DF-5220 |
Medium |
4.7 |
Use-after-free: execlists engine-init failure kfree()s the engine while engine->timeline.link is still on gt.timelines; the unwind's neighbor list_del writes kernel pointers into the freed chunk
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/i915_timeline.c:28 · pass 2 · new |
none |
inconclusive |
| DF-5462 |
Medium |
4.7 |
Lockless cam_extend array: ptopen/ptclose/ptstrategy/ptioctl UAF read races ptctor's cam_extend_set growth (kfree of ea->ps) — kernel pointer used from freed memory
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_pt.c:153 · pass 2 · new |
none |
not_reproduced |
| DF-5594 |
Medium |
4.7 |
devq lifetime is anchored to the sim refcount, not to the peripherals queued on it — devq freed with cam_ed allocation slots still queued (contradicts cam_sim.c's own comment)
Base system — bus / platform / cpu · sys/bus/cam/cam_sim.c:236 · pass 2 · new |
none |
untested |
| DF-2610 |
Low |
4.7 |
icmp6_redirect_input dereferences stale ip6/nd_rd pointers after IP6_EXTHDR_CHECK may reallocate the mbuf (latent UAF read)
Kernel memory corruption (UAF / OOB write) · sys/netinet6/icmp6.c:2160 · pass 2 · new |
none |
not_reproduced |
| DF-2638 |
Low |
4.7 |
hammer2_inode_chain_sync clears RESIZED/MODIFIED before the backend sync and never restores them on error — silent metadata loss
Hammer2 · sys/vfs/hammer2/hammer2_inode.c:1718 · pass 2 · new |
none |
untested |
| DF-2641 |
Low |
4.7 |
Latent unguarded chain->data derefs in xop_nrename/xop_unlink/chain_sync (539 lacks even the error check its siblings have; 670-678/1618 are NULL-offset WRITE variants after modify-succeeds-with-NULL)
Hammer2 · sys/vfs/hammer2/hammer2_xops.c:539 · pass 2 · new |
none |
untested |
| DF-2688 |
Low |
4.7 |
OOM kill block operates on bigproc without p_token and without liveness revalidation (TOCTOU vs concurrent exit)
Kernel memory corruption (UAF / OOB write) · sys/vm/vm_pageout.c:1826 · pass 2 · new |
none |
untested |
| DF-2901 |
Low |
4.7 |
hammer2 bulkfree scan uses kmem_alloc_swapbacked() result unchecked — NULL bitmap reaches bzero(NULL, size) panic in cbinfo_bmap_init; path is unprivileged via the ungated HAMMER2IOC_BULKFREE_SCAN/ASYNC ioctls
Hammer2 · sys/vfs/hammer2/hammer2_bulkfree.c:569 · pass 2 · new |
panic |
not_reproduced |
| DF-2939 |
Low |
4.7 |
dev_pager_dealloc runs cdev_pg_dtor before unlinking the object from dev_pager_object_list (and outside dev_pager_mtx): racing cdev_pager_allocate for the same handle clobbers the fresh object's si_object, making the later old_dev_pager_dtor KKASSERT(dev->si_object) a reachable panic
Base system — virtual memory · sys/vm/device_pager.c:211 · pass 2 · new |
panic |
not_reproduced |
| DF-2951 |
Low |
4.7 |
sysctl_hostname dereferences p->p_ucred up to four times with no p_spin/crhold and not via td_ucred — racy freed-credential read (cratom_proc can crfree the cred between load and deref), wild pr_host[256] copyout; DF-2868-class sibling with a jailed-root trigger
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_mib.c:221 · pass 2 · new |
none |
not_reproduced |
| DF-2992 |
Low |
4.7 |
flush_pagedep_deps: bwrite(inodedep->id_buf) dereferences id_buf after dropping the softdep lock — diverges from the file's own copy-first discipline (softdep_update_inodeblock), micro-window UAF read / NULL deref
Kernel memory corruption (UAF / OOB write) · sys/vfs/ufs/ffs_softdep.c:4655 · pass 2 · new |
panic |
not_reproduced |
| DF-3020 |
Low |
4.7 |
devfs_allocv failure ignored: devfs_spec_open and devfs_vop_nlookupdotdot use the vp unconditionally → vn_unlock(NULL)/NULL-fp panic under vnode exhaustion
Base system — filesystems · sys/vfs/devfs/devfs_vnops.c:535 · pass 2 · new |
none |
not_reproduced |
| DF-3021 |
Low |
4.7 |
devfs_fo_stat dereferences DEVFS_NODE(vp)->d_dev without the NULL guard getattr has → NULL-deref panic when the device node was torn down under an open fd
Base system — filesystems · sys/vfs/devfs/devfs_vnops.c:1432 · pass 2 · new |
none |
not_reproduced |
| DF-3025 |
Low |
4.7 |
tmpfs_nlink LINK_MAX check-then-increment race across different directories can push tn_links past LINK_MAX → KKASSERT panic (INVARIANTS) / nlink contract breach
Base system — filesystems · sys/vfs/tmpfs/tmpfs_vnops.c:1209 · pass 2 · new |
none |
not_reproduced |
| DF-3065 |
Low |
4.7 |
ext2_reallocblks feeds unclamped attacker-controlled pref (crafted i_db[]/indirect values via ext2_blkpref) into dtog() as the cg index — OOB heap reads of e2fs_gd[]/e2fs_maxcluster[]/e2fs_clustersum[] and wild device I/O (root-sysctl-gated)
Kernel memory leaks / uninitialized disclosure · sys/vfs/ext2fs/ext2_alloc.c:302 · pass 2 · new |
none |
not_reproduced |
| DF-3094 |
Low |
4.7 |
extendfile panics on FAT I/O error mid-extend (bread failure of the FAT block hits a KKASSERT instead of returning the error)
Base system — filesystems · sys/vfs/msdosfs/msdosfs_fat.c:400 · pass 2 · new |
none |
not_reproduced |
| DF-3196 |
Low |
4.7 |
nfs_nget_nonblock revalidation restart retains a stale vp across vput(vp); goto loop — any execution of the NFSTOV(np) != vp branch double-vputs the same vnode (refcount underflow → premature vnode recycling / assert panic); the adjacent retry path's own comment states the rule this restart forgot
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_node.c:271 · pass 2 · new |
none |
untested |
| DF-3231 |
Low |
4.7 |
Transaction starters ignore hammer_get_root_volume failure — trans->rootvol==NULL flows into ~25 ondisk derefs and hammer_rel_volume(NULL): kernel NULL-deref panic when root-volume-header reload I/O fails (KKASSERT is the only guard)
Base system — filesystems · sys/vfs/hammer/hammer_transaction.c:53 · pass 2 · new |
panic |
not_reproduced |
| DF-3302 |
Low |
4.7 |
add_mfc() uses mfc_find() result before taking mroute_token — UAF-write race against del_mfc() kfree
Kernel memory corruption (UAF / OOB write) · sys/net/ip_mroute/ip_mroute.c:976 · pass 2 · new |
none |
not_reproduced |
| DF-3435 |
Low |
4.7 |
check_keep_state dereferences an unchecked kmalloc(..., M_NULLOK) result - NULL-page write panic when the M_IPFW3_STATE type limit is reached
Base system — network stack · sys/net/ipfw3_basic/ip_fw3_state.c:319 · pass 2 · new |
none |
not_reproduced |
| DF-3467 |
Low |
4.7 |
check_uid/check_gid dereference an unreferenced, unlocked inpcb on the output path — use-after-free read race against in_pcbdetach()->kfree(inp)
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_layer4/ip_fw3_layer4.c:120 · pass 2 · new |
none |
not_reproduced |
| DF-3492 |
Low |
4.7 |
pppoe_ticker dereferences hook->private/sp->neg unconditionally and races ng_pppoe_disconnect freeing the session — NULL-deref/UAF panic (DF-3488 twin)
Kernel memory corruption (UAF / OOB write) · sys/netgraph/pppoe/ng_pppoe.c:1490 · pass 2 · new |
none |
untested |
| DF-3510 |
Low |
4.7 |
ng_socket: unsynchronized priv->node / priv->datasock reads in ngc_send/ngd_send/ngs_connect/ngs_disconnect/ngs_rcvdata race mtx-protected teardown — NULL-deref/UAF panic
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/socket/ng_socket.c:310 · pass 2 · new |
none |
not_reproduced |
| DF-3517 |
Low |
4.7 |
v1 ng_socket: ngsock/ngpcb lifetime entirely unsynchronized across socket threads, netisr and synchronous-dispatch contexts — NULL-deref/UAF race (DF-3510 ng7 twins, worse: no locks at all, refs-blind kfree)
Kernel memory corruption (UAF / OOB write) · sys/netgraph/socket/ng_socket.c:257 · pass 2 · new |
none |
untested |
| DF-3529 |
Low |
4.7 |
ng_ether control-plane teardown frees priv without synchronizing against BGL-held data-path callbacks (speculative UAF)
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ether/ng_ether.c:313 · pass 2 · new |
none |
untested |
| DF-3570 |
Low |
4.7 |
MP_ReadOtherFunPciEConfigSpace never reads RE_CSIDR (constant 0xffffffff return) -> NULL CMAC bus_space handle dereference at re(4) attach on RTL8168FP/RTL8117, plus PCIe config RMW clobbering on all CSIAR users
Base system — drivers · sys/dev/netif/re/re.c:31163 · pass 2 · new |
none |
not_reproduced |
| DF-3579 |
Low |
4.7 |
NULL-deref panic in iwn_ampdu_tx_done(): TX_DONE on an aggregation queue with no bound session (sc->qid2tap[qid]==NULL)
Base system — drivers · sys/dev/netif/iwn/if_iwn.c:3600 · pass 2 · new |
panic |
untested |
| DF-3581 |
Low |
4.7 |
Non-terminating RX loop in iwn_notif_intr(): 12-bit firmware closed_count vs 64-slot ring
Base system — drivers · sys/dev/netif/iwn/if_iwn.c:3900 · pass 2 · new |
dos |
untested |
| DF-3595 |
Low |
4.7 |
sc_curmac left NULL after failed band-switch/hw-reset while RUNNING: bwn_stop/bwn_detach/bwn_suspend NULL-deref panic
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn.c:4764 · pass 2 · new |
panic |
inconclusive |
| DF-3596 |
Low |
4.7 |
Unbounded device-register polling under BWN_LOCK in bwn_intr_pmq and bwn_core_start: permanent unkillable wlan-subsystem wedge
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn.c:5172 · pass 2 · new |
dos |
inconclusive |
| DF-3603 |
Low |
4.7 |
Unbounded exponential backtracking in hdaa_audio_trace_as_out/trace_dac DAC-assignment search — kernel livelock at attach/reconfig with hostile codec topology
Base system — drivers · sys/dev/sound/pci/hda/hdaa.c:3595 · pass 2 · new |
attach-time livelock with hostile codec |
untested |
| DF-3632 |
Low |
4.7 |
NVM sections parsed at fixed word offsets against firmware-returned lengths (exact-size kmallocs) - heap OOB reads incl. regulatory[0x507] LAR read requiring >=2576 bytes
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwm/if_iwm.c:2251 · pass 2 · new |
none |
inconclusive |
| DF-3655 |
Low |
4.7 |
NULL dereference via WARN_ON(!status->plane_count) immediately after WARN_ON(!status) in amdgpu_dm_atomic_commit_tail()
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm.c:4889 · pass 2 · new |
none |
inconclusive |
| DF-3662 |
Low |
4.7 |
wpi(4): divide-by-zero in wpi_get_power_index() — interpolate() divisor is the difference of two unvalidated int8 EEPROM TX-power sample bytes (wpi twin of DF-3578)
Base system — drivers · sys/dev/netif/wpi/if_wpi.c:3812 · pass 2 · new |
panic |
inconclusive |
| DF-3668 |
Low |
4.7 |
ix_txeof/ix_txgc loop on unvalidated NIC TX-head index: out-of-range write-back → infinite interrupt-context spin (KKASSERT panic on INVARIANTS, soft-lockup otherwise)
Base system — drivers · sys/dev/netif/ix/if_ix.c:2368 · pass 2 · new |
none |
inconclusive |
| DF-3674 |
Low |
4.7 |
mpt_scsi_reply_handler / mpt_scsi_tgt_reply_handler dereference req with no NULL check while mpt_intr dispatches handlers with req=NULL on out-of-range IOC context
Base system — drivers · sys/dev/disk/mpt/mpt_cam.c:2649 · pass 2 · new |
none |
inconclusive |
| DF-3680 |
Low |
4.7 |
igb_txeof/igb_txgc loop on unvalidated NIC TX-head index: out-of-range 32-bit write-back → infinite interrupt-context spin holding tx_serialize (KKASSERT panic on INVARIANTS after one ring wrap; permanent soft-lockup otherwise)
Base system — drivers · sys/dev/netif/igb/if_igb.c:2186 · pass 2 · new |
none |
inconclusive |
| DF-3707 |
Low |
4.7 |
vega10_get_sclk_od/get_mclk_od divide by zero when the golden top clock is 0 (crafted VBIOS) — world-readable sysfs trigger
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:4540 · pass 2 · new |
none |
inconclusive |
| DF-3714 |
Low |
4.7 |
em_tx_collect walks the TX ring toward an unvalidated TDH register value: out-of-range write-back -> all TX buffers freed (DMA-after-free) then KKASSERT panic / permanent serializer-held spin; same class in the 82547 TDT walk
Base system — drivers · sys/dev/netif/em/if_em.c:3284 · pass 2 · new |
none |
inconclusive |
| DF-3771 |
Low |
4.7 |
ahci_port_free() teardown use-after-free: armed per-CCB timeout callouts never cancelled and in-flight CCBs never failed before kfree
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/ahci/ahci.c:997 · pass 2 · new |
none |
not_reproduced |
| DF-3774 |
Low |
4.7 |
Unchecked NULL from ahci_ata_get_xfer() in ahci_set_feature() and ahci_read_log() (ahci.c instances of the DF-2420/DF-2423 class) -> kernel panic
Base system — drivers · sys/dev/disk/ahci/ahci.c:4057 · pass 2 · new |
none |
not_reproduced |
| DF-3810 |
Low |
4.7 |
skl_calc_wrpll_link divides by p0*p1*p2*5 with no zero guard — QDIV_MODE=1 with QDIV_RATIO=0 readback panics the kernel (#DE)
Base system — drivers · sys/dev/drm/i915/intel_ddi.c:1322 · pass 2 · new |
none |
untested |
| DF-3816 |
Low |
4.7 |
dev.mpr.N.dump_reqs sysctl: unlocked walk of live command/chain lists (UAF read) plus fixed 128-byte reads past device-sized frames
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr.c:1904 · pass 2 · new |
none |
inconclusive |
| DF-3831 |
Low |
4.7 |
No CCB timeout handler (callout block #if 0'd, 'XXX Need a timeout handler') plus ignored bus_dmamap error — lost commands wedge the SIM queue forever
Base system — drivers · sys/dev/disk/trm/trm.c:470 · pass 2 · new |
none |
inconclusive |
| DF-3833 |
Low |
4.7 |
Stale completion after bus-reset recovery: trm_Interrupt dereferences NULL pActiveDCB/pActiveSRB (KASSERT only guards INVARIANTS), and trm_SRBdone's going-list unlink walk has no termination
Base system — drivers · sys/dev/disk/trm/trm.c:1339 · pass 2 · new |
none |
inconclusive |
| DF-3842 |
Low |
4.7 |
Firmware ring indices used unmasked: iwi_tx_intr/iwi_rx_intr termination depends on a raw CSR-read index; an out-of-range value → non-terminating interrupt loop holding IWI_LOCK + wlan serializer (hard hang) and txq->queued int underflow
Base system — drivers · sys/dev/netif/iwi/if_iwi.c:1681 · pass 2 · new |
none |
inconclusive |
| DF-3855 |
Low |
4.7 |
ext-TMDS init parser desynchronisation: non-IGP case-0 record missing 'index += 4' — value/terminator bytes re-executed as MMIO commands
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_combios.c:2955 · pass 2 · new |
none |
inconclusive |
| DF-3866 |
Low |
4.7 |
emu_intr_unregister() uses caller-controlled slot index with no bounds check -> OOB read + four-field NULL write cluster at ihandler[-1] (NULLs sc->reg, sc->irq, sc->ih)
Kernel memory leaks / uninitialized disclosure · sys/dev/sound/pci/emu10kx.c:948 · pass 2 · new |
none |
inconclusive |
| DF-3867 |
Low |
4.7 |
emu_intr() NULL-check and indirect call of ihandler[].irq_func are unsynchronized against emu_intr_unregister() -> call through NULL function pointer (panic)
Base system — drivers · sys/dev/sound/pci/emu10kx.c:995 · pass 2 · new |
none |
untested |
| DF-3874 |
Low |
4.7 |
Firmware-controlled list counts drive walks past fixed arrays: pdlist->count>240, ld_count>64, evt count>15 -> heap OOB reads and console disclosure
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mfi/mfi.c:1331 · pass 2 · new |
leak |
inconclusive |
| DF-3877 |
Low |
4.7 |
MFI_DCMD_LD_DELETE post-check assumes the volume still exists: AEN-driven deletion during the DCMD round-trip hits KASSERT('volume dissappeared')/NULL->ld_dev deref
Base system — drivers · sys/dev/raid/mfi/mfi.c:2625 · pass 2 · new |
none |
inconclusive |
| DF-3884 |
Low |
4.7 |
drm_dp_mst_topology_mgr_destroy/suspend never flush tx_work: pending work can run after teardown (DF-3699 family)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dp_mst_topology.c:3254 · pass 2 · new |
race-window NULL panic / teardown UAF |
untested |
| DF-3897 |
Low |
4.7 |
WarpDrive validation accepts NumPhysDisks == 0: guaranteed integer divide-by-zero panic in mpssas_direct_drive_io on the first qualifying READ/WRITE
Base system — drivers · sys/dev/raid/mps/mps_sas.c:2190 · pass 2 · new |
none |
not_reproduced |
| DF-3918 |
Low |
4.7 |
Allocation-failure path in mpr_mapping_allocate_memory kfrees NULL (panic on DragonFly) / stale pointers; mpr_mapping_initialize return ignored by both callers
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_mapping.c:2177 · pass 2 · new |
logic/persistence corruption (3916/3917), error-path panic (3918), firmware-fed overread (3919) |
untested |
| DF-3926 |
Low |
4.7 |
siba_cc_power_init (reached unconditionally from siba_core_resume -> siba_cc_resume) dereferences scc->scc_dev with no NULL check - panic on resume when scan found no ChipCommon core
Base system — drivers · sys/dev/netif/bwn/siba/siba_core.c:1042 · pass 2 · new |
none |
not_reproduced |
| DF-3937 |
Low |
4.7 |
ena_com_admin_destroy() dereferences comp_ctx four lines before its NULL check — OOM-gated NULL-deref panic in the ena(4) attach path
Base system — drivers · sys/dev/virtual/amazon/ena/ena-com/ena_com.c:1581 · pass 2 · new |
none |
inconclusive |
| DF-4033 |
Low |
4.7 |
NULL radeon_encoder dereference family in connector property handlers and monitor-bpc path
Base system — drivers · sys/dev/drm/radeon/radeon_connectors.c:159 · pass 2 · new |
none |
not_reproduced |
| DF-4034 |
Low |
4.7 |
radeon_connector_get_edid dereferences ddc_bus->has_aux with ddc_bus==NULL (unprivileged GETCONNECTOR DoS)
Base system — drivers · sys/dev/drm/radeon/radeon_connectors.c:309 · pass 2 · new |
none |
not_reproduced |
| DF-4035 |
Low |
4.7 |
OOM NULL-deref family: unchecked drm_cvt_mode() results in radeon_tv_get_modes and radeon_add_common_modes
Base system — drivers · sys/dev/drm/radeon/radeon_connectors.c:549 · pass 2 · new |
none |
not_reproduced |
| DF-4040 |
Low |
4.7 |
polaris10 VCE/UVD/CAC level population NULL-derefs when the PowerPlay table omits usMMDependencyTableOffset / usVddcLookupTableOffset
Base system — drivers · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:712 · pass 2 · new |
none |
inconclusive fixed |
| DF-4074 |
Low |
4.7 |
NULL mm_dep_table dereference panics: V1 parser leaves the subtable NULL (usMMDependencyTableOffset==0) and vegam_update_uvd/vce_smc_table deref it unconditionally
Base system — drivers · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:340 · pass 2 · new |
none |
inconclusive |
| DF-4104 |
Low |
4.7 |
LVDS_Info parsed at unvalidated VBIOS data_offset: fixed 0x2C-byte struct reads run past the image-sized BIOS heap buffer (parse_data_header performs no bounds check; atom_context has no bios_size)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/atombios_encoders.c:2081 · pass 2 · new |
none |
untested |
| DF-4111 |
Low |
4.7 |
tl_start ignores tl_encap() failure: mbuf chain leak, NULL-mbuf BPF tap (panic with listener), non-READY list chained into TX
Base system — drivers · sys/dev/netif/tl/if_tl.c:1773 · pass 2 · new |
none |
untested |
| DF-4180 |
Low |
4.7 |
twe_done indexes twe_lookup[TWE_Q_LENGTH=255] with the controller-supplied 8-bit response_id — id 255 reads one pointer past the array (into twe_drive[0]) and dereferences it as a request (wild read + write)
Base system — drivers · sys/dev/raid/twe/twe.c:1199 · pass 2 · new |
none |
inconclusive |
| DF-4198 |
Low |
4.7 |
get_bytes_per_element() returns 0 for dm_444_8/dm_mono_8 and unhandled enum values -> integer divide-by-zero (#DE kernel panic) at get_surf_rq_param():641 for linear surfaces, dead-code #DE at :745, and dml_log2(0)=-127 UB-shift garbage programming
Base system — drivers · sys/dev/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c:44 · pass 2 · new |
dos |
inconclusive |
| DF-4200 |
Low |
4.7 |
NULL function-pointer call: encoder_funcs->detect invoked without check in amdgpu_connector_dp_detect DP-bridge branch — external (Travis/Nutmeg) encoders ship helper funcs without .detect
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_connectors.c:1386 · pass 2 · new |
none |
not_reproduced |
| DF-4201 |
Low |
4.7 |
NULL amdgpu_encoder / enc_priv dereference family after best_encoder() in the two scaling-mode property handlers and amdgpu_connector_get_monitor_bpc (radeon DF-4033 twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_connectors.c:139 · pass 2 · new |
none |
not_reproduced |
| DF-4202 |
Low |
4.7 |
Unchecked drm_cvt_mode() NULL results: mode->type write through NULL in amdgpu_connector_lcd_native_mode (:406) and drm_mode_probed_add(NULL) list_add in amdgpu_connector_add_common_modes (:460) (radeon DF-4035 twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_connectors.c:405 · pass 2 · new |
none |
not_reproduced |
| DF-4214 |
Low |
4.7 |
RAID reply-handler deregistration is not refcounted across mpt(4) controllers: detaching one controller strands the other's RAID actions (request leak → unkillable hang)
Base system — drivers · sys/dev/disk/mpt/mpt_raid.c:135 · pass 2 · new |
none |
untested |
| DF-4215 |
Low |
4.7 |
construct() OOM failure path dereferences NULL current_state / NULL dc->ctx in destruct() → GPU-probe NULL-deref panic instead of graceful fallback
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc.c:519 · pass 2 · new |
none |
untested |
| DF-4252 |
Low |
4.7 |
mprsas_fw_work IR-event early return leaks fw_event allocation and strands the startup refcount (discovery can never complete)
Base system — drivers · sys/dev/raid/mpr/mpr_sas_lsi.c:327 · pass 2 · new |
none |
untested |
| DF-4298 |
Low |
4.7 |
Lock-order inversion: kbd_token acquired before kb_lock in kbd.c cdev/register paths versus kb_lock before kbd_token in the syscons/vkbd ioctl paths (AB-BA deadlock)
Base system — drivers · sys/dev/misc/kbd/kbd.c:802 · pass 2 · new |
none |
untested |
| DF-4313 |
Low |
4.7 |
dcn10 construct() OOM error path: dcn10_pp_smu_create() result never checked — NULL pp_smu dereferenced via dcn_bw_notify_pplib_of_wm_ranges (probe-time panic); dcn10_create_resource_pool leaks the pool allocation on construct failure
Base system — drivers · sys/dev/drm/amd/display/dc/dcn10/dcn10_resource.c:1350 · pass 2 · new |
none |
inconclusive |
| DF-4336 |
Low |
4.7 |
twa CL: tw_cl_reset_ctlr has no entry guard — MPSAFE watchdog callout and passthru-timeout ioctl threads run the reset state machine concurrently
Base system — drivers · sys/dev/raid/twa/tw_cl_io.c:1049 · pass 2 · new |
none |
untested |
| DF-4354 |
Low |
4.7 |
sn_attach() ignores sn_activate() failure: rman_get_cpuid(NULL) guaranteed kernel panic during attach, plus register writes into ISA DMA-controller ports 0x00-0x0F
Base system — drivers · sys/dev/netif/sn/if_sn.c:164 · pass 2 · new |
none |
untested |
| DF-4367 |
Low |
4.7 |
Dispatch entries bypassing hwmgr->smu_lock (pp_get_power_profile_mode, pp_set_clockgating_by_smu, pp_dpm_powergate_gfx/mmhub/acp/sdma) race the smu_lock-protected pp_table reset path that frees the hwmgr backend
Base system — drivers · sys/dev/drm/amd/powerplay/amd_powerplay.c:873 · pass 2 · new |
Unlocked powerplay dispatch entries race the smu_lock-protected pp_table reset that frees the hwmgr backend: NULL-deref panic or freed-slab disclosure to an unprivileged reader. |
untested |
| DF-4385 |
Low |
4.7 |
oce_img_flashing_required bcmp's the image CRC at a fixed ~1.31 MB offset with no fw->datasize check — sibling OOB-read site missed by DF-1573
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/oce/oce_sysctl.c:458 · pass 2 · new |
none |
untested |
| DF-4392 |
Low |
4.7 |
mfi_process_fw_state_chg_isr dereferences sc->mfi_aen_cm before its NULL check - OCR with no AEN outstanding NULL-derefs (and uses an unvalidated pool index)
Base system — drivers · sys/dev/raid/mfi/mfi_tbolt.c:1298 · pass 2 · new |
The OCR completion path kfrees and indexes through mfi_aen_cm BEFORE the NULL guard |
untested |
| DF-4393 |
Low |
4.7 |
Unbounded firmware-status spin and 20-second busy-wait in the OCR path execute in the watchdog callout while holding mfi_io_lock - single stuck firmware wedges softclock and all card I/O
Base system — drivers · sys/dev/raid/mfi/mfi_tbolt.c:1275 · pass 2 · new |
The OCR path busy-waits 20 seconds then spins indefinitely on the RSR bit with no timeout; invoked from the reset holding mfi_io_lock in callout context - a firmware that never clears the bit parks the softclock thread forever with the lock held, permanently stalling all callouts system-wide |
untested |
| DF-4401 |
Low |
4.7 |
Unguarded tws_reset re-entry: unlocked state check races the TWS_RESET_START transition — KASSERT panic (INVARIANTS) or leaked xpt_freeze_simq count freezing the adapter forever (twa sibling DF-4336)
Base system — drivers · sys/dev/raid/tws/tws_cam.c:1203 · pass 2 · new |
dos |
inconclusive |
| DF-4403 |
Low |
4.7 |
Error paths permanently strand reserved request slots and hang CAM ccbs: AEN/param channels die; map-failed ccb is on no queue and can never be completed
Base system — drivers · sys/dev/raid/tws/tws_cam.c:807 · pass 2 · new |
dos |
inconclusive |
| DF-4426 |
Low |
4.7 |
UINT8 loop-variable wraparound in CMI enumeration causes infinite loop while holding the global hp lock (local DoS; also amplifies cmi_order[] overflow)
Base system — drivers · sys/dev/acpica/acpi_hp/acpi_hp.c:1145 · pass 2 · new |
dos |
not_reproduced |
| DF-4431 |
Low |
4.7 |
hubp1_program_tiling: log_2() of gfx9 tiling config uses ilog2, and ilog2(0)==-1 programs NUM_PIPES/NUM_BANKS/etc. to FIELD MAX (GPU-hang class, hw-config-gated)
Base system — drivers · sys/dev/drm/amd/display/dc/dcn10/dcn10_hubp.c:133 · pass 2 · new |
Misprogrammed DCSURF_ADDR_CONFIG (fields set to max) -> display-pipe hang on DCN1 hw with zeroed/garbage GB_ADDR_CONFIG readback |
untested |
| DF-4448 |
Low |
4.7 |
dce120 construct(): ipps/transforms NULL checks index slot [i] while allocation went to slot [j] — on Vega20 with harvested pipes the check trips on a never-written kzalloc slot and kills the entire display pool init (spurious false-failure)
Base system — drivers · sys/dev/drm/amd/display/dc/dce120/dce120_resource.c:1049 · pass 2 · new |
none |
untested |
| DF-4459 |
Low |
4.7 |
drm_gem_init: unchecked mm kzalloc (NULL-deref panic under kernel OOM), partial-failure leaks, and dangling dev->mm_private on the drm_ht_create failure path
Base system — drivers · sys/dev/drm/drm_gem.c:139 · pass 2 · new |
none |
inconclusive |
| DF-4498 |
Low |
4.7 |
sbsh: frames with total length 1-4 bytes cause an unbounded repack loop in encap_frame() - permanent kernel spin holding the interface serializer
Base system — drivers · sys/dev/netif/sbsh/if_sbsh.c:636 · pass 2 · new |
none |
inconclusive |
| DF-4501 |
Low |
4.7 |
radeon_uvd_send_msg NULL-deref panic: radeon_fence_ref(ib.fence) executed when radeon_ib_schedule() failed and left ib.fence NULL
Base system — drivers · sys/dev/drm/radeon/radeon_uvd.c:763 · pass 2 · new |
none |
untested |
| DF-4508 |
Low |
4.7 |
e1000_i210: e1000_release_swfw_sync_i210 acquires the SWSM hardware semaphore in an unbounded empty-body loop on every NVM/PHY release path — a stuck SMBI bit pins the calling thread forever
Base system — drivers · sys/dev/netif/ig_hal/e1000_i210.c:146 · pass 2 · new |
Calling thread pinned in-kernel forever on stuck SMBI/SWSM semaphore (surprise removal, wedged fw); interface wedged under serializer |
untested |
| DF-4509 |
Low |
4.7 |
twa(4): tw_cli_drain_aen_queue() re-queues a firmware-owned request into FREE_Q on ETIMEDOUT — request left dual-linked in BUSY_Q/PENDING_Q + FREE_Q (request handed out twice, internal DMA buffer re-assigned, crossed completions; third site of the DF-4335/DF-4274/DF-1814 class)
Base system — drivers · sys/dev/raid/twa/tw_cl_misc.c:353 · pass 2 · new |
none |
inconclusive |
| DF-4551 |
Low |
4.7 |
als_intr registered before channels exist (and kept after they are freed): IRQ in the attach/detach window runs chn_intr(NULL)/chn_intr(freed) — NULL-deref/UAF panic; als_resource_free releases the I/O region before tearing down the handler
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/als4000.c:648 · pass 2 · new |
none |
not_reproduced |
| DF-4558 |
Low |
4.7 |
MOD_UNLOAD ignores device_delete_child() failure and unconditionally kfrees the external softc while free_softc() can leave the cam_sim registered — post-kldunload use-after-free (lockmgr on freed+uninited state_lock) on the still-reachable SIM
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-cam.c:894 · pass 2 · new |
none |
not_reproduced |
| DF-4575 |
Low |
4.7 |
adw_idle_cmd_send: single RISC command mailbox shared across sim-lock, callout and lock-free ISR contexts without serialization — raced status word panics the kernel (or BDRs the wrong target); any chip wedge during timeout recovery ends in an unconditional panic after a 10-second interrupts-disabled spin
Base system — drivers · sys/dev/disk/advansys/adwlib.c:850 · pass 2 · new |
Kernel DoS: 10s interrupts-disabled spin then unconditional panic on raced mailbox (or wedged chip during timeout recovery); torn param can BDR the wrong target |
untested |
| DF-4602 |
Low |
4.7 |
Unconditional per-ioctl failure diagnostics + full chain rebuild per SETFMT/SPEED/CHANNELS give an unprivileged user an unbounded kernel console print rate; a one-shot guest-wide starvation/livelock was observed during such storms (inconclusive)
Base system — drivers · sys/dev/sound/pcm/feeder_chain.c:150 · pass 2 · new |
dos |
inconclusive |
| DF-4705 |
Low |
4.7 |
ar5210GetDiagState HAL_DIAG_EEREAD dereferences args==NULL when SIOCGATHDIAG omits ATH_DIAG_IN — kernel page-fault panic (DF-4670 family; sibling of DF-1819, distinct deref)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5210/ar5210_misc.c:633 · pass 2 · new |
none |
inconclusive |
| DF-4713 |
Low |
4.7 |
amdgpu_irq: amdgpu_irq_dispatch/amdgpu_irq_update call the optional .process/.set callbacks with no NULL guard while every sibling path validates them — dce_virtual registers a .process==NULL crtc_irq (num_types=7) at LEGACY/229, one stray IV away from a hardirq NULL-call panic
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_irq.c:436 · pass 2 · new |
none demonstrated (hardware-gated: guest has no AMD GPU, amdgpu never binds); realistic ceiling is a hardirq NULL function-pointer call => kernel panic (local DoS) if an IH vector ever reaches a .process-less registered source (dce_virtual crtc_irq, LEGACY/229, num_types=7) or if a num_types>0 .set-less source is ever registered and the GPU-reset resume helper runs |
untested fixed |
| DF-4720 |
Low |
4.7 |
XPT_ABORT/XPT_TERM_IO are decorative stubs and aac_timeout() only logs: timed-out CAM CCBs and their FIBs leak permanently until the command pool is exhausted (controller-wide I/O wedge)
Base system — drivers · sys/dev/raid/aac/aac_cam.c:526 · pass 2 · new |
none |
untested |
| DF-4730 |
Low |
4.7 |
apple_smc_sms_intr() unconditionally reads the ISA int-port via ASMC_INTPORT_READ while MMIO-backend devices have sc->sc_ioport == NULL — NULL resource deref → panic in interrupt context
Base system — drivers · sys/dev/apple/smc/smc_io.c:609 · pass 2 · new |
none |
untested |
| DF-4732 |
Low |
4.7 |
Sleepable lockmgr acquisition from the GPU interrupt dispatch path: DM_IRQ_TABLE_LOCK in amdgpu_dm_irq_immediate_work/amdgpu_dm_irq_schedule_work runs lockmgr(LK_EXCLUSIVE) inside the serialized DRM ithread (DragonFly port divergence)
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c:305 · pass 2 · new |
none |
untested |
| DF-4753 |
Low |
4.7 |
cxm_msp_* audio-decoder path runs with zero serialization: unprivileged ioctls sleep ~0.5s in cxm_msp_wait_for_lock while cxm_detach (kldunload) frees the softc — use-after-free (DF-4419 cxm twin; cxm_open is_opened TOCTOU folded)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/cxm/cxm_msp34xxx.c:588 · pass 2 · new |
panic |
inconclusive |
| DF-4770 |
Low |
4.7 |
intel_dp_mst_atomic_check dereferences drm_atomic_get_new_crtc_state() result without NULL check (latent NULL-deref panic, currently masked by earlier drm-core NULL sinks)
Base system — drivers · sys/dev/drm/i915/intel_dp_mst.c:123 · pass 2 · new |
panic |
untested |
| DF-4787 |
Low |
4.7 |
dal_gpio_service_open calls define_ddc_registers/define_hpd_registers on the create_*() result before the NULL check — deterministic NULL-deref panic when DDC/HPD pin allocation fails
Base system — drivers · sys/dev/drm/amd/display/dc/gpio/gpio_service.c:245 · pass 2 · new |
none |
untested |
| DF-4804 |
Low |
4.7 |
amdsbwd hardware watchdog is never deactivated on reboot/halt (nor on panic by default): shutdown handler #if 0'd and no framework hook fires — default-auto-armed 30s timer hard-resets machines mid-POST/dump and can reset-loop long-boot hosts
Base system — drivers · sys/dev/misc/amdsbwd/amdsbwd.c:112 · pass 2 · new |
dos |
untested |
| DF-4809 |
Low |
4.7 |
ata_ahci_reset() dereferences ch->dma unconditionally: NULL-pointer panic when ata_dmaalloc() failed at channel attach
Base system — drivers · sys/dev/disk/nata/chipsets/ata-ahci.c:466 · pass 2 · new |
none |
untested |
| DF-4821 |
Low |
4.7 |
radeon_ring_backup() uses the hardware rptr as the first ring index without ptr_mask — bogus GPU/writeback value gives a wild kernel OOB read (panic) during GPU reset
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_ring.c:299 · pass 2 · new |
panic |
untested |
| DF-4872 |
Low |
4.7 |
ar9287FillCapabilityInfo never enforces 2GHz-only Kiwi-2 silicon: GMODE divergence re-adds 11A/11NA wireless modes the EEPROM layer explicitly denies, and 5GHz channel edges are inherited — 5GHz channels admitted on 2GHz-only silicon (enabler of the DF-4793 NULL-pier panic)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar9002/ar9287_attach.c:433 · pass 2 · new |
none |
untested |
| DF-4878 |
Low |
4.7 |
dal_i2caux_release_engine(): unsynchronized engine state — engine->ddc NULL-wipe race ends in dal_ddc_close(NULL) NULL-deref panic (DAL2 mutex dropped) + wrong-owner pin close
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/i2caux.c:420 · pass 2 · new |
none |
untested |
| DF-4883 |
Low |
4.7 |
ASSERT-only domain validation in the DC fixed-point library: dc_fixpt_log diverges into an integer 0/0 (#DE, kernel panic) for nonpositive arguments; exp/mul/sqr/from_fraction overflow-and-UB siblings
Base system — drivers · sys/dev/drm/amd/display/dc/basics/fixpt31_32.c:378 · pass 2 · new |
none |
inconclusive |
| DF-4887 |
Low |
4.7 |
cik_ih: unvalidated device-DMA'd IH wptr — DF-4843/DF-4854 twin on the 16-byte CIK grid: a misaligned wptr permanently hard-spins the amdgpu interrupt thread, and the RB_OVERFLOW recovery preserves the bad phase so wptr∈[0xFFE4..0xFFEF]|OVERFLOW parks rptr at [0xFFF4..0xFFFF] → up to 12-byte OOB read past the zero-slack 64 KiB GTT ring BO
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/cik_ih.c:190 · pass 2 · new |
dos |
untested |
| DF-4897 |
Low |
4.7 |
complete_raid5_write() computes and commits RAID-4/5 parity from FAILED (undefined) first-phase reads — garbage parity written to stable storage; the corruption survives retries and later surfaces as silently corrupt recovery reads
Base system — drivers · sys/dev/raid/vinum/vinuminterrupt.c:194 · pass 2 · new |
persistent silent on-disk RAID-4/5 parity corruption from a single failed component read: complete_raid5_write() XORs undefined (non-zeroed malloc) buffer contents of failed reads into the parity block and writes it to stable storage; the corruption survives daemon retries (retry XORs against the poisoned parity) and later surfaces as corrupt reconstructed data with no error indication |
untested |
| DF-4979 |
Low |
4.7 |
SSIF READ_START retry loop retries ENXIO/EBUSY forever — unbounded kthread livelock DoS via stuck/malicious BMC (distinct site from DF-2026)
Base system — drivers · sys/dev/misc/ipmi/ipmi_ssif.c:186 · pass 2 · new |
dos |
untested |
| DF-4981 |
Low |
4.7 |
jmphy_attach 'no media present' branch skips mii_phy_add_media, so miibus_mediainit calls ifmedia_set on an EMPTY media list → DragonFly panic("ifmedia_set") at attach (DF-4930/DF-4957 family)
Base system — drivers · sys/dev/netif/mii_layer/jmphy.c:119 · pass 2 · new |
none |
not_reproduced |
| DF-5020 |
Low |
4.7 |
dp_enable_link_phy reassigns pipes to a possibly-NULL dp_clock_source and calls program_pix_clk through it unguarded — latent NULL function-pointer call on DCE112/DCE120/DCN10 whose resource constructs skip the dp_clock_source NULL check; planted NULL also dereferenced by dp_retrain_link_dp_test
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc_link_hwss.c:65 · pass 2 · new |
kernel UAF/NULL-deref per finding |
untested |
| DF-5028 |
Low |
4.7 |
NULL-pointer dereference of radeon_encoder->enc_priv in dce6_audio_get_pin() encoder walk (dce6_afmt.c:104-105)
Base system — drivers · sys/dev/drm/radeon/dce6_afmt.c:104 · pass 2 · new |
none |
not_reproduced |
| DF-5046 |
Low |
4.7 |
idad_detach frees the idad softc under open descriptors and in-flight bios — DF-5013/DF-5025 family detach UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ida/ida_disk.c:302 · pass 2 · new |
none |
untested |
| DF-5048 |
Low |
4.7 |
Bios submitted to a wedged IDA controller park forever — unkillable D-state issuers and unbounded bio_queue (DF-5026 pattern)
Base system — drivers · sys/dev/raid/ida/ida_disk.c:172 · pass 2 · new |
none |
untested |
| DF-5058 |
Low |
4.7 |
Bios submitted to a never-completing mlx controller park on mlx_bioq forever — unkillable D-state issuers and unbounded bioq growth (DF-5048/DF-5026 family, mlx instance)
Base system — drivers · sys/dev/raid/mlx/mlx_disk.c:156 · pass 2 · new |
none |
untested |
| DF-5115 |
Low |
4.7 |
ips_pci_detach clears sc->configured before ips_adapter_free can return EBUSY: one failed detach permanently corrupts the state machine — retried detaches silently succeed without ips_pci_free, leaving the adapter IRQ handler registered against softc freed at module-unload unit destruction (interrupt-context UAF read/write), and ips_pci_shutdown permanently skips the cache flush
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ips/ips_pci.c:202 · pass 2 · new |
none |
untested |
| DF-5136 |
Low |
4.7 |
vbt_sdvo_panel_type consumed as an unvalidated element index into the VBT BDB_SDVO_PANEL_DTDS table (dvo_timing + index) — root-set OOB read incl. negative offsets
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/i915_params.c:68 · pass 2 · new |
none |
untested |
| DF-5140 |
Low |
4.7 |
iwm_notification_wait_free() frees the wait structure with no drain of threads still sleeping in iwm_wait_notification(); the only guard is a KASSERT compiled out on non-INVARIANTS kernels, so a detach racing an in-flight waiter operates on freed memory
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/iwm/if_iwm_notif_wait.c:108 · pass 2 · new |
stalled control threads / detach-race UAF write |
untested |
| DF-5157 |
Low |
4.7 |
ath LED blink callout is re-armed by draining tx-completion tasks AFTER ath_stop()'s single cancel — no drain in ath_detach, callout fires into freed softc/HAL (use-after-free)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath/if_ath_led.c:166 · pass 2 · new |
none |
untested |
| DF-5212 |
Low |
4.7 |
cxm(4) REMOTE_GETKEY: cxm_ir_key() tsleeps 50 ms inside an unserialized ioctl while cxm_detach frees the softc and destroys sc->iicbus — use-after-free (IR-path twin of DF-4753/DF-3941 family)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/cxm/cxm_ir.c:114 · pass 2 · new |
panic |
inconclusive |
| DF-5280 |
Low |
4.7 |
pci_devq mutated without pci_token (pci_read_device/pci_freecfg) while PCIOCGETCONF walks the list under pci_token — use-after-free read race
Kernel memory corruption (UAF / OOB write) · sys/bus/pci/pci.c:576 · pass 2 · new |
VPD lookup always ENXIO (DF-5279) / devq UAF read race (DF-5280) / boot-time off-by-one (DF-5281) |
untested |
| DF-5283 |
Low |
4.7 |
cpu_mwait_cx_hint() NULL-pointer dereference when deep-sleep inhibition is active on a MONITOR-without-MWAIT-EXT CPU (latent kernel panic)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/machdep.c:1133 · pass 2 · new |
hard hang/panic in idle loop |
untested |
| DF-5323 |
Low |
4.7 |
Second M_WAITOK-sleep-reachable-inside-critical-section site: fw_bus_probe() holds crit_enter() across fw_bus_explore()'s device kmalloc(M_WAITOK|M_ZERO) (distinct from known DF-1021/fw_get_tlabel)
Base system — bus / platform / cpu · sys/bus/firewire/firewire.c:1239 · pass 2 · new |
sleep-inside-critical-section livelock hazard |
untested |
| DF-5341 |
Low |
4.7 |
mptable_map() adopts firmware pap/base_table_length with no 'PCMP' signature, no MPFPS/MPCTH checksum, and no PA-range validation — arbitrary-physical-address kernel read and parse at boot
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/mptable.c:512 · pass 2 · new |
none |
untested |
| DF-5480 |
Low |
4.7 |
EFI system-table / runtime-table ingestion is signature-only: th_crc32 never verified, efi_systbl_phys not range-checked before PHYS_TO_DMAP, st_rt never validated against the RT-mapped regions (firmware-fed ring-0 call targets), memory-map blob ingested without size cross-checks
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/efirt.c:399 · pass 2 · new |
none |
inconclusive |
| DF-5528 |
Low |
4.7 |
fwohci_pci_detach() swallows child-detach failure, force-deletes the firewire child and returns unconditional success — orphaned attached child + live callouts over unmapped BAR/freed DMA rings once a firewire-bus child vetoes detach (latent panic/UAF, lying kldunload)
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwohci_pci.c:350 · pass 2 · new |
none |
untested |
| DF-2844 |
Info |
4.7 |
vnode_pager_haspage() divides by mnt_stat.f_iosize without a zero guard (autofs leaves f_iosize=0)
Base system — virtual memory · sys/vm/vnode_pager.c:259 · pass 2 · new |
none |
untested |
| DF-3563 |
Info |
4.7 |
chacha_encrypt_bytes() leaves final partial-block plaintext in tmp[64] on the kernel stack (wg outbound packet tails via chachapoly)
Base system — crypto · sys/crypto/chacha20/chacha.c:116 · pass 2 · new |
latent stack plaintext residue (wg outbound tails) |
untested |
| DF-3571 |
Info |
4.7 |
Unbounded register busy-wait in re_reset() (MACFG_38/39/50-53) can hard-lock the kernel thread with the ifnet serializer held
Base system — drivers · sys/dev/netif/re/re.c:7678 · pass 2 · new |
hardware-gated thread+serializer hang |
untested |
| DF-3658 |
Info |
4.7 |
XPT_GET/SET_TRAN_SETTINGS index isp_devparam[tgt]/isp_dev_map[tgt] with unchecked CCB target_id (sym DF-3574 / ncr DF-3609 class)
Base system — drivers · sys/dev/disk/isp/isp_freebsd.c:4585 · pass 2 · new |
none |
not_reproduced |
| DF-3948 |
Info |
4.7 |
dce110_enable_per_frame_crtc_position_reset dereferences triggered_crtc_reset.event_source without NULL check (dead path today)
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:2320 · pass 2 · new |
none |
untested |
| DF-4886 |
Info |
4.7 |
phm_start_thermal_controller is the only phm_* dispatcher in hardwaremanager.c without PHM_FUNC_CHECK — unchecked hwmgr->adev (:238), hwmgr->hwmgr_func (:240) and stores through unchecked adev (:249-250): latent kernel NULL-deref (DF-1913/DF-4831 family)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/hardwaremanager.c:234 · pass 2 · new |
panic |
untested |
| DF-5198 |
Info |
4.7 |
fire_saver(): set_video_mode() return value ignored — on failure vid aliases the stale 32KB text window and the 64000-byte blit (repeated every 25 Hz tick) overruns it into 0xC0000+ (DF-5114/DF-5197 twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/fire/fire_saver.c:69 · pass 2 · new |
none |
not_reproduced |
| DF-5304 |
Info |
4.7 |
fwohci_irx_disable/fwohci_itx_disable free DMA buffers after a fixed 1-tick tsleep without confirming the OHCI context halted (device-write UAF window into freed memory)
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwohci.c:1269 · pass 2 · new |
device-driven heap UAF window on non-compliant hardware |
untested |
| DF-5524 |
Info |
4.7 |
Unchecked device_add_child in iicsmb_attach: a failed smbus allocation leaves sc->smbus NULL and the next slave-mode interrupt dereferences it via smbus_intr(NULL,...)
Base system — bus / platform / cpu · sys/bus/iicbus/iicsmb.c:156 · pass 2 · new |
NULL-deref panic on slave interrupt / system-down on stray event |
untested |
| DF-3586 |
Medium |
4.6 |
NULL-pointer dereference in si_dpm_enable when the VBIOS power table has no ATOM_PPLIB_CLASSIFICATION_BOOT state (boot_ps unvalidated)
Base system — drivers · sys/dev/drm/amd/amdgpu/si_dpm.c:6783 · pass 2 · new |
none |
inconclusive |
| DF-3587 |
Medium |
4.6 |
Wild OOB kernel read in btc_find_voltage when voltage-table count==0: entries[(u32)(count-1)] computes a ~31 GB wild index (0xffffffff)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/si_dpm.c:3214 · pass 2 · new |
none |
inconclusive |
| DF-3626 |
Medium |
4.6 |
Unbounded sink-driven go_again retry loop in intel_dp_check_mst_status livelocks the DP hotplug workqueue (MST HPD DoS)
Base system — drivers · sys/dev/drm/i915/intel_dp.c:4294 · pass 2 · new |
none |
inconclusive |
| DF-3654 |
Medium |
4.6 |
hpd_lock pairing asymmetry in handle_hpd_rx_irq(): link-type flip inside handler strands the lock (permanent hang) or releases an unheld lock
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm.c:1209 · pass 2 · new |
none |
inconclusive |
| DF-3682 |
Medium |
4.6 |
Divide-by-zero kernel panic in drm_gtf_mode_complex() via unvalidated secondary-GTF parameters from crafted EDID
Base system — drivers · sys/dev/drm/drm_edid.c:2066 · pass 2 · new |
panic |
not_reproduced |
| DF-3885 |
Medium |
4.6 |
drm_pick_crtcs exponential recursion: (CRTCs+1)^connectors kernel CPU burn holding fb_helper->lock via hub-inflated connector count
Base system — drivers · sys/dev/drm/drm_fb_helper.c:2422 · pass 2 · new |
dos |
inconclusive |
| DF-3888 |
Medium |
4.6 |
num_levels==0 count-1 underflow: kv_dpm_get_sclk (:2938) and kv_set_valid_clock_range (:1791/:1817) read levels[0xFFFFFFFF] ~48 GB out (panic; unprivileged ioctl trigger)
Base system — drivers · sys/dev/drm/amd/amdgpu/kv_dpm.c:2938 · pass 2 · new |
none |
inconclusive |
| DF-3889 |
Medium |
4.6 |
boot_ps NULL dereference in kv_dpm_disable (kv_update_current_ps(adev, NULL) -> kv_get_ps(NULL)) when the VBIOS power table has no BOOT-classified state — panic on suspend/detach
Base system — drivers · sys/dev/drm/amd/amdgpu/kv_dpm.c:1400 · pass 2 · new |
none |
inconclusive |
| DF-3988 |
Medium |
4.6 |
dc_link_dp_set_test_pattern NULL-pointer dereference on inactive pipe (pipes[i].stream->sink unchecked) plus wrong-object blank (pipes[0] instead of matched pipe_ctx) — sink-triggerable panic with no active stream
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc_link_dp.c:2605 · pass 2 · new |
none |
inconclusive |
| DF-4338 |
Medium |
4.6 |
atom_index_iio indexes the 10-entry static atom_iio_len[] with an unvalidated VBIOS opcode byte (OOB read of kernel .data) and the garbage stride drives an unbounded walk past the BIOS image at attach
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/atom.c:1254 · pass 2 · new |
none |
inconclusive |
| DF-4346 |
Medium |
4.6 |
amdgpu atom_index_iio indexes the 10-entry static atom_iio_len[] with an unvalidated VBIOS opcode byte (OOB read of kernel .data) and the garbage stride drives an unbounded walk past the BIOS image at attach (DF-4338's amdgpu twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/atom.c:1287 · pass 2 · new |
none |
inconclusive |
| DF-5248 |
Medium |
4.6 |
run_drain_fifo() unbounded: loop exit depends on device-controlled VLD bit and run_read() failure value 0xffffffff has that bit set - permanent task wedge and device_detach hang
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_run.c:2513 · pass 2 · new |
none |
inconclusive |
| DF-5261 |
Medium |
4.6 |
mtw_drain_fifo() is unbounded and exits only on a device-controlled VLD bit; mtw_read() failure stores 0xffffffff which HAS that bit set - permanent taskqueue wedge with the driver lock held
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_mtw.c:2173 · pass 2 · new |
none |
inconclusive |
| DF-3578 |
Low |
4.6 |
Division by zero in iwn4965_set_txpower(): interpolate() divisor from unvalidated EEPROM channel samples
Base system — drivers · sys/dev/netif/iwn/if_iwn.c:5707 · pass 2 · new |
panic |
inconclusive |
| DF-3777 |
Low |
4.6 |
jme_rxeof() descriptor consistency check admits (nsegs=0, pktlen=0): jme_rxpkt neither reposts the descriptor nor advances jme_rx_cons — unbounded in-handler spin (RX halt + serializer-held CPU softlock) on faulty/malicious controller writeback; the DF-3604-family heap-OOB is otherwise CLOSED by the :2513 cross-check
Base system — drivers · sys/dev/netif/jme/if_jme.c:2511 · pass 2 · new |
none |
inconclusive |
| DF-3815 |
Low |
4.6 |
mpr_resize_queues() divide-by-zero when firmware reports IOCMaxChainSegmentSize == 1
Base system — drivers · sys/dev/raid/mpr/mpr.c:399 · pass 2 · new |
none |
inconclusive |
| DF-4047 |
Low |
4.6 |
NULL command_table_helper callbacks called via BIOS-controlled command revisions on DCE10/11/11.2 (assign_control_parameter/transmitter_bp_to_atom/clock_source_id_to_ref_clk_src) + uninit crev steering in init_transmitter_control
Base system — drivers · sys/dev/drm/amd/display/dc/bios/command_table.c:204 · pass 2 · new |
panic |
inconclusive |
| DF-4078 |
Low |
4.6 |
Unbounded waits and ignored reinit errors: mps_wait_command(...,0) sleeps forever, mps_reinit ignores mps_reregister_events() failure leaving the simq frozen permanently
Base system — drivers · sys/dev/raid/mps/mps.c:353 · pass 2 · new |
none |
inconclusive |
| DF-4160 |
Low |
4.6 |
Unbounded PHY-reset poll in xe_mii_init(): card insertion / ifconfig up hangs the kernel in crit + serializer
Base system — drivers · sys/dev/netif/xe/if_xe.c:1637 · pass 2 · new |
dos |
untested |
| DF-5265 |
Low |
4.6 |
mtw_attach error path before MTW_LOCK jumps to 'detach:' which does MTW_UNLOCK of a never-acquired lock (plus usb_callout_drain of a never-initialized callout) - lockmgr assertion/panic on malformed USB device
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_mtw.c:541 · pass 2 · new |
none |
untested |
| DF-5299 |
Low |
4.6 |
usbd_get_phys() type-confuses the USB controller's newbus ivars as struct pci_devinfo * without any bus-type or NULL check — NULL/wild deref panic or limited kernel-heap leak into the V4L2 bus_info string on non-PCI controllers
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_device.c:2249 · pass 2 · new |
none |
inconclusive |
| DF-3251 |
Medium |
4.5 |
ext2 publish-before-validate (DF-3208 twin): all three post-ext2_ihashins error branches hand concurrent ext2_ihashget waiters half- or fully-initialized VBAD/VNON inodes, bypassing ext2_vinit's 'only unallocated inodes should be VNON' validation
Base system — filesystems · sys/vfs/ext2fs/ext2_ihash.c:1358 · pass 2 · new |
none |
untested |
| DF-4278 |
Medium |
4.5 |
R300_ZB_ZPASS_ADDR reloc patch records nothing and validates nothing: unvalidated u32 offset past the relocation BO becomes the GPU occlusion-counter write target (cross-BO GPU OOB increment-write)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r300.c:1073 · pass 2 · new |
none |
untested |
| DF-3615 |
Low |
4.5 |
Unvalidated hardware-fed ts_antenna indexes ast_ant_tx[8]/sc_ant_tx[8] in ath_tx_update_stats — latent OOB write with function-pointer-table and past-allocation reach (RX twin clamps after field crashes)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath/if_ath.c:4197 · pass 2 · new |
none |
inconclusive |
| DF-3925 |
Low |
4.5 |
hptmv_event_notify uses closed-blob param2 as MV_U8 channel index into sataEvents[8]/VDevices[8] with no bounds check (hardening)
Base system — drivers · sys/dev/raid/hptmv/entry.c:860 · pass 2 · new |
none |
not_reproduced |
| DF-4151 |
Low |
4.5 |
adv_send_scsi_queue computes the per-queue sense DMA bus address from unvalidated firmware-LRAM free_q_head — device-DMA write outside sense_buffers (write-side twin of DF-2412)
Base system — drivers · sys/dev/disk/advansys/advlib.c:1709 · pass 2 · new |
none |
inconclusive |
| DF-4422 |
Low |
4.5 |
Pixel-mode cursor draw functions lack the cursor_base < font_height bound the text path enforces: negative font-buffer index and negative video-memory offset
Kernel memory leaks / uninitialized disclosure · sys/dev/misc/syscons/scvgarndr.c:789 · pass 2 · new |
none |
untested |
| DF-4644 |
Low |
4.5 |
Adapter/vidsw/vidcdevsw array lifecycle races in fb.c: vid_realloc_array frees outside vga_token, tokenless readers, and a vid_register free-slot index race — UAF read/write window and cross-driver type confusion
Kernel memory corruption (UAF / OOB write) · sys/dev/video/fb/fb.c:76 · pass 2 · new |
none |
untested |
| DF-4651 |
Low |
4.5 |
fm801_pci_attach failure paths after mixer_init() leave a live 0666 /dev/mixerN over freed codec+softc — mixer-UAF family, 14th instance (DF-4156/4193/4299/4334/4344/4421/4457/4468/4504/4537/4550/4610)
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/fm801.c:620 · pass 2 · new |
none |
untested |
| DF-4738 |
Low |
4.5 |
cbb_pci_attach() CardBus secondary-bus allocator is a driver-static counter seeded at 2: collision-blind against firmware-assigned bus numbers (config-space aliasing of unrelated devices) and unbounded into the u_int8_t secbus/subbus (256 -> 0 'unconfigured' sentinel wrap)
Base system — drivers · sys/dev/pccard/pccbb/pccbb_pci.c:299 · pass 2 · new |
none |
not_reproduced |
| DF-5388 |
Low |
4.5 |
No bus arbitration across transactions: serializer is dropped during RX waits, so concurrent smbus callers interleave inside one transfer
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:197 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-3428 |
Medium |
4.4 |
generic_netmap_rxsync: 32-bit wrap in the ring->reserved sanitization drives the userspace-release loop index out of [0,lim] — unkillable kernel livelock holding the kring lock
Kernel memory corruption (UAF / OOB write) · sys/net/netmap/netmap_generic.c:641 · pass 2 · new |
none |
build_failed |
| DF-3646 |
Medium |
4.4 |
NULL-pointer dereference of rdev->pm.dpm.boot_ps when the VBIOS power table has no ATOM_PPLIB_CLASSIFICATION_BOOT state
Base system — drivers · sys/dev/drm/radeon/ci_dpm.c:3593 · pass 2 · new |
none |
inconclusive |
| DF-3647 |
Medium |
4.4 |
Integer divide-by-zero in spread-spectrum programming when the VBIOS SS rate is 0 (ci_calculate_mclk_params / ci_calculate_sclk_params)
Base system — drivers · sys/dev/drm/radeon/ci_dpm.c:2876 · pass 2 · new |
none |
inconclusive |
| DF-3648 |
Medium |
4.4 |
Integer divide-by-zero on zero MPLL/SPLL reference clock in ci_calculate_mclk_params (usMemoryReferenceClock stored unchecked)
Base system — drivers · sys/dev/drm/radeon/ci_dpm.c:2872 · pass 2 · new |
none |
inconclusive |
| DF-3651 |
Medium |
4.4 |
NULL-pointer dereference of rdev->pm.dpm.dyn_state.cac_tdp_table when the VBIOS has no PowerTune table: five unchecked sinks in ci_dpm.c
Base system — drivers · sys/dev/drm/radeon/ci_dpm.c:330 · pass 2 · new |
none |
inconclusive |
| DF-3741 |
Medium |
4.4 |
Zero-stride VBIOS record walks never terminate: radeon_atom_init_mc_reg_table and atom_lookup_voltage_object_v3 infinite loops
Base system — drivers · sys/dev/drm/radeon/radeon_atombios.c:4023 · pass 2 · new |
none |
inconclusive |
| DF-3742 |
Medium |
4.4 |
radeon_atom_get_clock_info stores VBIOS reference clocks unchecked — zero reference_freq reaches mode-set and DPM divisions (#DE panic)
Base system — drivers · sys/dev/drm/radeon/radeon_atombios.c:1146 · pass 2 · new |
none |
inconclusive |
| DF-3743 |
Medium |
4.4 |
Spread-spectrum rate parses to zero (raw 0, v2/v3 /=100 truncation, or ppll never assigns rate) and reaches DPM SS divisions (#DE panic)
Base system — drivers · sys/dev/drm/radeon/radeon_atombios.c:1546 · pass 2 · new |
none |
inconclusive |
| DF-3745 |
Medium |
4.4 |
Connector/encoder object-table walks dereference unvalidated VBIOS offsets, accumulate attacker sizes, and over-index ucMuxState[2] — OOB heap reads at connector init
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_atombios.c:537 · pass 2 · new |
none |
inconclusive |
| DF-3746 |
Medium |
4.4 |
VoltageObjectInfo walkers trust self-declared sizes — OOB heap reads, v1 stride bug, and ~130 KB unbounded leakage-array indices
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_atombios.c:3382 · pass 2 · new |
none |
inconclusive |
| DF-3837 |
Medium |
4.4 |
evergreen_ib_parse()/evergreen_vm_packet3_check(): no packet-fits-in-IB validation + fixed-offset payload reads without pkt->count checks → OOB reads up to 5 dwords past the IB sub-allocation + kernel-log disclosure oracle (DF-3592/DF-3764 family, pre-SI VM path)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/evergreen_cs.c:3519 · pass 2 · new |
none |
inconclusive |
| DF-3853 |
Medium |
4.4 |
Unchecked BIOS sysclk reference frequency (0) reaches calc_eng_mem_clock division — kernel divide-by-zero panic (local DoS) on PM transitions
Base system — drivers · sys/dev/drm/radeon/radeon_combios.c:741 · pass 2 · new |
none |
inconclusive |
| DF-3921 |
Medium |
4.4 |
hpt_async is an unconditional panic("Here") registered for AC_LOST_DEVICE - losing a disk (or any rescan of a lost target) panics the kernel
Base system — drivers · sys/dev/raid/hptmv/entry.c:2147 · pass 2 · new |
none |
not_reproduced |
| DF-4224 |
Medium |
4.4 |
Divide-by-zero (kernel panic) in drm_cvt_mode/drm_gtf_mode_complex via unvalidated cmdline @refresh — simple_strtol value wraps through unsigned int into int vfieldrate
Base system — drivers · sys/dev/drm/drm_modes.c:1465 · pass 2 · new |
dos |
untested |
| DF-2833 |
Low |
4.4 |
kern.syncdelay=SYNCER_MAXDELAY (32) accepted by sysctl → same-slot requeue livelock/deadlock of the syncer under sc_token (also kern.retrydelay<=0)
Base system — kernel core (kern) · sys/kern/vfs_sync.c:133 · pass 2 · new |
none |
not_reproduced fixed |
| DF-3266 |
Low |
4.4 |
pf_get_sport() infinite loop (netisr hang) when the NAT pool port range includes port 0 and all candidate ports are occupied
Base system — network stack · sys/net/pf/pf.c:3033 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3575 |
Low |
4.4 |
NULL CDB pointer dereference in sym_action() INQUIRY peek (CAM_CDB_POINTER with cdb_ptr == NULL)
Base system — drivers · sys/dev/disk/sym/sym_hipd.c:7526 · pass 2 · new |
none |
not_reproduced |
| DF-3599 |
Low |
4.4 |
cz_init_cp_jump_table: NULL mec2_fw dereference on Carrizo when the MEC2 firmware request fails
Base system — drivers · sys/dev/drm/amd/amdgpu/gfx_v8_0.c:1170 · pass 2 · new |
none |
inconclusive |
| DF-3825 |
Low |
4.4 |
asr(4) unbounded polled waits: wait-forever ASR_resetIOP retry spin inside a critical section (asr_timeout-reachable), zero-timeout ASR_initOutBound reply poll reachable from the I2OUSRCMD ioctl, deadline-less ASR_queue_s poll, PCATCH-less ioctl wait
Base system — drivers · sys/dev/raid/asr/asr.c:1254 · pass 2 · new |
none |
not_reproduced |
| DF-3841 |
Low |
4.4 |
Firmware-image walks lack bounds: iwi_load_ucode odd-size size_t underflow (1-2 → SIZE_MAX) walks kernel memory past the image under IWI_LOCK until a page fault; iwi_load_firmware command-block walk reads up to 7 bytes past the DMA buffer and emits up to ~2^19 MMIO command blocks from unvalidated len (DF-3573/3591/3631 firmware-trust family)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwi/if_iwi.c:2446 · pass 2 · new |
none |
inconclusive |
| DF-3924 |
Low |
4.4 |
fOsBuildSgl executes panic("physical address unsupported") on CAM_DATA_PHYS passthrough CCBs (reachable debug panic)
Base system — drivers · sys/dev/raid/hptmv/entry.c:2518 · pass 2 · new |
none |
not_reproduced |
| DF-3932 |
Low |
4.4 |
MLYIO_COMMAND size handling: DataTransferLength==INT_MIN negates/wraps into kmalloc(0xFFFFFFFF80000000, M_INTWAIT) — kernel panic (sink PROVEN on guest); plus unbounded wired allocation and ignored dmamap_load failure
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mly/mly.c:2889 · pass 2 · new |
panic |
inconclusive |
| DF-3949 |
Low |
4.4 |
MLX_NEXT_CHILD uses mlx_sysdrive[0] instead of the found [i] -> device_get_unit(NULL) panic after drive 0 is detached
Base system — drivers · sys/dev/raid/mlx/mlx.c:753 · pass 2 · new |
none |
inconclusive |
| DF-4058 |
Low |
4.4 |
hw.etX.tx_intr_nsegs sysctl accepts 0 (no handler) -> integer division by zero panic on the TX hot path (sc_tx / sc_tx_intr_nsegs); the two rx sysctls validate v>0 but the tx one is a raw SYSCTL_ADD_INT
Base system — drivers · sys/dev/netif/et/if_et.c:314 · pass 2 · new |
Privileged local user turns a tuning variable into an instant kernel panic of a production host. |
untested |
| DF-4062 |
Low |
4.4 |
ioctl transfer lengths are accepted unbounded and flow into kmalloc(M_WAITOK|M_ZERO)/copyin/copyout: native au_length is a 64-bit unsigned long (up to 2^64-1), the linux generic path's imax(inlen,outlen) sign-extends a negative int through amr_ioctl_buffer_length, and the linux PASS path allocates the card-supplied ap_data_transfer_length — oversized-allocation panic, unbounded kernel-memory pinning, and a round_page() wrap hazard in the oversized-slab path
Base system — drivers · sys/dev/raid/amr/amr.c:501 · pass 2 · new |
Privileged local user trivially panics/wedges the kernel (oversized-allocation path / round_page wrap) or pins tens of GB of zeroed kernel memory per ioctl; kmalloc happens before any controller access. |
untested |
| DF-4114 |
Low |
4.4 |
TBI-mode nge(4) deletes itself: sc->nge_miibus = dev passed to device_delete_child(dev, dev) — unbounded nge_detach recursion on detach, live-device kobj_delete UAF on attach failure
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/nge/if_nge.c:864 · pass 2 · new |
none |
untested |
| DF-4228 |
Low |
4.4 |
Root-writable debug.cyapa_*_freq sysctls feed an integer divide-by-zero in cyapa_poll_thread — deterministic kernel panic
Base system — drivers · sys/dev/smbus/cyapa/cyapa.c:227 · pass 2 · new |
none |
untested |
| DF-4324 |
Low |
4.4 |
ath_rate_get_static_rix() returns sc_rixmap[ucastrate] without validating the 0xff not-found sentinel — sn->static_rix=255 flows into rt->info[255] out-of-bounds reads (ni_txrate assignment and sample_stats sysctl print)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_rate/sample/sample.c:361 · pass 2 · new |
The comment claims the static rate index is only set on successful lookup, but the function returns the raw sc_rixmap value (memset 0xff, only filled for table rates) - a fixed rate matching a node-negotiated rate yet absent from the table yields static_rix=255 |
untested |
| DF-4351 |
Low |
4.4 |
Unconditional panic() on device-fed completion status bytes: q_status != QS_DONE (fires on legitimate QS_DONE|QS_DISC1) and unhandled host_stat codes (including codes the driver itself enumerates)
Base system — drivers · sys/dev/disk/advansys/advansys.c:1092 · pass 2 · new |
none |
untested |
| DF-4366 |
Low |
4.4 |
[amdgpu_pm.c] Deterministic kernel NULL-deref panic in the pp_power_profile_mode store: strsep returns NULL on un-delimited writes and isspace(*tmp_str) dereferences address 0; buf_cpy[128] also never NUL-terminated
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_pm.c:916 · pass 2 · new |
none |
inconclusive |
| DF-4481 |
Low |
4.4 |
complete_rqe() raw-plex error path never biodone's the user bio nor frees the request — permanent unkillable I/O hang plus locked-buf/pbuf leak per errored plex I/O
Base system — drivers · sys/dev/raid/vinum/vinuminterrupt.c:219 · pass 2 · new |
Permanent unkillable D-state I/O hang per errored raw-plex I/O plus locked-buf/pbuf leak; repeated errors exhaust the pbuf pool system-wide |
untested |
| DF-4554 |
Low |
4.4 |
amdgpu_debugfs: gpr_read VGPR path copies an UNINITIALIZED 4KB kmalloc to userspace when read_wave_vgprs is NULL (gfx_v7/v8) — independent of DF-1692; latent (dead code on DFly)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_debugfs.c:710 · pass 2 · new |
none |
not_reproduced |
| DF-4756 |
Low |
4.4 |
Mailbox handshake busy-waits up to 12 seconds with mdelay() inside the hard-IRQ .process handler — host/PF can stall the guest in interrupt context per message
Base system — drivers · sys/dev/drm/amd/amdgpu/mxgpu_vi.c:318 · pass 2 · new |
dos |
untested |
| DF-4911 |
Low |
4.4 |
dm_dp_aux_transfer discards dc_link_aux_transfer() return on both AUX write paths: failed DPCD/I2C writes reported as full success to the DRM DP core (MST sideband desync on failing/hostile DP hardware)
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c:105 · pass 2 · new |
none |
untested |
| DF-5313 |
Low |
4.4 |
daioctl(DAIOCTRIM) reads its 16-byte argument by direct user-pointer dereference with no copyin: kernel page fault (panic) on unmapped user pointer and user-memory race across the synchronous loop
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_da.c:423 · pass 2 · new |
stale-heap geometry disclosure (DF-5311) / permanent throttle underflow (DF-5312) / root-gated ioctl panic (DF-5313) |
untested |
| DF-5357 |
Low |
4.4 |
pccard_driver_added NULL-deref panic (pf->dev) after a CIS-error card, via kldload of any pccard client driver
Base system — bus / platform / cpu · sys/bus/pccard/pccard.c:1081 · pass 2 · new |
panic |
untested |
| DF-5086 |
Info |
4.4 |
AcpiOsFree forwards NULL to _kfree which panics ("trying to free NULL pointer") — latent NULL-free panic unique to the DragonFly port (Linux/FreeBSD kfree(NULL) are no-ops)
Base system — drivers · sys/dev/acpica/Osd/OsdMemory.c:73 · pass 2 · new |
per finding (panic-instead-of-degrade / latent NULL-free / DDC data corruption) |
untested |
| DF-5482 |
Info |
4.4 |
efi_enter()/efi_leave() are not re-entrancy-safe despite LK_CANRECURSE: nested entry clobbers the global efi_savevm and the outer efi_leave() then installs a NULL vmspace → pmap panic (latent landmine)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/efirt.c:352 · pass 2 · new |
none |
untested |
| DF-5257 |
Medium |
4.3 |
Remote net80211 taskqueue wedge: urtwn_tsf_task_adhoc sleeps hz*ni_intval*5/1000 with attacker-controlled IBSS beacon interval
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_urtwn.c:2414 · pass 2 · new |
deterministic 640KB/cycle kernel heap leak (DF-5255) / maddr list race panic (DF-5256) / remote taskqueue wedge (DF-5257) |
untested |
| DF-3211 |
Low |
4.3 |
md_lookup_swap() reads only the low 32 bits of the 64-bit NFSv3 file size into a signed int — swap files in the 2-4 GiB window produce negative swap_nblks (negative swap vnode size, ~2^64-block swaponvp claim); >=4 GiB sizes silently wrap
Base system — filesystems · sys/vfs/nfs/nfs_mountrpc.c:286 · pass 2 · new |
dos |
untested |
| DF-3751 |
Low |
4.3 |
((struct rt2860_node *)ni)->wcid read from nodes allocated as plain ieee80211_node (missing ic_node_alloc override)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ral/rt2860.c:919 · pass 2 · new |
none |
not_reproduced |
| DF-3752 |
Low |
4.3 |
Uninitialized retrycnt passed to ratectl on TX_FAILURE (amrr consumes it unconditionally)
Base system — drivers · sys/dev/netif/ral/rt2860.c:1129 · pass 2 · new |
none |
not_reproduced |
| DF-4006 |
Low |
4.3 |
fiji VCE/ACP/UVD level population NULL-derefs when the PowerPlay table omits usMMDependencyTableOffset
Base system — drivers · sys/dev/drm/amd/powerplay/smumgr/fiji_smumgr.c:1433 · pass 2 · new |
none |
inconclusive fixed |
| DF-4618 |
Low |
4.3 |
Uninterruptible multi-second busy-wait handshakes (VP0_LOW_SPINTMO=5,000,000 iterations) run under vpo_intr's critical section — malicious parallel-port device can pin a CPU against preemption per I/O
Base system — drivers · sys/dev/disk/vpo/immio.c:52 · pass 2 · new |
Multi-second per-CPU preemption blackout per I/O (missed ticks, soft lockups, uniprocessor system-wide stall); repeatable per command |
untested |
| DF-4881 |
Low |
4.3 |
onoe ath_rate_getxtxrates() stores hardware rate codes into ath_rc_series.rix (rate-table INDEX) fields and clobbers the correct series-0 index — wrong-entry rt->info[] derefs on every multi-rate-retry frame, heap OOB read on 11n HALs (DF-4850 amrr twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_rate/onoe/onoe.c:143 · pass 2 · new |
none |
inconclusive |
| DF-5032 |
Low |
4.3 |
ar5212InvalidGainReadback reads rfBuf[4..5] past the 3-word RF bank 7 of 2413/2316/2317/5413/2425/2417 radios (gain-readback OOB read, UB)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5212/ar5212_rfgain.c:140 · pass 2 · new |
none |
untested |
| DF-5039 |
Low |
4.3 |
os_map_pci_bar(): unchecked bus_alloc_resource() failure → rman_get_virtual(NULL) NULL-deref panic at device probe
Base system — drivers · sys/dev/raid/hpt27xx/hpt27xx_os_bsd.c:113 · pass 2 · new |
probe panic / shutdown-window UAF per finding |
untested |
| DF-3982 |
Medium |
4.2 |
Synchronous callout_stop() in the fwohci hard-interrupt path (mgm_ocb_timeout + CCB timeout sites) sleeps in 'costp' inside a critical section while the callback is INPROG — FireWire IRQ-thread stall plus softclock-wheel starvation cascade (DF-3715 family)
Base system — drivers · sys/dev/disk/sbp/sbp.c:1600 · pass 2 · new |
none |
inconclusive |
| DF-3090 |
Low |
4.2 |
NFSSVC_MNTD path type-confuses any filesystem's mount into struct nfsmount: VROOT check without vfstype check before VFSTONFS(vp->v_mount)
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_syscalls.c:175 · pass 2 · new |
none |
not_reproduced |
| DF-3299 |
Low |
4.2 |
carp_group_demote_adj()/carp_send_ad_all() execute outside netisr0: unsynchronized demotion counter, wrong-CPU callout arming, unlocked carpif_list walk
Kernel memory corruption (UAF / OOB write) · sys/netinet/ip_carp.c:2765 · pass 2 · new |
none |
not_reproduced |
| DF-3721 |
Low |
4.2 |
mxge_dma_test divides by the firmware-returned tick count with no zero check: kernel divide-by-zero panic at attach, at every open, and inside watchdog recovery
Base system — drivers · sys/dev/netif/mxge/if_mxge.c:459 · pass 2 · new |
none |
inconclusive |
| DF-3883 |
Low |
4.2 |
mst_primary NULL dereference race in drm_dp_mst_handle_up_req (send_up_ack_reply without lock or NULL check)
Base system — drivers · sys/dev/drm/drm_dp_mst_topology.c:2401 · pass 2 · new |
race-window NULL panic / teardown UAF |
untested |
| DF-3890 |
Low |
4.2 |
vid_mapping num_entries==0 count-1 underflow in kv_convert_vid2_to_vid7 — entries[0xFFFFFFFF] wild read (~+16 GiB) at boot via kv_construct_max_power_limits_table
Base system — drivers · sys/dev/drm/amd/amdgpu/kv_dpm.c:90 · pass 2 · new |
none |
inconclusive |
| DF-4268 |
Low |
4.2 |
wb_rxeof double-fetch of the device-written RX status word: check/use TOCTOU enabling a ~12KB OOB read copied into received packets (kernel heap info leak)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/wb/if_wb.c:983 · pass 2 · new |
none |
not_reproduced |
| DF-4653 |
Low |
4.2 |
puc_setup_intr publishes ihand before ihandarg: shared-IRQ event during child attach dispatches siointr with a NULL argument -> kernel NULL-deref panic (boot-time DoS)
Base system — drivers · sys/dev/misc/puc/puc.c:691 · pass 2 · new |
none |
not_reproduced |
| DF-4742 |
Low |
4.2 |
vce_v2_0: VCE bo sized exactly AMDGPU_VCE_FIRMWARE_OFFSET (256) bytes short — DATA window ends 256 bytes past vcpu_bo (DF-4524 v3 shape, v2 instance)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/vce_v2_0.c:424 · pass 2 · new |
none |
untested |
| DF-4832 |
Low |
4.2 |
uvd_v1_0_resume programs VCPU cache windows +128B into vcpu_bo at full sizes — STACK/SESSION window overruns the pinned bo by 128 bytes into the adjacent VRAM object (radeon UVD v1 twin of DF-4524/DF-4742)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/uvd_v1_0.c:121 · pass 2 · new |
none |
untested |
| DF-4988 |
Low |
4.2 |
Passthru SGL write destination steered by user-controlled gcmd->size byte: up to ~860 bytes of DMA descriptors written past the request's 1024-byte command packet into neighboring requests of the shared command pool
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws_user.c:126 · pass 2 · new |
none |
not_reproduced |
| DF-5229 |
Low |
4.2 |
uvd_v4_2_resume programs the VCPU cache-window chain +0x200 into vcpu_bo with a full-blob region0 size — chain ends 512 bytes past the pinned bo with the shipping CIK UVD firmware (radeon twin of DF-4832/DF-4524/DF-4742)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/uvd_v4_2.c:47 · pass 2 · new |
none |
untested |
| DF-5364 |
Low |
4.2 |
Unchecked kmalloc(M_NOWAIT) in CISTPL_FUNCID/CISTPL_CONFIG handlers: NULL-page write on allocation failure
Base system — bus / platform / cpu · sys/bus/pccard/pccard_cis.c:790 · pass 2 · new |
panic |
untested |
| DF-5767 |
Low |
4.2 |
ohci_controller_init latches unvalidated device-reported NDP (0-255) into sc_noport: kernel-driven OOB MMIO reads AND writes past the 0x100-byte OHCI register file on every roothub event (DF-5268/5282/5296 register-geometry family, OHCI variant)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/controller/ohci.c:286 · pass 2 · new |
none |
untested |
| DF-4985 |
Info |
4.2 |
dce110 command-table helper leaves five vtable members NULL while six consumer call sites invoke them unguarded — latent NULL-function-pointer call under ASIC/VBIOS generation mismatch (generation-locked, unreachable on genuine hardware)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/dce110/command_table_helper_dce110.c:332 · pass 2 · new |
none |
untested |
| DF-5000 |
Info |
4.2 |
dce80 command-table helper leaves 2 of 17 vtable members NULL (dc_clock_type_to_atom, transmitter_color_depth_to_atom); the SetPixelClock-V7 rev handlers call transmitter_color_depth_to_atom unguarded — latent NULL-function-pointer kernel panic under cross-generation VBIOS pairing on DCE8 silicon
Base system — drivers · sys/dev/drm/amd/display/dc/bios/dce80/command_table_helper_dce80.c:329 · pass 2 · new |
none |
untested |
| DF-3384 |
Medium |
4.1 |
BIOCFLUSH racing an in-flight bpfread() breaks the buffer-slot invariants: unconditional bd_fbuf = bd_hbuf after uiomove() loses the hold-buffer pointer (leak) and later leaves bd_sbuf == NULL so catchpacket() writes the bpf header through NULL
Base system — network stack · sys/net/bpf.c:439 · pass 2 · new |
none |
not_reproduced |
| DF-3470 |
Medium |
4.1 |
Ratectl registry has no in-use accounting: nrefs unload-guard exists but amrr/rssadapt/none never increment it — kldunload of an in-use algorithm leaves every vap->iv_rate dangling into unmapped module text; next ir_* indirect call (world-readable rate_stats sysctl, any TX, vap detach) executes unmapped code
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_ratectl.c:58 · pass 2 · new |
panic |
inconclusive |
| DF-3715 |
Medium |
4.1 |
em_stop() can deadlock the callout subsystem against em_timer/em_82547_move_tail (AB-BA): callout_stop() synchronously waits for a callback that is blocked on the serializer the stopper holds — mechanism PROVEN on the guest kernel via protocol-replica module (caller stuck D3/wchan 'costp')
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/em/if_em.c:2222 · pass 2 · new |
none |
inconclusive |
| DF-3717 |
Medium |
4.1 |
mxge_detach() AB-BA deadlock: synchronous callout_stop() under ifnet_serialize_all vs mxge_tick blocked on main_serialize — mechanism PROVEN on guest via protocol replica (wchan 'costp' + softclock 'slize' wedge)
Base system — drivers · sys/dev/netif/mxge/if_mxge.c:4463 · pass 2 · new |
none |
inconclusive |
| DF-3776 |
Medium |
4.1 |
jme_stop()/jme_miibus_statchg() vs jme_tick() AB-BA deadlock: blocking lwkt_serialize_enter in the self-re-arming CPU0-pinned tick callback vs synchronous callout_stop() under ifnet_serialize_all — permanent kernel wedge + CPU0 softclock callout blackout (DF-3715/3738/3760 family, jme(4) instance)
Base system — drivers · sys/dev/netif/jme/if_jme.c:2571 · pass 2 · new |
none |
inconclusive |
| DF-3806 |
Medium |
4.1 |
msk_stop()/msk_tick() AB-BA callout deadlock: synchronous callout_stop() under ifnet serializer vs blocking lwkt_serialize_enter in the self-re-arming non-MPSAFE tick (DF-3715/3738/3760/3776 family, if_msk.c instance)
Base system — drivers · sys/dev/netif/msk/if_msk.c:2936 · pass 2 · new |
none |
inconclusive |
| DF-3850 |
Medium |
4.1 |
dc_stop()/dc_tick() AB-BA callout deadlock: synchronous callout_stop under the ifnet serializer vs dc_tick's blocking lwkt_serialize_enter — and dc_stat_timer is a NON-MPSAFE callout, so the parked softclock thread also holds the Giant MP lock (DF-3715 family, dc(4) instance, mechanism PROVEN on guest)
Base system — drivers · sys/dev/netif/dc/if_dc.c:2725 · pass 2 · new |
none |
inconclusive |
| DF-3903 |
Medium |
4.1 |
sk_stop()/sk_tick()/sk_yukon_tick() AB-BA callout deadlock: synchronous callout_stop() under the card-global serializer vs blocking lwkt_serialize_enter in the self-re-arming NON-MPSAFE 1 Hz tick (DF-3715 family, 7th instance) — mechanism PROVEN on guest
Base system — drivers · sys/dev/netif/sk/if_sk.c:2782 · pass 2 · new |
none |
inconclusive |
| DF-3906 |
Medium |
4.1 |
xl_stop()/xl_stats_update() AB-BA callout deadlock: synchronous callout_stop() under the ifnet serializer vs blocking lwkt_serialize_enter in the self-re-arming NON-MPSAFE 1 Hz stats tick (DF-3715 family, 10th instance) — mechanism PROVEN on guest
Base system — drivers · sys/dev/netif/xl/if_xl.c:3169 · pass 2 · new |
none |
inconclusive |
| DF-3910 |
Medium |
4.1 |
Divide-by-zero panic in tonga_thermal_setup_fan_table from VBIOS fan temperatures (usTMed==usTMin or usTHigh==usTMed)
Base system — drivers · sys/dev/drm/amd/powerplay/smumgr/tonga_smumgr.c:2502 · pass 2 · new |
none |
inconclusive fixed |
| DF-3953 |
Medium |
4.1 |
boot_ps NULL dereference panic in kv_dpm_enable/kv_dpm_disable (plus kv_dpm_get_sclk ps_priv-NULL window) when VBIOS power table has no BOOT-classified state
Base system — drivers · sys/dev/drm/radeon/kv_dpm.c:1277 · pass 2 · new |
none |
inconclusive |
| DF-3954 |
Medium |
4.1 |
num_levels==0 (VBIOS state with no valid clock entries) -> levels[0xFFFFFFFF] wild read ~48 GiB out in kv_set_valid_clock_range / kv_dpm_get_sclk
Base system — drivers · sys/dev/drm/radeon/kv_dpm.c:1721 · pass 2 · new |
none |
inconclusive |
| DF-4022 |
Medium |
4.1 |
NULL-deref panic in rv770_dpm_enable->rv770_init_smc_table when the VBIOS power table has no BOOT-classified state (or ucStateEntrySize<=1) or dpm.init failed — rv770-family twin of DF-3972/3953/3754/3646/3610
Base system — drivers · sys/dev/drm/radeon/rv770_dpm.c:1901 · pass 2 · new |
none |
inconclusive |
| DF-4025 |
Medium |
4.1 |
re_stop(sc,TRUE) deadlocks against re_tick (callout_stop AB-BA under the driver serializer): permanent D3/'costp' wedge of two threads + CPU-wide callout starvation — re(4) instance of DF-3715(em); mechanism PROVEN on the guest via protocol replica
Base system — drivers · sys/dev/netif/re/if_re.c:1354 · pass 2 · new |
none |
inconclusive |
| DF-4064 |
Medium |
4.1 |
fxp_stop() callout_stop under the ifnet serializer can deadlock the callout subsystem against fxp_tick (AB-BA): non-MPSAFE fxp_stat_timer callback blocks in lwkt_serialize_enter and holds the MP lock while parked — DF-3715 (em) / DF-3738 (emx) family twin, mechanism PROVEN on the guest kernel via fxp-faithful replica
Base system — drivers · sys/dev/netif/fxp/if_fxp.c:399 · pass 2 · new |
none |
inconclusive fixed |
| DF-4109 |
Medium |
4.1 |
tl(4) callout AB-BA deadlock: tl_stats_update blocks on the driver serializer while tl_stop synchronously stops the timer under it (DF-3715 family)
Base system — drivers · sys/dev/netif/tl/if_tl.c:1671 · pass 2 · new |
none |
inconclusive |
| DF-4112 |
Medium |
4.1 |
nge(4) callout AB-BA deadlock: nge_tick blocks on the driver serializer while nge_stop synchronously stops the timer under it (DF-3715 family)
Base system — drivers · sys/dev/netif/nge/if_nge.c:1372 · pass 2 · new |
none |
inconclusive |
| DF-4118 |
Medium |
4.1 |
NULL boot_ps page-fault panic in cypress_dpm_enable when the VBIOS power table has no BOOT-classified state (or dpm.init failed) — cypress twin of DF-4022/DF-3972/DF-3754
Base system — drivers · sys/dev/drm/radeon/cypress_dpm.c:1839 · pass 2 · new |
none |
inconclusive |
| DF-4124 |
Medium |
4.1 |
NULL boot_ps page-fault panic in rv6xx_dpm_enable (:1577) / rv6xx_dpm_disable (:1640) when the VBIOS power table lacks a BOOT-classified state, has ucStateEntrySize<=1, or dpm.init failure was swallowed (rv6xx twin of DF-4022/DF-4118)
Base system — drivers · sys/dev/drm/radeon/rv6xx_dpm.c:1549 · pass 2 · new |
none |
inconclusive |
| DF-4125 |
Medium |
4.1 |
Dangling rdev->pm.dpm.ps after kfree-without-NULL in rv6xx_parse_power_table (:1908-1912) + swallowed rv6xx_dpm_init error -> double free in rv6xx_dpm_fini (:2106); ps_priv[0..i-1] leaked (rv6xx twin of DF-4023/DF-3973/DF-4117)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/rv6xx_dpm.c:1908 · pass 2 · new |
none |
inconclusive |
| DF-4290 |
Medium |
4.1 |
SS getters return success with speed_spectrum_rate == 0 — integer divide-by-zero panic in every clock-spread consumer
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:1248 · pass 2 · new |
none |
inconclusive |
| DF-4291 |
Medium |
4.1 |
atomctrl_get_voltage_evv dereferences hwmgr->dyn_state.vddc_dependency_on_sclk without a NULL check (parser contract allows NULL)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:1131 · pass 2 · new |
none |
inconclusive |
| DF-4626 |
Medium |
4.1 |
intel_dsi_vbt: MIPI sequence execution walks the kmemdup'd VBT blob with no end bound — heap OOB read via i2c/spi/delay/gpio elements and the operation loop
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_dsi_vbt.c:444 · pass 2 · new |
Kernel crash DoS in the panel-enable path (mode_config mutex held); OOB bytes only steer skip offsets - no direct info disclosure |
untested |
| DF-5457 |
Medium |
4.1 |
ichsmb_bread returns the raw slave-controlled block count (up to 255) while writing at most 32 bytes — /dev/smb SMB_BREAD then copies out up to 223 bytes of uninitialized kernel stack
Kernel memory leaks / uninitialized disclosure · sys/bus/smbus/ichsmb/ichsmb.c:424 · pass 2 · new |
none |
not_reproduced fixed |
| DF-2659 |
Low |
4.1 |
NULL column chain during sync: hammer2_inode_chain_and_parent() cannot return NULL (panics at chain->parent first) and synchro.c evaluates chain->bref.modify_tid BEFORE its own if (chain) guard
Hammer2 · sys/vfs/hammer2/hammer2_synchro.c:417 · pass 2 · new |
none |
untested |
| DF-2871 |
Low |
4.1 |
taskqueue_start_threads() swallows lwkt_create failures and always returns 0 (mirror of DF-2850 in subr_taskqueue.c)
Base system — kernel core (kern) · sys/kern/subr_taskqueue.c:594 · pass 2 · new |
none |
not_reproduced |
| DF-2899 |
Low |
4.1 |
fp_mmap maps VCHR vnodes without the revoked-device (v_rdev == NULL) guard its sibling kern_mmap has: NULL-deref panic in vm_mmap
Base system — kernel core (kern) · sys/kern/kern_fp.c:473 · pass 2 · new |
panic |
not_reproduced |
| DF-2967 |
Low |
4.1 |
wdog_disable() self-deadlocks/livelocks the panic path when panic strikes inside wdog_reset_all (WDOG_DISABLE_ON_PANIC)
Base system — kernel core (kern) · sys/kern/kern_wdog.c:96 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3069 |
Low |
4.1 |
Unbounded rterm/rte allocation during stage2 extended-range TERM collection — attacker-controlled kernel-heap exhaustion at mount ('This must fit in memory!' is not enforced)
Base system — filesystems · sys/vfs/hammer/hammer_recover.c:611 · pass 2 · new |
none |
not_reproduced |
| DF-3113 |
Low |
4.1 |
hammer_ioc_add_snapshot / hammer_ioc_del_snapshot return with hmp->snapshot_lock held when hammer_init_cursor fails — permanent snapshot-subsystem lockup on production kernels
Base system — filesystems · sys/vfs/hammer/hammer_ioctl.c:727 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3181 |
Low |
4.1 |
Failed non-forced umount of tmpfs has already truncated every regular file (destructive pre-vflush pass leaves a live mount with destroyed data)
Base system — filesystems · sys/vfs/tmpfs/tmpfs_vfsops.c:284 · pass 2 · new |
none |
not_reproduced |
| DF-3351 |
Low |
4.1 |
fs->backlogged never reset when pipe deletion purges child flowsets: p->sum accounting drift can reach exactly 0 → divide-by-zero panic in WF2Q drain
Base system — network stack · sys/net/dummynet/ip_dummynet.c:555 · pass 2 · new |
panic |
untested |
| DF-3403 |
Low |
4.1 |
slstart dereferences tp->t_slsc and the softc before acquiring any token (races slclose/SLIOCSUNIT)
Base system — network stack · sys/net/sl/if_sl.c:531 · pass 2 · new |
none |
not_reproduced |
| DF-3409 |
Low |
4.1 |
tuncreate() error path publishes a half-built softc via dev->si_drv1 (no rollback on if_alloc failure) — poisoned device node leads to later NULL-deref panic
Base system — network stack · sys/net/tun/if_tun.c:213 · pass 2 · new |
none |
not_reproduced |
| DF-3411 |
Low |
4.1 |
FIOSETOWN/TIOCSPGRP racing pgsigio: tun's async-SIGIO paths dereference a struct sigio that funsetown() frees outside sigio_token (generic UAF primitive, tun manifestation)
Kernel memory corruption (UAF / OOB write) · sys/net/tun/if_tun.c:536 · pass 2 · new |
none |
not_reproduced |
| DF-3437 |
Low |
4.1 |
Unlocked ss_ops dispatch TOCTOU: ieee80211_scan_timeout/iterate/flush check-then-call ss->ss_ops with no com lock while ieee80211_scan_vdetach's scan_detach kfrees ss->ss_priv (never cleared) - scan_age on freed table
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_scan.c:586 · pass 2 · new |
none |
untested |
| DF-3473 |
Low |
4.1 |
wlan_xauth: nrefs is never incremented (NULL ia_attach) so kldunload succeeds while VAPs hold raw iv_auth pointers — use-after-free of module text/data on later auth dispatch
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan_xauth/ieee80211_xauth.c:62 · pass 2 · new |
none |
inconclusive |
| DF-3488 |
Low |
4.1 |
pppoe_ticker NULL deref: queued callout item races ng_pppoe_disconnect freeing the session
Base system — network stack · sys/netgraph7/pppoe/ng_pppoe.c:1788 · pass 2 · new |
none |
not_reproduced |
| DF-3498 |
Low |
4.1 |
ng_fec_addport has no cross-bundle exclusivity: a second fec node can re-enslave an already-bundled interface, saving fec_if_input=ng_fec_input and clobbering IFP2NG — after either node's teardown the member is left with if_input=ng_fec_input over a NULL/freed IFP2NG (permanent RX blackhole or freed-node deref on every frame)
Base system — network stack · sys/netgraph7/ng_fec.c:384 · pass 2 · new |
none |
not_reproduced |
| DF-3509 |
Low |
4.1 |
ng_fec_rmnode vs ng_fec_tick callout race: callout_stop does not wait for an in-flight tick, IFF_RUNNING is never cleared on the rmnode path, and tick re-arms itself -> callout can fire on &priv->fec_timeout after kfree(priv)
Kernel memory corruption (UAF / OOB write) · sys/netgraph/fec/ng_fec.c:614 · pass 2 · new |
none |
not_reproduced |
| DF-3617 |
Low |
4.1 |
ath_intr TDMA/SWBA branch: unchecked TAILQ_FIRST(&ic->ic_vaps) and NULL vap->iv_tdma dereference in hard interrupt context (sc_tdma outlives the TDMA vap)
Base system — drivers · sys/dev/netif/ath/ath/if_ath.c:2186 · pass 2 · new |
none |
inconclusive |
| DF-3711 |
Low |
4.1 |
arcmsr_detach() tears down DMA pool and cdev before deregistering the CAM SIM; MSIX IRQ resources leaked
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/arcmsr/arcmsr.c:4632 · pass 2 · new |
detach-order UAF + IRQ resource leak |
untested |
| DF-3854 |
Low |
4.1 |
NULL-deref: 'rev = RBIOS8(offset)' executed before 'if (offset)' in radeon_combios_external_tmds_setup() IGP branch — panic when rdev->bios == NULL
Base system — drivers · sys/dev/drm/radeon/radeon_combios.c:2889 · pass 2 · new |
none |
inconclusive |
| DF-3870 |
Low |
4.1 |
bw_calcs_init: number_of_dram_channels = vram_width / dram_channel_width truncates to 0 -> #DE divisor family at boot/first modeset
Base system — drivers · sys/dev/drm/amd/display/dc/calcs/dce_calcs.c:2041 · pass 2 · new |
dos |
inconclusive |
| DF-3876 |
Low |
4.1 |
mfi_aen_setup dereferences log_state before the error check whose EBUSY/ENOMEM paths leave it NULL; attach caller discards the error via a comma-operator bug
Base system — drivers · sys/dev/raid/mfi/mfi.c:1020 · pass 2 · new |
none |
inconclusive |
| DF-3898 |
Low |
4.1 |
mps_detach_sas frees sassc/sim after a 30-second bounded wait even if the scanner thread never exited: use-after-free reads and writes from mpssas_scanner_thread
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps_sas.c:820 · pass 2 · new |
none |
not_reproduced |
| DF-3912 |
Low |
4.1 |
Non-terminating reply walk in mpt_recv_handshake_reply when IOC reports MsgLength < 2 (negative loop counter, attach-time livelock holding MPT_LOCK)
Base system — drivers · sys/dev/disk/mpt/mpt.c:1475 · pass 2 · new |
none |
not_reproduced fixed |
| DF-3942 |
Low |
4.1 |
Five spread-spectrum table parsers dereference unchecked GET_IMAGE() results — NULL-page kernel panic at amdgpu/dc attach from a crafted/corrupt VBIOS
Base system — drivers · sys/dev/drm/amd/display/dc/bios/bios_parser.c:1041 · pass 2 · new |
panic |
inconclusive |
| DF-3950 |
Low |
4.1 |
mlx_wait_command(): timeout counter never incremented -> unkillable infinite 1s-sleep loop on a never-completing command
Base system — drivers · sys/dev/raid/mlx/mlx.c:1655 · pass 2 · new |
none |
inconclusive |
| DF-3951 |
Low |
4.1 |
mlx_start() controller-wedge path leaks mlx_busycmds (slot never decremented) -> permanent EBUSY lockout of all controller I/O; WEDGED command's DMA map left loaded
Base system — drivers · sys/dev/raid/mlx/mlx.c:2107 · pass 2 · new |
none |
inconclusive |
| DF-3952 |
Low |
4.1 |
mlx_setup_dmamap() panics when DMA segments exceed firmware me_max_sg, while the buffer tag is hard-coded to MLX_NSEG=64 and MLX_COMMAND passthrough is uncapped
Base system — drivers · sys/dev/raid/mlx/mlx.c:2010 · pass 2 · new |
none |
inconclusive |
| DF-3956 |
Low |
4.1 |
vid_mapping_table num_entries==0 -> entries[0xFFFFFFFF] wild read (~16 GiB) in kv_convert_vid2_to_vid7 / kv_convert_vid7_to_vid2
Base system — drivers · sys/dev/drm/radeon/kv_dpm.c:566 · pass 2 · new |
none |
inconclusive |
| DF-3957 |
Low |
4.1 |
Unguarded enc_priv derefs (NULL-page read AND write) for LCD-supporting digital encoders when radeon_atombios_get_lvds_info returns NULL
Base system — drivers · sys/dev/drm/radeon/atombios_encoders.c:742 · pass 2 · new |
none |
inconclusive |
| DF-3968 |
Low |
4.1 |
rt2661_scan_end and rt2661_enable_tsf_sync dereference TAILQ_FIRST(&ic->ic_vaps) without a NULL check — panic racing last-vap destruction during the scanner's unlocked ic_scan_end call
Base system — drivers · sys/dev/netif/ral/rt2661.c:2747 · pass 2 · new |
none |
not_reproduced |
| DF-3969 |
Low |
4.1 |
rt2560_scan_end and rt2560_enable_tsf_sync dereference the vap without a NULL check — panic racing last-vap destruction during the scanner's unlocked ic_scan_end call (DF-3968 family)
Base system — drivers · sys/dev/netif/ral/rt2560.c:2277 · pass 2 · new |
none |
not_reproduced |
| DF-3997 |
Low |
4.1 |
NULL-pointer dereference in envy24_pci_attach error path when resource allocation fails before cfg initialization
Base system — drivers · sys/dev/sound/pci/envy24.c:2569 · pass 2 · new |
NULL deref (sc->cfg->codec) on the envy24 attach error path when resource allocation fails before envy24_init assigns sc->cfg — panic during device attach. Privileged/physical context only. Mirror of DF-1331 (envy24ht.c) on a distinct trigger path; distinct from DF-1322. |
untested |
| DF-4084 |
Low |
4.1 |
sis_stop uses callout_stop (which does not interlock against an in-progress callback) while sis_tick re-arms itself as its last serialized action: timer resurrects after stop; on detach the late tick dereferences freed softc/miibus/bus resources (use-after-free) — DF-3715 callout-discipline family
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/sis/if_sis.c:1123 · pass 2 · new |
none |
untested |
| DF-4103 |
Low |
4.1 |
Unguarded enc_priv (NULL) dereferences in all DIG setup paths when get_lcd_info()/get_dig_info() fail — NULL-page read at amdgpu attach and NULL-page WRITE on DPMS-ON (DF-3957 amdgpu twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/atombios_encoders.c:648 · pass 2 · new |
none |
untested |
| DF-4150 |
Low |
4.1 |
NULL deref of adev->pdev->bus->self in cik_pcie_gen3_enable/cik_program_aspm on root-bus CIK GPUs (upstream pci_is_root_bus guard #if 0'd out)
Base system — drivers · sys/dev/drm/amd/amdgpu/cik.c:1379 · pass 2 · new |
NULL deref of adev->pdev->bus->self in cik_pcie_gen3_enable/cik_program_aspm on root-bus CIK GPUs (upstream pci_is_root_bus guard #if 0'd out in the DFly port). Boot/attach DoS on server platforms / VM GPU passthrough topologies. |
untested |
| DF-4170 |
Low |
4.1 |
smu8_upload_pptable_to_smu NULL-derefs the four optional dependency tables (vdd_gfx/acp/uvd/vce) at driver load
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/smu8_hwmgr.c:443 · pass 2 · new |
none |
untested |
| DF-4186 |
Low |
4.1 |
radeon_device_init ignores radeon_doorbell_init() failure: NULL doorbell aperture WDOORBELL32 write (kernel panic) on CIK+ when BAR2 ioremap fails
Base system — drivers · sys/dev/drm/radeon/radeon_device.c:1422 · pass 2 · new |
none |
not_reproduced |
| DF-4212 |
Low |
4.1 |
mpt_raid_attach error paths: MPT_UNLOCK of a never-acquired lock (deterministic lockmgr panic) plus csa CCB leak
Base system — drivers · sys/dev/disk/mpt/mpt_raid.c:283 · pass 2 · new |
none |
untested |
| DF-4222 |
Low |
4.1 |
XPT_ABORT handler dereferences the ccb hcb back-pointer without a NULL check — aborting a completed/never-issued CCB panics
Base system — drivers · sys/dev/virtual/vmware/pvscsi/pvscsi.c:1395 · pass 2 · new |
none |
inconclusive |
| DF-4258 |
Low |
4.1 |
Unbounded PHY-reset busy-wait in my_autoneg_xmit — the driver's only hardware-wait loop without MY_TIMEOUT (spins forever in ioctl, boot, and ISR contexts holding ifp->if_serializer)
Base system — drivers · sys/dev/netif/my/if_my.c:377 · pass 2 · new |
none |
inconclusive |
| DF-4271 |
Low |
4.1 |
callout_stop() does not interlock with the self-serializing, self-re-arming wb_tick: repeating post-detach use-after-free of the softc
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/wb/if_wb.c:1590 · pass 2 · new |
none |
untested |
| DF-4277 |
Low |
4.1 |
cbb_resume() error path NULLs irq_res/base_res and returns ENOMEM but leaves the driver attached with stale intrhand — subsequent cbb_detach/cbb_suspend dereference the NULL resources (kernel panic)
Base system — drivers · sys/dev/pccard/pccbb/pccbb.c:1574 · pass 2 · new |
cbb_resume() error path NULLs irq_res/base_res and returns ENOMEM but leaves the driver attached with stale intrhand and a live event thread - subsequent cbb_detach/cbb_suspend dereference the NULL resources (kernel panic in rman/bus paths). |
untested |
| DF-4285 |
Low |
4.1 |
midi_init() error paths: lockuninit() on held locks (KKASSERT panic) and m->synth leaked on every destruction path
Base system — drivers · sys/dev/sound/midi/midi.c:371 · pass 2 · new |
attach-time kernel panic (KKASSERT in lockuninit on held lock) when MPU_INIT fails; permanent small heap leak of m->synth on all init-error and destroy paths |
inconclusive |
| DF-4292 |
Low |
4.1 |
atomctrl_get_smc_sclk_range_table dereferences the SMU_Info table pointer without a NULL check (companion guard to DF-1496)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:1374 · pass 2 · new |
none |
inconclusive |
| DF-4293 |
Low |
4.1 |
mixer_reinit and mixer_hwvol_* dereference the mixer cdev pointer with no NULL guards — NULL-deref panic on resume/reinit of a mixer-less pcm device
Base system — drivers · sys/dev/sound/pcm/mixer.c:828 · pass 2 · new |
none |
inconclusive |
| DF-4295 |
Low |
4.1 |
musycc_intr0 SHT event path dereferences sc->chan[ch] without the NULL check all sibling paths perform
Base system — drivers · sys/dev/misc/musycc/musycc.c:820 · pass 2 · new |
Interrupt-context NULL-deref write panic on stray MUSYCC event for an unconfigured channel. |
untested |
| DF-4296 |
Low |
4.1 |
musycc_attach error-path asymmetry leaves half-initialized csoftc on the global list; sibling-function attach then NULL-derefs
Base system — drivers · sys/dev/misc/musycc/musycc.c:1455 · pass 2 · new |
Boot-time NULL-deref write panic on a dual-function card whose f0 BAR/IRQ allocation fails. |
untested |
| DF-4301 |
Low |
4.1 |
vkbd_term() ignores kbd_unregister() failure and frees a possibly still-registered/busy keyboard (latent UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/vkbd/vkbd.c:848 · pass 2 · new |
none |
untested |
| DF-4333 |
Low |
4.1 |
hpt_add_disk_to_array divides by (bArnMember-1) unchecked - #DE panic on degenerate 1-member RAID1/5 arrays (sibling site of DF-1682)
Base system — drivers · sys/dev/raid/hptmv/gui_lib.c:935 · pass 2 · new |
none |
inconclusive |
| DF-4368 |
Low |
4.1 |
pp_get_current_clocks dereferences hwmgr->current_ps without the NULL guard its siblings have — NULL-deref panic after a failed pp_table reset, triggered via unprivileged DC modeset
Base system — drivers · sys/dev/drm/amd/powerplay/amd_powerplay.c:1051 · pass 2 · new |
pp_get_current_clocks NULL-deref panic after a failed (non-overflowing) pp_table reset leaves current_ps freed-and-NULL while dpm_enabled stays true; next unpriv modeset panics. |
untested |
| DF-4391 |
Low |
4.1 |
mfi_tbolt_alloc_cmd never checks request_desc_pool for NULL - failed kmalloc leads to near-NULL memset write on first command build
Base system — drivers · sys/dev/raid/mfi/mfi_tbolt.c:455 · pass 2 · new |
The alloc validates only the second kmalloc; if request_desc_pool fails while the pool-array succeeds, the setup loop stores cmd->request_desc = NULL + i and the first command issue memsets at address 16*index - a write into the unmapped low page, panicking the boot |
untested |
| DF-4394 |
Low |
4.1 |
hpt_shutdown_vbus frees LDM memory and adapter handles before terminating the vbus timer callout - pending timer callback runs on freed memory (shutdown-window UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptrr/hptrr_osm_bsd.c:363 · pass 2 · new |
In the shutdown path the vbus lock is dropped, LDM command pools are freed and adapter handles kfree'd, but the periodic vbus timer (callout_init_lk'd to the vbus lock, armed for background RAID/rebuild tasks) is only terminated AFTERWARDS |
untested |
| DF-4402 |
Low |
4.1 |
tws_cam_attach error symmetry: missing return after cam_sim_alloc failure -> xpt_bus_register(NULL) NULL-deref panic at attach; sc->sim left dangling/registered on later failure paths
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws_cam.c:173 · pass 2 · new |
dos |
inconclusive |
| DF-4408 |
Low |
4.1 |
acpi_asus_probe dereferences the unchecked INIT evaluation result: NULL Obj or non-String object -> NULL/wild pointer panic at probe
Base system — drivers · sys/dev/acpica/acpi_asus/acpi_asus.c:579 · pass 2 · new |
dos |
untested |
| DF-4409 |
Low |
4.1 |
acpi_asus_detach removes the wrong ACPI notify handler: installed handler survives detach and fires into freed/unloaded driver state (use-after-free)
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_asus/acpi_asus.c:810 · pass 2 · new |
dos |
untested |
| DF-4410 |
Low |
4.1 |
acpi_asus LED state changes run in an undrained AcpiOsExecute task: detach/led_destroy free the softc under a pending acpi_asus_led_task (UAF window)
Kernel memory corruption (UAF / OOB write) · sys/dev/acpi_asus/acpi_asus.c:922 · pass 2 · new |
dos |
untested |
| DF-4440 |
Low |
4.1 |
atmel_mxtread/write/kqfilter dereference softc without the NULL guard used by open/close/ioctl — NULL-deref (later UAF) via fd held across driver detach/kldunload
Kernel memory corruption (UAF / OOB write) · sys/dev/smbus/atmel_mxt/atmel_mxt.c:653 · pass 2 · new |
Deterministic NULL-deref panic (later UAF) via fd held across driver detach/kldunload; root-triggered self-DoS |
untested |
| DF-4453 |
Low |
4.1 |
RVDA raw-VBT mapping: pmap_mapbios() of unvalidated firmware-supplied 64-bit address / 32-bit length, re-leaked every resume (amplifies DF-1629; panic-capable KVA exhaustion)
Base system — drivers · sys/dev/drm/i915/intel_opregion.c:1029 · pass 2 · new |
none |
untested |
| DF-4528 |
Low |
4.1 |
Admin submission queue lock-protocol violation: every admin-queue caller submits without queue->lk, racing subq_tail/unsubmitted/ksubq[] against the admin thread
Base system — drivers · sys/dev/disk/nvme/nvme.c:446 · pass 2 · new |
none |
not_reproduced |
| DF-4538 |
Low |
4.1 |
cs4281_pci_detach() releases the BA0/BA1 register mappings before bus_teardown_intr - live ISR reads released mappings (shared-IRQ use-after-release window)
Base system — drivers · sys/dev/sound/pci/cs4281.c:893 · pass 2 · new |
none |
untested |
| DF-4539 |
Low |
4.1 |
cs4281 attach ignores mixer_init()/pcm_addchan() failures; half-attached device NULL-derefs in cs4281_pci_resume (ch->parent NULL and mixer_dev NULL)
Base system — drivers · sys/dev/sound/pci/cs4281.c:845 · pass 2 · new |
none |
untested |
| DF-4561 |
Low |
4.1 |
mpssas_ir_shutdown scans [start_idx, end_idx) over an inclusive IR map range: the last volume slot is never checked, so its SYSTEM_SHUTDOWN_INITIATED action (cache flush) is skipped
Base system — drivers · sys/dev/raid/mps/mps_sas_lsi.c:902 · pass 2 · new |
none |
untested |
| DF-4570 |
Low |
4.1 |
cancel_fake_irq() drains fake_irq before hangcheck, so a concurrently-running hangcheck callback can re-arm fake_irq after its drain — intel_breadcrumbs_fake_irq fires on a torn-down engine (use-after-free at driver unload)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_breadcrumbs.c:840 · pass 2 · new |
UAF read/write on freed engine->breadcrumbs at driver unload (panic-class; worse on heap reuse) |
untested |
| DF-4583 |
Low |
4.1 |
VBIOS-derived zero clocks reach dfixed_div as divisors in the rs690 watermark math: igp_sideport_mclk == 0 (no zero fallback) and ulDefaultEngineClock == 0 (pm.sclk == 0) → integer-divide-fault panic at boot/first bandwidth update
Base system — drivers · sys/dev/drm/radeon/rs690.c:371 · pass 2 · new |
panic |
inconclusive |
| DF-4590 |
Low |
4.1 |
tws_detach/attach-unwind never cancel driver callouts (reset_cb, reinit poll, per-req timeouts) before freeing the softc — use-after-free panic at driver unload during reset recovery
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws.c:346 · pass 2 · new |
none |
not_reproduced |
| DF-4627 |
Low |
4.1 |
mipi_exec_delay: VBT-controlled u32 delay becomes a kernel BUSY-WAIT of up to ~71.6 minutes per element (arbitrarily many elements)
Base system — drivers · sys/dev/drm/i915/intel_dsi_vbt.c:194 · pass 2 · new |
Display subsystem hang + CPU burn up to ~71.6 min per delay element, indefinitely across elements, holding KMS locks so all DRM clients block |
untested |
| DF-4656 |
Low |
4.1 |
Bounce-buffer kmalloc failure path drops the ccb (xpt_done never called) — CAM devq wedge under memory pressure
Base system — drivers · sys/dev/raid/twa/tw_osl_cam.c:336 · pass 2 · new |
Each occurrence permanently strands one CAM_SIM_QUEUED ccb / devq slot; repeated occurrences wedge all I/O to the controller until reboot |
untested |
| DF-4674 |
Low |
4.1 |
bktr_attach fail-path releases the memory BAR with resource type SYS_RES_IRQ — resource_list_release() panics ('can't find resource') on any attach failure after the BAR was mapped
Base system — drivers · sys/dev/video/bktr/bktr_os.c:331 · pass 2 · new |
bktr_attach fail-path releases the memory BAR with SYS_RES_IRQ type - resource_list_release panics on any attach failure after BAR mapping (availability only) |
untested |
| DF-4676 |
Low |
4.1 |
cmx_kqfilter is the only entry point without the sc==NULL||sc->dying guard — kevent() on a still-open fd after card detachment dereferences the NULL softc (kernel NULL-deref panic)
Base system — drivers · sys/dev/misc/cmx/cmx.c:663 · pass 2 · new |
cmx_kqfilter is the only entry without the sc/dying guard - kevent after card detach dereferences the NULL softc (panic). Root-only device + physical removal |
untested |
| DF-4706 |
Low |
4.1 |
amr_pci_attach registers amr_pci_intr before the state it consumes exists: a shared-IRQ-line interrupt in the attach window dispatches through the still-NULL sc->amr_get_work -> boot/kldload panic (setup-side mirror of DF-1820)
Base system — drivers · sys/dev/raid/amr/amr_pci.c:257 · pass 2 · new |
none |
inconclusive |
| DF-4718 |
Low |
4.1 |
aac_cam_fix_inquiry dereferences csio->data_ptr on successful INQUIRY completions without checking a data buffer exists (NULL-deref panic via no-data pass-through INQUIRY)
Base system — drivers · sys/dev/raid/aac/aac_cam.c:551 · pass 2 · new |
none |
untested |
| DF-4729 |
Low |
4.1 |
guc_ctl_log_params_flags() NULL-derefs guc->log.vma: the port compiled out intel_guc_log_create() but left its consumer live — instant kernel panic whenever GuC is enabled (enable_guc!=0) on Gen9+
Base system — drivers · sys/dev/drm/i915/intel_guc.c:294 · pass 2 · new |
none |
untested |
| DF-4734 |
Low |
4.1 |
ar5210ResetTxQueue dereferences ah_curchan (IEEE80211_IS_CHAN_TURBO) without the NULL guard both sibling HALs carry ("NB: chan can be NULL during attach")
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5210/ar5210_xmit.c:157 · pass 2 · new |
none demonstrated (hardware-gated: no Atheros NIC on guest; latent NULL-deref kernel panic if a DATA queue is reset before the first channel set) |
untested fixed |
| DF-4744 |
Low |
4.1 |
ahci_pci_detach() has no interlock against an in-flight ahci_intr() on another CPU — hardirq-context UAF of sc_ports[]/ap and access to the already-unmapped ABAR during controller detach
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/ahci/ahci_attach.c:589 · pass 2 · new |
none |
untested |
| DF-4755 |
Low |
4.1 |
VI mailbox RCV interrupt enabled before flr_work is initialized — host-timed mailbox interrupt executes schedule_work() on an uninitialized (zeroed) work_struct → NULL function-pointer call → kernel panic
Base system — drivers · sys/dev/drm/amd/amdgpu/mxgpu_vi.c:600 · pass 2 · new |
panic |
untested |
| DF-4793 |
Low |
4.1 |
ar9287SetPowerCalTable: non-2GHz channel leaves pCalBChans=NULL/numPiers=0, unguarded per-chain walk NULL-derefs pList[0] in ath_ee_getLowerUpperIndex (kernel panic)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar9002/ar9287_reset.c:68 · pass 2 · new |
panic |
untested |
| DF-4802 |
Low |
4.1 |
initsd() verify loop retries a mismatched block forever (verified=0 with no bound): unbounded uninterruptible kernel loop in the ioctl thread
Base system — drivers · sys/dev/raid/vinum/vinumrevive.c:526 · pass 2 · new |
none |
not_reproduced |
| DF-4805 |
Low |
4.1 |
Disable path programs count=0 + reload while leaving AMDSB_WD_RUN armed — kern.watchdog.period=0 (and panic-path wdog_disable) never actually disarm the watchdog; a zero count under RUN may act as a 1-tick fuse per SB RRG countdown semantics
Base system — drivers · sys/dev/misc/amdsbwd/amdsbwd.c:220 · pass 2 · new |
dos |
untested |
| DF-4818 |
Low |
4.1 |
nvme(4): unchecked MSI-X vector IRQ resource allocation (and missing pci_teardown_msix on fallback) leads to boot-time kernel panics
Base system — drivers · sys/dev/disk/nvme/nvme_attach.c:166 · pass 2 · new |
panic |
inconclusive |
| DF-4841 |
Low |
4.1 |
Post-reset OBFL refill is a no-op: sense_bufs[].posted survives the controller soft reset, so tws_init_obfl_q() re-run by the recovery path posts nothing — firmware left with zero response frames, permanent I/O hang after any successful reset recovery
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/tws/tws_hdm.c:118 · pass 2 · new |
none |
not_reproduced |
| DF-4923 |
Low |
4.1 |
tlphy_attach NULL-deref: strcmp(device_get_name(sibling), "tlphy") panics on not-yet-probed miibus sibling
Base system — drivers · sys/dev/netif/mii_layer/tlphy.c:175 · pass 2 · new |
none |
not_reproduced |
| DF-4950 |
Low |
4.1 |
ahci_os_stop_port(): unbounded thread-join with no escape converts any port-thread wedge into a permanent unkillable kernel hang at detach/shutdown, and tears down only 1 of the 3 per-port locks
Base system — drivers · sys/dev/disk/ahci/ahci_dragonfly.c:248 · pass 2 · new |
dos |
untested |
| DF-4975 |
Low |
4.1 |
AI (Vega20-class) mailbox RCV interrupt enabled before flr_work is initialized — host-timed IDH_FLR_NOTIFICATION executes schedule_work() on a zeroed work_struct → NULL function-pointer call → kernel panic (DF-4755 twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/mxgpu_ai.c:351 · pass 2 · new |
none |
untested |
| DF-4976 |
Low |
4.1 |
flr_work never drained at mailbox teardown — xgpu_ai_mailbox_put_irq() lacks cancel_work_sync(), so a queued FLR work survives put_irq and uses amdgpu_device after hw_fini/free (UAF at driver detach)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/mxgpu_ai.c:365 · pass 2 · new |
none |
untested |
| DF-4986 |
Low |
4.1 |
mfi_pci_attach failure unwind calls mfi_free() → callout_terminate() on a never-initialized watchdog callout → KKASSERT kernel panic (stock INVARIANTS GENERIC)
Base system — drivers · sys/dev/raid/mfi/mfi_pci.c:230 · pass 2 · new |
panic |
not_reproduced |
| DF-4990 |
Low |
4.1 |
SMB_TRANS copies out driver-updated rcount without re-clamping to SMB_MAXBLOCKSIZE — latent kernel stack OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/smbus/smb/smb.c:333 · pass 2 · new |
infoleak |
untested |
| DF-4994 |
Low |
4.1 |
vega10_smu_init error path kfrees hwmgr->smu_backend without NULLing it, defeating vega10_smu_fini's NULL guard — latent double-free/UAF (the vega12 DF-4951 shape)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vega10_smumgr.c:288 · pass 2 · new |
corruption |
untested |
| DF-5006 |
Low |
4.1 |
acpi_cpu has no detach teardown: ACPI notify handler (context=sc), both sysctl contexts and the sensordev outlive the softc → use-after-free on driver/module detach; attach error paths leak the same resources and orphan the first child
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_cpu.c:67 · pass 2 · new |
none |
untested |
| DF-5047 |
Low |
4.1 |
idad_dump reports fake success for controller-failed/rejected dump writes — ida_command discards the controller error status (DF-5027 pattern)
Base system — drivers · sys/dev/raid/ida/ida_disk.c:204 · pass 2 · new |
none |
untested |
| DF-5085 |
Low |
4.1 |
AcpiOsMapMemory's NULL failure contract is dead code — pmap_mapdev panics on KVA exhaustion instead of returning NULL, converting ACPICA's graceful AE_NO_MEMORY degradation into an unrecoverable kernel panic with firmware/AML-controlled mapping volume
Base system — drivers · sys/dev/acpica/Osd/OsdMemory.c:91 · pass 2 · new |
per finding (panic-instead-of-degrade / latent NULL-free / DDC data corruption) |
untested |
| DF-5093 |
Low |
4.1 |
wbsio_detach never detaches its lm child (no bus_generic_detach): driver unload leaves a dangling attached child with a live 5s sensor task, leaked rman ports, and a NULL-ops parent — any later bus op on the child is a guaranteed NULL-deref panic
Kernel memory corruption (UAF / OOB write) · sys/dev/powermng/wbsio/wbsio.c:268 · pass 2 · new |
guaranteed NULL-deref panic on later child bus op (root-gated) |
untested |
| DF-5121 |
Low |
4.1 |
MMalloc trace-table exhaustion (>=16384 outstanding allocations) makes Malloc() return NULL in VINUMDEBUG builds, breaking the never-fail M_WAITOK contract the vinum family relies on: unchecked callers NULL-deref (panic) and expand_table's CHECKALLOC longjmps out of crit_enter() without crit_exit() (leaked critical section)
Base system — drivers · sys/dev/raid/vinum/vinummemory.c:103 · pass 2 · new |
NULL-deref panic / leaked crit section in debug builds |
untested |
| DF-5127 |
Low |
4.1 |
rcu_expand() performs a blocking GFP_KERNEL (M_WAITOK) allocation inside crit_enter() — sleeping allocation in a critical section under memory pressure
Base system — drivers · sys/dev/drm/linux_rcu.c:139 · pass 2 · new |
latent crit-section rule violation under memory pressure |
untested |
| DF-5129 |
Low |
4.1 |
for_each_sg lacks Linux's `&& sg' termination guard — list shorter than sgmax turns into sg_next(NULL) NULL-deref panic
Base system — drivers · sys/dev/drm/linux_scatterlist.c:342 · pass 2 · new |
NULL-deref panic inside drm code on caller/list count mismatch |
untested |
| DF-5130 |
Low |
4.1 |
ecc_amd8000_attach() executes rdmsr(0x017B)/wrmsr(0x17B) with no CPU vendor/family validation — chipset-ID-gated #GP kernel panic (coretemp DF-4735 shape, hardware-gated)
Base system — drivers · sys/dev/misc/ecc/ecc_amd8000.c:165 · pass 2 · new |
attach-time #GP panic on platform-lying config |
untested |
| DF-5168 |
Low |
4.1 |
ips_ioctl_cmd ENOMEM error path feeds a failed bus_dmamem_alloc result into bus_dmamem_free — contigfree(NULL)/kfree(NULL) corrupts kernel_map or panics
Base system — drivers · sys/dev/raid/ips/ips_ioctl.c:116 · pass 2 · new |
kernel panic or kernel_map corruption on the ENOMEM path |
untested |
| DF-5179 |
Low |
4.1 |
atkbdattach() ignores IRQ resource allocation failure: NULL resource reaches nexus_setup_intr() which panics by design (boot-time kernel panic; BUS_SETUP_INTR result also unchecked, resource leaked)
Base system — drivers · sys/dev/misc/atkbd/atkbd_isa.c:120 · pass 2 · new |
panic |
untested |
| DF-5180 |
Low |
4.1 |
acpi_PkgStr() heap out-of-bounds read past non-NUL-terminated ACPI_TYPE_BUFFER: strlcpy()'s unbounded 'traverse rest of src' scan walks off the end of the ACPICA allocation for a zero-free firmware buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/acpica/acpi_package.c:86 · pass 2 · new |
dos |
untested |
| DF-5217 |
Low |
4.1 |
LKM screen-saver module unload TOCTOU: remove_scrn_saver samples scrn_blanked and clears current_saver with no syscons_lock, so kldunload can unmap the saver module (fade_saver et al.) while fade_saver is still executing — use-after-free of module text and statics
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/fade/fade_saver.c:46 · pass 2 · new |
kernel UAF executing freed module text - panic ceiling |
untested |
| DF-5339 |
Low |
4.1 |
probe/enumerate BSP-validation asymmetry: disabled second BSP entry passes probe then trips KKASSERT(!found_bsp) — deterministic boot-time kernel panic
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/mptable.c:684 · pass 2 · new |
none |
untested |
| DF-5340 |
Low |
4.1 |
mptable_search() base-memory scan: (base_memory-1)*1024 u_int underflow — KKASSERT(target!=0) boot panic (basemem==1) or mapdev scan of the top-of-4GB MMIO hole (basemem==0)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/mptable.c:284 · pass 2 · new |
none |
untested |
| DF-5342 |
Low |
4.1 |
u_int truncation of the HT-fixup APIC-id mask for ids 32..63 silently corrupts the fixup decision — phantom logical CPUs enumerated, boot hangs on smp_startup_mask
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/mptable.c:641 · pass 2 · new |
none |
untested |
| DF-5343 |
Low |
4.1 |
HT-fixup synthetic processor entries inherit the BSP flag: lapic_set_cpuid(0, fake_id) clobbers cpu_id_to_apic_id[0] and unconditionally burned CPU slots leave phantom SIPI targets — SMP boot hang / permanent cpu-0 IPI mis-addressing
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/mptable.c:702 · pass 2 · new |
none |
untested |
| DF-5345 |
Low |
4.1 |
Unguarded rdmsr() in print_svm_info()/print_vmx_info(): CPUID-advertised VMX/SVM without MSR backing → ring-0 #GP → boot-time kernel panic
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/identcpu.c:1279 · pass 2 · new |
none |
not_reproduced |
| DF-5346 |
Low |
4.1 |
SMAP missing from the hypervisor CPUID-feature blacklist in identify_cpu() despite the blacklist's own CR4-activation rule (buggy-hypervisor boot #GP)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/identcpu.c:668 · pass 2 · new |
none |
not_reproduced |
| DF-5366 |
Low |
4.1 |
cpu_id_to_apic_id[] lacks the -1 sentinel its sibling gets: lapic_config() initializes only apic_id_to_cpu_id[], so any unassigned forward mapping silently aliases APIC ID 0 (the BSP) — INIT/SIPI redirected at the running BSP
Base system — bus / platform / cpu · sys/platform/pc64/apic/lapic.c:1161 · pass 2 · new |
INIT IPI redirected at the BSP: boot-time reset loop / SMP bringup hang |
untested |
| DF-5622 |
Low |
4.1 |
Reachable KKASSERT panic on firmware-supplied _PSS control values: endpoint P-states are never validated and vid>=rvo is never checked before amd0f_set_fidvid asserts
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/cpufreq_machdep.c:92 · pass 2 · new |
none |
not_reproduced |
| DF-5729 |
Low |
4.1 |
lwkt_tokref published before initialization — DDB ps token walk (db_dump_td_tokens) can dereference a stale/NULL tr_tok (snapshot race)
Base system — kernel core (kern) · sys/kern/lwkt_token.c:660 · pass 2 · new |
none |
untested |
| DF-5771 |
Low |
4.1 |
mmc_rescan_cards()/mmc_delete_cards() free card ivars unconditionally after device_delete_child() — latent UAF if a child driver's detach ever fails
Kernel memory corruption (UAF / OOB write) · sys/bus/mmc/mmc.c:1728 · pass 2 · new |
slow kernel-memory leak (5770) / latent UAF (5771) |
untested |
| DF-3438 |
Info |
4.1 |
Cipher registry has no reference accounting: ieee80211_crypto_unregister clears ciphers[] while keys may still hold wk_cipher into the module; policy-module nrefs gate read unsynchronized at MOD_UNLOAD ('XXX disallow if references') - kldunload TOCTOU -> UAF
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_crypto.c:194 · pass 2 · new |
none |
untested |
| DF-4483 |
Info |
4.1 |
build_rq_buffer malloc-failure path frees the whole request via abortrequest() and then every caller frees it again (latent double-free / UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumrequest.c:833 · pass 2 · new |
Latent heap double-free/UAF landmine (dead today: Malloc is M_WAITOK and never fails) |
untested |
| DF-4825 |
Info |
4.1 |
clockmod: raw wrmsr(MSR_THERM_CONTROL) without a fault-safe accessor — a VMM advertising CPUID ACPI+TM but faulting the MSR turns the root sysctl select / kldunload raise-to-100% into an unrecoverable #GP panic (DF-4735 family)
Base system — drivers · sys/dev/powermng/clockmod/clockmod.c:472 · pass 2 · new |
panic |
untested |
| DF-4831 |
Info |
4.1 |
hwmgr_hw_fini dereferences hwmgr->hwmgr_func->backend_fini and hwmgr->pptable_func->pptable_fini guarded only by pm_en — no PHM_FUNC_CHECK-style preflight (same latent class as DF-1913 at unguarded sites)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/hwmgr.c:278 · pass 2 · new |
panic |
untested |
| DF-5096 |
Info |
4.1 |
parser2 command-table helper hub failure ignored leaves bp->cmd_helper NULL inside a 'successfully' constructed firmware BIOS parser — latent whole-vtable NULL function-pointer call (v2 twin of DF-5075)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/command_table_helper2.c:73 · pass 2 · new |
latent NULL vtable call (panic) |
untested |
| DF-5097 |
Info |
4.1 |
helper2 hub wires v1 dce80/dce110 tables into the atomfirmware parser; both tables lack transmitter_color_depth_to_atom and dc_clock_type_to_atom which command_table2.c invokes unguarded — latent NULL function-pointer kernel panic under cross-generation VBIOS pairing (parser2 twin of DF-5000/DF-4985)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/command_table_helper2.c:43 · pass 2 · new |
latent NULL vtable call (panic) |
untested |
| DF-5623 |
Info |
4.1 |
Unbounded FID/VID pending-bit spin and non-converging phase-2 ramp loop; success returned without completion verification
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/cpufreq_machdep.c:53 · pass 2 · new |
none |
not_reproduced |
| DF-3972 |
Medium |
4.0 |
NULL-pointer dereference of rdev->pm.dpm.boot_ps in btc_dpm_enable/btc_dpm_disable when the VBIOS power table has no BOOT-classified state (or ucStateEntrySize<=1) — btc twin of DF-3610/3646/3754/3953
Base system — drivers · sys/dev/drm/radeon/btc_dpm.c:2377 · pass 2 · new |
none |
inconclusive |
| DF-5298 |
Medium |
4.0 |
umass CBI command-phase error path leaks the in-flight CCB when the device detaches during teardown (sense-fetch fallback is a silent no-op) — unkillable hung process + leaked kernel CCB
Base system — bus / platform / cpu · sys/bus/u4b/storage/umass.c:1871 · pass 2 · new |
none (demonstrated); claimed impact: local DoS - permanent unkillable D-state process plus leaked kernel CCB, triggered by an evil CBI USB device that hangs an ADSC control transfer and then self-disconnects while a user I/O command is in flight |
not_reproduced |
| DF-2847 |
Low |
4.0 |
get_next_valid_apicid() -1 sentinel propagated unvalidated into get_cpuid_from_apicid(-1): OOB read of apic_id_to_cpu_id[-1] whose value is used, plus duplicate leaf member masks on exhaustion
Kernel memory leaks / uninitialized disclosure · sys/kern/subr_cpu_topology.c:100 · pass 2 · new |
none |
not_reproduced |
| DF-3848 |
Low |
4.0 |
bwn_phy_shm_tssi_read returns positive ENOENT but callers test < 0: the invalid-TSSI sentinel is treated as a valid TSSI of 2, so TX-power recalculation always 'adjusts' on fabricated data
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn_phy_g.c:3520 · pass 2 · new |
TX-power correctness (bwn) / mbuf leak + ring wedge (dc) |
untested |
| DF-3964 |
Low |
4.0 |
Firmware-fed get_config values consumed unvalidated: PAGE_SIZE*(max_sg_count-1) maxsize underflow and cam_simq_alloc(max_requests-1) with openings=-1
Base system — drivers · sys/dev/raid/hptiop/hptiop.c:1928 · pass 2 · new |
none |
not_reproduced |
| DF-4299 |
Low |
4.0 |
emu_pcm_attach late-failure path frees sc while the mixer remains registered: dangling mixer devinfo UAF on /dev/mixer ioctls and leaked AC97 submixer
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/emu10kx-pcm.c:1399 · pass 2 · new |
none |
untested |
| DF-4308 |
Low |
4.0 |
adwprocesserror() panics on firmware-defined host_status codes the switch does not enumerate (0x35/0x36/0x46/0x47) — target/device-triggerable kernel panic
Base system — drivers · sys/dev/disk/advansys/adwcam.c:1350 · pass 2 · new |
none |
untested |
| DF-4727 |
Low |
4.0 |
aureal: au_intr calls chn_intr(au->pch.channel) while the channel is NULL — NULL-deref panic in the snd_setup_intr-to-pcm_addchan window that au_testirq deliberately enables IRQs inside
Base system — drivers · sys/dev/sound/pci/aureal.c:411 · pass 2 · new |
none |
untested |
| DF-5040 |
Low |
4.0 |
vbus_ext->timer never drained before vbus_ext is freed at shutdown: os_timer_for_ldm() fires on freed memory (UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hpt27xx/hpt27xx_os_bsd.c:230 · pass 2 · new |
probe panic / shutdown-window UAF per finding |
untested |
| DF-5076 |
Low |
4.0 |
ipmi2_pci_probe is the only IPMI frontend probe lacking the ipmi_attached singleton guard — double attach (second cdev + second kthread, interleaved KCS/SMIC state machines) on dual-description BMC hardware
Base system — drivers · sys/dev/misc/ipmi/ipmi_pci.c:183 · pass 2 · new |
none |
untested |
| DF-5173 |
Low |
4.0 |
ata_jmicron_setmode reads PCI config from the ATA disk device instead of the controller — 80-pin cable detect and multifunction PATA routing are permanently dead
Base system — drivers · sys/dev/disk/nata/chipsets/ata-jmicron.c:161 · pass 2 · new |
UDMA over 40-pin cable / misrouted PATA setmode |
untested |
| DF-3974 |
Medium |
3.9 |
Wild OOB kernel read in btc_find_voltage when a VBIOS voltage object has zero LUT entries: entries[0xFFFFFFFF] ≈ +32 GiB past eg_pi (btc twin of DF-3649/3956)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/btc_dpm.c:1293 · pass 2 · new |
none |
inconclusive |
| DF-2742 |
Low |
3.9 |
diskstrategy consumes dp->d_slice with no serialization while disk_probe/disk_invalidate replace and free it — use-after-free window for every in-flight bio
Kernel memory corruption (UAF / OOB write) · sys/kern/subr_disk.c:1246 · pass 2 · new |
none |
not_reproduced |
| DF-3421 |
Low |
3.9 |
DDB 802.11 pretty-printers dump raw wlan key material (wk_key bytes) to the console — WEP/TKIP-temporal/CCM keys and RSC/TSC exposed in serial-console captures and crash-dump sessions
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_ddb.c:810 · pass 2 · new |
none |
untested |
| DF-3495 |
Low |
3.9 |
v1 ng_l2tp session-hook / hook-private use-after-free: unreferenced cached hook+hpriv used across a microsecond-scale window while concurrent rmhook frees both
Kernel memory corruption (UAF / OOB write) · sys/netgraph/l2tp/ng_l2tp.c:661 · pass 2 · new |
none |
not_reproduced |
| DF-4139 |
Low |
3.9 |
throw_rude_remark strcpy of up-to-1024-byte message into ioctl_reply->msg[256]: up to 4 bytes past the VINUM_CREATE ioctl allocation, plus dangling static ioctl_reply (UAF write)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumconfig.c:114 · pass 2 · new |
none |
untested |
| DF-4195 |
Low |
3.9 |
drm_connector_register() error path leaves dangling connector->kdev — latent double-free on re-register; unchecked kzalloc in the sysfs stub
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_connector.c:453 · pass 2 · new |
none |
untested |
| DF-4757 |
Low |
3.9 |
LLQ stride accounting inverted: ena_com.c:657 assigns descs_per_entry backwards (SINGLE→entry/64, MULTIPLE→1) contradicting the vendor enum semantics and ena_eth_com.c:202 — free-space guard undercounts for SINGLE-stride LLQs and lets the tail lap outstanding descriptors
Base system — drivers · sys/dev/virtual/amazon/ena/ena-com/ena_com.c:657 · pass 2 · new |
corruption |
untested |
| DF-5311 |
Low |
3.9 |
dagetcapacity() parses READ CAPACITY results from a non-zeroed heap buffer; short transfer with good status publishes stale kernel heap as disk geometry and prints it to console/dmesg
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_da.c:2199 · pass 2 · new |
stale-heap geometry disclosure (DF-5311) / permanent throttle underflow (DF-5312) / root-gated ioctl panic (DF-5313) |
untested |
| DF-5552 |
Low |
3.9 |
cam_devq reference count is a non-atomic int and cam_devq_release() frees the devq with active I/O after only a warning
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/cam_queue.c:242 · pass 2 · new |
none |
untested |
| DF-5595 |
Low |
3.9 |
sim refcount release protocol is non-atomic and unenforced — asymmetric with its atomic acquisition side and its cmpset-loop neighbors
Base system — bus / platform / cpu · sys/bus/cam/cam_sim.c:246 · pass 2 · new |
none |
untested |
| DF-4675 |
Info |
3.9 |
bktr_detach removes only the minor==unit device node — /dev/tunerN (unit+16) and /dev/vbiN (unit+32) persist after detach, stranding cdevs that reference the driver's dev_ops (use-after-free after module unload)
Kernel memory corruption (UAF / OOB write) · sys/dev/video/bktr/bktr_os.c:325 · pass 2 · new |
bktr_detach removes only minor==unit - /dev/tunerN and /dev/vbiN persist, stranding cdevs referencing the driver's dev_ops (dev_ops UAF after module unload) |
untested |
| DF-3634 |
Medium |
3.8 |
Uninitialized kernel stack variable leaked to userspace by i915_gem_mmap_ioctl() when size==0 (DragonFly-only path)
Privilege escalation (general use) · sys/dev/drm/i915/i915_gem.c:1923 · pass 2 · new |
leak |
inconclusive |
| DF-3975 |
Medium |
3.8 |
btc_dpm_get_sclk/btc_dpm_get_mclk dereference eg_pi->requested_rps.ps_priv (NULL) before the first power-state change — NULL-page read panic from display bandwidth updates
Base system — drivers · sys/dev/drm/radeon/btc_dpm.c:2802 · pass 2 · new |
none |
inconclusive |
| DF-5075 |
Low |
3.8 |
v1 command-table helper hub leaves bp->cmd_helper NULL for DCE12/DCN generations on the legacy BIOS parser; construct ignores the init failure — whole-vtable NULL dereference at the first BIOS command
Base system — drivers · sys/dev/drm/amd/display/dc/bios/command_table_helper.c:58 · pass 2 · new |
none |
untested |
| DF-5312 |
Low |
3.8 |
DA_CCB_TRIM completion decrements outstanding_cmds_rd that the TRIM start path never increments: permanent queue-accounting underflow disables read throttling and tps balancing
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_da.c:1327 · pass 2 · new |
stale-heap geometry disclosure (DF-5311) / permanent throttle underflow (DF-5312) / root-gated ioctl panic (DF-5313) |
untested |
| DF-3088 |
Low |
3.7 |
nfssvc_nfsd() socket-zap race after the pipelined dorec leaks the request's mbuf chain and client address (:576-579 frees only the descriptor)
Base system — filesystems · sys/vfs/nfs/nfs_syscalls.c:575 · pass 2 · new |
none |
not_reproduced |
| DF-2717 |
Medium |
3.6 |
nlookup_init()/retry path read the shared fd_ncdir/fd_nrdir nchandle pairs without fd_spin while sibling threads swap them — torn (ncp,mount) pairs and hold-after-drop race on the jail '..'-gate comparison inputs
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_nlookup.c:149 · pass 2 · new |
none |
not_reproduced |
| DF-5782 |
Medium |
3.6 |
i915: unauthenticated ADDFB2 u32 overflow in framebuffer offset/size math (tile-aligned geometry) defeats bo-bounds check — OOB GPU scanout read from wrong/adjacent GGTT mappings
Base system — drivers · sys/dev/drm/i915/intel_display.c:2398 · pass 2 · new |
none |
untested |
| DF-2760 |
Low |
3.6 |
sys_getsid reads the p_pgrp->pg_session chain without pt->p_token; sys_setpgid reads curp->p_session unlocked — lifetime race vs. enterpgrp's pgrp free (stale/UAF read)
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_prot.c:153 · pass 2 · new |
none |
not_reproduced |
| DF-3038 |
Low |
3.6 |
Crafted FAT32 root-alias directory entry (startCluster == pm_rootdirblk): rmdir() frees the live mount's root-directory cluster chain; rename() clobbers root entry 1 and creates a namespace cycle
Base system — filesystems · sys/vfs/msdosfs/msdosfs_vnops.c:1230 · pass 2 · new |
dos |
untested |
| DF-3066 |
Low |
3.6 |
ext2_nodealloccg honors unvalidated inode-bitmap content: reserved inode numbers (below first_ino) and tail bits (>= ipg) are allocatable — silent fs aliasing / OOB e2fs_gd[] reads via ino_to_cg
Kernel memory leaks / uninitialized disclosure · sys/vfs/ext2fs/ext2_alloc.c:1319 · pass 2 · new |
none |
not_reproduced |
| DF-3234 |
Low |
3.6 |
ext2_balloc's four newb > UINT_MAX EFBIG returns leak the just-allocated block and leave ip->i_blocks inflated — permanent unprivileged-triggerable space/accounting corruption on 64BIT filesystems
Base system — filesystems · sys/vfs/ext2fs/ext2_balloc.c:135 · pass 2 · new |
corruption |
not_reproduced |
| DF-3252 |
Low |
3.6 |
ffs_vget publishes the inode into the ihash before validating it: bread-failure branch hands concurrent ufs_ihashget waiters a VBAD zeroed-inode vnode as a successful VFS_VGET result (DF-3208 ufs twin)
Base system — filesystems · sys/vfs/ufs/ufs_ihash.c:1120 · pass 2 · new |
none |
untested |
| DF-3371 |
Low |
3.6 |
ip_setmoptions IP_ADD_MEMBERSHIP: route reference dropped before rt_ifp is read — use-after-release race and permanent rtentry leak when the lookup held the last reference
Kernel memory corruption (UAF / OOB write) · sys/netinet/ip_output.c:1658 · pass 2 · new |
none |
not_reproduced |
| DF-3624 |
Low |
3.6 |
Double free of adev->pm.dpm.ps on the kzalloc-failure path of ci_parse_power_table (plus leak of already-parsed ci_ps states)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/ci_dpm.c:5694 · pass 2 · new |
OOM-gated double free |
untested |
| DF-3642 |
Low |
3.6 |
e1000_erase_flash_bank_ich8lan: erase address accumulates across retries, retry budget never reset, and erase failure reported as success
Base system — drivers · sys/dev/netif/ig_hal/e1000_ich8lan.c:4779 · pass 2 · new |
none |
not_reproduced |
| DF-3734 |
Low |
3.6 |
Unchecked requested_clock_mode_index++ in r600 DYNPM downclock (discrete branch) — one-past-end OOB heap read of per-state clock_info
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r600.c:486 · pass 2 · new |
none |
inconclusive |
| DF-3747 |
Low |
3.6 |
sccngetch(): 8-byte size_t store through (size_t *)&fkeycp pun writes 4 bytes past a kernel static; unclamped bcopy into the 17-byte static fkey trusts the DF-1503-corruptible fkeytab length
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/syscons.c:2005 · pass 2 · new |
none |
inconclusive |
| DF-3809 |
Low |
3.6 |
port_to_transcoder[PORT_F] out-of-bounds read in intel_enable_ddi_hdmi on GEN9_BC DDI-F hardware leads to wild CHICKEN_TRANS MMIO write
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_ddi.c:3180 · pass 2 · new |
none |
untested |
| DF-4073 |
Low |
3.6 |
MVDD stash-table wrong-index reads: entries[count] instead of entries[level] (uninitialized voltage to SMU) and cross-table mvdd_voltage_table.entries[vdd_dep_on_mclk index] (true OOB read)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:459 · pass 2 · new |
none |
inconclusive |
| DF-4108 |
Low |
3.6 |
vge TX short-frame pad path writes one fragment past vge_frag[7] when a <60-byte frame arrives with 7 DMA segments (vge_encap maxsegs off-by-one corrupts the next descriptor's vge_sts/vge_ctl)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/vge/if_vge.c:676 · pass 2 · new |
none |
untested |
| DF-4115 |
Low |
3.6 |
hdac_stream_alloc() bidirectional-stream tag arithmetic: '+ num_oss' should be '- num_oss' (missing upstream FreeBSD fix) — wrong/colliding/negative HDA stream tags on BSS-capable controllers
Base system — drivers · sys/dev/sound/pci/hda/hdac.c:1933 · pass 2 · new |
none |
not_reproduced |
| DF-4206 |
Low |
3.6 |
vlv_dsi_init !fixed_mode error path kfrees a registered drm_connector without drm_connector_cleanup — dangling entry on mode_config.connector_list and in the mode-object IDR
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/vlv_dsi.c:1860 · pass 2 · new |
none |
untested |
| DF-4430 |
Low |
3.6 |
PRIME import on radeon: populate() ignores the sg table (fresh pages bound into the GART instead of the exporter's) and unpopulate() leaks those pages + TTM accounting per import cycle
Base system — drivers · sys/dev/drm/radeon/radeon_ttm.c:753 · pass 2 · new |
PRIME-imported (ttm_bo_type_sg) BOs on radeon: populate() ignores ttm->sg (slave branch #ifdef DUMBBELL_WIP, radeon_ttm.c:753-764) and allocates fresh pool pages that get bound into the GART — the GPU view of an imported dma-buf diverges from the exporter's memory (silent data corruption); unpopulate()'s live slave early-return (radeon_ttm.c:808,818-819) then never frees those pages or their ttm_mem_global accounting, so repeated import/destroy cycles leak wired kernel memory until GPU allocations fail (DoS). Sibling amdgpu_ttm.c:1268-1330 implements the slave path live. Reachability: cross-driver dma-buf import with radeon as importer (multi-GPU); same-driver self-import unaffected. |
untested |
| DF-4445 |
Low |
3.6 |
radeon_router_select_ddc_port/cd_port write an uninitialized kernel-stack byte into the DDC/CD router i2c mux registers when the router read fails (and leak it via DRM_DEBUG)
Base system — drivers · sys/dev/drm/radeon/radeon_i2c.c:1110 · pass 2 · new |
none |
untested |
| DF-4552 |
Low |
3.6 |
snd_vibes: sv_attach failure paths after mixer_init leave a live 0666 /dev/mixerN over released bus resources and leak sc (DF-4156 family, 12th instance)
Base system — drivers · sys/dev/sound/pci/vibes.c:789 · pass 2 · new |
none |
not_reproduced |
| DF-4588 |
Low |
3.6 |
ttm_bo_vm_fault_dfly maps SG (prime-imported) foreign pages through the card node — the upstream TTM_PAGE_FLAG_SG refusal is dead code in #if 0
Base system — drivers · sys/dev/drm/ttm/ttm_bo_vm.c:161 · pass 2 · new |
none |
untested |
| DF-4603 |
Low |
3.6 |
ttm_bo_util: stale/uninitialized old_copy.bus snapshot drives iounmap() decision in ttm_bo_move_memcpy out: cleanup — kernel mapping leak, and full VRAM-aperture unmap when old BO sits at aperture offset 0
Base system — drivers · sys/dev/drm/ttm/ttm_bo_util.c:344 · pass 2 · new |
Permanent kernel mapping leak per failed move (VA pressure DoS), or full VRAM-aperture unmap when old BO sits at aperture offset 0 -> kernel panic on next premapped access |
untested |
| DF-4799 |
Low |
3.6 |
ar9285: 2GHz-only Kite never enforced — 11A/11NA caps advertised (GMODE divergence), 5GHz channel edges inherited, nf_5g limits left zeroed: 5GHz channels selectable → radio programmed out of calibrated design with 2GHz-only data
Base system — drivers · sys/dev/netif/ath/ath_hal/ar9002/ar9285_attach.c:527 · pass 2 · new |
none |
untested |
| DF-4808 |
Low |
3.6 |
dce110 i2c hw engine: SW engine-state accounting desync corrupts subsequent DDC/I2C transactions — write-path buffer_used_bytes undercounts (length consumed before the accounting add), counters never reset on aborted-busy/release paths, and I2C_HW_BUFFER_SIZE=538 exceeds the 8-bit DC_I2C_INDEX/COUNT0 fields
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/dce110/i2c_hw_engine_dce110.c:334 · pass 2 · new |
none |
untested |
| DF-4816 |
Low |
3.6 |
iwm phy_db: unbounded firmware channel-group id → up-to-1 MiB M_WAITOK allocations from interrupt context + uint8_t truncation of the group-count upload loop (-1→255 hard-fails firmware load)
Base system — drivers · sys/dev/netif/iwm/if_iwm_phy_db.c:279 · pass 2 · new |
none |
untested |
| DF-4830 |
Low |
3.6 |
mdcreate unit allocation (mdunits++) is unlocked and non-atomic — concurrent mdopen auto-create of the top unit can mint two md devices with the same unit number/cdev name
Base system — drivers · sys/dev/disk/md/md.c:135 · pass 2 · new |
none |
untested |
| DF-5066 |
Low |
3.6 |
dm_target_autoload clobbers linker_file->userrefs with an unsynchronized '=1' store outside kld_lock, racing kldunload's locked read-modify-write — corrupts kld module pin accounting
Base system — drivers · sys/dev/disk/dm/dm_target.c:88 · pass 2 · new |
module pin-accounting corruption |
untested |
| DF-5087 |
Low |
3.6 |
Unsynchronized check-then-set of i2c_hw_buffer_in_use plus clear-before-release lets concurrent I2C acquisitions jointly drive the single shared DCE80 DC_I2C engine — interleaved single-instance MMIO and per-engine tracking divergence from the shared 144-byte HW data FIFO (dce80 twin of DF-5035/DF-4878; the DAL2 mutex was dropped in the port)
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/dce80/i2caux_dce80.c:139 · pass 2 · new |
per finding (panic-instead-of-degrade / latent NULL-free / DDC data corruption) |
untested |
| DF-5131 |
Low |
3.6 |
DF-4871 family completion: seven more callers discard radeon_sync_resv()/radeon_sync_rings() returns — GPU work submitted without its cross-ring dependency on sync failure (silent bo data corruption, racing VM page-table updates)
Base system — drivers · sys/dev/drm/radeon/radeon_sync.c:89 · pass 2 · new |
none |
untested |
| DF-5263 |
Low |
3.6 |
mtw_key_set defers the raw ieee80211_key pointer; pairwise keys are node-embedded, so a node freed before mtw_key_set_cb runs makes the callback read and WRITE (k->wk_pad) freed node memory
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_mtw.c:1940 · pass 2 · new |
none |
inconclusive |
| DF-5385 |
Low |
3.6 |
SMBus block write transmits the command byte where the byte-count belongs — every Block Write is malformed on the wire
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:381 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-5387 |
Low |
3.6 |
Read pipeline issues STOP despite SMB_TRANS_NOSTOP — chained I2C reads of >=2 bytes are terminated early
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:440 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-5389 |
Low |
3.6 |
RX-timeout error path leaks queued read allowances and stale FIFO state into the next same-slave transaction
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:448 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-5590 |
Low |
3.6 |
isavga_attach() discards both bus_alloc_resource() returns and never unwinds the legacy VGA window claims on attach failure — permanent orphaned rman claims (I/O 0x3c0-0x3df, MEM 0xa0000-0xbffff) on VGA-less BIOS boots, plus leaked cdev under FB_INSTALL_CDEV
Base system — bus / platform / cpu · sys/bus/isa/vga_isa.c:144 · pass 2 · new |
none |
inconclusive |
| DF-3043 |
Low |
3.5 |
nm_bioqlen async-bio accounting leaks on every nfs_startio early-exit and underflows on commit→write reissue — permanent mount-wide loss of NFS read-ahead from a malicious/oversized-server attribute claim (and loss of the async throttle)
Base system — filesystems · sys/vfs/nfs/nfs_bio.c:922 · pass 2 · new |
none |
not_reproduced |
| DF-4450 |
Low |
3.5 |
hsw_activate_psr2() overflows two 4-bit EDP_PSR2_CTL fields when sink reports max resync latency, corrupting IDLE_FRAME/FRAME_BEFORE_SU/TP2_TIME in one register write
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_psr.c:415 · pass 2 · new |
none |
inconclusive |
| DF-4876 |
Low |
3.5 |
afd_sense(): three uninitialized stack structs consumed as ATAPI response buffers — short (but successful) device responses publish stale kernel stack as disk geometry
Base system — drivers · sys/dev/disk/nata/atapi-fd.c:345 · pass 2 · new |
none |
untested |
| DF-4949 |
Low |
3.5 |
smic_read_byte() writes a status code (SC_SMS_RD_NEXT 0xC5) into the control register instead of the command code (CC_SMS_RD_NEXT 0x45) — out-of-spec SMIC control write on every reply byte
Base system — drivers · sys/dev/misc/ipmi/ipmi_smic.c:202 · pass 2 · new |
dos |
untested |
| DF-5175 |
Low |
3.5 |
CIS-controlled PCMCIA I/O window size drives out-of-allocation port I/O in ata_pccard_attach()
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-card.c:93 · pass 2 · new |
none |
untested |
| DF-5296 |
Low |
3.5 |
uhci_pci_attach never validates the device-reported I/O BAR size against the fixed 0x14-byte UHCI register file — every UREAD/UWRITE can touch port-I/O outside the allocated resource window (DF-5268/DF-5282 register-geometry class, UHCI variant)
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/controller/uhci.c:97 · pass 2 · new |
silent cross-device port-I/O register corruption (max 12B excursion) |
untested |
| DF-5495 |
Info |
3.5 |
uether_rxmbuf() sets m_len/m_pkthdr.len from the caller-supplied len with no clamp — the core's unguarded RX entry point (live instance: DF-5396 urndis 2-byte heap overflow)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/net/usb_ethernet.c:563 · pass 2 · new |
mbuf cluster leak per stall-unplug cycle (DF-5494) / future unguarded-len overflows (DF-5495) |
untested |
| DF-3503 |
Low |
3.4 |
ng7 ksocket rcvdata trusts foreign m_tag sockaddr with no m_tag_len validation: short tag (e.g. minted by ng_tag with user-configured cookie/type/len) makes stag->sa overlap OOB heap handed to sosend
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/ksocket/ng_ksocket.c:904 · pass 2 · new |
none |
not_reproduced |
| DF-4134 |
Low |
3.4 |
give_plex_to_volume() writes vol->plex[MAXPLEX] past the last member of struct volume when the 'too many plexes' throw is swallowed by disk-config mode
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumconfig.c:198 · pass 2 · new |
none |
untested |
| DF-4238 |
Low |
3.4 |
sdma_v3_0 reads SDMA firmware sub-header fields (ucode_version @+16, ucode_feature_version @+32, ucode_size_bytes @+24) after only the datasize==size_bytes equality check — heap OOB read on a short root-installed firmware blob (DF-1451/1553/1598 family, v3 variant)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/sdma_v3_0.c:312 · pass 2 · new |
none |
not_reproduced |
| DF-4342 |
Low |
3.4 |
ngmn_newhook channel aliasing: DOWN-state reuse admits two live hooks on one schan — silent second connect, blackholed peer hook, permanent nhooks drift blocking line reconfig, and a concurrent dual-teardown double-free enabler
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mn/if_mn.c:470 · pass 2 · new |
none |
inconclusive |
| DF-4600 |
Low |
3.4 |
renameobject() bcopy's full-width newname with no NUL-termination guard → unterminated drive/sd/plex/volume name fields persisted to config and fed to strcmp/strlen/log %s
Base system — drivers · sys/dev/raid/vinum/vinumioctl.c:740 · pass 2 · new |
Unterminated names in find_* strcmp walks, detachobject strlen/bcmp, log %s, and persisted config text (garbage-tailed lines re-parsed at next init); reads bounded within the struct by NULs in following state enums |
untested |
| DF-4880 |
Low |
3.4 |
amdgpu_ucode_print_* helpers read version-dispatched sub-header fields (up to +156) with no blob-size check: unconditional heap OOB read on short crafted firmware, distinct sink from DF-1946 that survives its proposed minimum-size fix
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_ucode.c:54 · pass 2 · new |
infoleak |
untested |
| DF-5104 |
Low |
3.4 |
led_attach() kstrdup()s the GPIOATTACH consumer arg without type validation: GPIO_TYPE_INT delivers an integer-as-pointer (arbitrary strlen/kstrdup deref) and the string variant is not termination-bounded (kernel-stack over-read)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/gpio/gpio_led.c:186 · pass 2 · new |
none |
inconclusive |
| DF-5166 |
Low |
3.4 |
disable_irq/enable_irq resolve the victim entry by irq number only, ignoring dev_id — teardown/re-setup of the wrong device's handler whenever two registrations share or numerically collide on an irq
Base system — drivers · sys/dev/drm/linux_irq.c:150 · pass 2 · new |
interrupt-context UAF / orphaned armed handler / cross-device teardown |
untested |
| DF-5290 |
Low |
3.4 |
pmap_extract()/pmap_kextract() return (va & PAGE_MASK)-only garbage as a physical address when the PT exists but the PTE is empty — vkernel twin of pc64 DF-5245; live caller kvm_access_check() (/dev/kmem) then accepts unmapped KVA and mem I/O lands on guest-physical page 0
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/pmap.c:662 · pass 2 · new |
wrong-page /dev/kmem I/O (DF-5290) / latent PDP reclaim (DF-5291) |
untested |
| DF-2874 |
Info |
3.4 |
DIOCRECLUSTER has no privilege check and the kernel services dmsg peers with root credentials (raw RW with proc0.p_ucred, no open/keyid validation on I/O) — gated today only by diskopen's SYSCAP_RESTRICTEDROOT
Privilege escalation (general use) · sys/kern/subr_diskiocom.c:108 · pass 2 · new |
none |
not_reproduced fixed |
| DF-3693 |
Medium |
3.3 |
hptv2 metadata writeback: 2048-byte write from a 512-byte heap buffer persists ~1.5KB of kernel heap to disk (OOB read -> durable info leak)
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/nata/ata-raid.c:1903 · pass 2 · new |
leak |
inconclusive |
| DF-3764 |
Medium |
3.3 |
r100_packet3_check() reads payload dwords at fixed offsets without validating pkt->count -> OOB read 1 dword past the IB chunk end + kernel-log disclosure (same class as DF-1492/DF-3592)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r100.c:1961 · pass 2 · new |
leak |
inconclusive |
| DF-4181 |
Medium |
3.3 |
radeon fb creation never validates framebuffer size against the backing GEM object — display engine scans out megabytes past a tiny BO, disclosing adjacent VRAM on the physical display
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_display.c:1294 · pass 2 · new |
none |
inconclusive |
| DF-2612 |
Low |
3.3 |
ses_getputstat swallows all decode/encode/device errors and returns 0 (fail-open enclosure status)
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_ses.c:1268 · pass 2 · new |
none |
inconclusive |
| DF-2637 |
Low |
3.3 |
Kernel heap-pointer (%p) kprintf disclosures on four hammer2 inode error paths
Hammer2 · sys/vfs/hammer2/hammer2_inode.c:711 · pass 2 · new |
none |
untested |
| DF-2669 |
Low |
3.3 |
kern_truncate()/kern_ftruncate() execute VFS_ACCOUNT() even when VOP_SETATTR failed — quota accounting corruption on failed truncates
Base system — kernel core (kern) · sys/kern/vfs_syscalls.c:4044 · pass 2 · new |
none |
untested |
| DF-2670 |
Low |
3.3 |
sys_fchown() performs VFS_ACCOUNT() ownership-transfer accounting against the cwd's mount instead of the chowned file's mount
Base system — kernel core (kern) · sys/kern/vfs_syscalls.c:3637 · pass 2 · new |
none |
untested |
| DF-2705 |
Low |
3.3 |
vop_write quota accounting skips unlinked-open files (va_nlink==0) → complete quota enforcement bypass when vfs.quota_enabled=1; concurrent-writer delta TOCTOU
Base system — kernel core (kern) · sys/kern/vfs_vopops.c:476 · pass 2 · new |
none |
untested |
| DF-2764 |
Low |
3.3 |
journal_setextattr journals the post-VOP, fully consumed uio: the extended-attribute VALUE is never recorded in the REDO stream
Base system — kernel core (kern) · sys/kern/vfs_jops.c:1027 · pass 2 · new |
none |
not_reproduced |
| DF-3059 |
Low |
3.3 |
dirfs_setattr chtimes ignores the requested va_atime/va_mtime entirely — lutimes(path, NULL) always stamps 'now' (utimensat/UTIME_OMIT semantics unimplemented)
Base system — filesystems · sys/vfs/dirfs/dirfs_vnops.c:556 · pass 2 · new |
none |
not_reproduced |
| DF-3126 |
Low |
3.3 |
cd9660_vptofh truncates ino_t into int ifid_ino; cd9660_vget_internal's int lbn truncation makes the volume-space bounds check bypassable — NFS file-handle inode aliasing and negative-offset bread on crafted/large images
Base system — filesystems · sys/vfs/isofs/cd9660/cd9660_vfsops.c:774 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3135 |
Low |
3.3 |
dirfs_open_helper tests the vnode type with a bitwise AND ('dnp->dn_type & VDIR') instead of equality — VBLK/VSOCK/VFIFO nodes are always opened with O_DIRECTORY and can never be opened at all; VLNK/VCHR take the plain-file path
Base system — filesystems · sys/vfs/dirfs/dirfs_subr.c:583 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3174 |
Low |
3.3 |
Trigger retry dedup re-acquires already-done requests: concurrent waiters burn their retry budget on stale errors and never re-notify automountd
Base system — filesystems · sys/vfs/autofs/autofs.c:324 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3636 |
Low |
3.3 |
Leftover debug kprintf("i915_gem_fault: CHUNKING PASS") on the unprivileged per-fault partial-view path enables console/log-flood DoS
Privilege escalation (general use) · sys/dev/drm/i915/i915_gem.c:2323 · pass 2 · new |
none |
untested |
| DF-3677 |
Low |
3.3 |
XPT_RESET_DEV on the RAID physical-disk bus resets the WRONG device: phydisk id passed unmapped as a bus-0 TMF TargetID (hits the RAID volume sharing the number)
Base system — drivers · sys/dev/disk/mpt/mpt_cam.c:3292 · pass 2 · new |
none |
inconclusive |
| DF-3694 |
Low |
3.3 |
VIA/JMicron metadata writeback persists low 32 bits of kernel device_t pointers into on-disk metadata
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/nata/ata-raid.c:2736 · pass 2 · new |
leak |
inconclusive |
| DF-3798 |
Low |
3.3 |
ena_start_xmit leaks the dequeued mbuf when ena_xmit_mbuf returns ENA_COM_NO_MEM or ENA_COM_NO_SPACE
Base system — drivers · sys/dev/virtual/amazon/ena/ena.c:2936 · pass 2 · new |
mbuf leak / latent NULL-deref |
untested |
| DF-3865 |
Low |
3.3 |
DragonFly rbtree shim never maintains rb_root_cached.rb_leftmost — amdgpu_vm_fini's mapping-cleanup loop is dead code, leaking every mapping left in the interval tree at teardown
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_vm.c:3325 · pass 2 · new |
none |
not_reproduced |
| DF-4016 |
Low |
3.3 |
mps_user_btdh off-by-one: target == max_devices passes the range check and reads mapping_table[max_devices] out of bounds
Base system — drivers · sys/dev/raid/mps/mps_user.c:2021 · pass 2 · new |
mps_user_btdh off-by-one: 'target > max_devices' lets TargetID==max_devices read mapping_table[max_devices] one 24-byte element past the heap allocation, disclosing a garbage u16 DevHandle to the operator-group caller. mpr twin at mpr_user.c:2240. |
untested |
| DF-4021 |
Low |
3.3 |
mpr_user_btdh off-by-one: target == max_devices passes the range check and reads mapping_table[max_devices] out of bounds (twin of DF-4016)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_user.c:2240 · pass 2 · new |
mpr_user_btdh off-by-one: 'target > max_devices' (mpr_user.c:2240) lets TargetID==max_devices read mapping_table[max_devices] one 24-byte element past the heap allocation (mpr_mapping.c:2140-2141 allocates exactly max_devices entries), disclosing a garbage u16 dev_handle (byte offset 18 of the element) to the operator-group caller (mpr_user.c:2245-2247, node mpr_user.c:205). Twin of DF-4016 (mps). |
untested |
| DF-4102 |
Low |
3.3 |
dce4/avivo_crtc_do_set_base pin the scanout BO then return -EINVAL on unsupported-format/tiling errors without unpinning — repeatable VRAM pin-exhaustion DoS
Base system — drivers · sys/dev/drm/radeon/atombios_crtc.c:1183 · pass 2 · new |
Local DoS by any DRM-master-holding user on radeon hardware: one ioctl per leaked pin; VRAM exhaustion degrades/deadlocks the display subsystem. |
untested |
| DF-4226 |
Low |
3.3 |
Output-lock asymmetry: radeon_legacy_tmds_int_commit and radeon_legacy_tv_dac_commit re-LOCK the output instead of unlocking after modeset
Base system — drivers · sys/dev/drm/radeon/radeon_legacy_encoders.c:768 · pass 2 · new |
none |
untested |
| DF-4239 |
Low |
3.3 |
Stale ata_xfer resid propagated to CAM for zero-data commands: sili_unload_prb skips the resid update when datalen==0 and the completion copies the previous command's resid into csio->resid
Base system — drivers · sys/dev/disk/sili/sili_cam.c:1667 · pass 2 · new |
Stale ata_xfer resid propagated to CAM for zero-data commands: sili_unload_prb skips the resid update when datalen==0 and the completion copies the previous command's resid into csio->resid. Operator-group /dev/passN users receive bogus resid for zero-data commands; can mask real short-transfer conditions. |
untested |
| DF-4264 |
Low |
3.3 |
AMDGPU_FENCE_TO_HANDLE default case returns -EINVAL without dma_fence_put — one leaked fence reference per invalid call
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_cs.c:1427 · pass 2 · new |
none |
untested |
| DF-4357 |
Low |
3.3 |
AMDGPU_INFO_READ_MMR_REG missing upstream count>128 cap: unprivileged unbounded kernel kmalloc (hardening / local DoS)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_kms.c:668 · pass 2 · new |
none |
inconclusive |
| DF-4428 |
Low |
3.3 |
smu10_hwmgr_backend_fini never frees vdd_dep_on_mclk — deterministic kernel heap leak per powerplay fini cycle on Raven APUs
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/smu10_hwmgr.c:443 · pass 2 · new |
none |
untested |
| DF-4470 |
Low |
3.3 |
Duplicate connector IDs in DRM_IOCTL_MODE_SETCRTC leak one drm_connector reference per duplicate slot per successful modeset — per-slot get vs per-connector binding, surplus dropped only on the fail path
Base system — drivers · sys/dev/drm/drm_crtc_helper.c:626 · pass 2 · new |
none |
inconclusive |
| DF-4471 |
Low |
3.3 |
__drm_helper_disable_unused_functions() NULLs crtc->primary->fb without dropping its framebuffer reference when invoked outside __drm_mode_set_config_internal — fb + GEM object reference leak per disabled unused CRTC
Base system — drivers · sys/dev/drm/drm_crtc_helper.c:189 · pass 2 · new |
none |
untested |
| DF-4517 |
Low |
3.3 |
DRM_CLOEXEC silently dropped when creating prime fds: dma_buf_fd() #if 0's the close-on-exec handling, so the validated flags lie to userspace and dma-buf fds leak across execve
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/drm_prime.c:669 · pass 2 · new |
Exported PRIME fds stay open across execve, contradicting the API contract Mesa/libdrm rely on; combined with DF-4429 (missing radeon verify_access) widens cross-user GPU buffer access |
untested |
| DF-4532 |
Low |
3.3 |
drm_open() calls device_busy() on every successful open but no live path ever calls device_unbusy(): drm device and its PCI parent wedged in DS_BUSY forever, permanently blocking device_detach()/kldunload
Base system — drivers · sys/dev/drm/drm_file.c:351 · pass 2 · new |
none |
inconclusive |
| DF-4533 |
Low |
3.3 |
drm_open() leaks one drm_device kref per successful open: the success-path drm_minor_acquire() is never balanced on any live path, so drm_dev_release()/drm_dev_fini() can never run and the whole drm_device leaks at unload
Base system — drivers · sys/dev/drm/drm_file.c:325 · pass 2 · new |
none |
inconclusive |
| DF-4571 |
Low |
3.3 |
missed_breadcrumb() forces intel_engine_dump() unconditionally ('|| 1' debug leftover): full engine state dump, including the DF-4246 over-reading ring hexdump path, runs from callout context every hangcheck period on production kernels
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_breadcrumbs.c:85 · pass 2 · new |
Unthrottled multi-KB engine dumps every ~1.5s per engine on production kernels (console/dmesg flood DoS + repeated DF-4246 over-read path in callout context) |
untested |
| DF-4794 |
Low |
3.3 |
ig_hal VF: E1000_VF_SET_MULTICAST_OVERFLOW flag overwritten before send — PF never told to enable multicast promiscuous, >30 joined groups silently dropped (IPv6 ND breakage)
Base system — drivers · sys/dev/netif/ig_hal/e1000_vf.c:421 · pass 2 · new |
none demonstrated at runtime (guest is virtio-only, no SR-IOV 82576/I350 VF device); logic-level proven from source ordering + upstream divergence: with >30 multicast groups joined on an igb VF interface, PF-visible msgbuf[0]=0x001E0003 loses E1000_VF_SET_MULTICAST_OVERFLOW (0x00800000), PF never enables multicast promiscuous, groups beyond the 30 forwarded hashes silently stop receiving (typ. IPv6 solicited-node/ND breakage) — availability only, no corruption/leak/privilege crossing |
untested |
| DF-4877 |
Low |
3.3 |
afd_strategy(): no mediasize bounds check and 32-bit LBA truncation — out-of-range raw I/O wraps onto low LBAs instead of failing
Base system — drivers · sys/dev/disk/nata/atapi-fd.c:241 · pass 2 · new |
none |
untested |
| DF-4924 |
Low |
3.3 |
gfx_v9_0 gfx-ring emit_frame_size never reserves the 2-dword TMZ FRAME_CONTROL emission that amdgpu_ib_schedule() makes unconditionally — ring-reservation overrun of 2 dwords for every CS job with num_ibs == 24 (mod 64)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_ib.c:159 · pass 2 · new |
dos |
not_reproduced |
| DF-5027 |
Low |
3.3 |
ipsd_dump() reports fake success when ips_poll_cmd never completes — copperhead crash dumps can be silently empty/corrupt (zeroed status defeats COMMAND_ERROR)
Base system — drivers · sys/dev/raid/ips/ips_disk.c:249 · pass 2 · new |
none |
not_reproduced |
| DF-5050 |
Low |
3.3 |
mfi_syspd_dump acknowledges every crash-dump chunk the firmware rejected — polled transport returns 0 for any firmware cmd_status, so a dead/removed system PD 'successfully' dumps an empty core (DF-5027 shape)
Base system — drivers · sys/dev/raid/mfi/mfi_syspd.c:287 · pass 2 · new |
none |
not_reproduced |
| DF-5215 |
Low |
3.3 |
LinuxKPI interval 'tree' is an O(N) circular list — sole live consumer (radeon_vm.c GEM_VA path) turns it into unprivileged per-ioctl O(N) kernel scans (O(N^2) mapping churn), a kernel-CPU amplification DoS on radeon-equipped systems
Base system — drivers · sys/dev/drm/linux_interval_tree.c:34 · pass 2 · new |
latent interior kfree / kernel-CPU amplification / unkillable spin |
untested |
| DF-5264 |
Low |
3.3 |
mtw_stop never frees the endpoint queues; the next mtw_init_locked memsets them via mtw_setup_tx_list, orphaning every queued mbuf and its ieee80211_node reference - repeatable kernel leak per down/up cycle
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_mtw.c:4647 · pass 2 · new |
none |
untested |
| DF-5279 |
Low |
3.3 |
pci_get_vpd_readonly_method(): missing break discards every successful keyword match — lookup always returns ENXIO
Base system — bus / platform / cpu · sys/bus/pci/pci.c:1354 · pass 2 · new |
VPD lookup always ENXIO (DF-5279) / devq UAF read race (DF-5280) / boot-time off-by-one (DF-5281) |
untested |
| DF-5417 |
Low |
3.3 |
Bind/xfer lifecycle leaks and open race: fw_bindadd failure leaks fwb+xfer, CBIND/close free binds without their xferlist xfers, fw_open si_drv1 check-then-set race double-allocates fw_drv1
Base system — bus / platform / cpu · sys/bus/firewire/fwdev.c:165 · pass 2 · new |
slow kernel memory exhaustion / open-race panic |
untested |
| DF-5743 |
Low |
3.3 |
netbt uses IFF_LINK0 (0x1000) as an mbuf direction tag, colliding with the core mbuf flag M_LASTFRAG — spurious flag propagation and spoofable SCM_HCI_DIRECTION metadata
Base system — network stack · sys/netbt/hci_socket.c:991 · pass 2 · new |
none |
untested |
| DF-5751 |
Low |
3.3 |
l2cap_connect() failure paths after l2cap_cid_alloc() leak the CID and leave the half-open channel on l2cap_active_list: socket bricked with EISCONN until close, CID space consumed
Base system — network stack · sys/netbt/l2cap_upper.c:166 · pass 2 · new |
CID-space exhaustion + bricked sockets on connect-failure |
untested |
| DF-2689 |
Info |
3.3 |
Leftover DEBUG kprintf in vm_pageout_scan_hold leaks kernel pointer and spams the console on the normal race-recovery path
Kernel memory leaks / uninitialized disclosure · sys/vm/vm_pageout.c:1932 · pass 2 · new |
none |
untested |
| DF-3397 |
Info |
3.3 |
rx-ring init writes 4-byte int into const uint16_t nr_buf_size, clobbering the adjacent flags field (latent — module does not build on DFly)
Kernel memory corruption (UAF / OOB write) · sys/net/netmap/netmap_mem2.c:974 · pass 2 · new |
none |
build_failed |
| DF-3868 |
Info |
3.3 |
Copy-paste bug: 'debug' hint clamps sc->mch_rec instead of sc->dbg_level (and clamping is skipped exactly when the hint exists)
Base system — drivers · sys/dev/sound/pci/emu10kx.c:3081 · pass 2 · new |
log-spam DoS (debug sysctl unbounded) |
untested |
| DF-4358 |
Info |
3.3 |
DF-added kprintf leaks drm_device/amdgpu_device kernel pointers to dmesg at amdgpu attach
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_kms.c:185 · pass 2 · new |
none |
untested |
| DF-4441 |
Info |
3.3 |
init_device kprintf leaks kernel heap pointer (COREBUF %p) to console — readable by unprivileged users via dmesg (security.unprivileged_read_msgbuf=1 default)
Privilege escalation (general use) · sys/dev/smbus/atmel_mxt/atmel_mxt.c:393 · pass 2 · new |
Kernel heap pointer disclosed to dmesg, readable unprivileged (security.unprivileged_read_msgbuf=1 default on guest) |
untested |
| DF-4514 |
Info |
3.3 |
ath_lna_div: dead ath_lna_div_ioctl() unconditionally copyouts uninitialized kernel heap (kmalloc without M_ZERO, executed even after EINVAL) - latent uninitialized-memory disclosure if ever wired up
Privilege escalation (general use) · sys/dev/netif/ath/ath/if_ath_lna_div.c:176 · pass 2 · new |
None in-tree (dead code); latent chosen-length uninitialized kernel-heap disclosure if ever wired to an ioctl (and the diag ioctls themselves lacked priv checks per DF-4320) |
untested |
| DF-4527 |
Info |
3.3 |
HAL_DIAG_ANI_CURRENT copies out the internal ar5212AniState including the kernel 'params' pointer: kernel heap address disclosure
Privilege escalation (general use) · sys/dev/netif/ath/ath_hal/ar5416/ar5416_ani.c:1081 · pass 2 · new |
none |
untested |
| DF-5318 |
Info |
3.3 |
USB_SETDEBUG sets the global usb_debug level with no privilege check (only FWRITE on the fd) — unprivileged kernel-log flooding / console spam
Privilege escalation (general use) · sys/bus/u4b/usb_generic.c:2084 · pass 2 · new |
latent unchecked index (DF-5317) / unpriv log flooding (DF-5318) |
untested |
| DF-3400 |
Low |
3.2 |
nat_del_dispatch never frees nat_state2 entries (free loops commented out): unbounded kernel heap leak on NAT delete/flush
Kernel memory leaks / uninitialized disclosure · sys/net/ipfw3_nat/ip_fw3_nat.c:806 · pass 2 · new |
none |
not_reproduced |
| DF-3670 |
Low |
3.2 |
ixgbe_calc_eeprom_checksum_generic walks NVM sections on NVM-resident pointer/length with u16 index and no word_size bounds — upstream FreeBSD fix missing in DragonFly
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ix/ixgbe_common.c:2214 · pass 2 · new |
none |
inconclusive |
| DF-3031 |
Low |
3.1 |
ext2_readdir never corrects the caller's cookie count — missing '*' deref leaves stale uninitialized kernel heap in the over-reported tail (VOP contract violation, latent NFS-facing heap disclosure)
Base system — filesystems · sys/vfs/ext2fs/ext2_vnops.c:262 · pass 2 · new |
none |
untested |
| DF-3097 |
Low |
3.1 |
HAMMER1 NFS file handles carry zero entropy and fhtovp honors an attacker-chosen as-of TID: forged handles yield historical-version reads and cross-PFS reach on exported mounts
Base system — filesystems · sys/vfs/hammer/hammer_vfsops.c:1083 · pass 2 · new |
none |
untested |
| DF-3209 |
Low |
3.1 |
xdr_string_encode() transmits 1-3 uninitialized kernel-heap bytes (XDR string padding) in every boot mountd MOUNT / NFS LOOKUP request
Kernel memory leaks / uninitialized disclosure · sys/vfs/nfs/nfs_mountrpc.c:143 · pass 2 · new |
leak |
untested |
| DF-3289 |
Low |
3.1 |
Uninitialized stack args[1] transmitted in mesh peer-timeout CLOSE action frames
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_mesh.c:3107 · pass 2 · new |
none |
not_reproduced fixed |
| DF-3313 |
Low |
3.1 |
ieee80211_recv_bar parses BAR control/sequence fields with no frame-length validation — OOB read of up to 10 bytes on short ctl frames
Kernel memory leaks / uninitialized disclosure · sys/netproto/802_11/wlan/ieee80211_ht.c:994 · pass 2 · new |
none |
untested |
| DF-3637 |
Low |
3.1 |
ath_tx_start() discards ath_tx_normal_setup() failure ('done: return 0') — leaked ath_buf + 10-descriptor slab + dmamap per failure, phantom TIM set, caller told success
Base system — drivers · sys/dev/netif/ath/ath/if_ath_tx.c:2021 · pass 2 · new |
none |
not_reproduced |
| DF-3638 |
Low |
3.1 |
ath_tx_retry_clone() dmasetup-failure path: node reference leaked forever and nbf recycled to the free list with bf_m dangling at a freed mbuf
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath/if_ath_tx.c:4568 · pass 2 · new |
none |
not_reproduced |
| DF-4210 |
Low |
3.1 |
Unknown-INOT-status path overwrites the parsed notify entry with a log string and then acknowledges the string instead of the queue entry (f/w gets garbage ack; original Immediate Notify never properly acknowledged)
Base system — drivers · sys/dev/disk/isp/isp_target.c:346 · pass 2 · new |
In the RQSTYPE_NOTIFY default case, ISP_SNPRINTF overwrites the just-parsed Immediate Notify entry in the 64-byte stack buffer with an ASCII log message; isp_print_bytes correctly prints the ORIGINAL entry (vptr) but isp_notify_ack(isp, local) is handed the STRING, which it re-parses as in_entry_t - na_iid/na_seqid/na_status/na_task_flags become ASCII bytes and that garbage Notify-Ack goes to the firmware queue |
untested |
| DF-4520 |
Low |
3.1 |
ar5212_xmit: halTotalQueues taken unclamped from EEPROM AR_EEP_MAXQCU (raw 5-bit, 1..31) bounds ah_txq[HAL_NUM_TX_QUEUES=10] in every queue op -> SetupTxQueue OOB scan read + OS_MEMZERO OOB heap write at attach on crafted EEPROM
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5212/ar5212_xmit.c:137 · pass 2 · new |
none demonstrated (latent OOB kernel heap read + OS_MEMZERO OOB write at driver attach if the NIC EEPROM encodes MAXQCU>10; crafted/corrupted EEPROM = physical-card trust boundary) |
untested fixed |
| DF-4548 |
Low |
3.1 |
ar5212_attach: halKeyCacheSize taken unclamped from 4-bit EEPROM AR_EEP_KCENTRIES field (up to 32768) vs 128-entry hardware key table; every HAL keycache bounds check validates against the attacker-controlled value
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_hal/ar5212/ar5212_attach.c:867 · pass 2 · new |
Latent MMIO OOB-write primitive (key-cache writes past the 128-entry table into PHY register space); defused in-tree by if_ath's ATH_KEYMAX=128 clamp |
untested |
| DF-2704 |
Low |
3.0 |
vop_cache_operate_ap dispatches through mnt_vn_journal_ops unsynchronized against journal_detach kfree (UAF window, root-gated)
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_vopops.c:1781 · pass 2 · new |
none |
untested |
| DF-2969 |
Low |
3.0 |
suspend_kproc holds no reference on the target thread: concurrent kthread_exit yields use-after-free read/write on a recycled struct thread and an unkillable sleeper
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_kthread.c:188 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3237 |
Low |
3.0 |
iod-shutdown race: after nfssvc_iod_stop1() the reader re-queues EINPROGRESS requests onto nm_reqtxq whose only consumer (the writer thread) is already exiting — stranded nfsreq + nfsm_info + bio (biodone never called, buffer stuck B_BUSY), queue head embedded in the nfsmount that nfs_unmount then frees
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_iod.c:134 · pass 2 · new |
leak |
not_reproduced |
| DF-3281 |
Low |
3.0 |
if_addgroup() increments ifg_refcnt outside ifgroup_lock and makes the find-or-create decision non-atomically: lost-update refcount -> premature group destruction with live references (latent UAF/double-free)
Kernel memory corruption (UAF / OOB write) · sys/net/if.c:1174 · pass 2 · new |
none |
not_reproduced |
| DF-3399 |
Low |
3.0 |
Unreplied cross-CPU netmsg in NAT state-add path leaks the kmalloc'd message when dispatch CPU diverges from hash CPU
Base system — network stack · sys/net/ipfw3_nat/ip_fw3_nat.c:385 · pass 2 · new |
none |
not_reproduced |
| DF-3401 |
Low |
3.0 |
ip_fw3_ctl_nat_get_record walks all CPUs' lockless per-CPU NAT state with no synchronization; break-vs-continue truncation
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_nat/ip_fw3_nat.c:585 · pass 2 · new |
none |
not_reproduced |
| DF-3496 |
Low |
3.0 |
ng_fec_rcvmsg treats msg->data as a NUL-terminated interface name with no arglen validation: ifunit()/kprintf %s walk past the ng_mesg allocation (unbounded heap OOB read)
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/ng_fec.c:1287 · pass 2 · new |
none |
not_reproduced |
| DF-3512 |
Low |
3.0 |
ng_bt3c_rcvmsg: NGM_BT3C_NODE_GET_QLEN reads msg->data without arglen validation — heap OOB read
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/bluetooth/drivers/bt3c/ng_bt3c_pccard.c:444 · pass 2 · new |
none |
inconclusive |
| DF-3542 |
Low |
3.0 |
Unsynchronized ng_eiface_next_unit++ races under concurrent node creation — duplicate interface unit numbers and duplicate if_xname ifnets
Privilege escalation (general use) · sys/netgraph/eiface/ng_eiface.c:127 · pass 2 · new |
state-confusion (duplicate if_xname) |
untested |
| DF-3659 |
Low |
3.0 |
ISP_TSK_MGMT ioctl (24xx) double-releases the FC scratch buffer on the error path, breaking fcbsy mutual exclusion
Base system — drivers · sys/dev/disk/isp/isp_freebsd.c:665 · pass 2 · new |
none |
not_reproduced |
| DF-3808 |
Low |
3.0 |
msk_txeof() consumes the device-supplied TX done index (0..4095) with no MSK_TX_RING_CNT (256) range check: bogus index force-completes the entire in-flight TX window (premature dmamap_unload/m_freem on HW_OWNER descriptors)
Base system — drivers · sys/dev/netif/msk/if_msk.c:2884 · pass 2 · new |
none |
untested |
| DF-3846 |
Low |
3.0 |
sc->map_update_cmd is never cleared: shutdown aborts a recycled pooled mfi command (stale handle → wrong-command abort)
Base system — drivers · sys/dev/raid/mrsas/mrsas.c:2953 · pass 2 · new |
none |
inconclusive |
| DF-3856 |
Low |
3.0 |
Unchecked BIOS shift counts >= 32 (C99 6.5.7p3 UB) used to build PLL/GPIO write masks at four parser sites
Base system — drivers · sys/dev/drm/radeon/radeon_combios.c:654 · pass 2 · new |
none |
inconclusive |
| DF-3994 |
Low |
3.0 |
Per-device runtime mutation of shared static radeon_asic dispatch tables (r420 'macs' hack, agp_disable gart swap, IGP clock-op NULL-out) — cross-device dispatch corruption in multi-GPU systems
Base system — drivers · sys/dev/drm/radeon/radeon_asic.c:167 · pass 2 · new |
none |
inconclusive |
| DF-4137 |
Low |
3.0 |
remove_plex_entry compacts vol->plex[] with a byte count instead of element count and decrements vol->plexes even when the plex was not found
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumconfig.c:1849 · pass 2 · new |
none |
untested |
| DF-4138 |
Low |
3.0 |
ppc_pc873xx_detect() static probe index never reset: OOB reads past pc873xx_basetab[] used as x86 I/O port numbers (wild inb/outb)
Kernel memory leaks / uninitialized disclosure · sys/dev/misc/ppc/ppc.c:435 · pass 2 · new |
none |
untested |
| DF-4243 |
Low |
3.0 |
dptexecuteccb send-failure path frees CCB without callout_stop — armed timeout later runs dpttimeout on stale/recycled state (premature abort of unrelated command / UAF read)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/dpt/dpt_scsi.c:778 · pass 2 · new |
none demonstrated on guest (no DPT hardware); source-level impact: stale armed callout after send failure → premature abort of unrelated command / UAF-read on recycled ccb |
untested |
| DF-4294 |
Low |
3.0 |
Kernel heap OOB read via non-NUL-terminated NGM_TEXT_CONFIG data fed to strcmp/strtol
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/musycc/musycc.c:1013 · pass 2 · new |
Kernel heap over-read (match/no-match oracle; page-fault panic at arena boundary). Root-only ngctl + LMC1504 hardware. |
untested |
| DF-4343 |
Low |
3.0 |
mn_attach error-path asymmetry: discarded BAR resource handles (unreleasable), leaking ENXIO returns, failures reported as attach success, dangling sc->node after node destruction, and no device_detach
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mn/if_mn.c:1302 · pass 2 · new |
none |
inconclusive |
| DF-4407 |
Low |
3.0 |
/dev/tpm open/read/write are not serialized across CPUs: TOCTOU on TPM_OPEN plus interleaved TIS transactions
Base system — drivers · sys/dev/crypto/tpm/tpm.c:1155 · pass 2 · new |
dos |
untested |
| DF-4443 |
Low |
3.0 |
vtpci_intr_alloc does not clamp the caller's interrupt count to the fixed vtpci_intr_res[1 + VIRTIO_MAX_VIRTQUEUES] array — latent out-of-bounds softc state write for any child driver requesting more than 9 vectors
Base system — drivers · sys/dev/virtual/virtio/pci/virtio_pci.c:460 · pass 2 · new |
Latent OOB write of softc state (rid ints + kernel pointers) past vtpci_intr_res[1+VIRTIO_MAX_VIRTQUEUES] for any child kmod requesting >9 vectors against a host advertising >=10 MSI-X vectors |
untested |
| DF-4519 |
Low |
3.0 |
ar5212_xmit: TX-queue index q validated only by inert HALASSERT in 5 exported xmit ops; signed q<0 bypasses the >=-only checks in Set/GetTxQueueProps down to &ah_txq[q]
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5212/ar5212_xmit.c:91 · pass 2 · new |
none demonstrated (no unprivileged trigger; latent OOB kernel read/write below ah_txq[0] via negative q, and UB 1<<q / unchecked MMIO addressing, if a future/buggy caller passes an unvalidated queue index) |
untested fixed |
| DF-4534 |
Low |
3.0 |
lptclose() frees the I/O buffers while the lptout callout and the registered interrupt handler can still execute (use-after-free read window)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/lpt/lpt.c:596 · pass 2 · new |
none |
not_reproduced |
| DF-4541 |
Low |
3.0 |
hptmv AS_INITIALIZE_START queues the rebuild DPC with flag VERIFY - initialization never starts; ioctl spins 80s and returns success
Base system — drivers · sys/dev/raid/hptmv/ioctl.c:572 · pass 2 · new |
none |
untested |
| DF-4595 |
Low |
3.0 |
read_drive_label() ignores the read_drive() error and classifies stale/uninitialized heap as a vinum label (wrong-drive acceptance)
Base system — drivers · sys/dev/raid/vinum/vinumio.c:332 · pass 2 · new |
none |
untested |
| DF-4616 |
Low |
3.0 |
ips_morpheus_poll deadline mixes units: seconds + watchdog ticks (10 s each) — crash-dump poll window 10× too short, dump proceeds on unacknowledged writes
Base system — drivers · sys/dev/raid/ips/ips.c:601 · pass 2 · new |
none |
not_reproduced |
| DF-4640 |
Low |
3.0 |
ips: attach-time info-command timeout unwind frees the DMA buffer and recycles the static command slot while the adapter still owns the command - device UAF write / zeroed-callback call on late completion
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ips/ips_commands.c:313 · pass 2 · new |
none |
not_reproduced |
| DF-4678 |
Low |
3.0 |
DragonFly devm_request_mem_region shim is conflict-blind: i915_adjust_stolen's stolen-region safety check is dead code (+ static-rman double-init/leak/off-by-one)
Base system — drivers · sys/dev/drm/i915/i915_gem_stolen.c:83 · pass 2 · new |
none |
untested |
| DF-4703 |
Low |
3.0 |
All fixed-struct VBIOS table reads discard the table size and dereference full structs after checking only frev/crev — OOB read past truncated ATOM data tables
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/ppatomfwctrl.c:61 · pass 2 · new |
Kernel reads past the declared table (potentially past the mapped ROM image): garbage power tables / fault-on-read near the ROM end |
untested |
| DF-4712 |
Low |
3.0 |
ar5211_xmit: signed q<0 bypasses the >=-only checks in ar5211Set/GetTxQueueProps down to &ah_txq[q], and five exported xmit ops gate q only via inert HALASSERT (ar5211 analogue of DF-4519)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5211/ar5211_xmit.c:84 · pass 2 · new |
none demonstrated (hardware-gated: no Atheros NIC on guest; latent kernel OOB read/write below ah_txq[10] via negative q through ar5211SetTxQueueProps/ar5211GetTxQueueProps and unchecked AR_Q* MMIO in five HALASSERT-gated ops, if a future/buggy caller ever passes an unvalidated queue index) |
untested fixed |
| DF-4717 |
Low |
3.0 |
XPT_RESET_DEV sends its FIB under a stale Header.Command (assigned only in the XPT_SCSI_IO case; aac_release_command() never clears it)
Base system — drivers · sys/dev/raid/aac/aac_cam.c:452 · pass 2 · new |
none |
untested |
| DF-4733 |
Low |
3.0 |
ar5210_xmit: signed q<0 bypasses the >=-only checks in ar5210Set/GetTxQueueProps down to ah_txq[3], and five exported xmit ops gate q only via inert HALASSERT (ar5210 twin of DF-4519/DF-4712)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5210/ar5210_xmit.c:40 · pass 2 · new |
none demonstrated (hardware-gated: no Atheros NIC on guest; latent kernel OOB read/write below ah_txq[3] via negative q through ar5210Set/GetTxQueueProps, plus unvalidated ah_txq[q] struct reads in five HALASSERT-gated ops, if a future/buggy caller ever passes an unvalidated queue index) |
untested fixed |
| DF-4798 |
Low |
3.0 |
ad_get_geometry: signed-int overflow in cylinders*heads*sectors (with u16→u8 truncation of IDENTIFY heads/sectors) corrupts total_secs into an ~16-EiB media-size lie
Base system — drivers · sys/dev/disk/nata/ata-disk.c:471 · pass 2 · new |
corruption |
untested |
| DF-4807 |
Low |
3.0 |
ppiwrite() ignores the uiomove() return value: stale ppi_buffer bytes are handed to byte_peripheral_write() and a fully-faulting write never terminates (DF-4535 lpt twin, IEEE1284-peripheral side)
Base system — drivers · sys/dev/misc/ppi/ppi.c:459 · pass 2 · new |
none |
not_reproduced |
| DF-4824 |
Low |
3.0 |
acpi_dock: _EJ0 executed even when device_detach() of an _EJD-dependent child fails — hardware ejected out from under still-attached drivers
Base system — drivers · sys/dev/acpica/acpi_dock/acpi_dock.c:300 · pass 2 · new |
none |
untested |
| DF-4855 |
Low |
3.0 |
agp_amd64_attach error exits leak the GATT and aperture resource and leave a live /dev/agpgart on an unattached device
Base system — drivers · sys/dev/agp/agp_amd64.c:183 · pass 2 · new |
none |
not_reproduced |
| DF-4972 |
Low |
3.0 |
Magic-map desync across master swap leaves dangling drm_file* in the old master's idr → use-after-free bit-write via DRM_IOCTL_AUTH_MAGIC
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_auth.c:65 · pass 2 · new |
corruption |
untested |
| DF-5068 |
Low |
3.0 |
psm_init_power_state_table() ignores get_pp_table_entry() failures: partially-initialized or all-zero power states enter the state table, become boot/current/request states, and are programmed to the SMC
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/pp_psm.c:76 · pass 2 · new |
none |
untested |
| DF-5153 |
Low |
3.0 |
dpt_pci_attach() bus_setup_intr failure leaves XPT-registered SIMs, CAM wildcard paths, shutdown_post_sync handler and softc (on global dpt_softcs) live against released io/irq — phantom CAM bus issues bus_space I/O through a released rman record and never completes a command
Base system — drivers · sys/dev/raid/dpt/dpt_pci.c:159 · pass 2 · new |
none |
untested |
| DF-5167 |
Low |
3.0 |
bt_pci_attach() error unwind after bus_setup_intr() failure leaves an XPT-registered SIM and wildcard CAM path live over released port/IRQ rman records, freed DMA memory and a lockuninit'd softc lock (UAF-class teardown)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/buslogic/bt_pci.c:152 · pass 2 · new |
none |
untested |
| DF-5210 |
Low |
3.0 |
iounmap() passes a raw physical address to pmap_change_attr(), which requires a kernel virtual address — RAM-backed ioremap cleanup writes PTEs via vtopte() on a user-range address (page-fault panic or silent cache-attr corruption)
Base system — drivers · sys/dev/drm/linux_iomapping.c:90 · pass 2 · new |
attach-time BUG_ON panic / user-range vtopte panic / 4B stack disclosure |
untested |
| DF-5461 |
Low |
3.0 |
ichsmb_wait timeout path never aborts the in-flight transaction — late completions are mis-attributed to the next command, which then returns success with stale/zero data
Base system — bus / platform / cpu · sys/bus/smbus/ichsmb/ichsmb.c:637 · pass 2 · new |
none |
untested |
| DF-5503 |
Low |
3.0 |
ioapic.c parameter-validation gap cluster: containment-defeating overlap test, npin<=0 acceptance, negative-pin acceptance in ioapic_gsi(), and completely unvalidated pin/select in ioapic_pin_prog() (the sink of the ABI entry points DF-5351/5352 proved callable on production kernels)
Base system — bus / platform / cpu · sys/platform/pc64/apic/ioapic.c:236 · pass 2 · new |
nested-GSI-space misrouting + undefined IOAPIC register programming |
untested |
| DF-5614 |
Low |
3.0 |
fwdma_map_cb and both fwdma load call sites discard the bus_dmamap_load error — stale/garbage bus_addr would be programmed into OHCI DMA registers (latent protocol violation)
Base system — bus / platform / cpu · sys/bus/firewire/fwdma.c:53 · pass 2 · new |
none demonstrated (latent): garbage/stale fwdma bus_addr programmed into OHCI registers on any bus_dmamap_load failure -> hardware DMA against a wild physical address; unreachable on the current x86_64 busdma because loads of bus_dmamem_alloc'd memory cannot fail there |
untested |
| DF-4776 |
Info |
3.0 |
am7990_start_locked feeds unbounded mbuf chains into lance_put, memcpy'ing into the fixed LEBLEN(1522)-byte TX slot with no clamp (dead-code twin of DF-4746)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/lnc/am7990.c:523 · pass 2 · new |
none |
untested |
| DF-5030 |
Info |
3.0 |
ar5312GetMacAddr dereferences unvalidated BSP board-config pointer; 6-byte wild read becomes the unprivileged-readable Wi-Fi MAC on a SoC build with corrupt flash board-data
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5312/ar5312_attach.c:289 · pass 2 · new |
none |
untested |
| DF-5193 |
Info |
3.0 |
drm_legacy_sg_free clears dev->sg before validating the handle: a failed match permanently orphans the sg allocation and a re-alloc stacks a second one on top (latent: root-only ioctl, no DRIVER_SG/LEGACY driver in tree)
Base system — drivers · sys/dev/drm/drm_scatter.c:131 · pass 2 · new |
none |
not_reproduced |
| DF-5194 |
Info |
3.0 |
Snake movement engine confines itself purely by wall-bounce arithmetic writing through the completely unchecked sc_vtb_putc: any geometry inconsistency escapes and runs away with unbounded 2-byte OOB writes marching through the video window into adjacent DMAP'd memory (latent — no in-tree path produces the trigger)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/snake/snake_saver.c:100 · pass 2 · new |
none |
not_reproduced |
| DF-5197 |
Info |
3.0 |
set_video_mode() return value ignored in warp_saver(): on failure vid aliases the stale 32KB text window and the 64000-byte bzero (plus every star write) overruns it by 32KB into 0xC0000+ (DF-5114 twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/warp/warp_saver.c:87 · pass 2 · new |
none |
untested |
| DF-5202 |
Info |
3.0 |
rain_saver(): set_video_mode() return value ignored — on blank-time failure vid aliases the stale 32KB text window and the 64000-byte bzero + drop-fill overrun it into 0xC0000+, while blanked/vid state commits as if the saver were active (DF-5114/DF-5197/DF-5198 family twin, last saver)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/rain/rain_saver.c:77 · pass 2 · new |
latent out-of-window MMIO writes + text-mode corruption |
untested |
| DF-3965 |
Low |
2.9 |
Per-I/O dmamap unloaded only on IOP_RESULT_SUCCESS: every BAD_TARGET/BUSY/FAIL/CHECK_CONDITION completion leaks a loaded map — deterministic per absent-target bus probe; bounce-pool exhaustion on bounce-capable configs
Base system — drivers · sys/dev/raid/hptiop/hptiop.c:2304 · pass 2 · new |
none |
not_reproduced |
| DF-4094 |
Low |
2.9 |
emu_pci_attach bad: path never unwinds emu_init — leaks ptb_pages/silent_page DMA and mpu401/midi state on every late attach failure
Base system — drivers · sys/dev/sound/pci/emu10k1.c:2114 · pass 2 · new |
none |
inconclusive |
| DF-4372 |
Low |
2.9 |
Unvalidated VBIOS container offsets in r600_parse_extended_power_table: usFanTableOffset and usExtendendedHeaderOffset dereferenced with no zero/bounds check — offset 0 aliases the PowerPlayInfo header as the extended header; out-of-range offsets read past the kmalloc'd BIOS image (sibling of DF-1566: offset provenance, not counts)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r600_dpm.c:893 · pass 2 · new |
none |
untested |
| DF-4573 |
Low |
2.9 |
DragonFly i915 PCI-ID matcher drops subvendor/subdevice criteria: Quanta entry hijacks all 0x016a IvyBridge devices into the headless num_pipes=0 info
Base system — drivers · sys/dev/drm/i915/i915_pci.c:810 · pass 2 · new |
Display loss (headless num_pipes=0 binding) on all non-Quanta 0x016a IvyBridge GT2-server iGPUs; no memory corruption |
untested |
| DF-4890 |
Low |
2.9 |
AcpiOsAcquireLock implements one ACPICA lock as two disjoint primitives (idle thread → spinlock, everyone else → lockmgr+crit): no mutual exclusion between the two holder classes across CPUs
Base system — drivers · sys/dev/acpica/Osd/OsdSynch.c:383 · pass 2 · new |
none |
not_reproduced |
| DF-5002 |
Low |
2.9 |
miibus leaks every kmalloc'd newbus ivars (mii_attach_args per PHY child, mii_probe_args per bus) — nothing in the kernel ever frees them
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mii_layer/mii.c:157 · pass 2 · new |
none |
untested |
| DF-5241 |
Low |
2.9 |
PROBE_INQUIRY_BASIC_DV1/DV2: bcmp() reads SHORT_INQUIRY_LENGTH (36) bytes from a kmalloc(additional_length+5) buffer as small as 6 bytes (device-controlled heap OOB read)
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/cam_xpt.c:5867 · pass 2 · new |
heap OOB comparison oracle / race-gated UAF / accounting drift |
untested |
| DF-3664 |
Info |
2.9 |
bge_nvram_getbyte leaks the NVRAM software-arbitration lock and leaves NVRAM access enabled on its two timeout paths (BCM5906)
Base system — drivers · sys/dev/netif/bge/if_bge.c:557 · pass 2 · new |
none |
inconclusive |
| DF-5095 |
Info |
2.9 |
MFI_DEBUG timeout helpers print kernel heap pointer (cm) and the pmap_kextract() physical address of command data buffers to the world-readable kernel message buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mfi/mfi_debug.c:197 · pass 2 · new |
KVA + physical address disclosure to dmesg in debug builds |
untested |
| DF-4081 |
Low |
2.8 |
oce_tso_setup reads ether/IP/TCP headers (incl. th->th_off at offset up to ~74) before its m_pullup — out-of-contract mbuf reads on split-header TSO frames (DF-1373 class, different function)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/oce/oce_if.c:1046 · pass 2 · new |
none |
untested |
| DF-4795 |
Low |
2.8 |
ig_hal VF: uninitialized kernel-stack tail (up to 50 of 60 payload bytes) of the E1000_VF_SET_MULTICAST mailbox message transmitted verbatim to the PF (hypervisor)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ig_hal/e1000_vf.c:403 · pass 2 · new |
none demonstrated at runtime (no SR-IOV VF hardware on guest); source-proven disclosure class: up to 50 of 60 payload bytes of the E1000_VF_SET_MULTICAST mailbox message are uninitialized guest kernel stack, copied verbatim by e1000_write_mbx_vf (e1000_mbx.c:476-477) into E1000_VMBMEM registers read by the host PF driver — guest-kernel-stack residue crosses the guest->host boundary on every multicast update with <30 groups (the common case) |
untested |
| DF-4974 |
Low |
2.8 |
radeon_crtc_cursor_set2 never validates cursor BO size — display engine scans up to 16 KiB from an arbitrarily small pinned BO (VRAM OOB read to screen)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_cursor.c:295 · pass 2 · new |
none |
not_reproduced |
| DF-5184 |
Low |
2.8 |
wake_up_bit() ignores the bit number and wakes exactly ONE waiter on the word: wrong-waiter wakeups and N-1 stranded waiters for N concurrent waiters on one word (no per-(addr,bit) wait identity, fanout of one)
Base system — drivers · sys/dev/drm/linux_wait.c:97 · pass 2 · new |
stolen wakeups / stranded waiters (bounded by timeouts + accidental 1Hz repoll) |
untested |
| DF-5537 |
Low |
2.8 |
vconswinch_intr() manipulates tty token-protected state (t_winsize write, t_pgrp read + pgsignal) without tp->t_token — INTR_MPSAFE lock-protocol violation with theoretical pgrp use-after-free (self-pinned in practice) and vconsole double-load vs vconsterm
Kernel memory corruption (UAF / OOB write) · sys/platform/vkernel64/platform/console.c:276 · pass 2 · new |
none |
not_reproduced |
| DF-5589 |
Info |
2.8 |
pmap_inval.c discards every madvise()/vmspace_mcontrol() return value — the vkernel's sole memory-revocation primitive fails silently, leaving stale (possibly writable) host PTEs with no diagnostic
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/pmap_inval.c:85 · pass 2 · new |
silent loss of the sole memory-revocation primitive on any future failure |
untested |
| DF-4418 |
Low |
2.7 |
e1000_phy_init_script_82541 ignores read_reg() failures and writes uninitialized kernel stack into PHY registers 0x2F5B / ANALOG_FUSE_CONTROL
Base system — drivers · sys/dev/netif/ig_hal/e1000_82541.c:1108 · pass 2 · new |
none |
untested |
| DF-3775 |
Low |
2.6 |
DHRS-error path bypasses the blockable filter: blocking error recovery (tsleep-based pwaits, 1-2s each) executes in hard-interrupt context
Base system — drivers · sys/dev/disk/ahci/ahci.c:2843 · pass 2 · new |
none |
not_reproduced |
| DF-4315 |
Low |
2.6 |
Device-fed rs_rate through rateCodeToIndex[] 0xFF sentinel indexes sc_hwmap[32] out of bounds in ath_rx_tap/dump/led paths (OOB read exported into radiotap)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath/if_ath_rx.c:577 · pass 2 · new |
none |
not_reproduced |
| DF-3753 |
Info |
2.6 |
Unvalidated legacy-rate lookups: rt2860_rates[255] OOB read (tx) and 1<<255 UB (set_basicrates, peer-fed rateset)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ral/rt2860.c:1576 · pass 2 · new |
none |
not_reproduced |
| DF-3107 |
Medium |
2.5 |
ufs_checkpath uses vp->v_mount after vput() drops intermediate ancestors to refcount 0 — read of unreferenced/reclaimable vnode on every deep rename (FreeBSD caches mp; ext2_checkpath shares the bug)
Kernel memory corruption (UAF / OOB write) · sys/vfs/ufs/ufs_lookup.c:1158 · pass 2 · new |
none |
not_reproduced |
| DF-4903 |
Medium |
2.5 |
Uninitialised heap read in k8_powernow_setperf tail leaks a heap word into world-readable sysctl machdep.powernow.frequency.current
Kernel memory leaks / uninitialized disclosure · sys/dev/powermng/powernow/powernow.c:223 · pass 2 · new |
leak |
untested |
| DF-2628 |
Low |
2.5 |
hammer2_vop_nlink increments meta.nlinks even when hammer2_dirent_create fails — permanent link-count inflation and unfreeable inodes
Hammer2 · sys/vfs/hammer2/hammer2_vnops.c:1599 · pass 2 · new |
none |
inconclusive |
| DF-2736 |
Low |
2.5 |
vop_compat_* '[diagnostic] EAGAIN on ncp %p %s' prints leak kernel namecache heap pointer and full pathname into the world-readable msgbuf
Base system — kernel core (kern) · sys/kern/vfs_default.c:200 · pass 2 · new |
none |
not_reproduced |
| DF-2758 |
Low |
2.5 |
FOFFSETLOCK acquisition sleeps uninterruptibly ('fpoff', no PCATCH): a thread wedged forever inside VOP_READ while holding the offset lock makes every other thread sharing the file descriptor permanently unkillable
Base system — kernel core (kern) · sys/kern/vfs_vnops.c:569 · pass 2 · new |
none |
not_reproduced |
| DF-2821 |
Low |
2.5 |
rand_thread_loop feeds generator output back into the entropy pools, so the MIN_POOL_SIZE reseed gate can be satisfied with zero external entropy
Base system — kernel core (kern) · sys/kern/kern_nrandom.c:886 · pass 2 · new |
none |
not_reproduced |
| DF-2858 |
Low |
2.5 |
shmfork() bcopy's transient shmmap_s reservation state: fork landing in shmat's blocked window permanently burns a SHMSEG slot in the child (and gives it an un-accounted, un-shmdt-able mapping)
Base system — kernel core (kern) · sys/kern/sysv_shm.c:646 · pass 2 · new |
none |
not_reproduced |
| DF-2868 |
Low |
2.5 |
ktrcanset dereferences targetp->p_ucred with no token and no crhold — freed-cred read / stale-cred TOCTOU racing cratom_proc (distinct from DF-0103's p_tracenode scope)
Base system — kernel core (kern) · sys/kern/kern_ktrace.c:668 · pass 2 · new |
none |
inconclusive |
| DF-2885 |
Low |
2.5 |
kern_ptrace permission checks dereference tracer (and pre-token target) ucred pointers with no crhold and no token — freed-cred read racing cratom_proc (DF-2868 analog in sys_process.c)
Kernel memory corruption (UAF / OOB write) · sys/kern/sys_process.c:153 · pass 2 · new |
panic |
not_reproduced |
| DF-2958 |
Low |
2.5 |
vkernel64: sys_umtx_sleep() calls blocking uservtophys() inside a critical section — LWKT lock-protocol violation
Base system — kernel core (kern) · sys/kern/kern_umtx.c:199 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-2991 |
Low |
2.5 |
handle_workitem_freeblocks: indir_trunc errors silently swallowed — '== 0' dead store makes allerror unreachable, I/O errors during indirect-block deallocation unreported
Base system — filesystems · sys/vfs/ufs/ffs_softdep.c:2168 · pass 2 · new |
corruption |
not_reproduced |
| DF-3032 |
Low |
2.5 |
ext2_rename strands IN_RENAME on the source directory inode when the step-3 relookup race resolves the from-name to a different inode — directory permanently unrenameable (EINVAL) until reclaim
Base system — filesystems · sys/vfs/ext2fs/ext2_vnops.c:1025 · pass 2 · new |
none |
untested |
| DF-3058 |
Low |
2.5 |
dirfs_fsync returns 0 unconditionally — host fsync() failures are swallowed (and the double-fsync retry stores -1, not errno)
Base system — filesystems · sys/vfs/dirfs/dirfs_vnops.c:588 · pass 2 · new |
none |
not_reproduced |
| DF-3105 |
Low |
2.5 |
64-bit file-offset truncation family beyond DF-0833: udf_read's int offset, and udf_strategy/udf_bmap passing off_t bio_offset/loffset into udf_bmap_internal's uint32_t parameter
Base system — filesystems · sys/vfs/udf/udf_vnops.c:399 · pass 2 · new |
none |
not_reproduced |
| DF-3110 |
Low |
2.5 |
hammer_cursor_moved_element() again2 parent-cursor leaf fixup indexes the PARENT's element array with the CHILD's index (oindex instead of pindex) — cursor->leaf left stale inside a node the cursor no longer references during rebalance
Base system — filesystems · sys/vfs/hammer/hammer_cursor.c:1054 · pass 2 · new |
corruption |
not_reproduced |
| DF-3129 |
Low |
2.5 |
msdosfs_sync() never initializes scaninfo.waitfor — uninitialized kernel-stack read gates metadata flushing: sync(2)/umount can silently skip fsync of dirty denodes depending on stack garbage's MNT_LAZY bit (DF-specific regression; NetBSD sets ctx.waitfor, FreeBSD passes waitfor directly)
Base system — filesystems · sys/vfs/msdosfs/msdosfs_vfsops.c:787 · pass 2 · new |
corruption |
not_reproduced |
| DF-3160 |
Low |
2.5 |
Lost-wakeup window in fifo_open() FIFO rendezvous (uninterlocked check-then-sleep, no re-check loop)
Base system — filesystems · sys/vfs/fifofs/fifo_vnops.c:295 · pass 2 · new |
none |
not_reproduced |
| DF-3186 |
Low |
2.5 |
hammer_upgrade_undo_4 never terminates on UNDO_ALIGN-incongruent alloc_offset in a crafted volume header — unkillable kernel thread under finalize_lock+sync_lock, mount hang and buffer-cache deadlock
Base system — filesystems · sys/vfs/hammer/hammer_undo.c:381 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3206 |
Low |
2.5 |
cd9660_lookup evaluates shared dp->i_ino after vn_unlock in the DOTDOT path: concurrent lookup race can bind '..' to an arbitrary concurrently-scanned inode (wrong parent vnode) or spuriously fail
Privilege escalation (general use) · sys/vfs/isofs/cd9660/cd9660_lookup.c:343 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3219 |
Low |
2.5 |
dirfs_root lazy-init of dmp->dm_root is completely unlocked — concurrent VFS_ROOT creates two root nodes, leaks a dnp + host fd forever, and can produce a mount root vnode without VROOT
Privilege escalation (general use) · sys/vfs/dirfs/dirfs_vfsops.c:234 · pass 2 · new |
denial_of_service |
not_reproduced |
| DF-3334 |
Low |
2.5 |
netmap_mem_get_info() failure paths unbalance the adapter refcount: premature free/UAF (NIOCREGIF) and pinned adapters (NIOCGINFO)
Kernel memory corruption (UAF / OOB write) · sys/net/netmap/netmap.c:1331 · pass 2 · new |
none |
untested |
| DF-3372 |
Low |
2.5 |
sockaddr address fields read before sa_len validation in tcp_usr_bind/connect/preconnect and tcp6_usr_bind/connect (DF-3357 family): heap read of recycled/uninitialized bytes past the user-copied length (v6: up to 16 bytes past the allocation)
Kernel memory leaks / uninitialized disclosure · sys/netinet/tcp_usrreq.c:316 · pass 2 · new |
none |
not_reproduced |
| DF-3484 |
Low |
2.5 |
Global netgraph queue (ngqbase/ngqlast/ngqfree) manipulated under crit_enter only: cross-CPU producer races silently orphan queue entries, permanently leaking their mbufs/messages and pinning node/hook references
Kernel memory corruption (UAF / OOB write) · sys/netgraph/netgraph/ng_base.c:1871 · pass 2 · new |
none |
not_reproduced |
| DF-3572 |
Low |
2.5 |
u32 wrap in SKL DDB data-rate sum over-subscribes the per-pipe DDB (un-carried Linux v5.2 u64 rework)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_pm.c:4216 · pass 2 · new |
none |
not_reproduced |
| DF-3675 |
Low |
2.5 |
Unsigned resid underflow: ccb->csio.resid = dxfer_len - IOC-supplied TransferCount with no clamp
Base system — drivers · sys/dev/disk/mpt/mpt_cam.c:3132 · pass 2 · new |
none |
inconclusive |
| DF-3723 |
Low |
2.5 |
ixgbe_checksum_ptr_x550/ixgbe_calc_checksum_X550 caller-buffer path: bounds checks ignore the ptr base offset (plus two off-by-ones) — OOB read past the caller's u16 buffer[]
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ix/ixgbe_x550.c:3358 · pass 2 · new |
none |
inconclusive |
| DF-3791 |
Low |
2.5 |
drm_atomic_helper: missing backport of upstream 022debad063e (drm_atomic_state->duplicated) — drm_atomic_helper_resume() fails after connector unregistered during suspend, leaving all outputs off
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/drm_atomic_helper.c:325 · pass 2 · new |
none |
inconclusive |
| DF-3803 |
Low |
2.5 |
mprsas_prepare_for_tm leaks one union ccb + cam_path per TM reuse/escalation during error recovery
Base system — drivers · sys/dev/raid/mpr/mpr_sas.c:3801 · pass 2 · new |
none |
not_reproduced |
| DF-3804 |
Low |
2.5 |
mprsas_check_eedp error path calls xpt_free_path(NULL) instead of freeing local_path: deterministic NULL-deref panic under memory pressure during EEDP probe
Base system — drivers · sys/dev/raid/mpr/mpr_sas.c:3689 · pass 2 · new |
none |
not_reproduced |
| DF-3826 |
Low |
2.5 |
e1000_read_i2c_byte_generic double-releases the SW_FW_SYNC PHY0 semaphore on retry exhaustion
Base system — drivers · sys/dev/netif/ig_hal/e1000_82575.c:3333 · pass 2 · new |
none |
not_reproduced |
| DF-3827 |
Low |
2.5 |
Unchecked PHY register read in e1000_set_d0_lplu_state_82575 writes stale data into IGP01E1000_PHY_PORT_CONFIG
Base system — drivers · sys/dev/netif/ig_hal/e1000_82575.c:833 · pass 2 · new |
none |
not_reproduced |
| DF-3829 |
Low |
2.5 |
NVMM VMX: host NMIs are silently swallowed while a vCPU is resident on the CPU
Privilege escalation (general use) · sys/dev/virtual/nvmm/x86/nvmm_x86_vmx.c:1252 · pass 2 · new |
none |
inconclusive |
| DF-3836 |
Low |
2.5 |
static channel-pointer in bwn_phy_lp_b2063_switch_channel defeats the not-found NULL check after the first call — unsupported channels silently re-program the previous channel's RF values and return success
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn_phy_lp.c:794 · pass 2 · new |
none |
inconclusive |
| DF-3840 |
Low |
2.5 |
1-byte use-after-free read in iwi_frame_intr(): frame->rate read at :1353 from the old rx mbuf after ieee80211_input() consumed/freed it (softled enabled by default)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/iwi/if_iwi.c:1353 · pass 2 · new |
none |
inconclusive |
| DF-3851 |
Low |
2.5 |
dc_setfilt_21143/dc_setfilt_xircom enqueue the setup frame with no TX-ring-space check: descriptor clobber of an in-flight frame, leaked mbuf, and tx_cnt overrun when invoked on a saturated ring
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/dc/if_dc.c:1055 · pass 2 · new |
TX-power correctness (bwn) / mbuf leak + ring wedge (dc) |
untested |
| DF-3896 |
Low |
2.5 |
Unsigned resid underflow on the DATA_UNDERRUN path: csio->resid = cm_length − IOC-supplied TransferCount with no clamp (DF-3675 twin)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mps/mps_sas.c:1943 · pass 2 · new |
none |
not_reproduced |
| DF-3901 |
Low |
2.5 |
Uninitialized stack consumed as SKL DRAM channel info when one MAD_DIMM channel register is blank: ch0/ch1 rank/is_16gb_dimm/symmetric_memory steer watermark config
Base system — drivers · sys/dev/drm/i915/i915_drv.c:1219 · pass 2 · new |
none |
inconclusive |
| DF-3916 |
Low |
2.5 |
_mapping_get_dpm_idx_from_id reads PhysicalIdentifier once from entry 0 and compares every DPM entry against it - DPM lookups only ever match entry 0
Base system — drivers · sys/dev/raid/mpr/mpr_mapping.c:547 · pass 2 · new |
logic/persistence corruption (3916/3917), error-path panic (3918), firmware-fed overread (3919) |
untested |
| DF-3917 |
Low |
2.5 |
_mapping_find_enc_map_space clears the WRONG enclosure entry (et_entry instead of enc_entry) and its overlap test is off by one, clearing a map slot beyond the enclosure
Base system — drivers · sys/dev/raid/mpr/mpr_mapping.c:1076 · pass 2 · new |
logic/persistence corruption (3916/3917), error-path panic (3918), firmware-fed overread (3919) |
untested |
| DF-3919 |
Low |
2.5 |
Event element-count walks (SAS topo PHY, PCIe topo Port, IR config Element) never validated against the event frame length
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_mapping.c:2917 · pass 2 · new |
logic/persistence corruption (3916/3917), error-path panic (3918), firmware-fed overread (3919) |
untested |
| DF-3936 |
Low |
2.5 |
ena_com_set_hash_function() inverts the device-support check — RSS hash function is never programmed on compliant devices
Base system — drivers · sys/dev/virtual/amazon/ena/ena-com/ena_com.c:2238 · pass 2 · new |
none |
inconclusive |
| DF-4010 |
Low |
2.5 |
mpr_user_pass_thru leaks the user-sized cm->cm_data buffer and its dmamap mapping across the timeout/reinit path
Base system — drivers · sys/dev/raid/mpr/mpr_user.c:1050 · pass 2 · new |
none |
untested |
| DF-4039 |
Low |
2.5 |
polaris10_get_dependency_volt_by_clk fall-through reads dep_table->entries[count] twice (one-entry heap OOB read)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:396 · pass 2 · new |
none |
inconclusive fixed |
| DF-4041 |
Low |
2.5 |
polaris10 MVDD stash walks: SmioTable2.Pattern reads entries[count] instead of entries[level]; populate_mvdd_value walks the raw MCLK-dependency count past mvdd_voltage_table.entries[32]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:657 · pass 2 · new |
none |
inconclusive fixed |
| DF-4065 |
Low |
2.5 |
Missing linux-4.19.y stable backports 690e0ec8e63d + 85636167e320: legacy ring buffers still allocated from stolen memory and BAR/WC-mapped on LLC platforms — upstream direction is 'never do this on LLC' (GPU hangs/wedges); DFly carries the pre-fix v4.19 state
Base system — drivers · sys/dev/drm/i915/intel_ringbuffer.c:1096 · pass 2 · new |
Missing linux-4.19.y stable backports 690e0ec8e63d + 85636167e320: legacy ring buffers still allocated from stolen memory and BAR/WC-mapped on LLC platforms (SNB+) - upstream direction 'never do this on LLC' (GPU hangs/wedges, display wedge/GPU DoS). |
untested |
| DF-4072 |
Low |
2.5 |
vegam_get_dependency_volt_by_clk reads dep_table->entries[count] one record past the heap allocation in the clock-above-max path
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:614 · pass 2 · new |
none |
inconclusive |
| DF-4087 |
Low |
2.5 |
_mapping_get_dpm_idx_from_id reads PhysicalIdentifier once from entry 0 and compares every DPM entry against it (mps twin of DF-3916)
Base system — drivers · sys/dev/raid/mps/mps_mapping.c:512 · pass 2 · new |
mapping-table/persistence corruption (logic class); no direct OOB |
untested |
| DF-4088 |
Low |
2.5 |
_mapping_find_enc_map_space clears the WRONG enclosure entry (et_entry instead of enc_entry), off-by-one overlap test, stale num_found (mps twin of DF-3917)
Base system — drivers · sys/dev/raid/mps/mps_mapping.c:802 · pass 2 · new |
mapping-table/persistence corruption (logic class); no direct OOB |
untested |
| DF-4089 |
Low |
2.5 |
Event element-count walks (SAS topo NumEntries, IR NumElements) never validated against the event frame length (mps twin of DF-3919)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps_mapping.c:2043 · pass 2 · new |
OOB read (up to ~2KB) of kmalloc'd event frame; garbage feeds mapping engine (kernleak-class, no direct copyout traced) |
untested |
| DF-4090 |
Low |
2.5 |
dpm_pg0_sz truncated to u16 in mps_mapping_initialize: DPM page silently under-read (benign direction proven)
Base system — drivers · sys/dev/raid/mps/mps_mapping.c:1721 · pass 2 · new |
none (functional: DPM entries silently under-read; overflow direction provably impossible) |
untested |
| DF-4091 |
Low |
2.5 |
dce110_timing_generator_validate_timing: h_blank/h_back_porch u32 underflow bypasses the minimum checks; no vertical checks -> program_blanking programs wrapped (masked) blanking registers
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_timing_generator.c:1109 · pass 2 · new |
none at kernel-memory level (masked MMIO only); worst case inconsistent CRTC timing -> display corruption/possible pipe hang on real ASIC |
untested |
| DF-4142 |
Low |
2.5 |
wi_rx_intr ignores wi_read_bap() failure: uninitialized kernel heap delivered to the 802.11 stack and BPF (device-fed info disclosure)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/wi/if_wi.c:1356 · pass 2 · new |
none |
inconclusive |
| DF-4166 |
Low |
2.5 |
get_gpio_i2c_info off-by-one (count < table_index) reads gpio_pin[count] one entry past the GPIO pin LUT
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/bios/bios_parser2.c:369 · pass 2 · new |
none |
untested |
| DF-4182 |
Low |
2.5 |
Page flip programmed on a CRTC disabled while the flip was pending wedges that CRTC's flip path forever (flip_status never leaves RADEON_FLIP_SUBMITTED) and leaks the vblank reference
Base system — drivers · sys/dev/drm/radeon/radeon_display.c:449 · pass 2 · new |
none |
inconclusive |
| DF-4197 |
Low |
2.5 |
DragonFly's Linux-compat seqlock shim returns the seqcount to its pre-write value — drm_vblank_count_and_time() accepts torn {count,time} vblank snapshots (protocol violation)
Base system — drivers · sys/dev/drm/drm_vblank.c:784 · pass 2 · new |
none |
inconclusive |
| DF-4213 |
Low |
2.5 |
16-byte memcpy out of the 4-byte MSG_RAID_ACTION_REPLY.ActionData publishes stale reply-frame bytes as RAID resync progress
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt_raid.c:578 · pass 2 · new |
none |
untested |
| DF-4265 |
Low |
2.5 |
amdgpu_cs_wait_all_fences reads fence->error after the final dma_fence_put (use-after-free read, value returned as ioctl errno)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_cs.c:1501 · pass 2 · new |
none |
untested |
| DF-4305 |
Low |
2.5 |
bfe_intr error reinit performs M_WAITOK cluster allocations in interrupt context: BFE_ISTAT_ERRORS -> bfe_init -> bfe_newbuf(init=1) sleeps in the ithread holding the serializer
Base system — drivers · sys/dev/netif/bfe/if_bfe.c:1186 · pass 2 · new |
none |
not_reproduced |
| DF-4306 |
Low |
2.5 |
e1000_release_swfw_sync_80003es2lan acquires the SWSM hardware semaphore in an unbounded empty-body loop on every PHY/KMRN/NVM release path: a stuck SMBI bit pins a CPU forever under the driver serializer
Base system — drivers · sys/dev/netif/ig_hal/e1000_80003es2lan.c:455 · pass 2 · new |
none |
not_reproduced |
| DF-4318 |
Low |
2.5 |
dce80/dce81/dce83_create_resource_pool leak struct dce110_resource_pool on construct() failure (missing kfree at 3 sites)
Base system — drivers · sys/dev/drm/amd/display/dc/dce80/dce80_resource.c:1052 · pass 2 · new |
none |
untested |
| DF-4360 |
Low |
2.5 |
cik_sdma_soft_reset(): stale status variable — SDMA1-busy test reads mmSDMA0_F32_CNTL instead of mmSRBM_STATUS2, so a hung SDMA1 is never soft-reset
Base system — drivers · sys/dev/drm/amd/amdgpu/cik_sdma.c:1089 · pass 2 · new |
none |
untested |
| DF-4377 |
Low |
2.5 |
probeCard() consumes uninitialized stack eeprom[256] when readEEProm() fails — uninitialized stack bytes reach console/dmesg and drive card identification (CWE-457)
Base system — drivers · sys/dev/video/bktr/bktr_card.c:605 · pass 2 · new |
none |
not_reproduced |
| DF-4386 |
Low |
2.5 |
hw.oceN.firmware_version served via unbounded strlen over a possibly non-NUL-terminated device string — adjacent softc bytes (incl. kernel pointers) leak to unprivileged readers
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/oce/oce_sysctl.c:84 · pass 2 · new |
none |
untested |
| DF-4387 |
Low |
2.5 |
Device-controlled sc->port_id indexes rxf.port[2]/port[4] unchecked in copy_stats_to_sc_be2/_be3 — OOB read past the stats DMA buffer into world-readable sysctls
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/oce/oce_sysctl.c:1151 · pass 2 · new |
none |
untested |
| DF-4389 |
Low |
2.5 |
hwmgr->num_vce_state_tables stored unclamped (up to 255) over vce_states[6] — fill loop silently skipped, count left as a loaded gun for any consumer
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/process_pptables_v1_0.c:1328 · pass 2 · new |
none |
untested |
| DF-4437 |
Low |
2.5 |
drm_mode_page_flip_ioctl: -EDEADLK retry bypasses the only drm_crtc_vblank_put - one vblank reference leaked per page flip that succeeds after a deadlock retry
Base system — drivers · sys/dev/drm/drm_plane.c:1053 · pass 2 · new |
none |
untested |
| DF-4455 |
Low |
2.5 |
Unbounded MMIO spin in PLL atomic-update wait can wedge the DRM master permanently
Base system — drivers · sys/dev/drm/radeon/radeon_legacy_crtc.c:219 · pass 2 · new |
Permanent DRM-stack DoS: modeset ioctl never returns, holding mode_config.mutex and modeset locks forever |
untested |
| DF-4477 |
Low |
2.5 |
ata_reinit drops a running request already flagged ATA_R_TIMEOUT after callout_cancel: request/buffer leak and permanently hung issuer
Base system — drivers · sys/dev/disk/nata/ata-all.c:205 · pass 2 · new |
none |
not_reproduced |
| DF-4490 |
Low |
2.5 |
acpi_ec: ec_sci_pend publication race between EcGpeHandler and EcGpeQueryHandler permanently disables the EC GPE (lost AcpiFinishGpe)
Base system — drivers · sys/dev/acpica/acpi_ec.c:693 · pass 2 · new |
Permanent loss of EC-mediated event delivery (lid suspend, brightness/hotkeys, battery/AC/thermal notifies) plus transient EC I/O degradation to polled mode; availability only |
untested |
| DF-4518 |
Low |
2.5 |
vke_stop() teardown can hang the vkernel forever: the tx cothread's done-fifo full-wait never checks the exit flag, and once the exit flag is set vke_tx_intr() refuses to drain the done fifo, so cothread_delete()'s pthread_join blocks while holding if_serializer
Base system — drivers · sys/dev/virtual/vkernel/net/if_vke.c:170 · pass 2 · new |
Unprivileged guest user permanently wedges the entire vkernel (ifconfig never returns; serializer held forever); vkernel-internal only, no host-kernel impact |
untested |
| DF-4522 |
Low |
2.5 |
amdgpu_vm_fini's DragonFly cleanup loop iterates on vm->va.rb_leftmost, which the DFly rb_root_cached shim never maintains: zero iterations, all leftover VM mappings leak
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_vm.c:3325 · pass 2 · new |
none |
inconclusive |
| DF-4523 |
Low |
2.5 |
cik_sdma_ib_test leaks the IB (SA-bo) and its fence on fence-wait error/timeout — reachable after every GPU reset while a DMA engine stays wedged
Base system — drivers · sys/dev/drm/radeon/cik_sdma.c:738 · pass 2 · new |
One IB SA-bo + fence leaked from the fixed-size shared pool per unrecovered reset cycle - progressive pool exhaustion = persistent GPU-accel DoS |
untested |
| DF-4524 |
Low |
2.5 |
vce_v3_0: VCE bo is sized exactly AMDGPU_VCE_FIRMWARE_OFFSET (256) bytes short — instance-1 DATA window extends past the end of vcpu_bo
Base system — drivers · sys/dev/drm/amd/amdgpu/vce_v3_0.c:427 · pass 2 · new |
Firmware writes to the last 256 bytes of the instance-1 DATA window land past vcpu_bo into the adjacent VRAM object (speculative - fw-defined; arithmetic overrun certain) |
untested |
| DF-4530 |
Low |
2.5 |
vchan_trigger commits trigger state and parent busy-list membership before chn_notify() and never rolls back on failure, enabling a masked-retry fake-running channel (per-fd wedge)
Base system — drivers · sys/dev/sound/pcm/vchan.c:144 · pass 2 · new |
Per-fd dsp wedge: fake-running channel blocks writes/poll until close; no memory-safety impact (busy-list membership stays paired) |
untested |
| DF-4572 |
Low |
2.5 |
exca_activate_resource leaks the ExCA window when parent BUS_ACTIVATE_RESOURCE fails (and exca_mem_map on its memreg error return) — permanent per-socket window-exhaustion DoS
Base system — drivers · sys/dev/pccard/exca/exca.c:250 · pass 2 · new |
none |
inconclusive |
| DF-4577 |
Low |
2.5 |
intel_hdcp_validate_v_prime KSV fill shift missing -sha_leftovers offset: repeater SHA-1 input stream corrupted for >=2 downstream devices, HDCP1.4 repeater authentication always fails (fail-closed)
Base system — drivers · sys/dev/drm/i915/intel_hdcp.c:221 · pass 2 · new |
none |
untested |
| DF-4581 |
Low |
2.5 |
radeon_cs_parser_relocs: unchecked radeon_vm_get_bos() NULL return silently drops the VM page directory and page tables from the CS reservation/validation set — submission then updates the PD/PTs without holding their reservations (stale-address GPU writes, unsynchronized eviction); latent on DF as compiled, live upstream and after any DF-4580-style allocator fix
Base system — drivers · sys/dev/drm/radeon/radeon_cs.c:190 · pass 2 · new |
none |
not_reproduced |
| DF-4632 |
Low |
2.5 |
radeon MST connector leaks one EDID blob per probe and per destroy (radeon_connector->edid overwritten without kfree)
Base system — drivers · sys/dev/drm/radeon/radeon_dp_mst.c:193 · pass 2 · new |
none |
untested |
| DF-4695 |
Low |
2.5 |
sili_pm_softreset: error = EBUSY on unexpected device-signature change is dead-stored by the unconditional error = 0 — signature-change detection never propagates, probe proceeds with stale at_type
Base system — drivers · sys/dev/disk/sili/sili_pm.c:489 · pass 2 · new |
none |
inconclusive |
| DF-4700 |
Low |
2.5 |
snd_hdspe PCM: hdspechan_setspeed() read-modify-writes the shared ctrl_register and writes CONTROL/FREQ registers without sc->lock — cross-channel lost-update race (interrupt storm / stuck-enable / rate scramble)
Base system — drivers · sys/dev/sound/pci/hdspe-pcm.c:486 · pass 2 · new |
none |
untested |
| DF-4745 |
Low |
2.5 |
am79900_start_locked lacks the upstream 'break' after detecting a LANCE-owned TMD: the driver dequeues an mbuf and overwrites the chip-owned descriptor and TX buffer on a ring desync
Base system — drivers · sys/dev/netif/lnc/am79900.c:540 · pass 2 · new |
none |
not_reproduced |
| DF-4775 |
Low |
2.5 |
am7990_start_locked lacks the 'break' after detecting a LANCE-owned TMD (dead-code twin of DF-4745): on a ring desync the driver dequeues an mbuf and overwrites the chip-owned descriptor and TX buffer
Base system — drivers · sys/dev/netif/lnc/am7990.c:503 · pass 2 · new |
dos |
untested |
| DF-4778 |
Low |
2.5 |
aux_engine: read-continuation chunks copied to buffer BASE instead of buffer+offset — corruption of partial-ACK DPCD/EDID read data (twin in dce_aux.c)
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/aux_engine.c:204 · pass 2 · new |
none |
untested |
| DF-4870 |
Low |
2.5 |
r600_dma_ib_test leaks the IB (suballocator chunk + fence ref) on both fence-wait error paths
Base system — drivers · sys/dev/drm/radeon/r600_dma.c:371 · pass 2 · new |
none |
untested |
| DF-4871 |
Low |
2.5 |
r600_copy_dma discards radeon_sync_resv()/radeon_sync_rings() failures — copy committed without cross-ring ordering, silent bo data corruption
Base system — drivers · sys/dev/drm/radeon/r600_dma.c:466 · pass 2 · new |
none |
untested |
| DF-4879 |
Low |
2.5 |
drm_mode_config_init() ignores drm_mode_create_standard_properties() failure: NULL standard properties are attached to every universal plane and atomic CRTC, turning any later unprivileged MODE_OBJ_GETPROPERTIES/GETCONNECTOR ioctl into a kernel NULL-deref panic (upstream-inherited, ENOMEM-gated)
Base system — drivers · sys/dev/drm/drm_mode_config.c:217 · pass 2 · new |
none |
untested |
| DF-4904 |
Low |
2.5 |
Unsigned underflow in Phase-1 voltage ramp writes malformed MSR 0xc0010041 (reserved bits 13-31 garbage, wrong VID)
Base system — drivers · sys/dev/powermng/powernow/powernow.c:231 · pass 2 · new |
dos |
untested |
| DF-4906 |
Low |
2.5 |
READ_PENDING_WAIT spins forever on a stuck FID/VID change-pending bit (unkillable kernel-context CPU spin)
Base system — drivers · sys/dev/powermng/powernow/powernow.c:84 · pass 2 · new |
dos |
untested |
| DF-4920 |
Low |
2.5 |
Recursive drm_modeset_lock (same ctx, same lock) double-tracks and corrupts ww accounting: ctx->acquired off-by-one leak, ww->ctx/ww->stamp cleared while lock still held, latent lockmgr LK_RELEASE panic
Base system — drivers · sys/dev/drm/drm_modeset_lock.c:269 · pass 2 · new |
none |
untested |
| DF-4969 |
Low |
2.5 |
adw_pci_attach error unwinds leak the PCI register/IRQ resources (and the softc on the setup() path): adw_free() never calls bus_release_resource
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/advansys/adw_pci.c:211 · pass 2 · new |
none |
untested |
| DF-4983 |
Low |
2.5 |
cxm_iic_detach()'s CXM_IVAR_IICBUS invalidation is a permanent no-op (write_ivar guard rejects clears) — cxm_softc.iicbus dangles to a destroyed device_t after iic teardown
Base system — drivers · sys/dev/video/cxm/cxm_i2c.c:229 · pass 2 · new |
none |
untested |
| DF-5035 |
Low |
2.5 |
Unsynchronized check-then-set of i2c_hw_buffer_in_use lets concurrent acquires on different DDC lines both hold the single shared DCE11 DC_I2C HW engine — acquire-side twin of DF-4878's release-side race (DAL2 mutex dropped)
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.c:105 · pass 2 · new |
none |
untested |
| DF-5089 |
Low |
2.5 |
si_copy_dma discards radeon_sync_resv()/radeon_sync_rings() returns — SI twin of DF-4871: DMA copy committed without cross-ring ordering on sync failure, silent bo data corruption
Base system — drivers · sys/dev/drm/radeon/si_dma.c:255 · pass 2 · new |
none |
untested |
| DF-5111 |
Low |
2.5 |
amdgpu_job_dependency busy-loops forever on amdgpu_vmid_grab allocation failure, wedging the ring scheduler kthread (unbounded spin + DRM_ERROR flood, ring-wide GPU DoS)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_job.c:193 · pass 2 · new |
sched kthread wedge / ring-wide GPU DoS |
untested |
| DF-5181 |
Low |
2.5 |
smu9_send_msg_to_smc/_with_parameter return 0 on SMC failure — vacuous caller error guards and stale smu9_get_argument() data flow into Vega10/12 powerplay decisions
Base system — drivers · sys/dev/drm/amd/powerplay/smumgr/smu9_smumgr.c:113 · pass 2 · new |
none |
untested |
| DF-5209 |
Low |
2.5 |
pci_iomap() treats pci_resource_start()'s -1 failure sentinel as a valid base and drives __ioremap_common() into BUG_ON panic for any absent/unallocatable BAR
Base system — drivers · sys/dev/drm/linux_iomapping.c:47 · pass 2 · new |
attach-time BUG_ON panic / user-range vtopte panic / 4B stack disclosure |
untested |
| DF-5219 |
Low |
2.5 |
i915_timeline kref leak: intel_engine_create_ring error path frees the ring without dropping its timeline reference
Base system — drivers · sys/dev/drm/i915/i915_timeline.c:1190 · pass 2 · new |
bounded refcount/heap leak (DF-5219) / latent API trap (DF-5221) |
untested |
| DF-5256 |
Low |
2.5 |
DragonFly-only unlocked traversal of ifp->if_multiaddrs in urtwn_set_multi (upstream takes if_maddr_rlock)
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_urtwn.c:4898 · pass 2 · new |
deterministic 640KB/cycle kernel heap leak (DF-5255) / maddr list race panic (DF-5256) / remote taskqueue wedge (DF-5257) |
untested |
| DF-5267 |
Low |
2.5 |
Five linear rate lookups store MTW_RIDX_MAX (36) on a miss; the stale ridx then indexes rt2860_rates[36] one element out of bounds (mtw_tx/mtw_set_tx_desc/tx_mgt/radiotap wt_rate)
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/wlan/if_mtw.c:1720 · pass 2 · new |
none |
untested |
| DF-5307 |
Low |
2.5 |
Unvalidated descriptor-field reads in the UVC entity/CS parsers: OOB reads up to 253 bytes past malformed descriptors
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/uvc/uvc_drv.c:1843 · pass 2 · new |
none |
untested |
| DF-5350 |
Low |
2.5 |
Unsynchronized dmat->segments allocation and map_count bookkeeping race under concurrent map create/destroy (double allocation leak; tag UAF window)
Kernel memory corruption (UAF / OOB write) · sys/platform/pc64/x86_64/busdma_machdep.c:337 · pass 2 · new |
device-side out-of-window DMA (DF-5349) / segments leak + tag UAF window (DF-5350) |
untested |
| DF-5373 |
Low |
2.5 |
bus_dmamap_create() leaks the freshly kmalloc'd map when the initial bounce-page allocation partially fails (vkernel never got pc64's kfree(*mapp) error-path fix)
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/busdma_machdep.c:333 · pass 2 · new |
device-visible corrupted S/G lists (DF-5371) / 72B-per-failure leak (DF-5373) |
untested |
| DF-5386 |
Low |
2.5 |
TX-abort condition is read and discarded — writes to a non-ACKing slave return SMB_ENOERR success (silent write failure)
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:482 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-5446 |
Low |
2.5 |
Unchecked 4-byte logical-ID read on short PNP_TAG_LOGICAL_DEVICE descriptor (≤3-byte overread; worst case past the allocation into uninitialized heap)
Kernel memory leaks / uninitialized disclosure · sys/bus/isa/pnp.c:444 · pass 2 · new |
none |
untested |
| DF-5609 |
Low |
2.5 |
cothread_create/cothread_delete ignore pthread_create()/pthread_join() failure — EAGAIN leaves a registered virtual interrupt, a NULL pthread handle, and undefined pthread_kill/pthread_join on it
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/cothread.c:110 · pass 2 · new |
vkernel host crash / dead device channel (5609); <=1s systimer latency spikes (5610) |
untested |
| DF-5610 |
Low |
2.5 |
cothread_wakeup()'s vktimer_ts zeroing races vktimer_thread's reload store — wakeup can be silently consumed, delaying all systimers by up to the stale sleep
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/cothread.c:188 · pass 2 · new |
vkernel host crash / dead device channel (5609); <=1s systimer latency spikes (5610) |
untested |
| DF-5685 |
Low |
2.5 |
Unchecked copyin() return in sys_vquotactl feeds an uninitialized stack plistref to _prop_object_copyin
Base system — kernel core (kern) · sys/kern/vfs_quota.c:345 · pass 2 · new |
see summary |
untested |
| DF-5705 |
Low |
2.5 |
_prop_stack_push growth allocation is kmalloc(M_WAITOK) executed while callers hold the global proplib dictionary lockmgr (release walk) or deep per-object mtx chains (equals walk); and push-failure handling is kernel-dead code since M_WAITOK never returns NULL on DragonFly
Base system — libkern / libprop / ddb · sys/libprop/prop_stack.c:50 · pass 2 · new |
none |
untested |
| DF-5783 |
Low |
2.5 |
i915: intel_legacy_cursor_update() fast path mutates the LIVE crtc state's active_planes before a fallible section — failure leaves state desynchronized from hardware (missing upstream c249c5f64336), breaking subsequent plane disable / IPS / watermark decisions
Base system — drivers · sys/dev/drm/i915/intel_display.c:13686 · pass 2 · new |
none |
untested |
| DF-2671 |
Info |
2.5 |
extattr syscalls copyin() the 255-byte attribute name without a NUL-termination guarantee (latent stack OOB read)
Kernel memory leaks / uninitialized disclosure · sys/kern/vfs_syscalls.c:5224 · pass 2 · new |
none |
untested |
| DF-2701 |
Info |
2.5 |
swap_pager_getpage() 20-second timeout diagnostic dereferences the completed-and-released pbuf (use-after-free read in console path)
Kernel memory corruption (UAF / OOB write) · sys/vm/swap_pager.c:1527 · pass 2 · new |
none |
untested |
| DF-3725 |
Info |
2.5 |
Dropped hardware-write statuses across X550 link/flow-control setup: failed MMIO/MDIO writes reported as success
Base system — drivers · sys/dev/netif/ix/ixgbe_x550.c:1812 · pass 2 · new |
none |
inconclusive |
| DF-3762 |
Info |
2.5 |
Unchecked PHY-read returns across the shared e1000 PHY layer: link-stall workaround branches on uninitialized kernel stack; power up/down write back a zeroed PHY_CONTROL after failed reads (DF-3643/DF-3671 family, e1000_phy.c instance)
Base system — drivers · sys/dev/netif/ig_hal/e1000_phy.c:3848 · pass 2 · new |
none |
inconclusive |
| DF-3770 |
Info |
2.5 |
DFly port deleted the upstream pre-B3 ValleyView PPGTT disable from intel_sanitize_enable_ppgtt: PPGTT enabled on silicon where upstream considers it broken (GPU fault/hang wedge on that hardware)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/i915_gem_gtt.c:136 · pass 2 · new |
none |
untested |
| DF-3927 |
Info |
2.5 |
siba_pcicore_intr computes 1 << tmp with MMIO-derived tmp up to 63 (SIBA_TPS_BPFLAG=0x3f) - UB shift silently enables the wrong INTR_MASK bit on pre-rev6 PCI cores
Base system — drivers · sys/dev/netif/bwn/siba/siba_core.c:2131 · pass 2 · new |
none |
not_reproduced |
| DF-3944 |
Info |
2.5 |
patch_gamut_packet_checksum reads up to 227 bytes past dc_info_packet.sb[32] (length byte unbounded; dead path — no producer ever sets gamut.valid)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/core/dc_resource.c:2152 · pass 2 · new |
none |
not_reproduced |
| DF-4176 |
Info |
2.5 |
iir_timeout is an empty handler: SCSI timeouts never abort, retry, or complete — timed-out CCBs hang forever; also references an undeclared identifier under GDT_DEBUG
Base system — drivers · sys/dev/raid/iir/iir.c:1472 · pass 2 · new |
gdtexecuteccb arms a callout to iir_timeout, but the body is only a GDT_DPRINTF - no abort, no retry, no xpt_done, no gdt_next |
untested |
| DF-4427 |
Info |
2.5 |
acpi_hp_notify / acpi_hp_evaluate_auto_on_off run without the hp serial lock, racing sysctl handlers on softc state
Base system — drivers · sys/dev/acpica/acpi_hp/acpi_hp.c:809 · pass 2 · new |
Lost/duplicated WLAN/BT/WWAN on-air notifications; benign int races, no memory-safety consequence |
untested |
| DF-4601 |
Info |
2.5 |
dce112_compressor_construct() init-order defects: dead 64-bit-bus LPT guard + lpt_channels_num wipe make program_lpt_control's only valid case unreachable (LPT armed with unprogrammed config)
Base system — drivers · sys/dev/drm/amd/display/dc/dce112/dce112_compressor.c:793 · pass 2 · new |
none |
not_reproduced |
| DF-4642 |
Info |
2.5 |
dce_virtual_pageflip dereferences crtcs[crtc_id] before the crtc_id bounds check (latent OOB read)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/dce_virtual.c:659 · pass 2 · new |
none |
untested |
| DF-4663 |
Info |
2.5 |
dc_link_ddc: unchecked dal_ddc_aux/i2c_payloads_create() NULL in dal_ddc_service_query_ddc_data — OOM-gated (and zero-size-deterministic) NULL deref
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc_link_ddc.c:578 · pass 2 · new |
none |
not_reproduced fixed |
| DF-4664 |
Info |
2.5 |
dc_link_ddc: uninitialized type2_dongle_buf[0x20] hexdumped to the kernel log on dongle-detect i2c failure (drm.debug)
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc_link_ddc.c:436 · pass 2 · new |
none |
not_reproduced fixed |
| DF-4728 |
Info |
2.5 |
aureal: AC97 codec access busy-spins up to ~0.35s per register and ignores timeouts (au_wrcd writes anyway and always returns success)
Base system — drivers · sys/dev/sound/pci/aureal.c:126 · pass 2 · new |
none |
untested |
| DF-4839 |
Info |
2.5 |
BUG_ON panic in drm_primary_helper_update() when SetPlane targets an enabled CRTC left with zero encoder-bound connectors
Base system — drivers · sys/dev/drm/drm_plane_helper.c:264 · pass 2 · new |
none |
not_reproduced |
| DF-5067 |
Info |
2.5 |
power_state_management() NULL-checks pcurrent then dereferences it anyway — latent kernel NULL-deref (NULL check at :249 routes control INTO the :254 deref; :248 passes the pointer to the smu8 backend unchecked)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/pp_psm.c:246 · pass 2 · new |
none |
untested |
| DF-5113 |
Info |
2.5 |
Spectral scan start/configure permanently disables radar-detect hardware with no restore path — latent DFS-detection bypass (ar5416DisableRadar clobbers AR_PHY_RADAR_0_ENA/EXT and AR_RX_FILTER; ar5416StopSpectralScan restores none of it)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5416/ar5416_spectral.c:47 · pass 2 · new |
none |
untested |
| DF-5137 |
Info |
2.5 |
kv_notify_message_to_smu() fail-open SMC response validation (radeon twin of DF-5132): timeout (0x0000) and all undefined response codes except 0x00FF/0x00FE reported as success to the entire kv_dpm power-management state machine
Base system — drivers · sys/dev/drm/radeon/kv_smc.c:44 · pass 2 · new |
none |
untested |
| DF-5148 |
Info |
2.5 |
AUX completion poll off-by-one: a DisplayPort AUX transaction that completes on the 1001st poll is misreported as -EIO (counter-keyed timeout gate)
Base system — drivers · sys/dev/drm/radeon/radeon_dp_auxch.c:152 · pass 2 · new |
none |
untested |
| DF-5163 |
Info |
2.5 |
amdgpu atom-i2c consumes the shared ATOM fb scratch after amdgpu_atom_execute_table released ctx->mutex — radeon's scratch_mutex fix was never ported; concurrent hw-DDC/atom-fb users deliver cross-contaminated DDC data
Base system — drivers · sys/dev/drm/amd/amdgpu/atombios_i2c.c:52 · pass 2 · new |
none |
not_reproduced |
| DF-5205 |
Info |
2.5 |
trinity_notify_message_to_smu() fail-open SMC response validation: timeout (0x0) and all undefined codes outside exactly 0xFF/0xFE reported as success to the entire trinity_dpm power-management state machine — and unlike the kv twin (DF-5137), the response register is read UNMASKED
Base system — drivers · sys/dev/drm/radeon/trinity_smc.c:43 · pass 2 · new |
none |
untested |
| DF-5206 |
Info |
2.5 |
trinity_acquire_mutex() is void and cannot fail: on SMC mutex-grant timeout it silently returns without the mutex, and all six live callers then program SMC registers assuming exclusivity against the SMC firmware
Base system — drivers · sys/dev/drm/radeon/trinity_smc.c:112 · pass 2 · new |
none |
untested |
| DF-5213 |
Info |
2.5 |
amdgpu df_v3_6_funcs is an incomplete vtable: .enable_ecc_force_par_wr_rmw left NULL (latent NULL function-pointer dereference for any Vega20 caller)
Base system — drivers · sys/dev/drm/amd/amdgpu/df_v3_6.c:108 · pass 2 · new |
none |
not_reproduced |
| DF-5216 |
Info |
2.5 |
interval_tree misuse degrades to unkillable kernel spin or panic instead of a bounded failure — remove() on an empty root NULL-derefs/panics, remove() of a non-member spins forever holding the caller's mutex, iter_next() on a detached sole node loops or returns the removed node
Base system — drivers · sys/dev/drm/linux_interval_tree.c:56 · pass 2 · new |
latent interior kfree / kernel-CPU amplification / unkillable spin |
untested |
| DF-5526 |
Info |
2.5 |
npxprobemask() lacks the mxcsr_mask==0 fallback guard: a non-conforming FXSAVE probe zeroes npx_mxcsr_mask, forcing MXCSR=0 (all SIMD FP exceptions unmasked) and a SIGFPE on every FPU restore
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/npx.c:134 · pass 2 · new |
MXCSR=0 + SIGFPE on every restore on mxcsr_mask==0 hardware |
untested |
| DF-5557 |
Info |
2.5 |
npxprobemask() lacks the mxcsr_mask==0 fallback guard — non-conforming FXSAVE probe zeroes npx_mxcsr_mask and SIGFPEs every FPU restore (DF-5526 twin)
Base system — bus / platform / cpu · sys/platform/vkernel64/x86_64/npx.c:112 · pass 2 · new |
none |
untested |
| DF-5591 |
Info |
2.5 |
vga_probe_unit() returns success without initializing the caller's video_adapter_t — uninit kernel-stack read programmed into world-readable devinfo resource ranges (structurally unreachable today: both sources gated on one 'optional vga')
Kernel memory leaks / uninitialized disclosure · sys/bus/isa/vga_isa.c:108 · pass 2 · new |
none |
not_reproduced |
| DF-5687 |
Info |
2.5 |
_prop_object_copyout() leaks the vm_mmap()ed user region when copyout() fails (no vm_map_remove on the error path)
Base system — libkern / libprop / ddb · sys/libprop/prop_kern.c:526 · pass 2 · new |
see summary |
untested |
| DF-3627 |
Low |
2.4 |
Uninitialized stack array consumed as eDP link rates when DP_SUPPORTED_LINK_RATES DPCD read fails
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_dp.c:3949 · pass 2 · new |
none |
inconclusive |
| DF-3971 |
Low |
2.4 |
rt2560 beacon-ring slot overwrite leaks the old beacon mbuf and its ieee80211_node reference on every re-entry to RUN — remotely pump-able in IBSS mode (ibss merge)
Base system — drivers · sys/dev/netif/ral/rt2560.c:820 · pass 2 · new |
none |
not_reproduced |
| DF-3985 |
Low |
2.4 |
sbp_free_target early-returns on luns==NULL without clearing target->fwdev or stopping scan_callout — target slots permanently leaked holding dangling fw_device pointers (8-round hostile device-churn attach DoS), plus freed-fw_device derefs via armed mgm/ccb timeouts
Base system — drivers · sys/dev/disk/sbp/sbp.c:1993 · pass 2 · new |
none |
inconclusive |
| DF-4077 |
Low |
2.4 |
Reply-frame range check off-by-one in mps_intr_locked accepts a ReplyFrameAddress exactly one frame past the end
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps.c:1541 · pass 2 · new |
none |
inconclusive |
| DF-4340 |
Low |
2.4 |
Undefined shift counts throughout the ATOM interpreter: VBIOS-controlled shift operands (up to 32-bit) used unmasked in shl/shr/shift ops, WS OR/AND masks and IIO masks
Base system — drivers · sys/dev/drm/radeon/atom.c:904 · pass 2 · new |
none |
inconclusive |
| DF-4348 |
Low |
2.4 |
Undefined shift counts throughout the amdgpu ATOM interpreter: VBIOS-controlled shift operands (up to full 32-bit) used unmasked in shl/shr/shift ops, WS OR/AND masks and IIO width masks (DF-4340's amdgpu twin)
Base system — drivers · sys/dev/drm/amd/amdgpu/atom.c:121 · pass 2 · new |
none |
inconclusive |
| DF-4556 |
Low |
2.4 |
dce_aux: partial AUX reads reassembled at buffer base instead of running offset — kernel-assembled DPCD/EDID corruption reported as SUCCEEDED
Base system — drivers · sys/dev/drm/amd/display/dc/dce/dce_aux.c:426 · pass 2 · new |
none |
not_reproduced |
| DF-4589 |
Low |
2.4 |
radeon_dp_link_train_cr has no overall iteration bound: malicious DP sink oscillating adjust-request voltage wedges the modeset path forever (unkillable task + display-subsystem lock)
Base system — drivers · sys/dev/drm/radeon/atombios_dp.c:684 · pass 2 · new |
none |
not_reproduced |
| DF-4637 |
Low |
2.4 |
vpoio: uninterruptible 500K/5M-iteration handshake spins (VP0_FAST_SPINTMO/VP0_LOW_SPINTMO) execute under vpo_intr's hard critical section — a malicious parallel-port ZIP emulator pins a CPU with interrupts masked for seconds per I/O (DF-4618 family, vpoio.c instance)
Base system — drivers · sys/dev/disk/vpo/vpoio.c:59 · pass 2 · new |
Multi-second uninterruptible CPU pin (preemption disabled, interrupts deferred) per I/O against a malicious/faulty parallel-port ZIP emulator; repeatable per command — local/physical DoS (DF-4618 family, vpio instance) |
untested |
| DF-4652 |
Low |
2.4 |
amdgpu_atombios_dp_link_train_cr has no overall iteration bound: malicious DP sink oscillating adjust-request voltage wedges the modeset path forever (unkillable task + mode_config.mutex) — line-for-line twin of DF-4589 (radeon)
Base system — drivers · sys/dev/drm/amd/amdgpu/atombios_dp.c:605 · pass 2 · new |
none |
not_reproduced |
| DF-4866 |
Low |
2.4 |
siba_bwn_attach leaks the MMIO BAR, bus mastering and scanned child devices when siba_core_attach fails (bad-SPROM CRC or unsupported-PHY child)
Base system — drivers · sys/dev/netif/bwn/siba/siba_bwn.c:151 · pass 2 · new |
leak |
untested |
| DF-5251 |
Low |
2.4 |
run_stop early-return on failed USB read skips run_unsetup_tx_list; the next run_init_locked->run_setup_tx_list memsets the queues, orphaning all queued mbufs and node references (repeatable leak with a wedged/hostile device)
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_run.c:6201 · pass 2 · new |
UAF read/write on freed node (DF-5250) / repeatable mbuf+node leak DoS (DF-5251) / device-bound 4B OOB (DF-5252) |
untested |
| DF-5270 |
Low |
2.4 |
Last-TRB TD Size clear masks only 4 of the 5 TDSZ field bits (15 instead of 31) — controller mispredicts remaining packets on large transfers
Base system — bus / platform / cpu · sys/bus/u4b/controller/xhci.c:2016 · pass 2 · new |
streams-endpoint transfer hangs (DF-5269) / large-transfer miscompletion (DF-5270) |
untested |
| DF-5359 |
Low |
2.4 |
uipaq_attach() stalls the per-bus USB explore thread up to ~12.8s on an unresponsive crafted device: 64 unconditional retries of the magic SET_CONTROL_LINE_STATE request (100ms timeout + ~100ms pause each) with no early-out on stall or device-removal, while holding the device enum lock
Base system — bus / platform / cpu · sys/bus/u4b/serial/uipaq.c:1129 · pass 2 · new |
none |
untested |
| DF-5400 |
Info |
2.4 |
mos_bulk_read_callback() passes actlen instead of pktlen to uether_rxbuf(): every received frame is delivered one byte too long (trailing device status byte included), and transfers of 2045..2056 bytes are silently dropped without ierrors
Base system — bus / platform / cpu · sys/bus/u4b/net/if_mos.c:817 · pass 2 · new |
non-canonical frames up the stack (DF-5400) / skewed interface stats (DF-5401) |
untested |
| DF-3440 |
Medium |
2.3 |
div_output() reads sin_port and sin_addr from an under-sized sendto() destination sockaddr (offsets 4..8 and 8..12 past a >=4-byte allocation); OOB bytes steer re-inject direction and the ipfw resume rule
Kernel memory leaks / uninitialized disclosure · sys/netinet/ip_divert.c:344 · pass 2 · new |
none |
not_reproduced |
| DF-2639 |
Low |
2.3 |
hammer2_inode_create_pfs never sets *errorp on its done2 error paths — snapshot ioctl silently succeeds without creating anything
Hammer2 · sys/vfs/hammer2/hammer2_inode.c:1042 · pass 2 · new |
none |
untested |
| DF-2699 |
Low |
2.3 |
vfs_setpublicfs() leaks nfs_pub.np_index when MNT_EXPUBLIC is re-exported on the same mount without MNT_DELEXPORT
Base system — kernel core (kern) · sys/kern/vfs_subr.c:2246 · pass 2 · new |
none |
not_reproduced |
| DF-2897 |
Low |
2.3 |
Mute-interaction bookkeeping leak: cnclose()'s muted early-return never clears cn_is_open/cn_phys_is_open, so un-muting re-opens the physical console device with a stale flags snapshot for a /dev/console that is no longer open
Base system — kernel core (kern) · sys/kern/tty_cons.c:403 · pass 2 · new |
none |
untested |
| DF-3130 |
Low |
2.3 |
update_mp() re-executed on every remount with fspec re-opens all four iconv handles without closing the previous ones — unbounded kernel-heap handle leak via mount -u loop (partial-failure remounts leak the replaced handles too)
Base system — filesystems · sys/vfs/msdosfs/msdosfs_vfsops.c:115 · pass 2 · new |
leak |
not_reproduced |
| DF-3166 |
Low |
2.3 |
Memory leaks on udf_mountfs/udf_vget error paths: s_table freed only in udf_unmount, unode->fentry leaked when udf_allocv fails
Base system — filesystems · sys/vfs/udf/udf_vfsops.c:393 · pass 2 · new |
leak |
not_reproduced |
| DF-3189 |
Low |
2.3 |
hammer_ioc_set_pseudofs clobbers hammer_load_pseudofs's error with copyin's result — every load failure (not just the intended ENOENT) is silently ignored and the PFS is created/saved anyway
Base system — filesystems · sys/vfs/hammer/hammer_pfs.c:117 · pass 2 · new |
corruption |
not_reproduced |
| DF-3214 |
Low |
2.3 |
netaddr_match union type confusion for RC_NAM cache entries: rc_nam pointer compared as an IPv4 address in the dup-reply-cache hash walk — spoofed-source false cache hits and a silently inert duplicate cache for non-AF_INET transports
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_srvcache.c:194 · pass 2 · new |
leak |
not_reproduced |
| DF-3262 |
Low |
2.3 |
O_VIA/O_RECV/O_XMIT name[IFNAMSIZ] not required NUL-terminated: kfnmatch() pattern scan runs past the 16-byte field toward/past the rule allocation
Kernel memory leaks / uninitialized disclosure · sys/net/ipfw/ip_fw2.c:1355 · pass 2 · new |
leak |
not_reproduced |
| DF-3497 |
Low |
2.3 |
ng_fec_delport (NGM_FEC_DEL_IFACE) frees the portlist entry without releasing the port's multicast list: M_DEVBUF ng_fec_mc structs and ifmultiaddr memberships leak on every add/del cycle
Base system — network stack · sys/netgraph7/ng_fec.c:514 · pass 2 · new |
none |
not_reproduced |
| DF-3532 |
Low |
2.3 |
ng7 ng_tty: ngt_open() error path leaks sc when ng_name_node() fails
Base system — network stack · sys/netgraph7/tty/ng_tty.c:212 · pass 2 · new |
none |
inconclusive |
| DF-3768 |
Low |
2.3 |
DFly i915 registers the GGTT aperture fictitious range on every attach but never unregisters it: unbounded fake-vm_page leak and permanent 8-slot exhaustion (graphics attach DoS) via kldload/kldunload cycles
Base system — drivers · sys/dev/drm/i915/i915_gem_gtt.c:2966 · pass 2 · new |
none |
not_reproduced |
| DF-3861 |
Low |
2.3 |
amdgpu_device_fini never unregisters the fictitious VRAM range — permanent vm_page-array leak and exhaustion of the 8-slot DRM shim (EBUSY driver-load DoS)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_device.c:2648 · pass 2 · new |
none |
not_reproduced |
| DF-3900 |
Low |
2.3 |
i915 MMIO register mapping (pmap_mapdev KVA + global iomap node + shim bus resource) leaked on every driver detach: the only pci_iounmap in the i915 tree is #if 0'd
Base system — drivers · sys/dev/drm/i915/i915_drv.c:1027 · pass 2 · new |
none |
inconclusive |
| DF-3970 |
Low |
2.3 |
rt2560_set_chan walks the unterminated RT2560_RF5222 channel table with no bound — OOB read across kernel .rodata and garbage RF programming when an admin-installed channel misses the table
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ral/rt2560.c:2210 · pass 2 · new |
none |
not_reproduced |
| DF-4061 |
Low |
2.3 |
Linux-compat ioctl 0xc0046d00 re-reads a 110-byte struct amr_linux_ioctl from the 4-byte IOC buffer: 106 bytes of uninitialized kernel stack become ali.inlen/outlen, opcode, mailbox and passthrough pointers that steer the whole management ioctl
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/amr/amr.c:765 · pass 2 · new |
Control-flow nondeterminism on a root-only surface: stack-garbage-chosen opcode/mailbox/pointers; no exfiltration (4-byte copy-back returns the user's own inlen). |
untested |
| DF-4135 |
Low |
2.3 |
config_volume readpol case falls through into setupstate: VF_CONFIG_SETUPSTATE silently set for every volume configured with readpol
Base system — drivers · sys/dev/raid/vinum/vinumconfig.c:1527 · pass 2 · new |
none |
untested |
| DF-4229 |
Low |
2.3 |
cyapawrite returns uninitialized stack `error` on zero-length writes — 32 bits of stale kernel stack leak to userland as the write() result/errno
Kernel memory leaks / uninitialized disclosure · sys/dev/smbus/cyapa/cyapa.c:791 · pass 2 · new |
none |
untested |
| DF-4257 |
Low |
2.3 |
my_softc my_ldata_ptr (2056-byte DMA descriptor block) is never freed: leaked on every detach and every failed attach
Base system — drivers · sys/dev/netif/my/if_my.c:870 · pass 2 · new |
none |
inconclusive |
| DF-4269 |
Low |
2.3 |
wb_attach() allocates sc->wb_ldata but wb_detach() frees never-assigned sc->wb_ldata_ptr: ~296KB contigmalloc leak per attach/detach cycle
Base system — drivers · sys/dev/netif/wb/if_wb.c:779 · pass 2 · new |
none |
untested |
| DF-4378 |
Low |
2.3 |
DRM minor teardown leaks the make_dev'd /dev/dri/cardN cdev and the kzalloc'd struct device; stale devfs node keeps si_drv1 dangling at the freed drm_device
Base system — drivers · sys/dev/drm/drm_drv.c:158 · pass 2 · new |
none |
not_reproduced |
| DF-4397 |
Low |
2.3 |
vega10_pp_tables_uninitialize leaks 8 subtable allocations (socclk/dcefclk/pixclk/phyclk/dispclk dep tables, pcie_table, valid_socclk_values, valid_dcefclk_values) on every pp_table reload / rebind cycle
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/vega10_processpptables.c:1183 · pass 2 · new |
none |
untested |
| DF-4553 |
Low |
2.3 |
snd_vibes: sv_suspend() aborts the playback channel with the record trigger (copy-paste) — playback DMA engine left enabled across suspend
Base system — drivers · sys/dev/sound/pci/vibes.c:644 · pass 2 · new |
none |
not_reproduced |
| DF-4579 |
Low |
2.3 |
radeon: every failed PCI attach leaks one drm_device — DragonFly's drm_get_pci_dev has its entire error unwind #if 0'd out (upstream err_free: drm_dev_unref deleted); radeon_pci_probe is the only live caller
Base system — drivers · sys/dev/drm/radeon/radeon_drv.c:318 · pass 2 · new |
dos |
not_reproduced |
| DF-4596 |
Low |
2.3 |
set_drive_parms() re-Mallocs drive->freelist without freeing the previous allocation (leak on drive re-init)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumio.c:143 · pass 2 · new |
none |
untested |
| DF-4597 |
Low |
2.3 |
daemon_save_config() 'break' aborts the whole drive loop on the first malformed drive — later drives' config replicas never updated
Base system — drivers · sys/dev/raid/vinum/vinumio.c:601 · pass 2 · new |
none |
untested |
| DF-4655 |
Low |
2.3 |
twa_bus_scan_cb frees the scan ccb with the wrong malloc type and leaks its timeout_ch callout on every bus scan
Base system — drivers · sys/dev/raid/twa/tw_osl_cam.c:578 · pass 2 · new |
Kernel heap leak (sizeof(struct callout) per bus scan, unbounded across controller resets) + wrong-type free corrupting malloc accounting |
untested |
| DF-4752 |
Low |
2.3 |
gpio_unregister leaks the master device's clone unit and the shared struct gpio_driver (unlinked while sibling controllers still reference it)
Base system — drivers · sys/dev/misc/gpio/gpio.c:587 · pass 2 · new |
leak |
inconclusive |
| DF-4860 |
Low |
2.3 |
Uninterruptible multi-second waits and 750ms DELAY busy-spin held under dev_lock in the TIS transmit path (DF-1991's TIS twin plus a new busy-wait): 40s unkillable sleep for user-chosen Create* command codes, ~750ms+ CPU spin per burst on a stalled TPM
Base system — drivers · sys/dev/crypto/tpm/tpm_tis.c:258 · pass 2 · new |
dos |
untested |
| DF-4989 |
Low |
2.3 |
TWS_IOCTL_GET_LOCK result dead-stored: time_remaining_msec/grant state computed but never copied back to the caller (advisory ioctl lock unobservable)
Base system — drivers · sys/dev/raid/tws/tws_user.c:280 · pass 2 · new |
none |
not_reproduced |
| DF-4991 |
Low |
2.3 |
SMB_BREAD passes uninitialized stack byte as the driver's requested count; user's clamped rcount never communicated
Kernel memory leaks / uninitialized disclosure · sys/dev/smbus/smb/smb.c:217 · pass 2 · new |
corruption |
untested |
| DF-5072 |
Low |
2.3 |
Silent 64→32-bit mb_lba truncation with a print-only beyond-EOF check in amr_bio_command — unchecked raw-disk I/O past 2TiB wraps to an attacker-chosen in-range LBA (firmware is the only backstop)
Base system — drivers · sys/dev/raid/amr/amr_disk.c:181 · pass 2 · new |
silent wrap-around write via raw device (root:operator) |
untested |
| DF-5255 |
Low |
2.3 |
640KB kernel heap leak per interface down/up cycle: urtwn_stop never frees rx/tx lists while urtwn_init unconditionally re-allocates (plus detach-side leak of queued TX mbufs)
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/wlan/if_urtwn.c:5268 · pass 2 · new |
deterministic 640KB/cycle kernel heap leak (DF-5255) / maddr list race panic (DF-5256) / remote taskqueue wedge (DF-5257) |
untested |
| DF-5295 |
Low |
2.3 |
rum_tx_mgt() pops a tx_free slot before validating the frame; the ENOENT/ENOBUFS error paths orphan it permanently - seven privileged raw injections wedge the TX path until down/up
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_rum.c:1478 · pass 2 · new |
none |
inconclusive |
| DF-3609 |
Info |
2.3 |
ncr(4): CCB-supplied target_id/target_lun index per-adapter arrays without bounds checks in ncr_action/ncr_get_nccb (and CDB/sense length fields are trusted for DMA sizing) — hardening against privileged CAM injection
Privilege escalation (general use) · sys/dev/disk/ncr/ncr.c:3874 · pass 2 · new |
privileged-only OOB read/hardening |
untested |
| DF-3625 |
Info |
2.3 |
ci_dpm_force_performance_level passes the AMD_DPM_FORCED_LEVEL_HIGH enum value (8) instead of the computed top index to ci_dpm_force_state_pcie
Base system — drivers · sys/dev/drm/amd/amdgpu/ci_dpm.c:4369 · pass 2 · new |
firmware-side misbehavior |
untested |
| DF-4254 |
Info |
2.3 |
IR_PHYSICAL_DISK OFFLINE branch uses '|= ~MPR_TARGET_FLAGS_RAID_COMPONENT' — sets every flag bit instead of clearing one (disk stays hidden; HDD not stopped at power-off)
Base system — drivers · sys/dev/raid/mpr/mpr_sas_lsi.c:598 · pass 2 · new |
none |
untested |
| DF-4261 |
Info |
2.3 |
acpi_pst_attach() leaks allocated _PCT bus resources (and burns cpu_next_rid) on every error path after acpi_pst_alloc_resource()
Base system — drivers · sys/dev/acpica/acpi_cpu_pstate.c:500 · pass 2 · new |
none |
untested |
| DF-4309 |
Info |
2.3 |
adw_free() never destroys the per-ACB buffer DMA maps — attach/detach asymmetry leaks up to 253 bus_dmamaps and destroys their tag while live
Base system — drivers · sys/dev/disk/advansys/adwcam.c:205 · pass 2 · new |
none |
untested |
| DF-4892 |
Info |
2.3 |
32-bit arithmetic overflow in SMU7 fan-RPM math: 60 * crystal_clock_freq * 10000 wraps for 25/27 MHz VI reference clocks, skewing reported fan speed (hwmon fan1_input/fan1_target) and the programmed TACH target by a constant ~7.09x (DF-4761 vega10 twin)
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/smu7_thermal.c:100 · pass 2 · new |
none |
untested |
| DF-4992 |
Info |
2.3 |
Failed SMB_READB/SMB_READW/SMB_PCALL transfers reported as success: copyout result clobbers the bus error
Base system — drivers · sys/dev/smbus/smb/smb.c:272 · pass 2 · new |
none |
untested |
| DF-4993 |
Info |
2.3 |
SMB_BWRITE silently truncates wcount to u_char: 256..1024-byte block writes are short/zero-length yet return success
Base system — drivers · sys/dev/smbus/smb/smb.c:305 · pass 2 · new |
none |
untested |
| DF-5081 |
Info |
2.3 |
mpu401_init/mpu401_uninit free the kobj without kobj_delete: parent mpu class reference never dropped, compiled kobj_ops pinned forever (unbounded refs++ on reattach)
Base system — drivers · sys/dev/sound/midi/mpu401.c:177 · pass 2 · new |
bounded leak + refcount imbalance |
untested |
| DF-5481 |
Info |
2.3 |
EFIIOC_GET_TABLE hands a kernel DMAP virtual address straight back to userspace: kernel-VA disclosure (root-only) and an ioctl ABI that can never work
Kernel memory leaks / uninitialized disclosure · sys/platform/pc64/x86_64/efirt.c:481 · pass 2 · new |
none |
inconclusive |
| DF-3490 |
Low |
2.1 |
sc_isoc_in_buffer mbuf leaked on device detach: unfreed SCO reassembly buffer enables repeatable mbuf exhaustion
Base system — network stack · sys/netgraph7/bluetooth/drivers/ubt/ng_ubt.c:683 · pass 2 · new |
none |
inconclusive |
| DF-4005 |
Low |
2.1 |
fiji_get_dependency_volt_by_clk fall-through reads dep_table->entries[count] twice (one-entry heap OOB read)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/smumgr/fiji_smumgr.c:406 · pass 2 · new |
none |
inconclusive fixed |
| DF-4873 |
Low |
2.1 |
ath_hal_9287EepromAttach error paths leak the malloc'd HAL_EEPROM_9287 (~1KB permanently per failed attach): three post-malloc returns never free or install it, and ar9287Attach's bad: unwind is a no-op for it
Base system — drivers · sys/dev/netif/ath/ath_hal/ar9002/ar9287_attach.c:239 · pass 2 · new |
none |
untested |
| DF-4970 |
Low |
2.1 |
xe_pccard_probe flag loop exits at the DINGO bit before testing it: case XE_CARD_TYPE_FLAGS_DINGO is dead code, CEM56/REM56 Dingo cards are never detected as Dingos
Base system — drivers · sys/dev/netif/xe/if_xe_pccard.c:311 · pass 2 · new |
none |
inconclusive |
| DF-3671 |
Info |
2.1 |
Unchecked PHY/EEPROM read return values across ixgbe_common.c: config decisions from uninitialized kernel stack, and a failed PHY read written back clobbering the AN advertisement (DF-3643 parity)
Base system — drivers · sys/dev/netif/ix/ixgbe_common.c:278 · pass 2 · new |
none |
inconclusive |
| DF-5489 |
Info |
2.1 |
uvisor VISOR GET_CONNECTION_INFORMATION: actlen never validated after a SHORT_XFER_OK control-IN into an 18-byte stack struct; the USB_DEBUG connection loop consumes uninitialized kernel stack when an evil Visor short-transfers (Info: no production consumer)
Base system — bus / platform / cpu · sys/bus/u4b/serial/uvisor.c:387 · pass 2 · new |
none |
untested |
| DF-3452 |
Low |
2.0 |
Uninitialized kernel stack transmitted in ipfw3 sync datagrams: ip_fw3_sync_send_state fills 3 of 9 fields of a stack struct and ships all of it over UDP (LATENT — subsystem dead by omission)
Kernel memory leaks / uninitialized disclosure · sys/net/ipfw3_basic/ip_fw3_sync.c:443 · pass 2 · new |
none |
not_reproduced |
| DF-3992 |
Low |
2.0 |
Ignored DPCD-read failures feed uninitialized stack into PSR/training-interval/dongle-quirk decisions; read_hpd_rx_irq_data uses a shared 'static' retval across links
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc_link_dp.c:50 · pass 2 · new |
Ignored DPCD-read failures feed uninitialized stack into PSR/training-interval/dongle-quirk decisions (comparison-only, no disclosure); shared 'static retval' across links in read_hpd_rx_irq_data. Low/hardening — PoC skipped per contract. |
untested |
| DF-4339 |
Low |
2.0 |
atom_parse name[512] stays unterminated when the VBIOS message is 511 printable bytes — %s leaks uninitialized kernel stack past the buffer to console/log; msg string also walked unbounded past the image
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/atom.c:1276 · pass 2 · new |
none |
inconclusive |
| DF-4347 |
Low |
2.0 |
amdgpu VBIOS part-number string walked and printed past the BIOS image: unvalidated idx deref + %s/strlcpy on an unterminated source (DF-4339's amdgpu twin, worse shape — no bounded copy at all)
Base system — drivers · sys/dev/drm/amd/amdgpu/atom.c:1340 · pass 2 · new |
none |
inconclusive |
| DF-5269 |
Low |
2.0 |
xhci_check_transfer decodes a stream ID from the completion TRB, which per xHCI spec carries none — completions on streams endpoints are silently dropped (transfer hang/DoS)
Base system — bus / platform / cpu · sys/bus/u4b/controller/xhci.c:908 · pass 2 · new |
streams-endpoint transfer hangs (DF-5269) / large-transfer miscompletion (DF-5270) |
untested |
| DF-5360 |
Low |
2.0 |
Poisoned num_iospace=16/num_memspace=8 persist on the queued cfe after the CFTABLE error-break and propagate via the default-copy
Kernel memory leaks / uninitialized disclosure · sys/bus/pccard/pccard_cis.c:916 · pass 2 · new |
none |
untested |
| DF-5494 |
Low |
2.0 |
uether_ifdetach() never drains ue->ue_rxq — mbufs enqueued by a final RX callback that exits without flushing leak with the freed softc (live path: cdce bulk-read tr_stall skips uether_rxflush)
Base system — bus / platform / cpu · sys/bus/u4b/net/usb_ethernet.c:576 · pass 2 · new |
mbuf cluster leak per stall-unplug cycle (DF-5494) / future unguarded-len overflows (DF-5495) |
untested |
| DF-5770 |
Low |
2.0 |
mmc_ivars leaked on every discovery error-break of a newly-allocated card (hostile card leaks ~0.7 kB per scan)
Base system — bus / platform / cpu · sys/bus/mmc/mmc.c:1440 · pass 2 · new |
slow kernel-memory leak (5770) / latent UAF (5771) |
untested |
| DF-3722 |
Info |
2.0 |
Uninitialized kernel stack shipped to the NIC: mxge_close's ETHERNET_DOWN cmd (12B), mxge_dummy_rdma (40 of 64 PIO bytes), mxge_load_firmware boot handoff (36 of 64 PIO bytes)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/mxge/if_mxge.c:3517 · pass 2 · new |
none |
inconclusive |
| DF-4049 |
Info |
2.0 |
dac1/2_output_control_v1 pass 3 uninitialized stack padding bytes as the VBIOS parameter space
Base system — drivers · sys/dev/drm/amd/display/dc/bios/command_table.c:1669 · pass 2 · new |
none |
untested |
| DF-4209 |
Info |
2.0 |
Memory leak of ee_modePowerArray5112[].pChannels allocations when legacyEepromReadContents fails after eepromAllocExpnPower5112 succeeded
Base system — drivers · sys/dev/netif/ath/ath_hal/ah_eeprom_v3.c:383 · pass 2 · new |
readEepromRawPowerCalInfo5112 allocates pChannels/pDataPerChannel per mode BEFORE eepromExpandPower5112 can fail |
untested |
| DF-2666 |
Low |
1.9 |
HAMMER2IOC_RECLUSTER leaks the holdfp() file reference on both error returns in hammer2_ioctl_recluster() — permanent struct-file pin per failing call
Hammer2 · sys/vfs/hammer2/hammer2_ioctl.c:212 · pass 2 · new |
none |
not_reproduced |
| DF-2696 |
Low |
1.9 |
sorecvtcp(): MSG_OOB + sio path is missing 'm = m_free(m)' — repeated sbappend of the same mbuf corrupts the caller's sockbuf (latent, currently unreachable)
Base system — kernel core (kern) · sys/kern/uipc_socket.c:1648 · pass 2 · new |
none |
not_reproduced |
| DF-2703 |
Low |
1.9 |
vop_mountctl/vop_markatime/vop_allocate dispatch with partially-uninitialized stack argument structs (a_vp pointer never set)
Base system — kernel core (kern) · sys/kern/vfs_vopops.c:1300 · pass 2 · new |
none |
untested |
| DF-2850 |
Low |
1.9 |
_gtaskqueue_start_threads() reports total thread-creation failure as success; queue teardown lifecycle is stubbed, so a half-born queue silently hangs all future drainers
Base system — kernel core (kern) · sys/kern/subr_gtaskqueue.c:455 · pass 2 · new |
none |
not_reproduced |
| DF-2945 |
Low |
1.9 |
acctwatch() ignores VFS_STATFS failure (decides suspend/resume on uninitialized stack statfs) and races vgone(): NULL v_mount dereference panic
Base system — kernel core (kern) · sys/kern/kern_acct.c:332 · pass 2 · new |
panic |
not_reproduced |
| DF-3019 |
Low |
1.9 |
devfs_spec_open clone path: EPERM/EBUSY early returns after securelevel/mounted-RW checks leak the cloned vp exclusively locked and referenced forever
Base system — filesystems · sys/vfs/devfs/devfs_vnops.c:1006 · pass 2 · new |
none |
not_reproduced |
| DF-3041 |
Low |
1.9 |
hammer_install_volume dereferences the volume header after brelse(bp) — buffer-cache use-after-release read of vol0_stat_bigblocks into f_blocks
Kernel memory corruption (UAF / OOB write) · sys/vfs/hammer/hammer_ondisk.c:250 · pass 2 · new |
none |
not_reproduced |
| DF-3125 |
Low |
1.9 |
cs_local/cs_disk charset names copied from mount args without NUL-termination — bounded kernel-stack OOB read in iconv_lookupcs strcmp
Kernel memory leaks / uninitialized disclosure · sys/vfs/isofs/cd9660/cd9660_vfsops.c:492 · pass 2 · new |
leak |
not_reproduced |
| DF-3200 |
Low |
1.9 |
u_char length truncation in struct devfs_rule: name[-1] heap over-read and cross-mount rule-clear confusion
Kernel memory leaks / uninitialized disclosure · sys/vfs/devfs/devfs_rules.c:200 · pass 2 · new |
none |
inconclusive fixed |
| DF-3420 |
Low |
1.9 |
pfi_update_status() snapshots the pf-group member-list head outside the ifgroup lock, then walks the snapshot — use-after-free of ifg_member under concurrent group churn
Kernel memory corruption (UAF / OOB write) · sys/net/pf/pf_if.c:722 · pass 2 · new |
none |
not_reproduced |
| DF-3430 |
Low |
1.9 |
ieee80211_create_wds leaks the find_node_locked reference in the 'station in use' branch: permanent node refcount inflation, un-reclaimable station table entries
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_wds.c:163 · pass 2 · new |
none |
untested |
| DF-3528 |
Low |
1.9 |
ng_car SHAPE timer leaks ng_item + node ref + hook ref on every armed-cancel (DF-3523 twin) - latent: module does not build on DragonFly
Base system — network stack · sys/netgraph7/ng_car.c:539 · pass 2 · new |
none |
build_failed |
| DF-3548 |
Low |
1.9 |
Lost-wakeup race in ngdread: NGDF_RWAIT armed after the empty-queue check
Base system — network stack · sys/netgraph7/ng_device.c:419 · pass 2 · new |
none |
untested |
| DF-3573 |
Low |
1.9 |
Legacy firmware size-check bypass (err clobbered by later request_firmware success) → heap OOB reads in constant-bounded legacy ucode loaders
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/cik.c:2080 · pass 2 · new |
none |
not_reproduced |
| DF-3583 |
Low |
1.9 |
RIO1 fast-post loop trusts card-controlled 8-bit rqs_seqno against a 15-element handle array — kernel stack OOB read up to 960 bytes
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/isp/isp.c:5093 · pass 2 · new |
none |
inconclusive |
| DF-3605 |
Low |
1.9 |
bce_nvram_read returns on first-dword read failure while still holding the NVRAM arbitration lock and access-enable, permanently bricking NVRAM access (driver re-attach fails)
Base system — drivers · sys/dev/netif/bce/if_bce.c:1726 · pass 2 · new |
NVRAM lock permanently held on error path |
untested |
| DF-3635 |
Low |
1.9 |
i915_gem_object_get_pages_phys() error path returns holding the GEM backing vm_object (leaked hold_count + LWKT token)
Base system — drivers · sys/dev/drm/i915/i915_gem.c:305 · pass 2 · new |
none |
untested |
| DF-3640 |
Low |
1.9 |
bnx_reset acquires the NVRAM software-arbitration lock (SET1/GNT1) and never releases it — flash arbiter held for the driver's lifetime, locking out APE/management firmware
Base system — drivers · sys/dev/netif/bnx/if_bnx.c:2585 · pass 2 · new |
none |
inconclusive |
| DF-3666 |
Low |
1.9 |
sumo_rlc_init() save/restore buffer object undersized for the ON/LN/TN packed format: deterministic 172-byte write past the 324-byte BO
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/evergreen.c:4155 · pass 2 · new |
none |
inconclusive |
| DF-3795 |
Low |
1.9 |
acpi_parse_prw() copies _PRW elements 0..N-3 instead of 2..N as wake power resources (off-by-two)
Base system — drivers · sys/dev/acpica/acpi.c:2963 · pass 2 · new |
none |
not_reproduced |
| DF-3847 |
Low |
1.9 |
int8_t overflow in bwn G-PHY tssi2dbm table generation: clamp to 128 wraps to -128 (DFly divergence; FreeBSD clamps to 127), corrupting TX-power control
Base system — drivers · sys/dev/netif/bwn/bwn/if_bwn_phy_g.c:230 · pass 2 · new |
TX-power correctness (bwn) / mbuf leak + ring wedge (dc) |
untested |
| DF-3905 |
Low |
1.9 |
sk_marv_miibus_writereg(): inverted YU_SMICR_BUSY check aborts the SMI write-completion wait the moment the write starts (vs correct msk_phy_writereg sibling)
Base system — drivers · sys/dev/netif/sk/if_sk.c:519 · pass 2 · new |
none |
inconclusive |
| DF-3913 |
Low |
1.9 |
intel_sdvo_connector_state leaked when intel_sdvo_connector_init() fails (all four SDVO connector init paths)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_sdvo.c:2481 · pass 2 · new |
none |
inconclusive |
| DF-4052 |
Low |
1.9 |
nfe_alloc_{rx,tx}_ring error paths destroy data[i].map in a loop over j — dmamap leak + tag map_count underflow on attach failure
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/nfe/if_nfe.c:1725 · pass 2 · new |
none |
untested |
| DF-4105 |
Low |
1.9 |
Fake-EDID record: kmalloc(128) with memcpy of only ucFakeEDIDLength bytes leaves an uninitialized heap tail that (a) drives an OOB multi-block walk in drm_edid_is_valid via uninit extensions byte and (b) on a ~1/256 checksum pass is exposed to unprivileged users via the EDID ioctl (kernel heap info leak)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/atombios_encoders.c:2171 · pass 2 · new |
none |
untested |
| DF-4126 |
Low |
1.9 |
Inverted error checks at rv6xx_dpm_init:1957-1969 — ATOM-computed SPLL/MPLL reference divider used only when radeon_atom_get_clock_dividers FAILS; default used on success (rv6xx twin of DF-3976/DF-4119)
Base system — drivers · sys/dev/drm/radeon/rv6xx_dpm.c:184 · pass 2 · new |
none |
inconclusive |
| DF-4128 |
Low |
1.9 |
Ignored rv6xx_convert_clock_to_stepping failures (:249-250, :302) leave uninitialized stepping structs programmed into the SPLL; VBIOS-driven step count overruns the SPLL step register block into VID_RT@0x6f8+
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/rv6xx_dpm.c:225 · pass 2 · new |
none |
inconclusive |
| DF-4141 |
Low |
1.9 |
ar5211GetLowerUpperValues with empty pier list: 2-byte OOB stack read below tempChannelList plus uninitialized stack read -> garbage regulatory edge-power selection (DF-3977 ar5211 twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5211/ar5211_reset.c:1818 · pass 2 · new |
none |
inconclusive fixed |
| DF-4177 |
Low |
1.9 |
isp_pci_mbxdma allocations are never torn down: error paths leak the master DMA tag, and a successful attach->detach leaks pcmd_pool, xflist, tgtlist, control queues, per-channel FC scratch, per-cmd maps and all DMA tags
Base system — drivers · sys/dev/disk/isp/isp_pci.c:1536 · pass 2 · new |
isp_pci_mbxdma creates the master tag and allocates pcmd_pool, xflist/tgtlist, control-space tag+memory, per-channel FC scratch, and per-cmd dma maps |
untested |
| DF-4204 |
Low |
1.9 |
ahci_cam_attach() error paths leak the cam_sim and leave the bus registered: ahci_cam_detach() no-ops because AP_F_CAM_ATTACHED is not yet set
Base system — drivers · sys/dev/disk/ahci/ahci_cam.c:137 · pass 2 · new |
ahci_cam_attach error paths leak the cam_sim and leave the bus registered: ahci_cam_detach no-ops because AP_F_CAM_ATTACHED is not yet set. Memory leak + stale registered CAM bus on any port whose device fails IDENTIFY at attach (hostile/misbehaving SATA device). Resource-exhaustion ceiling. |
untested |
| DF-4205 |
Low |
1.9 |
vlv_dsi_init dereferences vbt.dsi.config without a NULL/dangling guard — boot-time NULL-page panic or UAF read of freed config
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/vlv_dsi.c:1797 · pass 2 · new |
none |
untested |
| DF-4244 |
Low |
1.9 |
dpt_detect_cache log-page walk bounded by device-fed 16-bit page length (up to 65539) instead of the 512 bytes actually DMA'd — OOB kernel read up to ~64KB past the transfer buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/dpt/dpt_scsi.c:618 · pass 2 · new |
none demonstrated on guest (no DPT hardware); source-level impact: device-fed bound → OOB kernel read up to ~65KB past the 512-byte DMA buffer at attach (DoS ceiling, no disclosure) |
untested |
| DF-4246 |
Low |
1.9 |
linux_hexdump() memcmp over-reads up to rowsize-1 bytes past the kmalloc'd ring snapshot on the final partial row (print_request_ring path)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_engine_cs.c:1261 · pass 2 · new |
linux_hexdump() memcmp over-reads up to rowsize-1 (31) bytes past the kmalloc'd ring snapshot on the final partial row (print_request_ring path). Bytes never printed; only observable is a 1-bit equality oracle vs the previous row, on debugfs/error-capture paths. |
untested |
| DF-4248 |
Low |
1.9 |
V1-branch state-array off-by-one: entry_index == ucNumStates passes 'entry_index > ucNumStates' guard
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/processpptables.c:937 · pass 2 · new |
none |
untested |
| DF-4249 |
Low |
1.9 |
Extended-header fields read before/without usSize gating (usPPMTableOffset at +16 read before V5 check; overdrive limits read with no size gate)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/processpptables.c:1029 · pass 2 · new |
none |
untested |
| DF-4256 |
Low |
1.9 |
Unsigned underflow in agp_i915_get_stolen_size: sc->stolen wraps to ~0x3FFFFFFF entries when BIOS stolen size < GTT size, permanently breaking all GTT binds
Base system — drivers · sys/dev/agp/intel-gtt.c:696 · pass 2 · new |
none |
inconclusive |
| DF-4281 |
Low |
1.9 |
drm_addmap_core leaks vmalloc'd SHM maps, WC MTRR registrations and CONSISTENT DMA memory on error paths
Base system — drivers · sys/dev/drm/drm_bufs.c:338 · pass 2 · new |
Kernel memory / MTRR-slot exhaustion (DoS ceiling) by root on DRM machines. |
untested |
| DF-4353 |
Low |
1.9 |
dce110_create_resource_pool leaks the kzalloc'd pool on construct() failure (DF-4318 twin) and dc->fbc_compressor is never destroyed anywhere (dce110_compressor_destroy has zero callers)
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_resource.c:1373 · pass 2 · new |
none |
untested |
| DF-4359 |
Low |
1.9 |
cik_sdma_init_microcode() ignores amdgpu_ucode_validate() failure for all but the last SDMA instance — invalid firmware accepted (door to the DF-1553 sink that a DF-1553 fix does not close)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/cik_sdma.c:141 · pass 2 · new |
none |
untested |
| DF-4376 |
Low |
1.9 |
dce112_create_resource_pool leaks the kzalloc'd pool shell on construct() failure (DF-4318/4353 twin)
Base system — drivers · sys/dev/drm/amd/display/dc/dce112/dce112_resource.c:1353 · pass 2 · new |
none |
inconclusive |
| DF-4406 |
Low |
1.9 |
acpi_tz_attach: NULL-pointer panic when tz unit 0 fails acpi_tz_establish and a second thermal zone attaches (global sysctl tree dereferenced before creation)
Base system — drivers · sys/dev/acpica/acpi_thermal.c:237 · pass 2 · new |
NULL-pointer panic when thermal zone unit 0 fails acpi_tz_establish (malformed _ALx non-package) and a second zone attaches: the driver-global sysctl tree was never created (unit-0-only) but per-zone SYSCTL_ADD_NODE dereferences SYSCTL_CHILDREN(NULL) = raw (oid_ptr)->oid_arg1 -> NULL-page read panic at attach/boot. Firmware-gated boot-time DoS. |
untested |
| DF-4416 |
Low |
1.9 |
wa_init_mcr uses 1-based fls() to index the 0-based subslice_mask[] and (via intel_calculate_mcr_s_ss_select) programs invalid MCR slice steering on gen10/11
Base system — drivers · sys/dev/drm/i915/intel_workarounds.c:787 · pass 2 · new |
none |
untested |
| DF-4465 |
Low |
1.9 |
uguru_attach ENXIO paths leak the active IOPORT resource and permanently wedge the driver until reboot
Base system — drivers · sys/dev/powermng/uguru/uguru.c:757 · pass 2 · new |
Monitoring driver permanently dead until reboot after a failed attach (leaked active IOPORT range) |
untested |
| DF-4549 |
Low |
1.9 |
viapm_wait() off-by-one: while(count--)/if(!count) makes SMB_ETIMEOUT unreachable on genuine SMBus timeouts — silent success with stale register data — and spuriously fails last-poll successes
Base system — drivers · sys/dev/powermng/viapm/viapm.c:574 · pass 2 · new |
none |
not_reproduced |
| DF-4585 |
Low |
1.9 |
mpt_dump_sgl called with offset=0 on interior/stack pointers: bounded SGL walk overshoots the request area and hex-dumps up to ~504 bytes of kernel stack past an 8-byte stack union
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt_debug.c:604 · pass 2 · new |
Kernel stack contents (up to ~500 bytes) and adjacent request-area DMA data hex-dumped to console/dmesg - kernel info leak into logs; no write primitive |
untested |
| DF-4591 |
Low |
1.9 |
tws_attach error unwinds leak the MSI allocation and every post-probe resource (~200 KB heap + DMA maps/tags + scan_ccb per failed attach)
Base system — drivers · sys/dev/raid/tws/tws.c:309 · pass 2 · new |
none |
not_reproduced |
| DF-4609 |
Low |
1.9 |
Uninitialized stack value used as relocation buffer bound before session validation
Base system — drivers · sys/dev/drm/radeon/radeon_vce.c:561 · pass 2 · new |
Negligible standalone: a kernel-stack garbage value gates a bounds decision on an IB that is then rejected before any execution; hardening correctness issue |
untested |
| DF-4641 |
Low |
1.9 |
radeon: radeon_read_platform_bios maps a fixed 256KB window from the PCI ROM BAR base regardless of the BAR's actual decode size - kernel reads physical memory beyond the device's ROM aperture into rdev->bios (and maps device MMIO write-back cached)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_bios.c:130 · pass 2 · new |
Kernel-side over-read of adjacent physical space (other devices' MMIO / unclaimed holes) into rdev->bios, parsed by atom/combios; possible read side effects on neighboring devices; no userspace exposure path found |
untested |
| DF-4698 |
Low |
1.9 |
nvme(4): queue-creation retry (tryagain) re-allocates still-allocated subqueues — lockinit on a live lock and permanent leak of the previous queue's DMA maps + request array
Base system — drivers · sys/dev/disk/nvme/nvme_admin.c:453 · pass 2 · new |
none |
untested |
| DF-4699 |
Low |
1.9 |
nvme(4): admin commands have no completion timeout — a stalled controller permanently wedges the admin thread and hangs detach/shutdown in nvwend
Base system — drivers · sys/dev/disk/nvme/nvme_admin.c:104 · pass 2 · new |
none |
untested |
| DF-4701 |
Low |
1.9 |
twa(4): tw_cl_shutdown_ctlr tears the Common Layer down (active=FALSE, interrupts/watchdog gone, CL locks destroyed) before the firmware shutdown handshake result is known and returns that error — twa_detach then aborts, leaving a still-attached half-dead device whose I/O entry points have no ctlr->active gate, so subsequent unprivileged disk I/O is submitted to a dead controller until reboot
Base system — drivers · sys/dev/raid/twa/tw_cl_init.c:569 · pass 2 · new |
dos |
untested |
| DF-4731 |
Low |
1.9 |
DAL_VALID_IRQ_SRC_NUM off-by-one ('<=') accepts irq_source == DAL_IRQ_SOURCES_NUMBER — one-past-end index into amdgpu_dm_irq.c's handler-list tables
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c:176 · pass 2 · new |
none |
untested |
| DF-4737 |
Low |
1.9 |
amdpm_wait() off-by-one: while(count--)/if (!count) makes SMB_ETIMEOUT unreachable on genuine SMBus timeouts (silent success with stale HSTDATA/FIFO data) and spuriously fails completions on the final poll
Base system — drivers · sys/dev/powermng/amdpm/amdpm.c:302 · pass 2 · new |
none |
not_reproduced |
| DF-4739 |
Low |
1.9 |
alpm_wait() off-by-one: while(count--)/if(!count) makes SMB_ETIMEOUT unreachable on genuine SMBus timeouts (silent success with stale SMBHDATA/SMBHBLOCK data) and spuriously fails last-poll completions — DF-4737/DF-4549 family
Base system — drivers · sys/dev/powermng/alpm/alpm.c:297 · pass 2 · new |
none |
not_reproduced |
| DF-4767 |
Low |
1.9 |
PSP buffer objects (tmr_bo 3MB VRAM, asd_shared_bo 16KB, firmware.rbuf 4KB) leaked permanently when psp_hw_start fails during psp_load_fw
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_psp.c:379 · pass 2 · new |
none |
inconclusive |
| DF-4771 |
Low |
1.9 |
intel_dp_add_mst_connector error path leaks the allocated intel_connector (drm_connector_cleanup without kfree)
Base system — drivers · sys/dev/drm/i915/intel_dp_mst.c:473 · pass 2 · new |
leak |
untested |
| DF-4785 |
Low |
1.9 |
psp_v11_0_init_microcode never loads ASD firmware for Vega20 — generic psp_asd_load memcpy's from NULL and submits a zero-length LOAD_ASD whose PSP rejection is silently swallowed
Base system — drivers · sys/dev/drm/amd/amdgpu/psp_v11_0.c:95 · pass 2 · new |
none |
untested |
| DF-4815 |
Low |
1.9 |
mmhub_v1_0_init_cache_regs never programs mmVM_L2_CNTL3: BANK_SELECT/L2_CACHE_BIGK_FRAGMENT_SIZE computed into a stale CNTL2-derived local and silently dropped (the v5.2-era WREG32 line is missing from the DF vendor sync)
Base system — drivers · sys/dev/drm/amd/amdgpu/mmhub_v1_0.c:171 · pass 2 · new |
none at runtime on this guest (no AMD GPU present); code-level: GPU MMU L2 TLB geometry (BANK_SELECT/L2_CACHE_BIGK_FRAGMENT_SIZE) never programmed — functional misconfiguration on Vega10-class ASICs, VM-fault/hang class on translate_further (Raven2) parts; no attacker control, no demonstrated memory-isolation breach |
untested |
| DF-4859 |
Low |
1.9 |
Locality 0 never relinquished on any of the 9 error exits in tpmtis_transmit (TIS twin of DF-1990), leaving TPM arbitration claimed and imposing a fixed 750ms uninterruptible sleep on every later request
Base system — drivers · sys/dev/crypto/tpm/tpm_tis.c:407 · pass 2 · new |
dos |
untested |
| DF-4888 |
Low |
1.9 |
iir_pci_attach error paths: BAR0 resource never released (release commented out, rid clobbered) and bus_setup_intr failure skips iir_free()/XPT teardown — leaked DMA state + /dev/iirN + live registered CAM SIMs with no interrupt (CCBs hang forever)
Base system — drivers · sys/dev/raid/iir/iir_pci.c:195 · pass 2 · new |
Permanent BAR/rman resource leak on every failed attach; late attach failure (bus_setup_intr) leaks all DMA state and /dev/iirN and leaves XPT-registered CAM SIMs with no interrupt handler - CAM scan CCBs hang forever (iir_timeout empty), wedging the rescan thread; hardware-gated local DoS/resource exhaustion (CWE-404) |
untested |
| DF-4896 |
Low |
1.9 |
complete_rqe() volume error path abandons the request: recover_io() re-strategizes the bio but nobody frees the old request — full request leak (rq + rqg chain + malloc'd buffers + locked bufs) per errored volume I/O, unbounded with VF_RETRYERRORS; VOL[].active never decremented on error
Base system — drivers · sys/dev/raid/vinum/vinuminterrupt.c:223 · pass 2 · new |
kernel memory leak: full struct request + rqg chain + XFR_MALLOCED buffers + XFR_BUFLOCKED buf locks per errored volume I/O; unbounded with VF_RETRYERRORS (daemon retry loop leaks at I/O speed); VOL[].active counter also never decremented on the error path |
untested |
| DF-4917 |
Low |
1.9 |
vf2pf checksum computed over pf2vf_size bytes instead of sizeof(amdgim_vf2pf_info): out-of-bounds read of up to ~3 KB past the 1 KB vf2pf struct and a wrong protocol checksum published to the host
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_virt.c:383 · pass 2 · new |
none |
not_reproduced |
| DF-4941 |
Low |
1.9 |
if_ic: SIOCSIFMTU frees the working buffers before checking the replacement allocations — on kmalloc failure installs NULL ic_obuf/ic_ifbuf while still updating if_mtu
Base system — drivers · sys/dev/netif/ic/if_ic.c:211 · pass 2 · new |
none |
untested |
| DF-4953 |
Low |
1.9 |
aibs_attach ENXIO path leaks every kmalloc'd sc_asens_* sensor array — newbus never calls detach after a failed attach
Base system — drivers · sys/dev/acpica/aibs/atk0110.c:126 · pass 2 · new |
leak |
untested |
| DF-4954 |
Low |
1.9 |
mpr_pci_setup_interrupts: bus_setup_intr failure truncates msi_msgs before the failing queue, permanently leaking its allocated IRQ resource and (via ignored EBUSY) the MSI message
Base system — drivers · sys/dev/raid/mpr/mpr_pci.c:294 · pass 2 · new |
leak |
untested |
| DF-4955 |
Low |
1.9 |
mpr_pci_free destroys the parent DMA tag while mpr_free can early-return without freeing the 7 child tags — DragonFly's bus_dma_tag_destroy kfree()s unconditionally, permanently leaking all IOC-facts DMA memory and tags on a wedged-controller failed attach
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_pci.c:232 · pass 2 · new |
leak |
untested |
| DF-4982 |
Low |
1.9 |
iir(4) GDT_IOCTL_EVENT double-fetch: raced event_data.size drives memcmp/strcmp length against the 16-byte eu union in static ebuffer[] — OOB kernel read with byte-wise equality-oracle disclosure
Base system — drivers · sys/dev/raid/iir/iir_ctrl.c:324 · pass 2 · new |
none |
untested |
| DF-5014 |
Low |
1.9 |
Broken attach-unwind contract: mps_attach leaks sc->facts on early failures, and mps_pci_free destroys the parent DMA tag over live child tags when mps_free early-returns
Base system — drivers · sys/dev/raid/mps/mps_pci.c:234 · pass 2 · new |
none |
not_reproduced |
| DF-5074 |
Low |
1.9 |
acpi_acad has no device_detach: ACPI notify handler (context=dev), global ACPIIO_ACAD_GET_STATUS ioctl hook and hw.acpi.acline sysctl OID are never removed (DF-5006-family; resource pinning + blocked kldunload, structurally non-UAF today)
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_acad.c:76 · pass 2 · new |
none |
untested |
| DF-5091 |
Low |
1.9 |
acpi_button has no device_detach: ACPI notify handler and fixed-event handler (context = softc) plus PRW wake GPE are never removed — DF-5006/DF-5074 family teardown gap, guaranteed UAF if a detach path is ever added
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_button.c:74 · pass 2 · new |
none |
not_reproduced |
| DF-5094 |
Low |
1.9 |
Unbounded vinum daemon instantiation with pid-blind lazy abdication: N concurrent daemons, unkillable P_SYSTEM processes, non-unloadable module; the plausible multi-consumer daemonq double-free is refuted on current DragonFly only by incidental MP-token serialization
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumdaemon.c:83 · pass 2 · new |
none |
not_reproduced |
| DF-5103 |
Low |
1.9 |
apple_fw detach leaves softc-backed world-readable sysctls live while newbus frees the softc: use-after-free read window during driver unload
Kernel memory corruption (UAF / OOB write) · sys/dev/apple/fw/apple_fw.c:144 · pass 2 · new |
none |
untested |
| DF-5124 |
Low |
1.9 |
hyperv_md_uninit() frees the reference-TSC DMA page without quiescing in-flight hyperv_tc64 readers — shutdown-time use-after-free read window
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/hyperv/x86_64/hyperv_machdep.c:96 · pass 2 · new |
none |
untested |
| DF-5149 |
Low |
1.9 |
pps driver has no device_detach and never destroy_dev()s its make_dev() cdev (DF-5006/5074/5091 teardown-gap family): /dev/ppsN and the ppbus child are permanent, module unload is structurally blocked, and a future naive detach is an interrupt-context UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/pps/pps.c:103 · pass 2 · new |
none |
not_reproduced |
| DF-5150 |
Low |
1.9 |
acpi_lid has no device_detach: ACPI_DEVICE_NOTIFY handler installed with context = softc, AcpiOsExecute deferrals of that softc, and the enabled PRW wake GPE are never removed — DF-5091 (acpi_button) mirror; guaranteed UAF if a detach path is ever added
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_lid.c:66 · pass 2 · new |
none |
not_reproduced |
| DF-5158 |
Low |
1.9 |
ata_ati_chipinit AHCI fallback attaches with dangling ctlr->r_res2: double bus_release_resource (rman UAF/panic) at detach
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/chipsets/ata-ati.c:108 · pass 2 · new |
none |
untested |
| DF-5161 |
Low |
1.9 |
check_and_purge_safe_mem livelocks forever when the head allocation fails _free_safe_mem's integrity checks — unbounded kernel loop + tbridge_printf spam turns any detected corruption into a hard hang
Base system — drivers · sys/dev/misc/tbridge/safe_mem.c:171 · pass 2 · new |
kernel livelock + console spam (corruption-detector amplifier) |
untested |
| DF-5203 |
Low |
1.9 |
ipmi_smbus_probe is a second unguarded probe in the ipmi_attached singleton family (DF-5076 twin) — cross-bus/dual-smbus double attach yields second ssif kthread, second /dev/ipmiN cdev, and interleaved SSIF state machines on one BMC
Base system — drivers · sys/dev/misc/ipmi/ipmi_smbus.c:65 · pass 2 · new |
none |
untested |
| DF-5208 |
Low |
1.9 |
ata-isa registers no device_detach and never unwinds probe-allocated resources: kldunload of the ISA nata attachment permanently fails (ENXIO) and attach-failure leaks both IOPORT ranges with dangling softc pointers
Base system — drivers · sys/dev/disk/nata/ata-isa.c:102 · pass 2 · new |
permanent unload failure + per-boot resource leak |
untested |
| DF-5349 |
Low |
1.9 |
bus_dmamap_load() with map==NULL silently disables the tag's bounce constraints (segments returned with out-of-window addresses)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/busdma_machdep.c:601 · pass 2 · new |
device-side out-of-window DMA (DF-5349) / segments leak + tag UAF window (DF-5350) |
untested |
| DF-5404 |
Low |
1.9 |
sgread(): unlocked re-check of rdwr->state races another reader's remove+free — 4-byte use-after-free read and tsleep on a freed object (PINTERLOCKED pattern)
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_sg.c:795 · pass 2 · new |
UAF-read driven control flow (DF-5404) / uninit stack tag match (DF-5405) |
untested |
| DF-5440 |
Low |
1.9 |
targbhctor leaks the softc when targbhenlun() fails: cam_periph_alloc() never calls the dtor on ctor failure
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_targ_bh.c:384 · pass 2 · new |
per-attach softc leak + speculative abort-refusing-SIM UAF (DF-5440) / stale CTIO timeout (DF-5441) |
untested |
| DF-5566 |
Low |
1.9 |
iicbus ownership protocol is not MP-safe and mis-ordered: crit_enter gives no cross-CPU exclusion (double-acquire), and iicbus_release_bus frees the bridge resource BEFORE validating ownership (non-owner teardown of bridge exclusion)
Base system — bus / platform / cpu · sys/bus/iicbus/iiconf.c:100 · pass 2 · new |
double bus acquisition + non-owner bridge release -> external-device wire corruption |
untested |
| DF-5631 |
Low |
1.9 |
kqueue_del() latent API hazards: close-before-delete EV_DELETE-ENOENT panics the whole vkernel; delete-vs-dispatch kfree races kqueue_intr()'s in-flight udata dispatch on another vCPU host thread (zero in-tree callers today)
Kernel memory corruption (UAF / OOB write) · sys/platform/vkernel64/platform/kqueue.c:164 · pass 2 · new |
none |
untested |
| DF-3336 |
Info |
1.9 |
netmap kqfilter stub: f_event always returns 0 and no knlist bookkeeping — kevent() on netmap fds sleeps forever
Base system — network stack · sys/net/netmap/netmap.c:1516 · pass 2 · new |
none |
untested |
| DF-3475 |
Info |
1.9 |
ng_callout() discards the previously-armed callout item (oitem) - re-arming a pending ng_callout leaks the item and permanently pins the node reference (unfreeable node, unloadable type)
Base system — network stack · sys/netgraph7/netgraph/ng_base.c:3266 · pass 2 · new |
none |
untested |
| DF-3560 |
Info |
1.9 |
Camellia derived key material (subL/subR schedule words, setup192's kk[32] 256-bit folded key, Encrypt/DecryptBlock tmp[4] block data) left unzeroed on the kernel stack; keyBitLength validated only by KASSERT, so production kernels would run on an uninitialized subkey schedule if a non-{128,192,256} length ever reached these entry points (latent: no in-tree callers)
Base system — crypto · sys/crypto/camellia/camellia.c:445 · pass 2 · new |
latent key-material stack residue + KASSERT-only key-size validation |
untested |
| DF-3574 |
Info |
1.9 |
sym_action2 XPT_GET/SET_TRAN_SETTINGS index np->target[]/lunmp[] without the bounds check used by every other sym entry point
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/sym/sym_hipd.c:8048 · pass 2 · new |
none |
not_reproduced |
| DF-3580 |
Info |
1.9 |
Inverted privilege-check result in iwn_cdev_ioctl(): returns success when caps_priv_check fails
Privilege escalation (general use) · sys/dev/netif/iwn/if_iwn.c:5157 · pass 2 · new |
none |
untested |
| DF-3724 |
Info |
1.9 |
ixgbe_led_{on,off}_t_X550em: PHY read return ignored — uninitialized stack word masked and written back to the X557 LED provisioning register
Base system — drivers · sys/dev/netif/ix/ixgbe_x550.c:4551 · pass 2 · new |
none |
inconclusive |
| DF-3726 |
Info |
1.9 |
Unvalidated vf index in exported MDD/anti-spoof ops: register-index OOB MMIO read-modify-write if wired with vf beyond the register window
Base system — drivers · sys/dev/netif/ix/ixgbe_x550.c:1078 · pass 2 · new |
none |
inconclusive |
| DF-3792 |
Info |
1.9 |
drm_atomic_helper: unguarded CRTC helper-callback dereferences — funcs->dpms/funcs->commit fallbacks and crtc->helper_private derefs lack the NULL guards of upstream fe6169288666/26564be56d73
Base system — drivers · sys/dev/drm/drm_atomic_helper.c:461 · pass 2 · new |
none |
inconclusive |
| DF-3947 |
Info |
1.9 |
dce110_apply_ctx_to_hw error return leaves VBIOS scratch critical-state latched
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:2028 · pass 2 · new |
none |
untested |
| DF-3979 |
Info |
1.9 |
dcn10_enable_per_frame_crtc_position_reset dereferences triggered_crtc_reset.event_source without NULL check (dead path; DF-3948 twin)
Base system — drivers · sys/dev/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c:1406 · pass 2 · new |
none |
inconclusive |
| DF-3980 |
Info |
1.9 |
hwss1_plane_atomic_disconnect: `opp` NULL check placed after its own dereferences (dead guard, latent NULL-deref)
Base system — drivers · sys/dev/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c:889 · pass 2 · new |
none |
inconclusive |
| DF-4175 |
Info |
1.9 |
iir_init: bus_dmamap_create failure returns with sc_init_level one step low — iir_free skips kfree(sc_gccbs) and every created gc_dmamap (permanent leak on failed attach)
Base system — drivers · sys/dev/raid/iir/iir.c:261 · pass 2 · new |
The gccb init loop returns 1 immediately when bus_dmamap_create fails, but sc_init_level is only bumped to 5 after the whole loop |
untested |
| DF-4420 |
Info |
1.9 |
oce_mq_create leaks the CQ (struct + 4KiB DMA ring) on two of its three error paths
Base system — drivers · sys/dev/netif/oce/oce_queue.c:675 · pass 2 · new |
none |
untested |
| DF-4434 |
Info |
1.9 |
sln_attach() error path after bus_setup_intr failure leaks the 64KB contigmalloc RX ring and skips ifmedia cleanup
Base system — drivers · sys/dev/netif/sln/if_sln.c:221 · pass 2 · new |
64KB contiguous kernel memory + media list leaked per failed attach (bus_setup_intr failure) |
untested |
| DF-4435 |
Info |
1.9 |
dce120 TG: CRTC_REG_UPDATE_4/_5 pass pair-count 3 - 4th/5th CRTC register-field updates silently dropped (CRTC_TEST_PATTERN_COLOR_FORMAT / VRES / RAMP0_OFFSET never programmed)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dce120/dce120_timing_generator.c:57 · pass 2 · new |
none |
untested |
| DF-4462 |
Info |
1.9 |
Duplicate sensordev xname "dieN" installs two ksensordevs on one shared hw.sensors.dieN node; the CCD device's first sensor sysctl is silently dropped
Base system — drivers · sys/dev/powermng/amdtemp/amdtemp.c:706 · pass 2 · new |
Broken sensors(4) presentation on Zen: missing/shadowed temperature readings; no security impact |
untested |
| DF-4487 |
Info |
1.9 |
acpi_video_vo_init kmalloc without M_ZERO leaves vo_sysctl_tree uninitialized when the global video sysctl node is absent
Base system — drivers · sys/dev/acpica/acpi_video/acpi_video.c:513 · pass 2 · new |
none |
untested |
| DF-4598 |
Info |
1.9 |
vinum_scandisk() strncpy leaves partname[64] unterminated → stack OOB read in find_drive_by_dev()'s %s copy
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/vinum/vinumio.c:723 · pass 2 · new |
none |
untested |
| DF-4708 |
Info |
1.9 |
amr_sglist_map's 'goto retry' leaks the rejected s/g table (and its dmamap) unboundedly and all three PCI-setup bus_dmamap_load results are ignored — load failure yields a 0/garbage bus address handed to the controller
Base system — drivers · sys/dev/raid/amr/amr_pci.c:592 · pass 2 · new |
none |
inconclusive |
| DF-4761 |
Info |
1.9 |
32-bit arithmetic overflow in fan-RPM math: 60 * crystal_clock_freq * 10000 wraps for any realistic xclk, skewing reported fan speed and the programmed TACH target
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/vega10_thermal.c:115 · pass 2 · new |
none |
untested |
| DF-4817 |
Info |
1.9 |
iwm phy_db: latent heap OOB read in channel_id_to_txp() — entry size never checked before reading txp_chg->max_channel_idx (path currently dead in-tree)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwm/if_iwm_phy_db.c:382 · pass 2 · new |
none |
untested |
| DF-4822 |
Info |
1.9 |
DCE110v output regamma mode is never engaged: configure_regamma_mode() discards its computed value and writes literal 0, and dce110_opp_set_regamma_mode_v() is an empty stub — user-programmed gamma LUT silently bypassed
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_opp_regamma_v.c:111 · pass 2 · new |
none |
untested |
| DF-4833 |
Info |
1.9 |
iwm_send_cmd: TFD TB0 length field (12 bits of hi_n_len) truncates to zero at the wide-header max-payload boundary (hdrlen+paylen==4096)
Base system — drivers · sys/dev/netif/iwm/if_iwm_util.c:220 · pass 2 · new |
none |
untested |
| DF-4845 |
Info |
1.9 |
lnc PCI detach/unwind never frees the ifmedia list — small kernel memory leak per attach/detach cycle
Base system — drivers · sys/dev/netif/lnc/if_lnc_pci.c:449 · pass 2 · new |
leak |
untested |
| DF-4851 |
Info |
1.9 |
aac(4) MSI tunable registered under mfi(4)'s name (hw.mfi.msi.enable) and NOMSI attach path permanently mutates the shared boot global
Base system — drivers · sys/dev/raid/aac/aac_pci.c:61 · pass 2 · new |
none |
untested |
| DF-4856 |
Info |
1.9 |
agp_amd64_via_init reads AGP3_VIA_ATTBASE as the RMW source for the AGP3_VIA_GARTCTRL write (register confusion, upstream-inherited)
Base system — drivers · sys/dev/agp/agp_amd64.c:460 · pass 2 · new |
none |
not_reproduced |
| DF-4865 |
Info |
1.9 |
cpuctl microcode update paths ignore rdmsr_safe() failures: uninitialized stack rev0/rev1/res printed to dmesg and used in the success decision
Privilege escalation (general use) · sys/dev/misc/cpuctl/cpuctl.c:289 · pass 2 · new |
none |
not_reproduced |
| DF-4867 |
Info |
1.9 |
ata_intel_chipinit treats any non-ENXIO return of ata_ahci_chipinit as AHCI success, leaving ctlr->setmode NULL — latent boot-time NULL function-pointer call (fragile error contract)
Base system — drivers · sys/dev/disk/nata/chipsets/ata-intel.c:177 · pass 2 · new |
none |
not_reproduced |
| DF-4869 |
Info |
1.9 |
mps table layer latent defects: live-but-gated unvalidated NumEntries walk in mps_print_evt_sas, three dead config-page printers, sentinel+1 fallback in mps_describe_table
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps_table.c:63 · pass 2 · new |
none |
untested |
| DF-4894 |
Info |
1.9 |
aac_disk_dump() consumes firmware-derived geometry with no host-side clamping: max_sectors==0 → infinite dump loop, sg_tablesize/SectorCount/BlockNumber overflow the FIB fields, single static dump map shared across controllers
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/aac/aac_disk.c:256 · pass 2 · new |
dos |
untested |
| DF-4898 |
Info |
1.9 |
vpo_cam_rescan leaks struct cam_path on every attach and the entire CCB (incl. its callout) when xpt_create_path fails
Base system — drivers · sys/dev/disk/vpo/vpo.c:176 · pass 2 · new |
leak |
untested |
| DF-4899 |
Info |
1.9 |
vpo_attach error paths never unwind the low-level attach (leaked microseq buffer, ppbus left claimed) and the driver has no detach method at all
Base system — drivers · sys/dev/disk/vpo/vpo.c:139 · pass 2 · new |
leak |
untested |
| DF-4935 |
Info |
1.9 |
dc_clock_type_to_atom default path leaves *atom_clock_type uninitialized yet returns true — latent uninitialized-stack feed into ATOM setdceclock params (v2)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/dce112/command_table_helper2_dce112.c:333 · pass 2 · new |
none |
untested |
| DF-4937 |
Info |
1.9 |
dc_clock_type_to_atom default leaves *atom_clock_type uninitialized while returning true (v1 twin; consumer command_table.c:2432/2440)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/dce112/command_table_helper_dce112.c:333 · pass 2 · new |
none |
untested |
| DF-4938 |
Info |
1.9 |
gfx sw_init partial-failure strands KIQ/MQD kernel BOs and mqd_backup allocations — teardown unreachable behind the IP-block status.sw gate
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_gfx.c:282 · pass 2 · new |
leak |
untested |
| DF-4968 |
Info |
1.9 |
radeon GART table_size / num_gpu_pages 32-bit wrap (DF-2012's radeon twin): missing table-size invariant check in radeon_gart_table_{vram,ram}_alloc — power-of-two validator currently forces the wrapped size to 0 (fail-closed)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_gart.c:68 · pass 2 · new |
none |
untested |
| DF-4973 |
Info |
1.9 |
dm_target_dmirror.c is non-compiling, unregistered dead scratch code whose latent memory-safety bugs must never be revived — delete it
Base system — drivers · sys/dev/disk/dm/dmirror/dm_target_dmirror.c:135 · pass 2 · new |
none |
not_reproduced |
| DF-4980 |
Info |
1.9 |
amdgpu_atomfirmware_allocate_fb_scratch: VBIOS-controlled unclamped kernel allocation (up to 64MB) whose -ENOMEM return the caller ignores — NULL-scratch write chain on Linux-semantics allocators, dead error branch on DragonFly
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_atomfirmware.c:99 · pass 2 · new |
none |
untested |
| DF-5023 |
Info |
1.9 |
Unsupported 5GHz radio revision check fails open: the HAL_ENOTSUPP/goto bad rejection is compiled out unless AH_DEBUG (FreeBSD-inherited)
Base system — drivers · sys/dev/netif/ath/ath_hal/ar9001/ar9160_attach.c:228 · pass 2 · new |
fail-open hardware validation, no memory-safety impact |
untested |
| DF-5038 |
Info |
1.9 |
Unsupported 5GHz radio-revision check fails open: the HAL_ENOTSUPP/goto bad rejection is compiled out unless AH_DEBUG — ar9160/DF-5023 family twin, shipped fail-open empirically confirmed
Base system — drivers · sys/dev/netif/ath/ath_hal/ar9001/ar9130_attach.c:197 · pass 2 · new |
none |
untested |
| DF-5092 |
Info |
1.9 |
ar5416GpioCfgInput drives the GPIO pin (DRV_ALL) instead of tristating it, contradicting its own contract — BT-coexistence sense pins get actively driven by the MAC
Base system — drivers · sys/dev/netif/ath/ath_hal/ar5416/ar5416_gpio.c:146 · pass 2 · new |
none |
untested |
| DF-5098 |
Info |
1.9 |
oce_page_list() hardcodes the segment bound (8) regardless of the destination pages[] array size — latent out-of-bounds write into RQ/CQ mailbox commands
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/oce/oce_util.c:258 · pass 2 · new |
latent intra-object OOB write into fw command struct |
untested |
| DF-5100 |
Info |
1.9 |
joyread() busy-loops inside crit_enter() against WALL-CLOCK nanotime() — backward clock steps extend an interrupt-masked critical section beyond the configured timeout
Base system — drivers · sys/dev/misc/joy/joy.c:190 · pass 2 · new |
none |
inconclusive |
| DF-5114 |
Info |
1.9 |
dragon_saver(): set_video_mode() return value ignored — on failure vid would alias the wrong (32KB text) video window and 64000-byte writes would overrun it
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/syscons/dragon/dragon_saver.c:192 · pass 2 · new |
latent out-of-window video MMIO writes (not kernel RAM) |
untested |
| DF-5117 |
Info |
1.9 |
dmamap_cb continues after bus_dmamap_load error: uninitialized-stack segments[0].ds_addr is stored into dcons_paddr and published to FireWire peers via the ConfigROM, the stack bytes are printed to console, and dcons_conf is wired for a map whose load failed — while dcons_crom_attach ignores all three bus_dma return values and reports success
Base system — drivers · sys/dev/misc/dcons/dcons_crom.c:130 · pass 2 · new |
none |
untested |
| DF-5132 |
Info |
1.9 |
amdgpu_kv_notify_message_to_smu() fail-open response validation: SMC timeout (0x00) and undefined response codes 0x02–0xFD all reported as success to the whole kv_dpm power-management state machine
Base system — drivers · sys/dev/drm/amd/amdgpu/kv_smc.c:40 · pass 2 · new |
none |
untested |
| DF-5139 |
Info |
1.9 |
iwm_abort_notification_waits() is dead code: the device stop path never aborts pending notification waits, leaving control threads stalled for their full bounded timeout during firmware stop/restart
Base system — drivers · sys/dev/netif/iwm/if_iwm_notif_wait.c:155 · pass 2 · new |
stalled control threads / detach-race UAF write |
untested |
| DF-5187 |
Info |
1.9 |
Default drm_gem_dumb_map_offset returns offsets from the wrong address space: vma-manager address without DRM_GEM_MAPPING_KEY bits, unresolvable by (and unrelated to) the hash-key mmap path — latent (no in-tree caller)
Base system — drivers · sys/dev/drm/drm_dumb_buffers.c:123 · pass 2 · new |
none |
untested |
| DF-5204 |
Info |
1.9 |
Terminal-emulator registry sc_term_list is completely unsynchronized and the SCTERM_MODULE MOD_UNLOAD refcount gate is TOCTOU — kldload/kldunload of any third-party scterm module races console ioctls into list corruption or a call into freed module text
Base system — drivers · sys/dev/misc/syscons/scterm.c:74 · pass 2 · new |
list corruption / freed-text call on third-party module races |
untested |
| DF-5211 |
Info |
1.9 |
arch_io_reserve_memtype_wc() persists 4 uninitialized stack bytes into the global memrange descriptor array (mr_owner[4..7] leak into MEMRANGE_GET)
Base system — drivers · sys/dev/drm/linux_iomapping.c:111 · pass 2 · new |
attach-time BUG_ON panic / user-range vtopte panic / 4B stack disclosure |
untested |
| DF-5214 |
Info |
1.9 |
Dead dal_hw_factory_destroy() kfrees a struct hw_factory that only ever exists embedded in gpio_service — latent interior-pointer kfree (heap corruption) one future caller away
Base system — drivers · sys/dev/drm/amd/display/dc/gpio/hw_factory.c:101 · pass 2 · new |
latent interior kfree / kernel-CPU amplification / unkillable spin |
untested |
| DF-5231 |
Info |
1.9 |
Dead FPGA-diagnostics GPIO factory installs an all-NULL function table (create_* and both omitted define_*_registers): activating the unreachable MAXIMUS env would turn every GPIO operation into an immediate NULL function-pointer call
Base system — drivers · sys/dev/drm/amd/display/dc/gpio/diagnostics/hw_factory_diag.c:45 · pass 2 · new |
latent NULL-call panic / latent invalid-state commit |
untested |
| DF-5232 |
Info |
1.9 |
dml_init_instance's default branch is ASSERT-only (WARN_ON) and still commits lib->project with no bounding box loaded — latent invalid-state commit that would feed the dml_log2(0)/÷0 family (DF-4198) if a future project value lacks a case
Base system — drivers · sys/dev/drm/amd/display/dc/dml/display_mode_lib.c:38 · pass 2 · new |
latent NULL-call panic / latent invalid-state commit |
untested |
| DF-5234 |
Info |
1.9 |
u32-domain shift truncation: PF_LFB_SIZE<<24 wraps before assignment to u64 node_segment_size, silently mis-placing XGMI VRAM/GART apertures
Base system — drivers · sys/dev/drm/amd/amdgpu/gfxhub_v1_1.c:47 · pass 2 · new |
GPU VM faults / device hang from mis-placed apertures (config DoS) |
untested |
| DF-5281 |
Info |
1.9 |
Vendor-database parsing: ksscanf("%80[^\n]") writes up to 81 bytes into kmalloc(80); pci_load_vendor_data() writes one byte past the preloaded blob
Kernel memory corruption (UAF / OOB write) · sys/bus/pci/pci.c:3243 · pass 2 · new |
VPD lookup always ENXIO (DF-5279) / devq UAF read race (DF-5280) / boot-time off-by-one (DF-5281) |
untested |
| DF-5310 |
Info |
1.9 |
usb_open() error path after successful fifo opens leaks the opened fifos with curr_cpd pointing at the freed cpd (latent — devfs_set_cdevpriv cannot currently fail for a fresh open)
Base system — bus / platform / cpu · sys/bus/u4b/usb_dev.c:929 · pass 2 · new |
single-owner invariant break -> concurrent-setup entry / latent dangling curr_cpd |
untested |
| DF-5325 |
Info |
1.9 |
ukbd_detach mutates keyboard state and runs ukbd_interrupt() without kb_lock, racing the kb_lock-holding ukbd_timeout callout (in-bounds input-ring corruption on unplug)
Base system — bus / platform / cpu · sys/bus/u4b/input/ukbd.c:1380 · pass 2 · new |
garbled trailing key events / ring count drift at unplug |
untested |
| DF-5331 |
Info |
1.9 |
DELAY()/DRIVERSLEEP() do not clamp negative arguments — sign-extended usec count converts to a multi-hour (i8254: multi-month) kernel spin
Base system — bus / platform / cpu · sys/platform/pc64/isa/clock.c:386 · pass 2 · new |
none |
untested |
| DF-5391 |
Info |
1.9 |
No suspend/resume/shutdown lifecycle: controller loses programming across D3 while softc caches stay 'valid' — bus dead until reload
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:493 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-5435 |
Info |
1.9 |
uvc_buf_queue_req_bufs signed count handling is undefined behavior: count >= 0x80000000 makes int num negative, skipping the clamp and the allocation loop and reading an uninitialized mem (currently compiled to ENOMEM by the shipping gcc - fragile across toolchains)
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/uvc/uvc_buf.c:540 · pass 2 · new |
toolchain-fragile UB, no exploit on the shipping build |
untested |
| DF-5460 |
Info |
1.9 |
ichsmb_attach error path: lockuninit() runs while the interrupt handler is still registered and the smbus child is leaked
Base system — bus / platform / cpu · sys/bus/smbus/ichsmb/ichsmb.c:132 · pass 2 · new |
none |
untested |
| DF-5508 |
Info |
1.9 |
amdsmb_detach performs no quiesce/drain: kldunload during an in-flight locked transaction runs lockuninit() on a held lock and releases the I/O resource under live bus_read_1/bus_write_1 (post-release port I/O)
Base system — bus / platform / cpu · sys/bus/smbus/amdsmb/amdsmb.c:175 · pass 2 · new |
none |
untested |
| DF-5529 |
Info |
1.9 |
urio_open: WRITE-stall flag updated without sc_lock, asymmetric with the locked READ path
Base system — bus / platform / cpu · sys/bus/u4b/storage/urio.c:404 · pass 2 · new |
lost stall-flag update -> one spurious/missed ENDPOINT_HALT clear |
untested |
| DF-5567 |
Info |
1.9 |
cb_dumpdata() has no PAGE_SIZE floor on di->maxiosize: a dump device reporting sub-page si_iosize_max makes chunk==0 and spins the panic-dump loop forever (no dump, silent hang)
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/dump_machdep.c:168 · pass 2 · new |
silent panic-dump hang (no dump produced) |
untested |
| DF-5569 |
Info |
1.9 |
DRIVERSLEEP() divides by its usec argument with no zero guard (latent SIGFPE/vkernel panic; vkernel64-only divergence from pc64 twin)
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/systimer.c:358 · pass 2 · new |
none |
untested |
| DF-5598 |
Info |
1.9 |
ichsmb_pci.c:175 wires device_detach to ichsmb_detach, whose device_delete_child() return is ignored (ichsmb.c:689) before unconditional IRQ/IO teardown, lockuninit and a lying success return — DF-5544/5545/5562 family residue, with every deterministic veto path caught one level up (live-proven)
Base system — bus / platform / cpu · sys/bus/smbus/ichsmb/ichsmb_pci.c:175 · pass 2 · new |
none |
not_reproduced |
| DF-5599 |
Info |
1.9 |
ichsmb_pci_attach requests 16 I/O-port bytes first while ICH5+ SMBus BARs decode 32 (64 under QEMU's ICH9) — on BIOS-unprogrammed BARs pci_alloc_map under-claims the decode window, leaving the tail bytes hardware-decoded but rman-unowned
Base system — bus / platform / cpu · sys/bus/smbus/ichsmb/ichsmb_pci.c:233 · pass 2 · new |
none |
not_reproduced |
| DF-5606 |
Info |
1.9 |
sys/bus/u4b/net/ruephy.c is an unbuildable FreeBSD-API orphan — dead weight carrying latent ifm_cur derefs (:145/:231) and the DF-5001 empty-media branch
Base system — bus / platform / cpu · sys/bus/u4b/net/ruephy.c:47 · pass 2 · new |
none |
not_reproduced |
| DF-5644 |
Info |
1.9 |
smbus_attach mints NAMELESS wildcard children that any driver on the smbus bus with probe priority above GENERIC can hijack: ipmi_smbus_probe's unconditional BUS_PROBE_DEFAULT claims the generic child and all 96 address children — /dev/smbN disappears on non-SSIF platforms and duplicates the SSIF instance on SSIF ones
Base system — bus / platform / cpu · sys/bus/smbus/smbus.c:94 · pass 2 · new |
none |
untested |
| DF-5645 |
Info |
1.9 |
smbus_attach's probe loop issues 96 real bus transactions without smbus_request_bus — the only transaction issuer in the tree that bypasses the sc->owner arbitration, interleaving with owned transfers on re-attach and stalling attach by ~96 controller timeouts
Base system — bus / platform / cpu · sys/bus/smbus/smbus.c:95 · pass 2 · new |
none |
untested |
| DF-5651 |
Info |
1.9 |
Proximity-domain association compares only ProximityDomainLo (ignores ProximityDomainHi and x2APIC entries): NUMA domains >= 256 silently dropped or organized under the wrong socket
Base system — bus / platform / cpu · sys/platform/pc64/acpica/acpi_srat.c:112 · pass 2 · new |
none |
untested |
| DF-5665 |
Info |
1.9 |
FRAMEUP stack-walk heuristic accepts non-monotonic, non-aligned, non-contained frame pointers — a stale kernel %rbp can fault the chain walk (panic-only ceiling; hardening)
Kernel memory corruption (UAF / OOB write) · sys/cpu/x86_64/misc/ktr.c:67 · pass 2 · new |
none |
untested |
| DF-4442 |
Info |
1.8 |
Partially uninitialized mxt_message on the poller stack when the (device-controlled) T5 object size < 7 — stale stack bytes flow into the user-visible ELO pressure field
Base system — drivers · sys/dev/smbus/atmel_mxt/atmel_mxt.c:917 · pass 2 · new |
1-2 stale kernel stack bytes per ELO packet to the (already-root) device reader; physical-attacker/lying-device model |
untested |
| DF-5319 |
Info |
1.8 |
usbd_req_get_ss_hub_descriptor() length formula overflows struct usb_hub_ss_descriptor by one byte for nports >= 249 — latent kernel stack OOB write with device-controlled byte
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_request.c:1503 · pass 2 · new |
1-byte device-controlled write past 42-byte stack struct if a caller forwards unclamped bNbrPorts |
untested |
| DF-5491 |
Info |
1.8 |
pcfintr slave-transmitter paths transmit an uninitialized kernel stack byte onto the I2C wire (callbacks never fill *buf)
Kernel memory leaks / uninitialized disclosure · sys/bus/iicbus/pcf/pcf.c:396 · pass 2 · new |
leak |
untested |
| DF-3817 |
Low |
1.6 |
mpr_log_evt_handler(): hexdump() of firmware event data with unvalidated device-controlled length — OOB heap read to console
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr.c:2065 · pass 2 · new |
none |
inconclusive |
| DF-3977 |
Low |
1.6 |
ar5212GetLowerUpperValues with an empty pier list: uninitialized stack read (lp[0]) and 2-byte OOB stack read (ep[-1]) below tempChannelList, garbage pier selection can program MAX_RATE_POWER
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5212/ar5212_reset.c:2639 · pass 2 · new |
none |
inconclusive |
| DF-4119 |
Low |
1.6 |
Inverted error check at cypress_dpm_init:2059-2064 — ATOM engine-PLL reference divider used only on FAILURE (consuming uninitialized stack dividers), default used on success (cypress twin of DF-3976)
Base system — drivers · sys/dev/drm/radeon/cypress_dpm.c:2059 · pass 2 · new |
none |
inconclusive |
| DF-5448 |
Info |
1.6 |
uvscom_cfg_read_status() returns uninitialized kernel stack on a successful SHORT vendor control-IN (core actlen=NULL zeroization gap) — return value currently has no consumer, latent DF-1062/2342-class hazard
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/serial/uvscom.c:719 · pass 2 · new |
none |
untested |
| DF-3008 |
Low |
1.1 |
devfs_find_device_by_devid/name return an unreferenced cdev: kern.devname sysctl (CTLFLAG_ANYBODY) reads si_name after devfs_lock is released — microscopic UAF-read race
Kernel memory corruption (UAF / OOB write) · sys/vfs/devfs/devfs_core.c:910 · pass 2 · new |
none |
not_reproduced |
| DF-3070 |
Low |
1.1 |
tmpfs_reg_resize shrink path: 'aobj->size = osize;' self-assignment dead store — swap-object size never reduced after truncate (intended aobj->size = nsize), live upstream
Base system — filesystems · sys/vfs/tmpfs/tmpfs_subr.c:1055 · pass 2 · new |
none |
not_reproduced |
| DF-4234 |
Low |
1.1 |
txp(4) attach error path leaks the 256 rb_sd swdesc kmallocs (and ifmedia) when txp_alloc_rings() fails after allocating them
Base system — drivers · sys/dev/netif/txp/if_txp.c:948 · pass 2 · new |
none |
untested |
| DF-5080 |
Low |
1.1 |
nsgphy_status() decodes 1000baseT from link-partner advertisement with no master/slave resolution (MSR) or fault (MSF) check, and never reports the master role (IFM_ETH_MASTER)
Base system — drivers · sys/dev/netif/mii_layer/nsgphy.c:252 · pass 2 · new |
none |
untested |
| DF-5286 |
Low |
1.1 |
samount parses uninitialized 8192-byte heap as block-limits geometry on short READ BLOCK LIMITS
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/scsi/scsi_sa.c:1961 · pass 2 · new |
none |
untested |
| DF-3155 |
Info |
1.1 |
ext2 extent-block checksum verification is dead code — metadata_csum extent blocks are never verified, and the dormant implementation carries a latent eh_max-driven ~768KB OOB read
Kernel memory leaks / uninitialized disclosure · sys/vfs/ext2fs/ext2_csum.c:375 · pass 2 · new |
none |
untested |
| DF-3451 |
Info |
1.1 |
mld6_stop_listening mutates the shared MLD link-local templates without mld6_token — torn if_index embedding mis-addresses MLD DONE and mis-classifies the all-nodes comparison
Base system — network stack · sys/netinet6/mld6.c:163 · pass 2 · new |
none |
untested |
| DF-3602 |
Info |
1.1 |
Off-by-one OOB write in hdaa_pcmchannel_setup: ch->io[ret] = -1 with ret==16 writes one nid_t past nid_t io[16]
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pci/hda/hdaa.c:5317 · pass 2 · new |
intra-object clobber only |
untested |
| DF-3643 |
Info |
1.1 |
Unchecked PHY-read return values: EEE decisions and PHY write-backs derived from uninitialized kernel stack
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ig_hal/e1000_ich8lan.c:959 · pass 2 · new |
none |
not_reproduced |
| DF-3767 |
Info |
1.1 |
r100_cs_parse_packet0() off-by-one: '> n' boundary check permits auth[n] read one u32 past the const reg_safe_bm arrays
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r100.c:1392 · pass 2 · new |
4-byte rodata OOB read, decision-only |
untested |
| DF-3769 |
Info |
1.1 |
gen8 3-level PPGTT insert walk advances pdpe with a 4-level wrap constant and no runtime bound: latent OOB walk of the 4-entry pdp->page_directory array (GEM_BUG_ON-only guard is a production no-op)
Base system — drivers · sys/dev/drm/i915/i915_gem_gtt.c:1075 · pass 2 · new |
none |
untested |
| DF-4158 |
Info |
1.1 |
vga_save_font()/vga_load_font() accept unvalidated glyph index and count — latent OOB-write API one caller away from a real overflow (user-influenced mouse_char already flows in)
Base system — drivers · sys/dev/video/fb/vga.c:1189 · pass 2 · new |
none |
not_reproduced |
| DF-4446 |
Info |
1.1 |
DF i2c bit-bang shim is a stub: every radeon bit-banged DDC/EDID transfer is a silent no-op and the whole bit-bang block of radeon_i2c.c is dead code on DragonFly (display detection dead on default config)
Base system — drivers · sys/dev/drm/radeon/radeon_i2c.c:955 · pass 2 · new |
none |
untested |
| DF-4460 |
Info |
1.1 |
drm_gem_create_mmap_offset_size commits the offset-hash + unr publication before drm_vma_offset_add() is known to succeed — failure leaves on_map=true, a leaked idx/hash slot, and a retry that reports success with no vma offset
Base system — drivers · sys/dev/drm/drm_gem.c:566 · pass 2 · new |
none |
inconclusive |
| DF-4480 |
Info |
1.1 |
ac97mix_setrecsrc() reads codec->mix[AC97_MIXER_SIZE] out of bounds when src==0 (latent OOB read; guard present upstream)
Kernel memory leaks / uninitialized disclosure · sys/dev/sound/pcm/ac97.c:1069 · pass 2 · new |
Latent intra-object OOB read (mix[25] lands on name[]) + bogus RECSEL write + nonexistent 1<<25 return bit; no userspace disclosure |
untested |
| DF-4540 |
Info |
1.1 |
cxm_tuner: terminator-less band_codes[3]/assignments[17] walks rely on unenforced min_freq/min_channel data invariants - latent rodata OOB-read hazard (Info hardening)
Kernel memory leaks / uninitialized disclosure · sys/dev/video/cxm/cxm_tuner.c:725 · pass 2 · new |
none |
untested |
| DF-4586 |
Info |
1.1 |
mmgrab() reference on the DF shim mm_struct is a no-op (mm_count is never read) and dev_priv->mm_structs is keyed by the recyclable current->mm pointer: i915_mm_struct.mm dangles after process exit and can silently alias a later process
Base system — drivers · sys/dev/drm/i915/i915_gem_userptr.c:322 · pass 2 · new |
None in shipped config (all stale-pointer derefs compiled out; unpriv creates fail, root objects inert); becomes a genuine UAF/wrong-mm registration when the notifier path is ever enabled |
untested |
| DF-4645 |
Info |
1.1 |
ix(4): ixgbe_read_posted_mbx/ixgbe_write_posted_mbx dispatch to ops.read/ops.write with unvalidated size, bypassing the clamp/reject enforced only by the non-posted wrappers - latent OOB-to-MMIO and cross-VF-mailbox-spill footgun (twin of DF-4638 e1000_mbx); no in-tree caller (DF ix has no SR-IOV)
Base system — drivers · sys/dev/netif/ix/ixgbe_mbx.c:224 · pass 2 · new |
none |
untested |
| DF-4810 |
Info |
1.1 |
AHCI command-table stride ATA_AHCI_CT_SIZE=256 != sizeof(struct ata_ahci_cmd_tab)=1152: latent work-area OOB the moment per-tag (NCQ) dispatch is enabled
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/chipsets/ata-ahci.c:254 · pass 2 · new |
none |
untested |
| DF-4838 |
Info |
1.1 |
amdgpu userptr MMU-notifier invalidation is entirely absent on DragonFly — inert shim, compiled-out amdgpu_mn.c, stubbed userptr ioctl; latent missed-invalidation GPU DMA + stripped-kref p->mn UAF if ever enabled
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_mn.c:72 · pass 2 · new |
none |
untested |
| DF-4885 |
Info |
1.1 |
check_atom_bios() off-by-one/out-of-bounds heap reads: size==0x49 accepted then bios[0x49] read; memcmp reads up to 4 bytes past the buffer when bios_header_start+4==size
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_bios.c:52 · pass 2 · new |
none |
untested |
| DF-5060 |
Info |
1.1 |
feed_format_init does not bound the channel count: AFMT's 5-bit channel field (up to 31) flows unvalidated into ialign/oalign sizing against a reservoir and FXDIV table dimensioned for SND_CHN_MAX — latent reservoir heap overflow and snd_fxdiv_table OOB index if a >SND_CHN_MAX-channel format ever becomes constructible
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pcm/feeder_format.c:48 · pass 2 · new |
none (latent invariant gap) |
untested |
| DF-5088 |
Info |
1.1 |
Generic I2C HW-engine chunker boundary defects (max_payload_size==1 degenerate unaddressed 1-byte chunking; zero-length payload rejected FAILED_INCOMPLETE; timeout omits in-batch stretch accounting) — latent only: unit proven dead code in this port
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/i2c_generic_hw_engine.c:96 · pass 2 · new |
none |
untested |
| DF-5112 |
Info |
1.1 |
amdgpu_job_submit_direct latches *fence before checking the error and leaks ownership on failure — any future caller passing an uninitialized fence gets a wild kref_get (latent error-contract defect)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_job.c:164 · pass 2 · new |
latent (no affected caller) |
untested |
| DF-5122 |
Info |
1.1 |
Legacy drm_mmap: page/byte hash-key mismatch, user-token double-count in phys math, and PFN/byte return-convention inconsistency — three latent memory-safety-relevant defects in a d_mmap handler unreachable on every shipped driver
Base system — drivers · sys/dev/drm/drm_vm.c:91 · pass 2 · new |
none |
untested |
| DF-5123 |
Info |
1.1 |
hypercall_md() inline asm violates the GCC extended-asm contract: %r8 lifetime unprotected between two separate asm statements; indirect-call asm missing caller-saved-register and memory clobbers
Base system — drivers · sys/dev/virtual/hyperv/x86_64/hyperv_machdep.c:80 · pass 2 · new |
none |
not_reproduced |
| DF-5144 |
Info |
1.1 |
Latent unchecked res_pool deref + lockless dc->current_state walk in dead public API dc_plane_get_status()
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/core/dc_surface.c:115 · pass 2 · new |
none |
untested |
| DF-5159 |
Info |
1.1 |
WARN-only range guards: silent 64-bit wrap in bw_int_to_fixed_nonconst / bw_frc_to_fixed / bw_mul in release builds
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/calcs/bw_fixed.c:49 · pass 2 · new |
none |
untested |
| DF-5178 |
Info |
1.1 |
hw_hpd set_config consumes the gpio_config_data tagged union's hpd member without validating config_data->type
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/gpio/hw_hpd.c:107 · pass 2 · new |
latent tagged-union confusion -> masked MMIO mis-programming |
untested |
| DF-5190 |
Info |
1.1 |
vmalloc()/vzalloc()/vmalloc_user() implemented over plain kmalloc violate the Linux vmalloc contract (page alignment, whole-page ownership) that the legacy DRM _DRM_SHM userspace-mapping path is written against
Base system — drivers · sys/dev/drm/linux_vmalloc.c:107 · pass 2 · new |
none |
untested |
| DF-5192 |
Info |
1.1 |
Latent negative LRU bucket index in i915_gem_batch_pool_get walks cache_list[-1] and would write through the live engine pointer
Base system — drivers · sys/dev/drm/i915/i915_gem_batch_pool.c:95 · pass 2 · new |
none |
untested |
| DF-5225 |
Info |
1.1 |
amdgpu_vf_error: signed int ring counters used as modulo array indices — 2^31 wrap is signed-overflow UB and produces negative array-index writes (contained inside amdgpu_vf_error_buffer)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_vf_error.c:42 · pass 2 · new |
latent signed-overflow UB, in-struct self-corruption ceiling |
untested |
| DF-5233 |
Info |
1.1 |
Underlay OPP constructed with NULL regs/shift/mask while every entry in its vtable dereferences regs (latent NULL-deref held off only by four cross-file invariants)
Base system — drivers · sys/dev/drm/amd/display/dc/dce110/dce110_opp_v.c:39 · pass 2 · new |
latent NULL-deref panic (display DoS) |
untested |
| DF-5300 |
Info |
1.1 |
usbd_get_iface() computes udev->ifaces + iface_index before bounds-checking iface_index — out-of-bounds/NULL-base pointer arithmetic on every call with an unvalidated index
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/usb_device.c:2237 · pass 2 · new |
latent OOB pointer arithmetic |
untested |
| DF-5671 |
Info |
1.1 |
Stale MP-safety comment asserts a locked OR where code emits a nonlocked OR — safety of gd_spending pend path rests on an unstated, unenforced per-CPU invariant
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/ipl_funcs.c:45 · pass 2 · new |
latent silently-lost soft interrupts if a future cross-CPU caller follows the comment |
untested |
| DF-5401 |
Info |
1.0 |
mos_intr_callback() counts every interrupt as an output error and copies 4 bytes regardless of actlen into a never-used local; mos_bulk_write_callback() double-counts opackets
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/net/if_mos.c:862 · pass 2 · new |
non-canonical frames up the stack (DF-5400) / skewed interface stats (DF-5401) |
untested |
| DF-3119 |
Low |
0.8 |
flush_group refcount asymmetry: critical-error unmount decrements flg->refs for inodes that never incremented it
Base system — filesystems · sys/vfs/hammer/hammer_flusher.c:423 · pass 2 · new |
corruption |
not_reproduced |
| DF-5245 |
Low |
0.8 |
pmap_extract() KVA path duplicates the PDE validity check instead of testing the PTE — bogus non-zero physical address returned for unmapped kernel VA
Base system — bus / platform / cpu · sys/platform/pc64/x86_64/pmap.c:1739 · pass 2 · new |
none |
untested |
| DF-2737 |
Info |
0.8 |
Dead code: oid_running is never incremented, so the dynamic-oid teardown drain and CTLFLAG_DYING machinery in sysctl_remove_oid_locked are inoperative
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_sysctl.c:384 · pass 2 · new |
none |
untested |
| DF-2743 |
Info |
0.8 |
DIOCGKERNELDUMP handled with no privilege check or FWRITE gate (diskioctl → disk_dumpconf → set_dumper)
Privilege escalation (general use) · sys/kern/subr_disk.c:1186 · pass 2 · new |
none |
not_reproduced |
| DF-3007 |
Info |
0.8 |
devfs_uninit passes NULL to devfs_msg_send — two stores through address 0 (latent; SYSUNINIT path dead on stock configs)
Base system — filesystems · sys/vfs/devfs/devfs_core.c:1112 · pass 2 · new |
none |
not_reproduced |
| DF-3417 |
Info |
0.8 |
in6_ifdetach purges UDP moptions from udbinfo[0] only while the IPv4 twin loops all CPUs — latent purge gap that is complete today only by accident of udp6_attach()
Kernel memory corruption (UAF / OOB write) · sys/netinet6/in6_ifattach.c:818 · pass 2 · new |
none |
not_reproduced |
| DF-3522 |
Info |
0.8 |
ng_bridge constructor leaks priv->tab on ng_make_node_common failure
Base system — network stack · sys/netgraph/bridge/ng_bridge.c:310 · pass 2 · new |
none |
untested |
| DF-3549 |
Info |
0.8 |
sys/netgraph7/ng_device.c is non-compilable, unwired FreeBSD-port rot (all defects latent)
Base system — network stack · sys/netgraph7/ng_device.c:107 · pass 2 · new |
none |
untested |
| DF-3641 |
Info |
0.8 |
bnx_sysctl_std_refill iterates bnx_tx_ringcnt while indexing bnx_rx_ret_ring[] (copy-paste): only min(tx,rx) rings updated; validation admits negative cntmax that wraps uint16_t
Base system — drivers · sys/dev/netif/bnx/if_bnx.c:5684 · pass 2 · new |
none |
inconclusive |
| DF-3652 |
Info |
0.8 |
ci_populate_mvdd_value returns -EINVAL unconditionally, so MemoryACPILevel.MinMvdd is always programmed as 0 (inverted success return)
Base system — drivers · sys/dev/drm/radeon/ci_dpm.c:2355 · pass 2 · new |
none |
inconclusive |
| DF-3678 |
Info |
0.8 |
mpt_cam detach/error-path memory leaks: els_cmd_ptrs and tgt_cmd_ptrs never freed; sas_portinfo freed without its phy_info array
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt_cam.c:1198 · pass 2 · new |
none |
inconclusive |
| DF-3799 |
Info |
0.8 |
validate_tx_req_id inverts its NULL check — latent NULL-pointer dereference on out-of-range TX req_id (currently dead code)
Base system — drivers · sys/dev/virtual/amazon/ena/ena.c:1181 · pass 2 · new |
mbuf leak / latent NULL-deref |
untested |
| DF-4011 |
Info |
0.8 |
COMPAT_FREEBSD32 shim is un-compilable dead code (undefined com/td, 5-arg mpr_ioctl call); native RAID_ACTION ioctl referenced but unimplemented
Base system — drivers · sys/dev/raid/mpr/mpr_user.c:2425 · pass 2 · new |
none |
untested |
| DF-4017 |
Info |
0.8 |
COMPAT_FREEBSD32 shim is un-compilable dead code (5-arg mps_ioctl call, undefined td); RAID_ACTION32 dispatches to an unimplemented ioctl (DF-4011 mpr twin)
Base system — drivers · sys/dev/raid/mps/mps_user.c:2303 · pass 2 · new |
COMPAT_FREEBSD32 shim is un-compilable dead code (5-arg mps_ioctl call, undefined td; RAID_ACTION32 dispatches to an unimplemented ioctl). DF-4011 mpr twin. |
untested |
| DF-4230 |
Info |
0.8 |
cyapa_poll_thread reads uninitialized regs.stat when the device stops responding — garbage byte drives the bootloader re-init state machine
Base system — drivers · sys/dev/smbus/cyapa/cyapa.c:1164 · pass 2 · new |
none |
untested |
| DF-4436 |
Info |
0.8 |
drm_universal_plane_init: format_count>64 error path returns without kfree(format_types) or drm_mode_object_unregister - leaks the format array and the mode-object id
Base system — drivers · sys/dev/drm/drm_plane.c:193 · pass 2 · new |
none |
untested |
| DF-4464 |
Info |
0.8 |
uguru_attach sensor-table loop writes sc_sensors[i] with no bound against UGURU_MAX_SENSORS — one table entry away from a kernel softc overflow
Kernel memory corruption (UAF / OOB write) · sys/dev/powermng/uguru/uguru.c:867 · pass 2 · new |
Latent softc overflow if a future board table exceeds 27 entries (max today 26) |
untested |
| DF-4489 |
Info |
0.8 |
apply_below_the_range: BTR min-frame-duration clamp uses wrong constant (dead clamp) and computed min_frame_duration_in_ns is never used
Base system — drivers · sys/dev/drm/amd/display/modules/freesync/freesync.c:250 · pass 2 · new |
none |
untested |
| DF-4716 |
Info |
0.8 |
b_cmd=READ guard tests m.flags for XFR_BAD_SUBDISK (never set there) - condition always true, intended bad-subdisk exclusion dead
Base system — drivers · sys/dev/raid/vinum/vinumraid5.c:536 · pass 2 · new |
none |
untested |
| DF-4959 |
Info |
0.8 |
's' blink-spec filter typo `*s <= 'u'` accepts arbitrary characters as no-op blink ticks
Base system — drivers · sys/dev/misc/led/led.c:169 · pass 2 · new |
none |
untested |
| DF-4960 |
Info |
0.8 |
led_set() always returns 0, swallowing ENOMEM/ENOENT from its own logic (DragonFly divergence)
Base system — drivers · sys/dev/misc/led/led.c:273 · pass 2 · new |
none |
untested |
| DF-4998 |
Info |
0.8 |
hw.ral.msi_disable tunable is a no-op: pci_alloc_1intr receives ~ral_msi_disable, which is truthy for BOTH tunable values (0 → ~0 == 0xFFFFFFFF, 1 → ~1 == 0xFFFFFFFE)
Base system — drivers · sys/dev/netif/ral/if_ral_pci.c:254 · pass 2 · new |
none |
untested |
| DF-4999 |
Info |
0.8 |
generic_reg_wait prototype in dm_services.h transposes (mask, shift) vs the definition in dc_helper.c (shift, mask) — latent API-contract trap in the DC register hub; harmless today only because REG_WAIT's FN() expansion happens to match the definition
Base system — drivers · sys/dev/drm/amd/display/dc/dc_helper.c:254 · pass 2 · new |
none |
untested |
| DF-5018 |
Info |
0.8 |
radeon_mn.c is dead code: excluded from the kernel and module builds, gated on CONFIG_MMU_NOTIFIER which is never defined, and not even compilable if re-enabled (missing mmu_notifier compat API)
Base system — drivers · sys/dev/drm/radeon/radeon_mn.c:90 · pass 2 · new |
none |
untested |
| DF-5031 |
Info |
0.8 |
it_generic_stemp missing (int8_t) sign conversion for IT87xx temperature registers — sub-zero temps published as +128..+255 C under hw.sensors
Base system — drivers · sys/dev/powermng/it/it.c:259 · pass 2 · new |
none |
untested |
| DF-5063 |
Info |
0.8 |
Plexno()/Sdno() decode composite plex/sd minors with no bounds checks and propagate the -1 empty-slot sentinel into PLEX[-1].sdnos[] (latent heap OOB read; unreachable on current DragonFly)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/vinum/vinumutil.c:261 · pass 2 · new |
none |
not_reproduced fixed |
| DF-5221 |
Info |
0.8 |
framebuffer_alloc() ignores its size argument and never sets info->par — silent API-contract divergence, latent NULL-deref for any caller passing size > 0
Base system — drivers · sys/dev/drm/linux_compat.c:58 · pass 2 · new |
bounded refcount/heap leak (DF-5219) / latent API trap (DF-5221) |
untested |
| DF-5222 |
Info |
0.8 |
drm_legacy_ioremapfree() bypasses linuxkpi iounmap() bookkeeping, permanently orphaning iomap_list entries for any kmod user of the exported wrapper pair
Base system — drivers · sys/dev/drm/drm_memory.c:76 · pass 2 · new |
none |
untested |
| DF-5243 |
Info |
0.8 |
PROBE_SERIAL_NUM_0 frees the VPD page list with the wrong malloc type (M_DEVBUF vs allocation's M_CAMXPT) — per-type accounting corruption, upstream-corrected divergence
Base system — bus / platform / cpu · sys/bus/cam/cam_xpt.c:5922 · pass 2 · new |
heap OOB comparison oracle / race-gated UAF / accounting drift |
untested |
| DF-5379 |
Info |
0.8 |
sub_info->uvc_info stores a dangling pointer to the initializer's dead stack frame (persisted in the heap control graph)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/uvc/uvc_ctrls.c:776 · pass 2 · new |
latent stale-stack back-pointer in the heap control graph |
untested |
| DF-5405 |
Info |
0.8 |
sgread() parses the 12-byte old-sg header with uiomove short-copy semantics: pack_len/reply_len/pack_id left uninitialized when read() supplies fewer than 12 bytes
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_sg.c:764 · pass 2 · new |
UAF-read driven control flow (DF-5404) / uninit stack tag match (DF-5405) |
untested |
| DF-5549 |
Info |
0.8 |
fwmem_close() frees the fwmem_softc with the wrong malloc type (M_FW instead of M_FWMEM) — per-type accounting corruption and wrong-type-free panic on debug builds
Base system — bus / platform / cpu · sys/bus/firewire/fwmem.c:272 · pass 2 · new |
corrupted per-type malloc accounting + wrong-type-free panic on debug builds |
untested |
| DF-5621 |
Info |
0.8 |
MWAIT hint word validated only in its low 8 bits: firmware-supplied GAS.Address bits 31:8 forwarded verbatim to MWAIT EAX (and GAS BitWidth vendor field never checked)
Base system — bus / platform / cpu · sys/platform/pc64/acpica/acpi_cstate_machdep.c:50 · pass 2 · new |
forward-compat: reserved hint bits to MWAIT EAX |
untested |
| DF-5670 |
Info |
0.8 |
hw.acpi.reset_video is a dead control surface: sysctl+tunable registered by acpi_machdep_init() with zero consumers anywhere in sys/
Base system — bus / platform / cpu · sys/platform/pc64/acpica/acpi_machdep.c:49 · pass 2 · new |
none |
untested |
| DF-5010 |
Info |
0.7 |
vkd_doio performs COW-mode bcopy and backing-file pread/pwrite with no driver-level bounds check on bio_offset/b_bcount (latent arbitrary-offset OOB R/W in the vkernel host process; shielded today only by dscheck)
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/vkernel/disk/vdisk.c:311 · pass 2 · new |
none |
not_reproduced fixed |
| DF-5252 |
Info |
0.7 |
RT5592 TXWI extension: usbd_copy_in reads sizeof(data->desc)+4 bytes from data->desc, a 4-byte out-of-bounds read of desc/ridx/padding into the USB transfer buffer
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/wlan/if_run.c:3089 · pass 2 · new |
UAF read/write on freed node (DF-5250) / repeatable mbuf+node leak DoS (DF-5251) / device-bound 4B OOB (DF-5252) |
untested |
| DF-4050 |
Info |
0.6 |
adjust_display_pll_v2 never executes the AdjustDisplayPll BIOS command (missing EXEC block; always returns BP_RESULT_FAILURE)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/bios/command_table.c:1457 · pass 2 · new |
none |
untested |
| DF-5398 |
Info |
0.6 |
urndis(4) control channel: RNDIS response length never validated against the actually-transferred size — short device responses are parsed from stale sc_response_buf bytes
Base system — bus / platform / cpu · sys/bus/u4b/net/if_urndis.c:418 · pass 2 · new |
none |
untested |
| DF-5412 |
Info |
0.6 |
udav_csr_read1() ignores control-request failure and returns an uninitialized stack byte (error-masking; device-register influence only)
Base system — bus / platform / cpu · sys/bus/u4b/net/if_udav.c:395 · pass 2 · new |
none |
untested |
| DF-4612 |
Low |
0.5 |
init_transmitter_control steers handler selection with uninitialized stack byte crev when the VBIOS master command table lacks dig1transmittercontrol (v2 twin of DF-4047's sub-issue)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/command_table2.c:178 · pass 2 · new |
none |
not_reproduced |
| DF-4905 |
Low |
0.5 |
k8pnow_states() PST walk advances past BIOS_START+BIOS_LEN with no bounds check (kernel OOB read of physical 0x100000+ region)
Kernel memory leaks / uninitialized disclosure · sys/dev/powermng/powernow/powernow.c:187 · pass 2 · new |
none |
untested |
| DF-5615 |
Low |
0.5 |
fwdma_malloc leaks the dma_tag when bus_dmamem_alloc fails (the '/* XXX destory tag */' path)
Base system — bus / platform / cpu · sys/bus/firewire/fwdma.c:94 · pass 2 · new |
none (resource leak): one leaked bus_dma_tag (M_DEVBUF) per failed attach-time allocation; attach-error path only |
untested |
| DF-2613 |
Info |
0.5 |
Off-by-one bounds checks in SES config-page walkers allow a 1-byte heap over-read past the 8192-byte sdata allocation
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/scsi/scsi_ses.c:1311 · pass 2 · new |
none |
inconclusive |
| DF-2934 |
Info |
0.5 |
m_tag_alloc silently truncates int len/type into uint16_t m_tag_len/m_tag_id (allocation sized from untruncated len)
Base system — kernel core (kern) · sys/kern/uipc_mbuf2.c:260 · pass 2 · new |
none |
untested |
| DF-3123 |
Info |
0.5 |
Latent memory-safety landmines in the #if 0-disabled htree write path: heap memcpy overflow in the index-node split, four double-brelse-after-bwrite error paths, u16 reclen truncation at 64K blocks
Kernel memory corruption (UAF / OOB write) · sys/vfs/ext2fs/ext2_htree.c:457 · pass 2 · new |
none |
untested |
| DF-3151 |
Info |
0.5 |
HAMMER1 mirror ioctls: mirror->size unchecked in mirror_read and mirror->count unchecked in both paths — wrap-around bypasses the ubuf bounds guards (hardening)
Base system — filesystems · sys/vfs/hammer/hammer_mirror.c:156 · pass 2 · new |
none |
untested |
| DF-3210 |
Info |
0.5 |
krpc_call() early-error exits (socreate/sosetopt/sobind) leak the caller's request mbuf chain — md_mount/md_lookup_swap's 'message already freed' ownership contract is false
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_mountrpc.c:146 · pass 2 · new |
none |
untested |
| DF-3306 |
Info |
0.5 |
expire_upcalls() frees bw meters without unschedule_bw_meter() — latent UAF (FreeBSD divergence; unreachable today)
Kernel memory corruption (UAF / OOB write) · sys/net/ip_mroute/ip_mroute.c:1423 · pass 2 · new |
none |
untested fixed |
| DF-3335 |
Info |
0.5 |
netmap_sw_to_nic() double-increments k1: skipped TX rings and latent kring array OOB (currently dead code — no d_poll wired)
Base system — network stack · sys/net/netmap/netmap.c:798 · pass 2 · new |
none |
untested |
| DF-3461 |
Info |
0.5 |
in6_cksum() sanity check bypassable via u32 wrap of off+len — intended panic becomes uncontrolled NULL-mbuf deref (defense-in-depth)
Base system — network stack · sys/netinet6/in6_cksum.c:116 · pass 2 · new |
none |
untested |
| DF-3616 |
Info |
0.5 |
ath_txq_addholdingbuf bounds check is off-by-one and comes after the OOB pointer arithmetic; ath_freebuf has no check at all (hardening, currently unreachable)
Base system — drivers · sys/dev/netif/ath/ath/if_ath.c:4805 · pass 2 · new |
latent one-past-end hardening |
untested |
| DF-3676 |
Info |
0.5 |
mpt_abort_target_ccb indexes trt[target_lun] without the MPT_MAX_LUNS guard every sibling target-mode XPT case has
Base system — drivers · sys/dev/disk/mpt/mpt_cam.c:4631 · pass 2 · new |
none |
inconclusive |
| DF-3899 |
Info |
0.5 |
Off-by-one bounds guard in intel_get_shared_dpll_id(): one-past-end pointer passes the WARN_ON and yields id == num_shared_dpll (== 7 == I915_NUM_PLLS on ICL)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_dpll_mgr.c:103 · pass 2 · new |
none |
untested |
| DF-4096 |
Info |
0.5 |
emupchan_init feeds unchecked emu_valloc() NULL into emu_vinit() — latent NULL-pointer write (unreachable today, hardening)
Base system — drivers · sys/dev/sound/pci/emu10k1.c:824 · pass 2 · new |
none |
inconclusive |
| DF-4131 |
Info |
0.5 |
Latent stack OOB in oce_mbox_eqd_modify_periodic: protocol array delay[8] filled from set_eqd[OCE_MAX_EQ=32] with num<=neqs
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/oce/oce_mbox.c:1795 · pass 2 · new |
none |
not_reproduced |
| DF-4282 |
Info |
0.5 |
drm addmap/rmmap validation gaps: unbounded _DRM_SCATTER_GATHER offset, x86_64 offset-wrap check compiled out, RM_MAP lacks ownership check (tokens enumerable via unauthenticated GET_MAP)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_bufs.c:311 · pass 2 · new |
Currently theoretical (root-gated creation + the mmap units mismatch). |
untested |
| DF-4361 |
Info |
0.5 |
select_tuner() bounds the tuners[] dispatch index on the upper side only — a negative tuner type would install an out-of-bounds TUNER pointer that consumers dereference (hardening)
Base system — drivers · sys/dev/video/bktr/bktr_tuner.c:710 · pass 2 · new |
none |
untested |
| DF-4380 |
Info |
0.5 |
Latent divide-by-zero family in calculate_ss: unvalidated ATOMBIOS spread-spectrum percentage_divider and modulation rate (runtime-dead today: ENABLE_SS hardcoded 0)
Base system — drivers · sys/dev/drm/amd/display/dc/dce/dce_clock_source.c:673 · pass 2 · new |
Latent divide-by-zero family in calculate_ss (percentage_divider / modulation rate / modulation_time); runtime-dead today because ENABLE_SS is hardcoded 0 at both producers. Becomes a crafted-VBIOS #DE panic the moment any change wires it. |
untested |
| DF-4384 |
Info |
0.5 |
set_dp_phy_pattern_80bit_custom() ignores param->custom_pattern_size and unconditionally reads pattern[0..9] (latent OOB read for future callers)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dcn10/dcn10_link_encoder.c:292 · pass 2 · new |
none |
untested |
| DF-4411 |
Info |
0.5 |
acpi_asus probe/attach mutate the driver-wide static model tables (non-idempotent probe, cross-device state contamination)
Base system — drivers · sys/dev/acpi/acpi_asus/acpi_asus.c:682 · pass 2 · new |
none |
untested |
| DF-4417 |
Info |
0.5 |
wal_add() leaks the previous W/A list kmalloc on every growth step; i915_wa_lists are never freed at teardown
Base system — drivers · sys/dev/drm/i915/intel_workarounds.c:604 · pass 2 · new |
none |
untested |
| DF-4449 |
Info |
0.5 |
audio_regs[] defines 6 entries while res_cap.num_audio=7 — the audio_regs[6] global-out-of-bounds fixed upstream in 44035ec2fde1 (KASAN, Cc:stable) is latent in this tree behind resource_construct's pipe_count cap
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/dce120/dce120_resource.c:322 · pass 2 · new |
none |
untested |
| DF-4456 |
Info |
0.5 |
radeon_crtc_do_set_base returns 'false' (0 = success) for unsupported framebuffer formats
Base system — drivers · sys/dev/drm/radeon/radeon_legacy_crtc.c:417 · pass 2 · new |
Swallowed error masquerading as success: unsupported fb formats leave stale scanout while modeset reports success (DRM-master-only) |
untested |
| DF-4463 |
Info |
0.5 |
amdtemp_detach frees dev.cpu.N.temperature oids but leaves dangling entries in the cpu devices' sysctl contexts — latent UAF if a cpu device is ever destroyed
Kernel memory corruption (UAF / OOB write) · sys/dev/powermng/amdtemp/amdtemp.c:674 · pass 2 · new |
Latent UAF/double-free if a cpu device is ever hot-destroyed (none today - pc64 cpu driver has no detach) |
untested |
| DF-4491 |
Info |
0.5 |
intel_lvds_init failed-path leaks the DDC EDID buffer when the EDID yields no preferred mode
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_lvds.c:1001 · pass 2 · new |
One-shot boot/probe-time heap leak (up to EDID size) per driver load; no unpriv trigger, no disclosure, no UAF |
untested |
| DF-4499 |
Info |
0.5 |
amdgpu_amdkfd_gfx_v9: entire KFD backend is dead code on DragonFly; latent unvalidated vmid/engine_id index arithmetic should be guarded before KFD is ever enabled
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c:291 · pass 2 · new |
none |
inconclusive |
| DF-4536 |
Info |
0.5 |
lptopen() sc_state check-then-set TOCTOU: double open double-allocates and the paired closes double-free the same M_LPT chunks (latent — currently blocked by the mplock)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/lpt/lpt.c:458 · pass 2 · new |
none |
not_reproduced |
| DF-4547 |
Info |
0.5 |
lm_setup_sensors copies sensor tables into sc->sensors[WB_MAX_SENSORS] with no bound — w83792d_sensors already fills all 19 slots, zero headroom
Base system — drivers · sys/dev/powermng/lm/lm78.c:635 · pass 2 · new |
None today; defense-in-depth against a future one-row change turning into a softc overflow at probe |
untested |
| DF-4565 |
Info |
0.5 |
Unbounded flicker-filter table index: a non-matching flicker_removal exits the SLOPE_limit walk at i==5 and reads SLOPE_value[5]/YCOEF_value[5]/YCOEF_EN_value[5] one element past the [5]-element arrays (latent; unreachable with current constants)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_legacy_tv.c:647 · pass 2 · new |
none |
untested |
| DF-4574 |
Info |
0.5 |
adw_init_chip restores the saved BIOS image to the wrong LRAM base (addr + ADW_MC_BIOSLEN instead of addr + ADW_MC_BIOSMEM): 16-byte-shifted restore smears BIOS bytes over microcode variables 0x9C/0x9E and leaves 0x40-0x4F clobbered
Base system — drivers · sys/dev/disk/advansys/adwlib.c:353 · pass 2 · new |
Chip-internal correctness only: WDTR_ABLE/SDTR_ABLE LRAM words hold arbitrary BIOS bytes (negotiation misbehavior on the EEPROM-checksum-failed corner); no host memory impact |
untested |
| DF-4578 |
Info |
0.5 |
amdgpu_amdkfd_gfx_v8: entire gfx8 KFD backend is dead code on DragonFly; latent unvalidated vmid/engine index arithmetic and a structurally broken fw-NULL guard must be fixed before KFD is ever enabled (gfx8 sibling of DF-4499)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c:245 · pass 2 · new |
none |
untested |
| DF-4638 |
Info |
0.5 |
ig_hal: e1000_read_posted_mbx/e1000_write_posted_mbx dispatch to ops.read/ops.write with unvalidated size, bypassing the read-clamp/write-reject discipline enforced only by the non-posted wrappers — latent kernel-buffer overflow footgun on the live igb VF path
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ig_hal/e1000_mbx.c:246 · pass 2 · new |
none demonstrated (latent API hazard): any future caller of e1000_read_posted_mbx/e1000_write_posted_mbx with size > mbx->size(16) gets an unconditional ops.read/ops.write word loop (e1000_mbx.c:476-477/:512-513/:697-698/:736-737) past its buffer — kernel overflow with PF-controlled mailbox content plus OOB E1000_VMBMEM register-array writes; all in-tree callers (e1000_vf.c) pass sizes 1/2/3/16 with correctly sized buffers |
untested |
| DF-4667 |
Info |
0.5 |
phm_wait_on_indirect_register passes (mask, value) to a (index, value, mask) callee — swapped arguments, currently latent
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/smu_helper.c:145 · pass 2 · new |
None today (sole caller passes value==mask); future callers could silently skip an SMC firmware-ready handshake and race the microcontroller during clock/voltage programming |
untested |
| DF-4736 |
Info |
0.5 |
Unguarded indirect call dev_priv->display.hpd_irq_setup() in intel_hpd_irq_handler() — guard asymmetry vs sibling call sites (latent NULL function-pointer call)
Base system — drivers · sys/dev/drm/i915/intel_hotplug.c:483 · pass 2 · new |
none |
inconclusive |
| DF-4768 |
Info |
0.5 |
psp_cmd_submit_buf ignores resp.status when ucode==NULL: TMR/ASD firmware-command failures silently reported as success
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_psp.c:137 · pass 2 · new |
none |
inconclusive |
| DF-4788 |
Info |
0.5 |
Pin-busy bitmap indexed by unbounded caller-supplied en (no number_of_pins bound) and NULL create_generic/sync/gsl creator pointers invoked unchecked
Base system — drivers · sys/dev/drm/amd/display/dc/gpio/gpio_service.c:197 · pass 2 · new |
none |
untested |
| DF-4790 |
Info |
0.5 |
Dead debug printers carry latent unvalidated firmware-count walks (mpr_print_evt_sas NumEntries) and mpr_describe_table relies on a sentinel-plus-one convention
Base system — drivers · sys/dev/raid/mpr/mpr_table.c:72 · pass 2 · new |
none |
untested |
| DF-4811 |
Info |
0.5 |
aps(4): every EC transaction reads and writes I/O port 0x161F, one byte past the driver's allocated 0x1f-byte I/O resource window
Kernel memory corruption (UAF / OOB write) · sys/dev/powermng/aps/aps.c:102 · pass 2 · new |
none |
untested |
| DF-4861 |
Info |
0.5 |
aux_engine_funcs.configure left NULL in the dce110 (and all i2caux) AUX engine func tables while dal_i2caux_configure_aux() invokes it unconditionally — latent NULL function-pointer call/panic on any future caller
Base system — drivers · sys/dev/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c:451 · pass 2 · new |
panic |
untested |
| DF-4912 |
Info |
0.5 |
Dead stats module carries a latent length==0 off-by-one OOB write in mod_stats_update_event (unsigned wrap → event_string[-1]); module not compiled by any Makefile — delete it or guard
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/modules/stats/stats.c:316 · pass 2 · new |
none |
untested |
| DF-4921 |
Info |
0.5 |
Parent-NIC softc type confusion gated only by 4-char driver-name prefix (strncmp "stge")
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mii_layer/ip1000phy.c:435 · pass 2 · new |
none |
untested |
| DF-4934 |
Info |
0.5 |
dig_encoder_sel_to_atom computes a 7-case DIGx_SEL mapping then discards it with literal 'return 0' (v2; DF-2030 dce110 sibling; upstream-intended dead code)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/dce112/command_table_helper2_dce112.c:151 · pass 2 · new |
none |
untested |
| DF-4936 |
Info |
0.5 |
dig_encoder_sel_to_atom dead 7-case mapping discarded by literal 'return 0' (v1 twin of DF-2030; file additionally dead code on genuine hardware)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/dce112/command_table_helper_dce112.c:151 · pass 2 · new |
none |
untested |
| DF-4984 |
Info |
0.5 |
dm_target_snapshot.c is unbuildable, unregistered dead scratch code (dmirror DF-4973 class) whose latent memory-safety bugs — cross-allocator free, double-unbusy refcount underflow, NULL-deref panics — must never be revived; delete it
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/snapshot/dm_target_snapshot.c:113 · pass 2 · new |
none |
untested |
| DF-5034 |
Info |
0.5 |
acquire_i2c_hw_engine bounds the signed gpio_ddc_line index only above (line < GPIO_DDC_LINE_COUNT): GPIO_DDC_LINE_UNKNOWN(-1) passes and reads i2c_hw_engines[-1] — dce110 twin of DF-2089 (dce80)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.c:98 · pass 2 · new |
none |
untested |
| DF-5036 |
Info |
0.5 |
Shared family constructor dal_i2caux_dce110_construct indexes callee-static 6-entry hw_ddc_lines[]/hw_aux_lines[] with the caller-supplied num_i2caux_inst — latent OOB read feeding an unbounded engine-array store (DF-4827 pattern-class)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/i2caux/dce110/i2caux_dce110.c:127 · pass 2 · new |
none |
untested |
| DF-5056 |
Info |
0.5 |
smu10_smu_init error paths leave hwmgr->smu_backend dangling after kfree(priv)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/smu10_smumgr.c:232 · pass 2 · new |
latent UAF hardening, not reachable in this tree |
untested |
| DF-5069 |
Info |
0.5 |
psm_fini_power_state_table() leaves hwmgr->boot_ps and hwmgr->uvd_ps dangling into the freed hwmgr->ps allocation (incomplete cleanup; no in-tree reader today)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/pp_psm.c:102 · pass 2 · new |
none |
untested |
| DF-5090 |
Info |
0.5 |
ruephy(4) is dead code — unreachable carrier of the DF-4930/DF-5001 empty-media attach-panic leaf and the DF-1973/DF-5009 service KASSERT: no NIC driver named "rue" exists anywhere in DragonFly
Base system — drivers · sys/dev/netif/mii_layer/ruephy.c:96 · pass 2 · new |
none |
untested |
| DF-5099 |
Info |
0.5 |
joyopen/joyclose exclusive-access protocol is an unsynchronized check-then-set (TOCTOU) — EBUSY exclusivity bypass
Base system — drivers · sys/dev/misc/joy/joy.c:159 · pass 2 · new |
none |
inconclusive |
| DF-5101 |
Info |
0.5 |
Non-atomic, lock-inconsistent dev->irq_enabled test-and-clear in drm_irq_uninstall() admits double driver->irq_uninstall()+free_irq(), amplified into a double-free/UAF by the unlocked free_irq() lookup in linux_irq.c
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_irq.c:182 · pass 2 · new |
none |
inconclusive |
| DF-5107 |
Info |
0.5 |
SCDC channel-lock status bit macros use '<' comparison instead of '<<' shift: SCDC_CH0_LOCK==0 (channel-0 lock undetectable) and SCDC_CH_LOCK_MASK aliases SCDC_CLOCK_DETECT
Base system — drivers · sys/dev/drm/drm_scdc_helper.c:53 · pass 2 · new |
latent logic defect (no runtime path) |
untested |
| DF-5118 |
Info |
0.5 |
dcons_crom_detach half-clears the global dcons state: dcons_conf->dma_map is left dangling over the map freed two lines below and the module-static dcons_paddr is never reset — a re-attached dcons_crom takes dmamap_cb's 'already set' early-return and can never re-wire dcons_conf, silently disabling DMA sync; memory safety rests solely on the incidental dma_tag!=NULL guards in dcons_os.c
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/dcons/dcons_crom.c:199 · pass 2 · new |
none |
untested |
| DF-5142 |
Info |
0.5 |
rdrand_rng() asm loop body executes unconditionally: called with limit < 8 it writes a full 8-byte RDRAND word past a sub-word buffer
Kernel memory corruption (UAF / OOB write) · sys/dev/crypto/rdrand/rdrand.c:85 · pass 2 · new |
latent 8B OOB write on sub-word call / GPU DMA to aliased physical pages |
untested |
| DF-5147 |
Info |
0.5 |
dcons_load_buffer() trusts wire-supplied buffer geometry (sizes, ooffset/ioffset, 24-bit pos) with an overflowable sum-only sanity check — latent arbitrary-displacement OOB read/write primitive in dead-but-exported module code
Kernel memory leaks / uninitialized disclosure · sys/dev/misc/dcons/dcons.c:158 · pass 2 · new |
none |
untested |
| DF-5152 |
Info |
0.5 |
if_ath_alq.c is permanently dead, unportable code on DragonFly: ATH_DEBUG_ALQ cannot be defined (commented out of sys/conf/options) and the FreeBSD ALQ API it calls does not exist anywhere in the DF tree
Base system — drivers · sys/dev/netif/ath/ath/if_ath_alq.c:58 · pass 2 · new |
none as shipped (dead TU); latent strncpy/race on revival |
untested |
| DF-5154 |
Info |
0.5 |
dpt_pci_attach() leaks parent_dmat on the dpt_init() failure path — dpt_free()'s init_level ladder never destroys the DMA tag the PCI glue created
Base system — drivers · sys/dev/raid/dpt/dpt_pci.c:143 · pass 2 · new |
none |
untested |
| DF-5156 |
Info |
0.5 |
orm_identify() scan loop lacks a MAX_ROMS bound — latent OOB write past the 196-byte orm_softc (rid[]/res[]), currently unreachable because the ISA bus caps SYS_RES_MEMORY rids at ISA_NMEM=8
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/orm/orm.c:104 · pass 2 · new |
none |
not_reproduced |
| DF-5164 |
Info |
0.5 |
amdgpu_atombios_i2c_channel_trans() omits the memset of its stack atom parameter block that its sibling performs — defense-in-depth only, every struct byte is coincidentally initialized today
Base system — drivers · sys/dev/drm/amd/amdgpu/atombios_i2c.c:169 · pass 2 · new |
none |
untested |
| DF-5172 |
Info |
0.5 |
vcdstrategy() drains sc->bio_queue with no serialization — latent SMP TAILQ-corruption race, shielded today only by the missing D_MPSAFE flag (mplock)
Base system — drivers · sys/dev/virtual/vkernel/cdrom/vcd.c:148 · pass 2 · new |
none |
not_reproduced |
| DF-5174 |
Info |
0.5 |
Operator-precedence bug in ata_jmicron_ident: '& 0x02 >> 1' tests bit 0 of config 0x40 instead of bit 1, misidentifying the AHCI/PATA function in the device description
Base system — drivers · sys/dev/disk/nata/chipsets/ata-jmicron.c:58 · pass 2 · new |
cosmetic description string |
untested |
| DF-5185 |
Info |
0.5 |
intel_guc_ads_create() indexes GUC_MAX_ENGINES_NUM-sized ADS arrays with sparse engine->guc_id (hw_id) values — out-of-bounds array writes on ICL (VCS3_HW=6) that stay contained inside the packed blob but corrupt GuC-visible ADS fields
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_guc_ads.c:109 · pass 2 · new |
corrupted GuC-visible ADS data (in-blob contained) |
untested |
| DF-5188 |
Info |
0.5 |
ttm_agp_backend.c is dead, un-compilable, half-ported code: excluded from every build (CONFIG_AGP undefined, asm/agp.h absent), and if ever revived carries a pages-vs-bytes 8x under-allocation (heap OOB write in the bind loop) plus an agp_memory leak on every AGP rebind
Base system — drivers · sys/dev/drm/ttm/ttm_agp_backend.c:60 · pass 2 · new |
latent: 8x heap OOB write of page pointers + unbounded kernel agp_memory leak, only if the dead backend is ever compiled in; impact today: none |
untested |
| DF-5196 |
Info |
0.5 |
warp_update() wrap off-by-one: star coordinate resting at exactly SCRW*SCRH is drawn and erased at vid[64000], one byte past the 64000-byte M_VGA_CG320 framebuffer (`>` must be `>=`)
Base system — drivers · sys/dev/misc/syscons/warp/warp_saver.c:72 · pass 2 · new |
none |
untested |
| DF-5236 |
Info |
0.5 |
Dead FPGA-diagnostics hw_translate installs an all-NULL function table: activating the MAXIMUS env would turn every GPIO id/offset translation into an immediate NULL function-pointer call (kernel panic)
Base system — drivers · sys/dev/drm/amd/display/dc/gpio/diagnostics/hw_translate_diag.c:34 · pass 2 · new |
none |
untested |
| DF-5237 |
Info |
0.5 |
cxm_dec_fw: decoder firmware download loop bounded by hardcoded CXM_FW_SIZE (256 KiB), not the actual object size — latent kernel-rodata OOB read if an under-length firmware blob is compiled in
Kernel memory leaks / uninitialized disclosure · sys/dev/video/cxm/cxm_dec_fw.c:3 · pass 2 · new |
latent rodata OOB read DMA'd to card, no userspace read-back |
untested |
| DF-5238 |
Info |
0.5 |
cxm_enc_fw: encoder firmware download loop bounded by hardcoded CXM_FW_SIZE (256 KiB), not the actual object size — latent kernel-rodata OOB read if an under-length firmware blob is compiled in (twin of the dec_fw finding)
Kernel memory leaks / uninitialized disclosure · sys/dev/video/cxm/cxm_enc_fw.c:3 · pass 2 · new |
latent rodata OOB read DMA'd to card, no userspace read-back |
untested |
| DF-5275 |
Info |
0.5 |
cdregister(): cdchanger leaked on camq_init failure and callout_init'd twice
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_cd.c:889 · pass 2 · new |
CCB double-release UAF family (DF-5274) / attach-time leak (DF-5275) |
untested |
| DF-5291 |
Info |
0.5 |
pmap_growkernel() PDP-page branch allocates the new PDP page without vm_page_wire() (the PT branch wires at :1623) — latent reclaim/reuse of an installed kernel page-directory-pointer page
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/pmap.c:1584 · pass 2 · new |
wrong-page /dev/kmem I/O (DF-5290) / latent PDP reclaim (DF-5291) |
untested |
| DF-5317 |
Info |
0.5 |
USB_FS_STOP bounds-checks u.pstop->ep_index but indexes the array with u.pstart->ep_index — latent union type-confusion, benign only because the two ioctl structs are layout-identical
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_generic.c:1435 · pass 2 · new |
latent unchecked index (DF-5317) / unpriv log flooding (DF-5318) |
untested |
| DF-5354 |
Info |
0.5 |
usb_temp_setup() correctness silently depends on scratch-union aliasing between temp_setup and hw_ep_scratch
Base system — bus / platform / cpu · sys/bus/u4b/template/usb_template.c:1247 · pass 2 · new |
refactor trap toward bad free / leak on gadget bring-up |
untested |
| DF-5390 |
Info |
0.5 |
write_started/read_started are never set anywhere — the compensating-STOP cleanup in set_slave_addr() is dead code
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_iic.c:266 · pass 2 · new |
SMBus/I2C wire-protocol corruption + silent write failures + lifecycle gaps |
untested |
| DF-5429 |
Info |
0.5 |
Error-path hygiene in the v4l2 layer: uninitialized return + leak in uvc_v4l2_reg(), ignored devfs_set_cdevpriv() in open(), NULL check after deref in query_cap()
Base system — bus / platform / cpu · sys/bus/u4b/uvc/uvc_v4l2.c:125 · pass 2 · new |
defensive correctness only |
untested |
| DF-5441 |
Info |
0.5 |
EVPD/page-code INQUIRY branch leaves descr->timeout stale (zero on first use), feeding an unconfigured timeout into the CTIO engine
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_targ_bh.c:565 · pass 2 · new |
per-attach softc leak + speculative abort-refusing-SIM UAF (DF-5440) / stale CTIO timeout (DF-5441) |
untested |
| DF-5469 |
Info |
0.5 |
lookup_res() rid aliasing: truncating PCI_RID2BAR division maps unaligned rids (0x11-0x13, 0x15-0x17, ...) onto the lower BAR's cache slot while bus_alloc_resource programs config space at the unaligned offset
Base system — bus / platform / cpu · sys/bus/pci/vga_pci.c:425 · pass 2 · new |
rid-alias cache confusion for buggy children |
untested |
| DF-5492 |
Info |
0.5 |
Lifecycle gaps: no device_detach/shutdown methods, probe-time resource allocation, and an attach-error path that permanently leaks the ISA ioport and IRQ
Base system — bus / platform / cpu · sys/bus/iicbus/pcf/pcf.c:104 · pass 2 · new |
none |
untested |
| DF-5499 |
Info |
0.5 |
passioctl() runs CAMIOCOMMAND/CAMGETPASSTHRU with no periph reference across unkillable blocking sleeps (DF-5273/5327 hold/release class) — latent premature-free/UAF, currently blocked by devfs close semantics
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_pass.c:451 · pass 2 · new |
none |
not_reproduced |
| DF-5570 |
Info |
0.5 |
ECAM config path (pciereg_cfgread/cfgwrite) skips the bytes and alignment validation that the type-1 path enforces
Base system — bus / platform / cpu · sys/bus/pci/x86_64/pci_cfgreg.c:322 · pass 2 · new |
device-semantics confusion from a future buggy kernel caller; confined to ECAM mapping |
untested |
| DF-5572 |
Info |
0.5 |
Latent NULL-deref and wrong-CPU binding in #ifdef notyet cpu attachment block (cpu_add_child)
Base system — bus / platform / cpu · sys/bus/pci/x86_64/legacy.c:316 · pass 2 · new |
wrong per-CPU device binding + NULL-write panic path when the block is ever enabled |
untested |
| DF-5585 |
Info |
0.5 |
Quirk-apply fast path trusts stale pf / absent pf_last state: latent write-after-free and NULL-page write (unreachable with the shipped quirk table)
Kernel memory corruption (UAF / OOB write) · sys/bus/pccard/pccard_cis_quirks.c:248 · pass 2 · new |
none |
untested |
| DF-5602 |
Info |
0.5 |
ig4_pci.c:236's device_detach wiring runs child deletion with unchecked device_delete_child() returns (ig4_iic.c:690/694) before unconditional IRQ/MSI/MMIO teardown and a success return — DF-5544/5545/5562/5598 detach-veto family residue; every deterministic veto is caught one level up in this driver
Base system — bus / platform / cpu · sys/bus/smbus/ichiic/ig4_pci.c:207 · pass 2 · new |
none |
untested |
| DF-5612 |
Info |
0.5 |
cpu_enable_intr()/cpu_unmask_all_signals() implement 'restore' as an absolute sigsetmask(0) — not nesting-aware, silently discards any unrelated blocked-signal state
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/machintr.c:198 · pass 2 · new |
none |
untested |
| DF-5666 |
Info |
0.5 |
bios_sigsearch(9) performs no parameter sanity checks: paralen==0 infinite-loops the kernel, negative paralen/sigofs walk outside the BIOS scan window (kernel-API robustness; all in-tree callers pass constants)
Kernel memory leaks / uninitialized disclosure · sys/platform/pc64/x86_64/bios.c:67 · pass 2 · new |
none |
untested |
| DF-5731 |
Info |
0.5 |
db_get_value/db_put_value clamp only the upper bound of size; size<=0 would wrap to a huge size_t in db_read_bytes/db_write_bytes' unbounded byte loops
Base system — libkern / libprop / ddb · sys/ddb/db_access.c:79 · pass 2 · new |
none |
not_reproduced |
| DF-5778 |
Info |
0.5 |
Unguarded irq_service->funcs->to_dal_irq_source double-deref in dal_irq_service_to_irq_source plus swallowed dal_irq_service_construct failure — latent NULL-deref hardening gap (asymmetric with every sibling entry point)
Base system — drivers · sys/dev/drm/amd/display/dc/irq/irq_service.c:53 · pass 2 · new |
none |
untested |
| DF-5224 |
Info |
0.4 |
bios_get_image()'s sole bounds check uses wrap-prone 32-bit offset+size arithmetic — the security boundary for ~60 DC atombios parser dereferences, latent only (all current callers are uint16-bounded)
Base system — drivers · sys/dev/drm/amd/display/dc/bios/bios_parser_helper.c:40 · pass 2 · new |
none |
untested |
| DF-5438 |
Info |
0.4 |
ulpt_status_callback consumes uninitialized stack byte on short/zero-length status replies
Base system — bus / platform / cpu · sys/bus/u4b/serial/ulpt.c:295 · pass 2 · new |
none |
untested |
| DF-5439 |
Info |
0.4 |
ulpt_attach alt-setting scan counter is uint8_t and wraps with >=256 crafted alt settings
Base system — bus / platform / cpu · sys/bus/u4b/serial/ulpt.c:523 · pass 2 · new |
none |
untested |
| DF-4496 |
Low |
0.3 |
dce100_create_resource_pool leaks the kzalloc'd dce110_resource_pool shell on construct() failure (DF-4318/4353/4376 family twin)
Base system — drivers · sys/dev/drm/amd/display/dc/dce100/dce100_resource.c:1060 · pass 2 · new |
none |
untested |
| DF-3986 |
Info |
0.3 |
cayman_init dropped the CP1/CP2 compute-ring r600_ring_init (upstream and in-tree si.c/cik.c have them) — cayman_cp_resume evaluates order_base_2(0) = 1UL<<64 (C-standard UB) and programs CP_RB1/CP_RB2_CNTL/BASE with zero-size garbage on every Cayman/Aruba boot
Base system — drivers · sys/dev/drm/radeon/ni.c:2434 · pass 2 · new |
none |
inconclusive |
| DF-3976 |
Info |
0.2 |
Inverted error check at btc_dpm_init:2602-2607 — ATOM-computed PLL reference divider used only when radeon_atom_get_clock_dividers FAILS (default used on success)
Base system — drivers · sys/dev/drm/radeon/btc_dpm.c:2602 · pass 2 · new |
none |
inconclusive |
| DF-3943 |
Medium |
— |
resource_build_scaling_params INTERLACE error path leaves shared dc_stream_state.dst.height permanently doubled (compounding signed-int overflow)
Base system — drivers · sys/dev/drm/amd/display/dc/core/dc_resource.c:1119 · pass 2 · new |
none |
not_reproduced |
| DF-3984 |
Medium |
— |
ORB completion matching is address-only across ORB recycling (OCB_MATCH compares just orb_lo): a replayed management-ORB status completes the recycled *command* ORB — premature bus_dmamap_unload while device DMA is live and spurious CAM_REQ_CMP with device-chosen status
Base system — drivers · sys/dev/disk/sbp/sbp.c:156 · pass 2 · new |
none |
inconclusive |
| DF-5328 |
Medium |
— |
cam_periph_alloc() TOCTOU on the borrowed cam_periph_find() pointer: camperiphfree() can free the periph between find() and the unlocked NEW_DEV_FOUND/deferred_callback stores (UAF write), and camperiphfree() can see the flag set before deferred_callback is published (NULL/garbage indirect call)
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/cam_periph.c:176 · pass 2 · new |
none observed; code-level race: camperiphfree() can free the periph inside cam_periph_alloc()'s find()->flag-store window (UAF write into freed 144B M_CAMPERIPH chunk) and camperiphfree() can observe CAM_PERIPH_NEW_DEV_FOUND with an unpublished NULL deferred_callback -> indirect call through NULL/garbage |
not_reproduced |
| DF-3060 |
Low |
— |
dm_fdlist is walked and mutated without any common lock — nresolve's TAILQ scan races dirfs_node_setpassive's INSERT/REMOVE on SMP vkernels (UAF / list corruption)
Kernel memory corruption (UAF / OOB write) · sys/vfs/dirfs/dirfs_vnops.c:171 · pass 2 · new |
none |
not_reproduced |
| DF-4957 |
Low |
— |
ciphy_attach 'no media present' branch skips all media setup → miibus_mediainit calls ifmedia_set on an EMPTY list → panic("ifmedia_set") at attach on zero-capability PHYs
Base system — drivers · sys/dev/netif/mii_layer/ciphy.c:142 · pass 2 · new |
none |
not_reproduced |
| DF-5410 |
Low |
— |
vkernel sys_sigreturn calls npxpop without the critical section pc64 has — FPU ownership-state race (npxdna panic class)
Base system — bus / platform / cpu · sys/platform/vkernel64/x86_64/cpu_regs.c:447 · pass 2 · new |
none |
not_reproduced |
| DF-5084 |
Info |
— |
radeon_agp.c treats ai_aperture_size (BYTES tree-wide) as MB — dead <32MB guard and 2^20x-inflated gtt_size mis-programming GPU DMA-window registers; latent only (IS_ENABLED(CONFIG_AGP)==0, code not compiled on DragonFly)
Base system — drivers · sys/dev/drm/radeon/radeon_agp.c:152 · pass 2 · new |
none |
not_reproduced |
| DF-5199 |
Info |
— |
virtual_str_enc_funcs omits 7 ops (NULL) — latent NULL function-pointer call if virtual encoders are ever instantiated and cross-assigned to a real DP stream by the allocator fallback
Base system — drivers · sys/dev/drm/amd/display/dc/virtual/virtual_stream_encoder.c:77 · pass 2 · new |
latent NULL function-pointer call (DoS) |
untested |
| DF-5207 |
Info |
— |
virtual_lnk_enc_funcs omits 6 of 18 link_encoder_funcs ops (NULL) with 4 unguarded call sites in DC core — latent NULL function-pointer call if virtual link encoders are ever instantiated
Base system — drivers · sys/dev/drm/amd/display/dc/virtual/virtual_link_encoder.c:86 · pass 2 · new |
latent NULL function-pointer call (DoS) |
untested |
| DF-2597 |
High |
9.8 |
TCP-MD5 signature option writes past 40-byte opt[] stack buffer with no bounds check
Kernel memory corruption (UAF / OOB write) · sys/netinet/tcp_output.c:186 |
none |
not_reproduced |
| DF-0285 |
High |
9.6 |
Remote heap overflow via unchecked MESHID IE length in ieee80211_parse_meshid
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_mesh.c:3456 |
Real CWE-787 heap OOB write in ieee80211_parse_meshid (up to 223 bytes) reaching ni_mltimer.toc / ni_mlhtimer.toc pointers => indirect function-pointer-control RCE surface on wifi-equipped MBSS hosts; NOT runtime-triggerable on this KVM guest (no wifi hardware). |
inconclusive |
| DF-0326 |
High |
9.6 |
Remote heap overflow: unbounded SSID IE length copied into fixed 32-byte ni_essid
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_node.c:815 |
remote-unauth heap overflow corrupting ni_chan kernel pointer (panic/DoS minimum, code-exec plausible) on any host with a wlan radio |
inconclusive |
| DF-0428 |
High |
9.6 |
pfsync_input has no source/peer authentication: any on-link host can inject/modify/destroy pf state
Base system — network stack · sys/net/pf/if_pfsync.c:462 |
none |
not_reproduced |
| DF-0430 |
High |
9.3 |
Unauthenticated PFSYNC_ACT_CLR/DEL/DEL_C let attacker mass-destroy arbitrary pf state across all CPUs
Base system — network stack · sys/net/pf/if_pfsync.c:542 |
none |
not_reproduced |
| DF-1125 |
High |
9.3 |
Unvalidated FCP response/sense lengths cause OOB read from kernel stack (24XX FC)
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/isp/isp.c:5225 |
leak |
inconclusive |
| DF-0704 |
Medium |
9.1 |
No length validation on received sync datagram — type field and struct casts read without bounds check
Base system — network stack · sys/net/ipfw3_basic/ip_fw3_sync.c:208 |
none |
not_reproduced |
| DF-1215 |
Critical |
8.8 |
METEORSVIDEO ioctl allows unprivileged user to set arbitrary physical DMA target address
Kernel memory corruption (UAF / OOB write) · sys/dev/video/bktr/bktr_core.c:1396 |
none |
inconclusive fixed |
| DF-1020 |
High |
8.8 |
OOB heap write in fw_bus_explore_callback via wire-controlled ongoaddr jump -> arbitrary 4-byte kernel heap write
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/firewire.c:1500 |
none |
inconclusive |
| DF-0269 |
Medium |
8.8 |
Stack buffer overflow in sppp_print_bytes: VLA sized len but hexncpy writes 3*len bytes
Kernel memory corruption (UAF / OOB write) · sys/net/sppp/if_spppsubr.c:5290 |
panic |
not_reproduced |
| DF-0289 |
Medium |
8.8 |
UAF/TOCTOU: mesh route pointers returned unreferenced, forward_to_gates drops lock mid-traversal
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_mesh.c:230 |
dos |
inconclusive |
| DF-1264 |
High |
8.4 |
drm_dp_sideband_msg_build OOB write past chunk[48] via unchecked msg_len from malicious MST hub
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/drm_dp_mst_topology.c:349 |
none |
inconclusive |
| DF-1314 |
High |
8.4 |
Off-by-one in btc_copy_vbios_mc_reg_table + post-write bound check in btc_set_mc_special_registers allows VBIOS OOB write
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/btc_dpm.c:1992 |
none |
inconclusive |
| DF-1319 |
High |
8.4 |
Unvalidated DisplayPort lane_count from malicious sink causes stack buffer overflow in link-training helpers
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/core/dc_link_dp.c:1784 |
none |
inconclusive |
| DF-2348 |
High |
8.4 |
Dead unsigned comparison in mmap bounds check enables vtophys on wrapped ZERO_LENGTH_PTR - kernel panic / kernel memory exposure
Base system — bus / platform / cpu · sys/bus/u4b/uvc/uvc_buf.c:107 |
none |
not_reproduced |
| DF-2349 |
High |
8.4 |
uint8_t loop index in uvc_ctrl_init_dev wraps causing infinite loop and heap OOB write on malicious UVC descriptor
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/uvc/uvc_ctrls.c:925 |
none |
not_reproduced |
| DF-2354 |
High |
8.4 |
txd_fw buffer overflow on firmware load: 6 KB heap overflow + kernel-memory leak to USB device on every attach
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_mtw.c:128 |
none |
not_reproduced |
| DF-2511 |
High |
8.4 |
Unvalidated controller-supplied subq_id and cmd_id in nvme_poll_completions enable kernel memory corruption
Privilege escalation (general use) · sys/dev/disk/nvme/nvme.c:713 |
none |
not_reproduced |
| DF-1526 |
Medium |
8.3 |
Missing RX length bounds check + RX buffer-length misprogramming allows heap OOB read (info leak) and DMA heap overflow write
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/sf/if_sf.c:883 |
none |
inconclusive |
| DF-0291 |
High |
8.2 |
Heap buffer overflow in setmlme_assoc_adhoc: unbounded im_ssid_len into 32-byte buffers
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_ioctl.c:1568 |
none observed; real shipped heap+stack overflow (CWE-787) unreachable on this guest (no wlan vap possible without a wifi radio) |
inconclusive |
| DF-0401 |
High |
8.2 |
Heap buffer overflow via unchecked slot->len in VALE bridge forwarding: pkt_copy up to 65536 bytes into 2048-byte buffer
Kernel memory corruption (UAF / OOB write) · sys/net/netmap/netmap_vale.c:988 |
none |
not_reproduced |
| DF-0429 |
High |
8.2 |
Unauthenticated PFSYNC_ACT_UREQ forces victim to multicast entire pf state table: info disclosure + amplification DoS
Base system — network stack · sys/net/pf/if_pfsync.c:900 |
none |
not_reproduced |
| DF-0775 |
High |
8.2 |
RPC reply verifier-length integer overflow corrupts XDR cursor — wild pointer deref client kernel panic or OOB read
Kernel memory leaks / uninitialized disclosure · sys/vfs/nfs/nfs_socket.c:1498 |
none |
not_reproduced |
| DF-0640 |
High |
8.1 |
ieee80211_crypto_decap minimum-length check ignores hdrlen, causing OOB-read panic or mbuf double-free on crafted encrypted frames
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_crypto.c:587 |
none |
inconclusive |
| DF-1123 |
High |
8.1 |
iwn_rx_done: missing upper bound on firmware-reported frame length allows OOB read of RX mbuf / kernel heap
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwn/if_iwn.c:3069 |
none |
not_reproduced |
| DF-1131 |
High |
8.1 |
Heap OOB read in bwn_dma_rxeof via dr_rx_bufsize / descriptor bytecnt mismatch
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/bwn/bwn/if_bwn.c:5486 |
leak |
inconclusive |
| DF-1849 |
Medium |
8.1 |
Missing upper-bound on attacker-controlled stream.len in fwe_as_input causes kernel heap over-read
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/fwe/if_fwe.c:560 |
none |
inconclusive |
| DF-1213 |
High |
8.0 |
Unchecked device-supplied rxcd->rxd_idx yields OOB array access, OOB write, arbitrary-pointer write, and infinite loop in RX completion path
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/vmware/vmxnet3/if_vmx.c:2056 |
none |
not_reproduced |
| DF-0320 |
Medium |
8.0 |
Reorder buffer rxa_m[] mutated without dedicated lock: RX races timer flush and ADDBA re-init (double-free/UAF)
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_ht.c:780 |
none |
inconclusive |
| DF-0509 |
High |
7.8 |
Stack buffer overflow in ng_ksocket_sockaddr_unparse via negative pathlen (PF_LOCAL sun_len underflow)
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ksocket/ng_ksocket.c:323 |
none |
not_reproduced |
| DF-0641 |
High |
7.8 |
Heap buffer overflow in smb_smb_ssnsetup: ntencpass sized from charset-converted password but filled from original
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/netproto/smb/smb_smb.c:255 |
none |
not_reproduced |
| DF-1057 |
High |
7.8 |
FW_ASYREQ bcopy writes payload into &xfer->send.payload (field address) instead of xfer->send.payload (buffer) — kernel heap overflow + arbitrary kfree
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwdev.c:558 |
none |
inconclusive |
| DF-1058 |
High |
7.8 |
fw_write trusts user-supplied stream packet header len field for the second uiomove, overflowing the per-packet DMA buffer slot
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwdev.c:388 |
none |
inconclusive |
| DF-1065 |
High |
7.8 |
Use-after-free: bq->mem kfree()d while userspace mmap mappings keep wired fictitious pages pointing at freed memory
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/uvc/uvc_buf.c:517 |
none |
not_reproduced |
| DF-1081 |
High |
7.8 |
Stack buffer overflow in cue_setmulti: 8-byte hashtbl for 64-byte (512-bit) multicast hash table
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/net/if_cue.c:273 |
none |
inconclusive |
| DF-1141 |
High |
7.8 |
Heap OOB write in ci_setup_default_dpm_tables: VBIOS counts index fixed-size dpm_levels[8] unchecked
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/ci_dpm.c:3593 |
none |
inconclusive |
| DF-1142 |
High |
7.8 |
Heap OOB write in ci_populate_smc_uvd/vce/acp/samu_level: VBIOS counts index fixed-size SMU7 level arrays unchecked
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/ci_dpm.c:2791 |
none |
inconclusive |
| DF-1146 |
High |
7.8 |
i915_gem_fault missing offset bounds check and partial-view GMADR misaddressing allow cross-object GPU memory read/write
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/i915_gem.c:2229 |
none |
inconclusive |
| DF-1149 |
High |
7.8 |
Heap OOB write in SMC UVD/VCE/ACP/SAMU level population from unbounded VBIOS count
Kernel memory corruption (UAF / OOB write) · DF-1142 · sys/dev/drm/radeon/ci_dpm.c:2689 |
none |
inconclusive |
| DF-1150 |
High |
7.8 |
Heap OOB write in ci_setup_default_dpm_tables from unbounded VBIOS dependency-table counts
Kernel memory corruption (UAF / OOB write) · DF-1141 · sys/dev/drm/radeon/ci_dpm.c:3514 |
none |
inconclusive |
| DF-1166 |
High |
7.8 |
Heap overflow: unvalidated VBIOS dependency-table count overflows fixed dpm_levels[8] array
Kernel memory corruption (UAF / OOB write) · DF-1141 · sys/dev/drm/amd/powerplay/hwmgr/smu7_hwmgr.c:784 |
none |
inconclusive |
| DF-1258 |
High |
7.8 |
Heap buffer overflow / OOB read in MFI_CMD and MFI_LINUX_CMD_2 ioctl SGL copyin/copyout
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mfi/mfi.c:2991 |
none |
inconclusive |
| DF-1271 |
High |
7.8 |
VBIOS-controlled mm_dep_table->count drives heap overflow in UVD/VCE/ACP SMC level population
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/tonga_smumgr.c:1320 |
none |
inconclusive |
| DF-1272 |
High |
7.8 |
VBIOS-controlled vdd_dep_on_sclk->count drives Sclk_voltageOffset[8] heap overflow in clock stretcher
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/tonga_smumgr.c:1584 |
none |
inconclusive |
| DF-1311 |
High |
7.8 |
Stack buffer overflow in hptiop_post_req_itl via unbounded SG list (HIGH_MEM_ACESS path)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptiop/hptiop.c:2417 |
none |
inconclusive |
| DF-1337 |
High |
7.8 |
Use-after-free write in chn_write (and UAF read in chn_read) via concurrent chn_resizebuf during uiomove unlock window
Kernel memory corruption (UAF / OOB write) · sys/dev/sound/pcm/channel.c:502 |
none |
inconclusive |
| DF-1340 |
High |
7.8 |
u32 integer overflow in CB/DB size validation bypasses BO bounds check (cross-process GPU memory corruption)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_cs.c:430 |
none |
inconclusive |
| DF-1342 |
High |
7.8 |
r600_dma_cs_parse reads/writes IB past length_dw (heap OOB via DMA packet at IB end)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_cs.c:2388 |
none |
inconclusive |
| DF-1574 |
High |
7.8 |
Heap OOB write in get_vddc_lookup_table: buffer sized for 16 records, loop walks up to 255
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/process_pptables_v1_0.c:168 |
none |
not_reproduced |
| DF-1586 |
High |
7.8 |
Heap buffer overflow in get_vddc_lookup_table: VBIOS ucNumEntries > caller max_levels writes ~2.5KB past slab object
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/vega10_processpptables.c:1026 |
none |
not_reproduced |
| DF-2002 |
High |
7.8 |
Forged EEPROM pwr_table_offset drives OOB write/read in ar9280AdjustPDADCValues (up to 64KB overflow)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_hal/ar9002/ar9280_olc.c:188 |
none |
inconclusive |
| DF-2028 |
High |
7.8 |
Use-after-free in GDT_IOCTL_GENERAL: framework-owned ioctl buffer queued into persistent controller command list
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/iir/iir_ctrl.c:245 |
none |
inconclusive |
| DF-2079 |
High |
7.8 |
Heap buffer overflow and NULL-deref panic in SPECTRAL_CONTROL_GET_PARAMS due to missing output-buffer validation
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath/if_ath_spectral.c:218 |
none |
not_reproduced |
| DF-2145 |
High |
7.8 |
tasklet_kill() UAF: premature return on TASKLET_STATE_RUN leaves dangling tasklet_entry; runner later calls t->func from freed memory
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/linux_tasklet.c:170 |
none |
not_reproduced |
| DF-2174 |
High |
7.8 |
__hrtimer_task epilogue unconditionally clears timer->active, defeating callback re-arm via hrtimer_start + NORESTART — un-cancellable timer leading to UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/linux_hrtimer.c:59 |
none |
not_reproduced |
| DF-2268 |
High |
7.8 |
Heap overflow in cdcheckmedia via uncapped num_entries from device-supplied TOC header
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_cd.c:2947 |
none |
not_reproduced |
| DF-2282 |
High |
7.8 |
sgopen: missing cam_periph_acquire plus cam_periph_unlock-without-lock in securelevel path (lock/refcount protocol violation; refcount underflow->UAF)
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_sg.c:385 |
none |
not_reproduced |
| DF-2332 |
High |
7.8 |
Stack buffer overflow in ACPI SMBus handler: buf[32] vs attacker-controlled GSBUS lengths
Kernel memory corruption (UAF / OOB write) · sys/bus/smbus/smbacpi/smbacpi.c:137 |
none |
not_reproduced |
| DF-2339 |
High |
7.8 |
Stack buffer overflow in cue_setmulti: 8-byte hashtbl indexed by 9-bit multicast hash
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/net/if_cue.c:273 |
none |
not_reproduced |
| DF-2346 |
High |
7.8 |
SIGIO recipient stored as raw unreferenced struct proc pointer -> use-after-free after fork+exit
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_dev.c:1044 |
none |
not_reproduced |
| DF-2389 |
High |
7.8 |
Heap out-of-bounds write in link_add_prs when _PRS has more IRQ resources than _CRS
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_pci_link.c:311 |
none |
not_reproduced |
| DF-2424 |
High |
7.8 |
Scatter/gather list heap overflow via unchecked sglist_cnt in amdexecutesrb
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/amd/amd.c:455 |
none |
not_reproduced |
| DF-2428 |
High |
7.8 |
Unbounded scatter/gather copy in btexecuteccb via CAM_SCATTER_VALID CCB (kernel heap overflow)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/buslogic/bt.c:1055 |
none |
not_reproduced |
| DF-2488 |
High |
7.8 |
Stack buffer overflow/underflow in ad_describe via unterminated model string
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-disk.c:508 |
none |
not_reproduced |
| DF-2493 |
High |
7.8 |
Heap OOB write via untrusted disk index in ATA RAID metadata parsers
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-raid.c:1806 |
none |
not_reproduced |
| DF-2506 |
High |
7.8 |
TOCTOU kernel heap overflow in acd_send_cue via double-fetch of cuesheet->len
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-cd.c:1318 |
none |
not_reproduced |
| DF-2510 |
High |
7.8 |
Unbounded ctlr->channels (up to 32) overflows the fixed interrupt[8] dispatch array causing heap OOB write of function pointers and OOB indirect call on every IRQ
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/chipsets/ata-ahci.c:101 |
none |
not_reproduced |
| DF-2517 |
High |
7.8 |
Heap OOB write in sili_load_prb_callback via unchecked DMA segment count
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/sili/sili.c:1245 |
none |
not_reproduced |
| DF-2520 |
High |
7.8 |
trm_ExecuteSRB writes user-controlled nseg SG entries into 32-slot pSRBSGL with no bounds check (heap overflow via pass(4) SCATTER_VALID)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:433 |
none |
not_reproduced |
| DF-2521 |
High |
7.8 |
trm_action bcopy of user-controlled cdb_len into 12-byte CmdBlock overflows the SRB struct (pass(4)-reachable heap corruption)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:591 |
none |
not_reproduced |
| DF-0547 |
Medium |
7.8 |
Ring cleanup trusts userspace-writable buf_idx: cross-adapter double-free / buffer aliasing
Kernel memory corruption (UAF / OOB write) · sys/net/netmap/netmap_mem2.c:995 |
corruption |
not_reproduced |
| DF-1235 |
Medium |
7.8 |
Unbounded CDB copy into fixed 12-byte SRB CmdBlock overwrites SRB pointer fields
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:591 |
none |
not_reproduced |
| DF-1454 |
Medium |
7.8 |
Heap buffer overflow in autosense data copy - missing sizeof(sense_data) bound
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/vmware/pvscsi/pvscsi.c:945 |
none |
inconclusive |
| DF-1848 |
Medium |
7.8 |
ENA LLQ bounce buffer heap overflow via descs_left_in_line u16 underflow when device reports descs_num_before_header=0
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/amazon/ena/ena-com/ena_eth_com.c:150 |
none |
inconclusive |
| DF-2469 |
Medium |
7.8 |
Heap OOB write via double channel offset in RQCS_SYNCXFER_FAILED status handler
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/isp/isp.c:6282 |
none |
not_reproduced |
| DF-2492 |
Medium |
7.8 |
OOB heap write/read on interrupt[8] via AHCI controllers with >8 ports
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-pci.c:383 |
none |
not_reproduced |
| DF-1122 |
High |
7.6 |
iwn_rx_compressed_ba: unbounded firmware-supplied qid indexes sc->txq[] and sc->qid2tap[] (OOB array access / NULL deref)
Base system — drivers · sys/dev/netif/iwn/if_iwn.c:3251 |
none |
not_reproduced |
| DF-2490 |
High |
7.6 |
ATAPI PIO heap OOB write/read via unvalidated device byte_count_limit
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-lowlevel.c:329 |
none |
not_reproduced |
| DF-0349 |
Medium |
7.6 |
Heap OOB read in PREQ processing when target count is 0: (ndest-1) wraps to SIZE_MAX
Kernel memory leaks / uninitialized disclosure · sys/netproto/802_11/wlan/ieee80211_hwmp.c:311 |
none |
inconclusive |
| DF-0220 |
High |
7.5 |
Predictable RNG: /dev/urandom+getrandom+kern.random return deterministic ChaCha20 keystream (zero key) before first reseed
Base system — kernel core (kern) · sys/kern/subr_csprng.c:84 |
none |
not_reproduced |
| DF-0580 |
High |
7.5 |
ieee80211_defrag UAF/dangling-pointer: DragonFly m_cat frees fragment but code reads wh + m_pkthdr.len after
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_input.c:248 |
none |
inconclusive |
| DF-0703 |
High |
7.5 |
Unauthenticated remote mbuf leak in edge socket handler — kernel memory exhaustion DoS
Base system — network stack · sys/net/ipfw3_basic/ip_fw3_sync.c:203 |
none |
not_reproduced |
| DF-2358 |
High |
7.5 |
Heap buffer overflow in USB RX frame aggregation: m_getcl (2048B) too small for m_copydata of up to 4080B
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_run.c:3013 |
none |
not_reproduced |
| DF-0533 |
Medium |
7.5 |
Unsynchronized sc->inq/outq between hard ISR and netgraph forward: ifqueue corruption / UAF
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/bluetooth/drivers/bt3c/ng_bt3c_pccard.c:939 |
none |
not_reproduced |
| DF-0581 |
Medium |
7.5 |
ieee80211_setup_rates trusts attacker-controlled IE length byte: latent heap overflow of rs_rates[15]
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_input.c:420 |
none |
not_reproduced |
| DF-1060 |
Medium |
7.5 |
FW_GTPMAP bcopy uses (topology_map->crc_len + 1) * 4 as the count — attacker-inflatable crc_len drives unbounded kernel OOB read up to 256 KB
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwdev.c:657 |
none |
not_reproduced |
| DF-1249 |
Medium |
7.5 |
dc_pnic_rx_bug_war unbounded copy loop - heap overflow if PNIC produces >5 descriptors
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/dc/if_dc.c:2414 |
none |
inconclusive |
| DF-1828 |
Medium |
7.5 |
isc_soc: NULL deref in so_input pdu_alloc retry uses M_NOWAIT (comment says WAITOK) derefs unchecked
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/isc_soc.c:540 |
none |
inconclusive |
| DF-1236 |
High |
7.4 |
Extended SCSI message accumulation overflows MsgInBuf[6] into adjacent SRB pointer fields
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:1832 |
none |
not_reproduced |
| DF-1292 |
Medium |
7.4 |
Multi-segment RX reads overwritten descriptor length inflating m_len past cluster for heap OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/age/if_age.c:2025 |
none |
not_reproduced |
| DF-1385 |
Medium |
7.4 |
Unbounded VBIOS num_entries overflows MC register table arrays (heap OOB write+read)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/cypress_dpm.c:1052 |
none |
inconclusive |
| DF-1950 |
Medium |
7.4 |
Unvalidated ACPI ATRM return object: heap buffer overflow and type confusion in amdgpu_atrm_call
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_bios.c:264 |
none |
inconclusive |
| DF-2418 |
Medium |
7.4 |
Missing bounds check on device-controlled NCQ error tag causes OOB heap read/write in ahci_port_read_ncq_error
Privilege escalation (general use) · sys/dev/disk/ahci/ahci.c:3488 |
none |
not_reproduced |
| DF-2484 |
Medium |
7.4 |
Untrusted IOC-supplied indices index raid_disks/raid_volumes arrays without bounds checks (OOB read/write of kernel heap)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_raid.c:414 |
none |
not_reproduced |
| DF-1211 |
High |
7.3 |
OOB read+write in NCQ error recovery: device-controlled log-page tag indexes ap_ccbs without bounds check
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/ahci/ahci.c:3488 |
none |
not_reproduced |
| DF-2322 |
High |
7.3 |
MSI-X table/PBA out-of-bounds MMIO read/write - no BAR-size validation against attacker-controlled table offset and vector count
Kernel memory corruption (UAF / OOB write) · sys/bus/pci/pci.c:1431 |
none |
not_reproduced |
| DF-2356 |
High |
7.3 |
Unvalidated hardware frame length in RX path allows heap OOB read / uninitialized-memory info leak
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_rum.c:1146 |
none |
not_reproduced |
| DF-2401 |
High |
7.3 |
intel_gtt_insert_page swaps index/physical arguments causing out-of-bounds MMIO write into Intel IGD BAR
Kernel memory corruption (UAF / OOB write) · sys/dev/agp/intel-gtt.c:1407 |
none |
not_reproduced |
| DF-0275 |
Medium |
7.2 |
Heap buffer overflow in WPA/RSN IE construction: variable-length IE written into fixed sizeof(ieee80211_ie_wpa)=100 slot
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_output.c:1976 |
none |
inconclusive |
| DF-0243 |
High |
7.1 |
size_t underflow in exec_shell_imgact when argv[0] longer than interpreter+fname -> kernel panic
Base system — kernel core (kern) · sys/kern/imgact_shell.c:117 |
none |
not_reproduced |
| DF-0700 |
High |
7.1 |
User-controlled edge count not bounded by MAX_EDGES — BSS OOB array write/read on centre_socks[]
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_basic/ip_fw3_sync.c:143 |
none |
not_reproduced |
| DF-0701 |
High |
7.1 |
Heap OOB read in centre_conf bcopy and integer overflow in size math
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_basic/ip_fw3_sync.c:143 |
none |
not_reproduced |
| DF-0702 |
High |
7.1 |
Mbuf UAF and double-free in centre_test and send_state (sosend consumes m, loop reuses it)
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_basic/ip_fw3_sync.c:339 |
none |
not_reproduced |
| DF-0766 |
High |
7.1 |
Heap OOB write in NFS READDIR/READDIRPLUS reply XDR null-padding — missing nfsm_clget before pad bytes
Kernel memory corruption (UAF / OOB write) · sys/vfs/nfs/nfs_serv.c:3186 |
none |
not_reproduced fixed |
| DF-1227 |
High |
7.1 |
mpr_build_nvme_prp heap overflow via unbounded user-supplied data length
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr.c:2716 |
none |
not_reproduced |
| DF-1257 |
High |
7.1 |
Integer overflow in offset+size bounds check in amdgpu_vm_bo_map enables OOB read in amdgpu_vm_bo_split_mapping
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_vm.c:2509 |
none |
not_reproduced |
| DF-1259 |
High |
7.1 |
mfi_stp_cmd unbounded SGE loop corrupts mfi_softc via size-[2] DMA arrays and STP frame SGL
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mfi/mfi.c:2724 |
none |
inconclusive |
| DF-1333 |
High |
7.1 |
OOB kernel-heap read in rv7xx_parse_power_table via unchecked VBIOS offsets and indices
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/rv770_dpm.c:2284 |
none |
inconclusive |
| DF-2133 |
High |
7.1 |
dma_buf_export never initializes dmabuf->resv; GPU drivers dereference uninitialized slab pointer
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/linux_dma-buf.c:127 |
none |
not_reproduced |
| DF-2175 |
High |
7.1 |
hrtimer_start_range_ns corrupts per-CPU systimer queue by calling systimer_init_oneshot without prior systimer_del on an armed timer
Base system — drivers · sys/dev/drm/linux_hrtimer.c:115 |
none |
not_reproduced |
| DF-2221 |
High |
7.1 |
ar5312EepromRead unbounded OOB read of radio-config buffer reachable from unprivileged SIOCGATHDIAG/HAL_DIAG_EEREAD
Privilege escalation (general use) · sys/dev/netif/ath/ath_hal/ar5312/ar5312_eeprom.c:37 |
none |
not_reproduced |
| DF-2419 |
High |
7.1 |
Heap OOB write in INQUIRY handler when data buffer length is 36..95 bytes
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/ahci/ahci_cam.c:1109 |
none |
not_reproduced |
| DF-0286 |
Medium |
7.1 |
Missing length validation in mesh action frame handler: OOB read of stale data
Kernel memory leaks / uninitialized disclosure · sys/netproto/802_11/wlan/ieee80211_mesh.c:2548 |
none |
inconclusive |
| DF-0352 |
Medium |
7.1 |
RANN frame overwrites shared global ieee80211_hwmp_rannint without lock: remote timer DoS
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_hwmp.c:1971 |
none |
not_reproduced |
| DF-0706 |
Medium |
7.1 |
Lifecycle bugs in edge handler cleanup: UAF on replaced edge_sock, missing thread join, soclose+sofree imbalance
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_basic/ip_fw3_sync.c:195 |
none |
not_reproduced |
| DF-0998 |
Medium |
7.1 |
TX DMA buffer overflow: mbuf length clamp uses wrong bound -> 16-byte heap overwrite past USB DMA buffer
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_rum.c:438 |
none |
not_reproduced |
| DF-1181 |
Medium |
7.1 |
Heap OOB read via unbounded vddInd/vddciInd/mvddInd into 4-8 entry voltage lookup tables
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:635 |
none |
not_reproduced |
| DF-1197 |
Medium |
7.1 |
2-byte DMA heap overflow in RX buffer when ETHER_ALIGN applied with jumbo MTU
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/emx/if_emx.c:2918 |
none |
not_reproduced |
| DF-1237 |
Medium |
7.1 |
Scatter/gather list copy has no bounds vs the 32-entry hardware SG buffer
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/trm/trm.c:434 |
none |
not_reproduced |
| DF-1476 |
Medium |
7.1 |
intel_gtt_insert_page swaps (index, physical) arguments to install_gtt_pte causing OOB MMIO write past BAR
Kernel memory corruption (UAF / OOB write) · sys/dev/agp/intel-gtt.c:1406 |
none |
inconclusive |
| DF-1545 |
Medium |
7.1 |
amdgpu atom WS operand index unbounded + NULL deref when ws=0
Base system — drivers · sys/dev/drm/amd/amdgpu/atom.c:222 |
none |
inconclusive |
| DF-1551 |
Medium |
7.1 |
snread subtracts 6 from u_short packet_length with no bound: underflow -> ~63KB kernel heap overflow write via insw into 2KB mbuf cluster
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/sn/if_sn.c:983 |
none |
inconclusive |
| DF-1755 |
Medium |
7.1 |
radeon_vce: integer overflow in create-message size calculation bypasses reloc buffer checks
Base system — drivers · sys/dev/drm/radeon/radeon_vce.c:602 |
none |
inconclusive |
| DF-0315 |
High |
7.0 |
UAF: wg_peer_destroy frees peer struct while data-plane paths access it without peer refcount
Kernel memory corruption (UAF / OOB write) · sys/net/wg/if_wg.c:589 |
none observed dynamically (code-level UAF real + reachable; race window too narrow to panic on GENERIC) |
not_reproduced |
| DF-0590 |
High |
7.0 |
No serialization of bridge state in legacy netgraph/ng_bridge -- UAF in rehash, OOB heap write in GET_TABLE, deterministic KASSERT panics under concurrent traffic
Kernel memory corruption (UAF / OOB write) · sys/netgraph/bridge/ng_bridge.c:297 |
none |
not_reproduced |
| DF-0636 |
High |
7.0 |
ngt_rcvdata accesses sc->outq without tp->t_token, racing with ngt_start (UAF write / queue corruption)
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/tty/ng_tty.c:591 |
none |
not_reproduced |
| DF-0691 |
High |
7.0 |
UAF / NULL-deref in mld6_input MLD_LISTENER_QUERY path: in6m freed concurrently with iteration
Kernel memory corruption (UAF / OOB write) · sys/netinet6/mld6.c:280 |
none |
not_reproduced |
| DF-0692 |
High |
7.0 |
UAF and unsynchronized list walk in mld6_fasttimeo over global in6_multihead
Kernel memory corruption (UAF / OOB write) · sys/netinet6/mld6.c:382 |
panic |
not_reproduced |
| DF-1136 |
High |
7.0 |
Heap buffer overflow in si_get_svi2_voltage_table: unbounded VBIOS count writes past entries[32]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/si_dpm.c:3939 |
panic |
inconclusive |
| DF-1246 |
High |
7.0 |
OOB read/write past IB end in evergreen_dma_cs_parse (no per-packet bounds check)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/evergreen_cs.c:2806 |
none |
not_reproduced |
| DF-1298 |
High |
7.0 |
Heap overflow in update_slot_layout_info via unchecked ucConnNum
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/bios/bios_parser.c:2634 |
none |
inconclusive |
| DF-1305 |
High |
7.0 |
Heap OOB write/read in kv_init_graphics_levels via unbounded VBIOS table->count
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/kv_dpm.c:2355 |
none |
inconclusive |
| DF-2146 |
High |
7.0 |
Unbounded AUX reply byte count causes heap/stack OOB write on DP read path
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_dp_auxch.c:177 |
none |
not_reproduced |
| DF-2292 |
High |
7.0 |
FW_ASYREQ response copy overflows ioctl heap buffer with oversized FireWire RRESB payload
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwdev.c:576 |
none |
not_reproduced |
| DF-0420 |
Medium |
7.0 |
Use-after-free in SCO reassembly buffer: sc_isoc_in_buffer retains dangling pointer after mbuf forwarded and realloc fails
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/bluetooth/drivers/ubt/ng_ubt.c:1103 |
none |
inconclusive |
| DF-0497 |
Medium |
7.0 |
TOCTOU use-after-free on rtentry in ng_btsocket_l2cap_raw_bind: releases rt_lock before storing pointer
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/bluetooth/socket/ng_btsocket_l2cap_raw.c:690 |
none |
inconclusive |
| DF-0513 |
Medium |
7.0 |
Rule-chain mutation raced with concurrent ip6_fw_chk: lock-free linked list + kfree after crit_exit -> UAF on SMP
Kernel memory corruption (UAF / OOB write) · sys/net/ip6fw/ip6_fw.c:856 |
none |
not_reproduced |
| DF-0521 |
Medium |
7.0 |
PF_LOCAL sockaddr unparse: sun_len<pathoff underflow -> giant bcopy -> stack OOB write (DF-0509 v1 twin)
Kernel memory corruption (UAF / OOB write) · sys/netgraph/ksocket/ng_ksocket.c:321 |
none |
not_reproduced |
| DF-0586 |
Medium |
7.0 |
Lockless global hci_pcb list allows use-after-free during concurrent socket teardown and packet tap
Kernel memory corruption (UAF / OOB write) · sys/netbt/hci_socket.c:87 |
none |
not_reproduced |
| DF-0589 |
Medium |
7.0 |
sc->outq mbuf-queue race in ng_h4: IF_DEQUEUE in ng_h4_start (tty ctx) vs IF_DRAIN in disconnect/shutdown (netgraph ctx); NG_H4_LOCK is only per-CPU crit_enter
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/bluetooth/drivers/h4/ng_h4.c:88 |
dos |
inconclusive |
| DF-0793 |
Medium |
7.0 |
Use-after-free in async TRIM block-free path — ffs_blkfree stores unrefcounted mount/device pointers in deferred task
Kernel memory corruption (UAF / OOB write) · sys/vfs/ufs/ffs_alloc.c:1633 |
none |
not_reproduced fixed |
| DF-1061 |
Medium |
7.0 |
LOOPRECOVER watchdog silently abandons unacknowledged TLB invalidations, enabling stale-TLB use-after-free / info leak
Kernel memory corruption (UAF / OOB write) · sys/platform/pc64/x86_64/pmap_inval.c:342 |
none |
not_reproduced |
| DF-1153 |
Medium |
7.0 |
Use-after-free in reset functions when kzalloc fails (connector/crtc/plane state)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm.c:3190 |
none |
not_reproduced |
| DF-1289 |
Medium |
7.0 |
Heap overflow of CAM CCB sense buffer from untrusted controller sense_length
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mly/mly.c:2356 |
none |
not_reproduced |
| DF-1290 |
Medium |
7.0 |
OOB write on softc via untrusted controller event channel/target/lun in mly_process_event
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mly/mly.c:1323 |
none |
not_reproduced |
| DF-1295 |
Medium |
7.0 |
METEORSSIGNAL stores curproc without reference causing UAF via fork+child-exit+DMA interrupt
Kernel memory corruption (UAF / OOB write) · sys/dev/video/cxm/cxm.c:2760 |
none |
not_reproduced |
| DF-1366 |
Medium |
7.0 |
Heap OOB write in vegam_populate_smc_vce_level via unbounded VBIOS mm_dep_table->count
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:1217 |
none |
inconclusive |
| DF-1367 |
Medium |
7.0 |
Heap OOB write in vegam_populate_smc_uvd_level via unbounded VBIOS mm_dep_table->count
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:1330 |
none |
inconclusive |
| DF-1368 |
Medium |
7.0 |
Heap+stack OOB writes in clock-stretcher/AVFS/CAC table population via unbounded VBIOS counts
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/vegam_smumgr.c:515 |
none |
inconclusive |
| DF-1411 |
Medium |
7.0 |
DragonFly sysctl backlight handler captures drm_connector_state before modeset lock enabling UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_panel.c:1815 |
none |
inconclusive |
| DF-1417 |
Medium |
7.0 |
Stack/heap OOB write in smu8_get_clock_by_type from unbounded SMU-derived count
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/hwmgr/smu8_hwmgr.c:1613 |
none |
inconclusive |
| DF-1449 |
Medium |
7.0 |
Unvalidated firmware-reported disk/volume indices cause OOB read/write in RAID arrays and IOC page buffers
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_raid.c:414 |
none |
inconclusive |
| DF-1546 |
Medium |
7.0 |
Unclamped csio->sense_len handed to firmware allows DMA overflow of the per-queue sense slot
Base system — drivers · sys/dev/disk/advansys/advansys.c:559 |
none |
inconclusive |
| DF-1758 |
Medium |
7.0 |
ips: card-controlled command_id indexes commandarray[] with no bounds check in both interrupt handlers -> OOB r/w + indirect call
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ips/ips.c:567 |
none |
inconclusive |
| DF-1911 |
Medium |
7.0 |
r600_audio_update_hdmi walks encoder_list without mode_config.mutex: UAF race vs modeset/unload (variant of DF-1878)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r600_hdmi.c:114 |
none |
inconclusive |
| DF-1959 |
Medium |
7.0 |
dm_dev_remove use-after-free via concurrent stale-pointer removal: operator-group to root heap corruption
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/dm_dev.c:305 |
none |
inconclusive |
| DF-1979 |
Medium |
7.0 |
queue_work broken lock protocol on multi-CPU bound workqueues enables cross-CPU STAILQ corruption and arbitrary code execution
Base system — drivers · sys/dev/drm/linux_workqueue.c:119 |
none |
inconclusive |
| DF-2331 |
Medium |
7.0 |
Slave-controlled block-read byte count overflows sc->block_data[32] in interrupt handler
Kernel memory corruption (UAF / OOB write) · sys/bus/smbus/ichsmb/ichsmb.c:575 |
none |
not_reproduced |
| DF-2382 |
Medium |
7.0 |
Heap OOB write in cmi_order[] when BIOS exposes >128 CMI instances
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_hp/acpi_hp.c:1165 |
none |
not_reproduced |
| DF-2414 |
Medium |
7.0 |
Unvalidated 32-bit ccb_index read from firmware LRAM used to index ccb_infos[] in all four halt/error paths - OOB read NULL-deref and UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/advansys/advlib.c:1005 |
none |
not_reproduced |
| DF-2453 |
Medium |
7.0 |
Use-after-free of target config in async read iodone callback
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/flakey/dm_target_flakey.c:303 |
none |
not_reproduced |
| DF-2474 |
Medium |
7.0 |
Stale base pointer in isp_pci_mbxdma error cleanup causes kernel-heap multi-free
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/isp/isp_pci.c:1600 |
none |
not_reproduced |
| DF-2585 |
Medium |
7.0 |
Cluster array chain pointer read without cluster_spin UAF in volume dedup check
Hammer2 · sys/vfs/hammer2/hammer2_flush.c:1377 |
none |
not_reproduced |
| DF-1157 |
High |
6.8 |
wpi_rx_done reads tail and sets m_len from unbounded firmware head->len (OOB read / heap info-leak)
Kernel memory leaks / uninitialized disclosure · DF-1123 · sys/dev/netif/wpi/if_wpi.c:1959 |
none |
not_reproduced |
| DF-2288 |
High |
6.8 |
OOB heap write/read in fw_bus_explore_callback via unchecked ongoaddr (Config ROM parser)
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/firewire.c:1500 |
none |
not_reproduced |
| DF-2289 |
High |
6.8 |
Off-by-one stack buffer overflow in crom_next depth guard allows kernel stack corruption
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwcrom.c:115 |
none |
not_reproduced |
| DF-2359 |
High |
6.8 |
urtwn_efuse_read_data writes past sc->rom union on malicious USB device (RTL8188EU extended efuse header)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/wlan/if_urtwn.c:1756 |
none |
not_reproduced |
| DF-1097 |
Medium |
6.8 |
Use-after-free and double-free in fwohci_pci_add_child error path
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwohci_pci.c:442 |
none |
not_reproduced |
| DF-1268 |
Medium |
6.8 |
Unbounded VBIOS dependency-table counts overflow fixed SMU7 level arrays (heap OOB write)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/kv_dpm.c:2422 |
none |
inconclusive |
| DF-1280 |
Medium |
6.8 |
Unchecked reply-frame offset in mpt_intr allows OOB read/write via malicious IOC address reply
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt.c:734 |
none |
inconclusive |
| DF-1369 |
Medium |
6.8 |
Unvalidated SMID in reply descriptors allows NULL-deref panic and heap OOB write
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps.c:1515 |
none |
inconclusive |
| DF-1439 |
Medium |
6.8 |
Wild OOB read in sumo_get_valid_engine_clock when num_max_dpm_entries==0
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/sumo_dpm.c:1037 |
none |
inconclusive |
| DF-1478 |
Medium |
6.8 |
my_rxeof trusts 12-bit NIC-reported FLNG as m_len without MCLBYTES bound: OOB heap read past RX mbuf cluster
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/my/if_my.c:1105 |
none |
inconclusive |
| DF-1490 |
Medium |
6.8 |
epic_rx_done trusts 16-bit NIC-reported rxlength as m_len with no MCLBYTES bound: large OOB heap read past RX mbuf cluster (16-bit field)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/tx/if_tx.c:557 |
none |
inconclusive |
| DF-1517 |
Medium |
6.8 |
Missing MCLBYTES upper-bound check on rs->rs_datalen allows kernel heap OOB read via crafted wifi frames
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath/if_ath_rx.c:730 |
none |
inconclusive |
| DF-1519 |
Medium |
6.8 |
lge_rxeof trusts 16-bit NIC-reported LGE_RXBYTES as m_len / m_devget copy length without jumbo-buffer bound: up to 56519-byte OOB heap read
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/lge/if_lge.c:870 |
none |
inconclusive |
| DF-1554 |
Medium |
6.8 |
Malformed _CST leaving cst_cx_count==0 causes heap OOB read in set_lowest_oncpu and NULL-deref panic in idle
Kernel memory leaks / uninitialized disclosure · sys/dev/acpica/acpi_cpu_cstate.c:530 |
none |
inconclusive |
| DF-1562 |
Medium |
6.8 |
pcn_rxeof trusts 16-bit NIC-reported pcn_rxlen as m_len without MCLBYTES bound: OOB heap read past RX mbuf cluster
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/pcn/if_pcn.c:798 |
none |
inconclusive |
| DF-1566 |
Medium |
6.8 |
VBIOS-controlled VCE/UVD clock-info index and table entry counts are not bounds-checked against the BIOS allocation
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/r600_dpm.c:1104 |
none |
inconclusive |
| DF-1590 |
Medium |
6.8 |
Firmware-supplied req_id used as heap index without bounds check in tws_intr_resp and tws_err_complete
Base system — drivers · sys/dev/raid/tws/tws_cam.c:1163 |
none |
inconclusive |
| DF-1767 |
Medium |
6.8 |
virtqueue: host-controlled descriptor indices used without runtime bounds validation cause OOB read/write
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/virtio/virtio/virtqueue.c:559 |
none |
inconclusive |
| DF-1868 |
Medium |
6.8 |
Write-reply stack buffer overflow in aux_engine: missing bound of returned_byte against reply_data[16]
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/i2caux/aux_engine.c:415 |
none |
inconclusive |
| DF-1882 |
Medium |
6.8 |
Stack buffer overflow/underflow in ad_describe via non-NUL-terminated IDENTIFY model string
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/ata-disk.c:508 |
none |
inconclusive |
| DF-1938 |
Medium |
6.8 |
Write-reply stack overflow via self-comparing bound check in read_channel_reply (cross-file with DF-1868)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/i2caux/dce110/aux_engine_dce110.c:316 |
none |
inconclusive |
| DF-2005 |
Medium |
6.8 |
Missing VBIOS PowerPlay table size validation allows heap out-of-bounds read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega12_processpptables.c:64 |
none |
inconclusive fixed |
| DF-2518 |
Medium |
6.8 |
OOB array access in sili_port_read_ncq_error via unchecked device-supplied NCQ tag
Base system — drivers · sys/dev/disk/sili/sili.c:2138 |
none |
not_reproduced |
| DF-0720 |
High |
6.7 |
Heap overflow: if_alloc(IFT_PPP) returns sizeof(struct ifnet) but sppp_attach writes ~1KB of struct sppp fields past allocation (LATENT — file does not compile)
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ng_sppp.c:249 |
none |
not_reproduced |
| DF-1066 |
High |
6.7 |
Integer underflow in uvc_buf_queue_mmap_locked bounds check allows out-of-bounds vtophys() when buf_size==0
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/uvc/uvc_buf.c:107 |
none |
not_reproduced |
| DF-1133 |
High |
6.7 |
RLC v2.0 firmware register-list parsing: integer-overflow kmalloc + unvalidated offset to heap OOB write/read
Kernel memory corruption (UAF / OOB write) · DF-1119 · sys/dev/drm/amd/amdgpu/gfx_v8_0.c:1106 |
panic |
inconclusive |
| DF-1134 |
High |
6.7 |
cz_init_cp_jump_table: unvalidated jt_offset/jt_size from CE/PFP/ME/MEC fw headers -> OOB read/write
Kernel memory corruption (UAF / OOB write) · DF-1119 · sys/dev/drm/amd/amdgpu/gfx_v8_0.c:1301 |
panic |
inconclusive |
| DF-1163 |
High |
6.7 |
gfx_v7_0_init_cp_pg_table heap OOB write+read via untrusted firmware jt_offset/jt_size
Kernel memory corruption (UAF / OOB write) · DF-1134 · sys/dev/drm/amd/amdgpu/gfx_v7_0.c:3788 |
none |
inconclusive |
| DF-1203 |
High |
6.7 |
Heap buffer overflow via unbounded ucStateEntrySize in power table parsing
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/ni_dpm.c:3927 |
none |
not_reproduced |
| DF-1578 |
High |
6.7 |
Heap OOB write in INQUIRY/READ_CAPACITY/SERVICE_ACTION_IN via missing dxfer_len check
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptrr/hptrr_osm_bsd.c:582 |
none |
not_reproduced |
| DF-0402 |
Medium |
6.7 |
Direct kernel dereference of user-controlled pointer in netmap_bdg_learning before copyin: panic or kernel info-leak
Kernel memory corruption (UAF / OOB write) · sys/net/netmap/netmap_vale.c:994 |
none |
not_reproduced |
| DF-1112 |
Medium |
6.7 |
Stack buffer overflow: 32-byte buf[] used for GSBUS BLOCK/BYTES transfers carrying up to 255 bytes
Kernel memory corruption (UAF / OOB write) · sys/bus/smbus/smbacpi/smbacpi.c:137 |
none |
not_reproduced |
| DF-1178 |
Medium |
6.7 |
gfx_v9_1_parse_ind_reg_list: stack buffer overflow on indirect_start_offsets[10] and unbounded OOB read
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/gfx_v9_0.c:2084 |
none |
not_reproduced |
| DF-1185 |
Medium |
6.7 |
arcmsr_iop_message_xfer writes up to 1031 bytes past CAM buffer sized by dxfer_len
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/arcmsr/arcmsr.c:2604 |
none |
not_reproduced fixed |
| DF-1221 |
Medium |
6.7 |
RX completion req_id used as array index without bounds check (OOB read/write/panic)
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/amazon/ena/ena.c:1450 |
none |
inconclusive |
| DF-1703 |
Medium |
6.7 |
iscsi initiator: heap overflow in i_send via 32-bit truncation in i_prepPDU + signedness bypass of maxBurstLength check
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/iscsi.c:454 |
none |
inconclusive |
| DF-1820 |
Medium |
6.7 |
amr_pci: UAF - amr_pci_intr runs concurrently with DMA/command teardown in amr_pci_free; interrupt torn down LAST
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/amr/amr_pci.c:477 |
none |
inconclusive |
| DF-1854 |
Medium |
6.7 |
Heap overflow in psp_asd_load via unvalidated asd_ucode_size from PSP ASD firmware header
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_psp.c:254 |
none |
inconclusive |
| DF-1933 |
Medium |
6.7 |
Heap OOB write in dm_target_stripe_table via unchecked ksnprintf return value
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/striped/dm_target_striped.c:208 |
none |
inconclusive |
| DF-1955 |
Medium |
6.7 |
Unvalidated firmware header fields enable heap overflow in ASD firmware copy (variant of DF-1838/1854/1875)
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/psp_v10_0.c:142 |
none |
inconclusive |
| DF-1320 |
High |
6.6 |
Unbounded MST stream_count overflows stack work_table[MAX_CONTROLLER_NUM=6] in update_mst_stream_alloc_table
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/display/dc/core/dc_link.c:2322 |
none |
inconclusive |
| DF-1242 |
Medium |
6.6 |
mrsas_complete_cmd uses firmware-controlled SMID to index mpt_cmd_list with no bounds check
Base system — drivers · sys/dev/raid/mrsas/mrsas.c:1269 |
none |
not_reproduced |
| DF-1243 |
Medium |
6.6 |
mrsas_get_pd_list indexes local_pd_list with unchecked firmware deviceId (OOB write)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mrsas/mrsas.c:3307 |
none |
not_reproduced |
| DF-1244 |
Medium |
6.6 |
mrsas_get_ld_list indexes ld_ids with unchecked firmware targetId (OOB byte write)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mrsas/mrsas.c:3380 |
none |
inconclusive |
| DF-1528 |
Medium |
6.6 |
Missing bounds validation of firmware-provided RAID map fields enables out-of-bounds kernel reads on every IO
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mrsas/mrsas_fp.c:1428 |
none |
inconclusive |
| DF-1561 |
Medium |
6.6 |
Stack OOB write in pp_dpm_get_pp_num_states when hwmgr->num_ps > 16
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/amd_powerplay.c:619 |
none |
inconclusive |
| DF-2507 |
Medium |
6.6 |
afd_reinit frees softc without destroying disk orphaning device with freed memory (UAF / NULL-deref panic)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-fd.c:176 |
none |
not_reproduced |
| DF-2464 |
Low |
6.6 |
Unbounded attacker-controlled buffer offset (bo) in so_recv douio path - latent OOB write + latent stack overflow same defect class as scsi_decap DF-2458
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/isc_soc.c:450 |
none |
not_reproduced |
| DF-0489 |
High |
6.5 |
Neighbor Advertisement handler leaks route refcount on every received NA: remote kernel memory exhaustion DoS
Kernel memory corruption (UAF / OOB write) · sys/netinet6/nd6_nbr.c:734 |
none |
not_reproduced |
| DF-0542 |
High |
6.5 |
inquiry_result: unbounded variable-length loop reads past mbuf end -> remote kernel panic
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/bluetooth/hci/ng_hci_evnt.c:380 |
none |
not_reproduced |
| DF-0543 |
High |
6.5 |
num_compl_pkts: unbounded variable-length loop reads past mbuf end -> remote kernel panic
Base system — network stack · sys/netgraph7/bluetooth/hci/ng_hci_evnt.c:884 |
none |
not_reproduced |
| DF-0558 |
High |
6.5 |
hci_event_num_compl_pkts: unbounded variable-length loop NO per-iteration bounds check -> remote kernel panic
Base system — network stack · sys/netbt/hci_event.c:376 |
panic |
inconclusive |
| DF-0559 |
High |
6.5 |
hci_event_inquiry_result/rssi_result: unbounded loops guarded only by KKASSERT -> remote kernel panic on short data
Base system — network stack · sys/netbt/hci_event.c:447 |
panic |
inconclusive |
| DF-0627 |
High |
6.5 |
mbuf chain leak on duplicate SMB responses enables remote kernel memory-exhaustion DoS
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/netproto/smb/smb_iod.c:363 |
none |
not_reproduced |
| DF-0654 |
High |
6.5 |
ccmp_setkey omits AES key load for software-decrypt-only keys, causing kernel panic on first received frame
Base system — network stack · sys/netproto/802_11/wlan_ccmp/ieee80211_crypto_ccmp.c:140 |
none |
inconclusive |
| DF-1315 |
High |
6.5 |
Guest debug registers (DR7/DR0-3) persist in host after #VMEXIT enabling host kernel panic via injected HW breakpoint
Base system — drivers · sys/dev/virtual/nvmm/x86/nvmm_x86_svm.c:1492 |
none |
inconclusive |
| DF-2274 |
High |
6.5 |
Divide-by-zero panic from device-supplied block size of zero in READ CAPACITY
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_da.c:2280 |
none |
not_reproduced |
| DF-2475 |
High |
6.5 |
Wrong bitmask includes task-attribute bits in IU length causing ATIO queue consumer to skip 128-960 entries on any non-SIMPLE FCP_CMND (remote DoS)
Base system — drivers · sys/dev/disk/isp/isp_target.c:174 |
none |
not_reproduced |
| DF-0287 |
Medium |
6.5 |
Divide-by-zero kernel panic in mesh_airtime_calc via ni_txrate==0
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_mesh.c:3366 |
none |
inconclusive |
| DF-0560 |
Medium |
6.5 |
hci_event_command_compl: reads status byte beyond asserted length -> short-event remote kernel panic
Kernel memory leaks / uninitialized disclosure · sys/netbt/hci_event.c:299 |
panic |
not_reproduced |
| DF-0562 |
Medium |
6.5 |
STEPBY unsigned underflow via missing minimum length check (LMI_MIN_LENGTH defined never used) — ng7 twin of DF-0554
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/lmi/ng_lmi.c:88 |
leak:3 |
not_reproduced |
| DF-0611 |
Medium |
6.5 |
Missing IP-length / ip_hl validation in ng_nat_rcvdata allows OOB access of mbuf trailing area
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/ng_nat.c:690 |
none |
not_reproduced |
| DF-0625 |
Medium |
6.5 |
Missing arglen check in L2CA_Ping path causes kernel heap OOB read leaked to remote Bluetooth peer
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/bluetooth/l2cap/ng_l2cap_main.c:438 |
none |
not_reproduced |
| DF-0673 |
Medium |
6.5 |
nbssn_recv() leaks mbufs on receive error and on inner-loop sbinit() reset
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/netproto/smb/smb_trantcp.c:345 |
none |
not_reproduced |
| DF-0946 |
Medium |
6.5 |
swp_pager_meta_build while-loop can free a swap block assigned by a concurrent meta_build for the same pindex
Base system — virtual memory · sys/vm/swap_pager.c:2385 |
none |
not_reproduced |
| DF-0999 |
Medium |
6.5 |
Divide-by-zero in rum_set_sleep_time when associated AP advertises zero beacon interval
Base system — bus / platform / cpu · sys/bus/u4b/wlan/if_rum.c:2184 |
none |
not_reproduced |
| DF-1515 |
Medium |
6.5 |
Kernel divide-by-zero via zero dst_height/dst_width in overlay PUT_IMAGE ioctl
Base system — drivers · sys/dev/drm/i915/intel_overlay.c:916 |
none |
inconclusive |
| DF-2538 |
Medium |
6.5 |
Unbounded loop in inquiry_result causes kernel panic from malformed HCI Inquiry Result event
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/bluetooth/hci/ng_hci_evnt.c:387 |
none |
not_reproduced |
| DF-0525 |
High |
6.4 |
ng_fec_tick iterates port list with no list lock: UAF race vs addport/delport
Kernel memory corruption (UAF / OOB write) · sys/netgraph/fec/ng_fec.c:579 |
none |
not_reproduced |
| DF-0662 |
High |
6.4 |
Use-after-free of readq: ng_device_disconnect frees readq BEFORE destroy_dev drains in-flight operations
Kernel memory corruption (UAF / OOB write) · sys/netgraph/ng_device.c:407 |
none |
not_reproduced |
| DF-1018 |
High |
6.4 |
TRIM req->data heap buffer overflow and OOB DMA read via device-controlled small secsize
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_da.c:1341 |
none |
inconclusive |
| DF-1109 |
High |
6.4 |
TOCTOU double-fetch of d->nmsgs in I2CRDWR ioctl causes kernel heap buffer overflow
Kernel memory corruption (UAF / OOB write) · sys/bus/iicbus/iic.c:291 |
none |
not_reproduced |
| DF-1346 |
High |
6.4 |
Heap buffer overflow populating VCE/UVD SMC levels from unclamped VBIOS MM-dependency count
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/powerplay/smumgr/polaris10_smumgr.c:1298 |
none |
inconclusive |
| DF-0246 |
Medium |
6.4 |
UAF: eventhandler dispatch traverses entry list without token while deregister frees entries
Kernel memory corruption (UAF / OOB write) · sys/kern/subr_eventhandler.c:114 |
panic |
not_reproduced |
| DF-0296 |
Medium |
6.4 |
UAF/cross-node races: peer hooks/nodes dereferenced without reference or peer-token
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/netgraph/ng_base.c:1092 |
dos |
inconclusive |
| DF-0306 |
Medium |
6.4 |
UAF in add_bw_upcall: mfc pointer used after mroute_token released across blocking kmalloc
Kernel memory corruption (UAF / OOB write) · sys/net/ip_mroute/ip_mroute.c:2285 |
none |
inconclusive |
| DF-0355 |
Medium |
6.4 |
nd6_sysctl_drlist/prlist iterate defrouter/prefix lists without nd6_mtx: UAF race with RA processing
Kernel memory corruption (UAF / OOB write) · sys/netinet6/nd6.c:2168 |
none |
inconclusive |
| DF-0503 |
Medium |
6.4 |
ifnet locking-contract violation and stored-pointer UAF on member ifnets
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ng_fec.c:370 |
none |
not_reproduced |
| DF-0664 |
Medium |
6.4 |
Missing locking enables TOCTOU/UAF races across all device entry points
Kernel memory corruption (UAF / OOB write) · sys/netgraph/ng_device.c:309 |
none |
not_reproduced |
| DF-1045 |
Medium |
6.4 |
UAF in targclose: softc kfree'd before periph's targdtor runs
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_target.c:207 |
none |
not_reproduced |
| DF-1071 |
Medium |
6.4 |
pnp_create_devices mis-tracks scanning, causing heap OOB reads (and possible bogus device creation) from crafted PnP resource data
Kernel memory leaks / uninitialized disclosure · sys/bus/isa/pnp.c:378 |
none |
not_reproduced |
| DF-1228 |
Medium |
6.4 |
Untrusted SMID used directly as sc->commands[] index in mpr_intr_locked (OOB array access)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr.c:2336 |
none |
not_reproduced |
| DF-1294 |
Medium |
6.4 |
Unvalidated hypervisor sq_idx causes stack buffer overflow in ena_com_ind_tbl_convert_from_device
Kernel memory corruption (UAF / OOB write) · sys/dev/virtual/amazon/ena/ena-com/ena_com.c:1156 |
none |
not_reproduced |
| DF-1404 |
Medium |
6.4 |
cuse_server_open re-initializes global cuse_global_mtx on every /dev/cuse open corrupting lock state
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/cuse/cuse.c:791 |
none |
inconclusive |
| DF-1502 |
Medium |
6.4 |
Lockless driver: interrupt handler races musycc_disconnect freeing tx/rx descriptor rings (UAF)
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/musycc/musycc.c:1341 |
none |
inconclusive |
| DF-1759 |
Medium |
6.4 |
isc_sm: UAF on pduq via early pdu_free in proc_out racing network-stack ext_free
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/isc_sm.c:530 |
none |
inconclusive |
| DF-1768 |
Medium |
6.4 |
ar9285_reset: unbounded EEPROM pwrPdg delta causes OOB write/read in GetGainBoundariesAndPdadcs
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_hal/ar9002/ar9285_reset.c:650 |
none |
inconclusive |
| DF-1829 |
Medium |
6.4 |
isc_soc: UAF/refcount corruption via ext_free on reused pduq in send path
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/isc_soc.c:70 |
none |
inconclusive |
| DF-1961 |
Medium |
6.4 |
Heap OOB write in k8pnow_decode_pst insertion sort from unchecked pst->n_states>16
Kernel memory corruption (UAF / OOB write) · sys/dev/powermng/powernow/powernow.c:99 |
none |
inconclusive |
| DF-2397 |
Medium |
6.4 |
Unsynchronized AGP memory list operations enable UAF and TAILQ corruption via concurrent ioctls
Kernel memory corruption (UAF / OOB write) · sys/dev/agp/agp.c:455 |
none |
not_reproduced |
| DF-2499 |
Medium |
6.4 |
TAILQ_FOREACH iterates pending_hcbs while freeing each element use-after-free read
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-cam.c:885 |
none |
not_reproduced |
| DF-2500 |
Medium |
6.4 |
In-flight ATA requests outlive atapi_cam_detach atapi_cb dereferences freed hcb and destroyed state_lock
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-cam.c:619 |
none |
not_reproduced |
| DF-2536 |
Medium |
6.4 |
Sense-data bcopy length not capped to 32-byte sense buffer kernel heap overflow into CCB and info leak of request-frame memory
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_cam.c:3141 |
none |
not_reproduced |
| DF-0593 |
Info |
6.4 |
Latent UAF: fairq_class_destroy does not clear dangling pif_default pointer (currently unreachable via pf ioctls)
Kernel memory corruption (UAF / OOB write) · sys/net/altq/altq_fairq.c:428 |
none |
not_reproduced |
| DF-0546 |
High |
6.3 |
OOB read of lut[] in netmap_mem_ofstophys: page-padding offset maps garbage physical page into userspace
Kernel memory leaks / uninitialized disclosure · sys/net/netmap/netmap_mem2.c:165 |
none |
not_reproduced |
| DF-0974 |
Medium |
6.3 |
xpt_action_sasync_cb use-after-free race on async_node during concurrent register/unregister
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/cam_xpt.c:3353 |
none |
not_reproduced |
| DF-1194 |
Medium |
6.3 |
ciss_init_logical and ciss_free use <= instead of < on ciss_logical array (off-by-one OOB write/kfree)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/ciss/ciss.c:1376 |
none |
not_reproduced |
| DF-1285 |
Medium |
6.3 |
INQUIRY/READ_CAPACITY/SERVICE_ACTION_IN write fixed-size replies past dxfer_len-sized kernel buffer
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/hptmv/entry.c:2709 |
none |
inconclusive |
| DF-1287 |
Medium |
6.3 |
siba_pci_sprom dispatches rev-4/5/8 parsers without validating buffer size -> heap OOB read past 64-word buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/bwn/siba/siba_core.c:1398 |
none |
inconclusive |
| DF-1362 |
Medium |
6.3 |
Integer overflow in mps_diag_read_buffer bounds check enables OOB read of fw_diag_buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps_user.c:1583 |
none |
inconclusive |
| DF-1398 |
Medium |
6.3 |
Off-by-one OOB array access in all four remove_*_entry functions (index==allocated passes > check)
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumconfig.c:1728 |
none |
inconclusive |
| DF-1418 |
Medium |
6.3 |
OOB read of vddc_dependency_on_sclk via clock_info_index clamped to wrong bound
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/smu8_hwmgr.c:1350 |
none |
inconclusive |
| DF-1475 |
Medium |
6.3 |
Off-by-one in PhyNum target-ID fallback allows heap OOB write past sassc->targets[]
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_sas_lsi.c:885 |
none |
inconclusive |
| DF-1814 |
Medium |
6.3 |
tw_cl_init: tw_cli_init_connection double-inserts req into FREE_Q on any submit_and_poll failure -> queue list corruption
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/twa/tw_cl_init.c:677 |
none |
inconclusive |
| DF-1039 |
Medium |
6.2 |
NULL-pointer dereference in pccard_child_pnpinfo_str via pccard_safe_quote when CIS lacks VERS_1 tuple
Base system — bus / platform / cpu · sys/bus/pccard/pccard.c:1020 |
none |
not_reproduced |
| DF-1238 |
Medium |
6.2 |
trm_Reselect walks circular DCB ring with no termination -> infinite loop / hard system hang
Base system — drivers · sys/dev/disk/trm/trm.c:2426 |
none |
not_reproduced |
| DF-1377 |
Medium |
6.2 |
btccbptov() performs no bounds validation on HBA-supplied ccb_addr yielding OOB kernel read/write
Base system — drivers · sys/dev/disk/buslogic/bt.c:104 |
none |
inconclusive |
| DF-1438 |
Medium |
6.2 |
Divide-by-zero in sumo_calculate_bsp via BIOS-controlled high_clk=0
Base system — drivers · sys/dev/drm/radeon/sumo_dpm.c:323 |
none |
inconclusive |
| DF-2317 |
Medium |
6.2 |
Kernel panic on attacker-controlled SD CSD structure version
Base system — bus / platform / cpu · sys/bus/mmc/mmc.c:1056 |
none |
not_reproduced |
| DF-2350 |
Medium |
6.2 |
Heap OOB read in VS frame descriptor interval parsing leaks kernel memory to userspace
Privilege escalation (general use) · sys/bus/u4b/uvc/uvc_drv.c:1900 |
none |
not_reproduced |
| DF-2411 |
Medium |
6.2 |
Divide-by-zero DoS in XPT_CALC_GEOMETRY via target-controlled block_size
Base system — drivers · sys/dev/disk/advansys/advansys.c:441 |
none |
not_reproduced |
| DF-1210 |
High |
6.1 |
OOB read in DHRS error path: device-controlled CCS slot index dereferenced before bounds check
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/ahci/ahci.c:2843 |
none |
not_reproduced |
| DF-2024 |
High |
6.1 |
size_t underflow in multi-part SSIF write loop causes unbounded kernel heap OOB read and DoS
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/ipmi/ipmi_ssif.c:126 |
none |
inconclusive |
| DF-2083 |
High |
6.1 |
Heap OOB read in amdgpu_dm_set_degamma_lut via unvalidated blob size before __is_lut_linear
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_color.c:259 |
none |
not_reproduced |
| DF-0997 |
Medium |
6.1 |
RX mbuf m_len set from device-controlled RX descriptor flags with no bound -> heap OOB read
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/wlan/if_rum.c:1146 |
none |
not_reproduced |
| DF-1012 |
Medium |
6.1 |
Heap underflow write when VS_COLORFORMAT or VS_FRAME precedes any VS_FORMAT descriptor
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/uvc/uvc_drv.c:1836 |
none |
not_reproduced |
| DF-1182 |
Medium |
6.1 |
Heap OOB read via unbounded state-record clock indices into VBIOS dep tables
Kernel memory leaks / uninitialized disclosure · DF-1168 · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:2970 |
none |
not_reproduced |
| DF-1200 |
Medium |
6.1 |
Unvalidated BIOS-supplied array indices in power table parsing cause OOB heap reads
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_atombios.c:606 |
none |
not_reproduced |
| DF-1223 |
Medium |
6.1 |
XPT_RESET_DEV frees a regular-pool command via the high-priority free list, corrupting command-pool accounting
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_sas.c:3351 |
none |
inconclusive |
| DF-1253 |
Medium |
6.1 |
uint32 underflow in vega20_apply_clocks_adjust_rules: count-1 wraps to UINT32_MAX causing ~4B-iteration heap OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega20_hwmgr.c:3035 |
none |
inconclusive |
| DF-1388 |
Medium |
6.1 |
Integer overflow in find_section block-size check lets crafted MIPI block bypass bounds for OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_bios.c:96 |
none |
inconclusive |
| DF-1389 |
Medium |
6.1 |
get_lvds_dvo_timing uses unvalidated VBT offsets for pointer arithmetic -> OOB read of LVDS timing
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_bios.c:169 |
none |
inconclusive |
| DF-1390 |
Medium |
6.1 |
parse_mipi_config indexes config[6]/pps[6] with panel_type 0-15 -> OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_bios.c:808 |
none |
inconclusive |
| DF-1456 |
Medium |
6.1 |
Signed integer overflow in drm_mode_set_crtcinfo via unchecked vscan on user-supplied modes
Base system — drivers · sys/dev/drm/drm_modes.c:1102 |
none |
inconclusive |
| DF-1547 |
Medium |
6.1 |
Missing cdb_len bounds check overflows the LRAM queue slot (sibling of DF-1356)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/advansys/advansys.c:548 |
none |
inconclusive |
| DF-1576 |
Medium |
6.1 |
Unbounded VBIOS USHORT offsets and VCE state indices cause OOB reads from BIOS image
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/process_pptables_v1_0.c:283 |
none |
inconclusive |
| DF-1810 |
Medium |
6.1 |
cardbus_cis: unbounded CIS tuple-chain offset allows OOB read past BAR mapping (panic / MAC info leak)
Kernel memory leaks / uninitialized disclosure · sys/dev/pccard/cardbus/cardbus_cis.c:451 |
none |
inconclusive |
| DF-1920 |
Medium |
6.1 |
Cleanup loop reads uninitialized stack arrays when copyin fails partway through SGL loop
Base system — drivers · sys/dev/raid/mrsas/mrsas_ioctl.c:153 |
none |
inconclusive |
| DF-0660 |
High |
6.0 |
Heap memory corruption: kfree() on mbuf data pointer in ng_device_rcvdata
Base system — network stack · sys/netgraph/ng_device.c:363 |
none |
not_reproduced |
| DF-1250 |
High |
6.0 |
Hardcoded EDID parser over-reads rdev->bios by up to 32KB, disclosing kernel heap via EDID ioctl
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_combios.c:373 |
none |
not_reproduced |
| DF-1251 |
High |
6.0 |
Unbounded BIOS table-walking loops read past rdev->bios and drive WREG32 with OOB garbage
Base system — drivers · sys/dev/drm/radeon/radeon_combios.c:291 |
none |
not_reproduced |
| DF-2570 |
High |
6.0 |
Heap corruption: kfree() of in-mbuf data pointer on every received packet in ng_device_rcvdata
Base system — network stack · sys/netgraph/ng_device.c:363 |
none |
not_reproduced |
| DF-0276 |
Medium |
6.0 |
Wrong-pointer kfree in DIOCADDADDR error path frees framework pointer instead of allocated pooladdr
Base system — network stack · sys/net/pf/pf_ioctl.c:2160 |
none |
not_reproduced |
| DF-1408 |
Medium |
6.0 |
Integer underflow / unbounded num_indices in multiple ATOM table parsers yields OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_atombios.c:106 |
none |
inconclusive |
| DF-1733 |
Medium |
6.0 |
vinumio: heap overflow on drivelist[] - DRIVEPARTS(35) smaller than loop bound (240) in vinum_scandisk
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumio.c:678 |
none |
inconclusive |
| DF-1750 |
Medium |
6.0 |
amdgpu_acpi: heap OOB read and ACPI_OBJECT type confusion in 4 ATIF/ATCS return buffer parsers
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_acpi.c:199 |
none |
inconclusive |
| DF-1769 |
Medium |
6.0 |
ar9285_reset: chain loop uses AR5416_MAX_CHAINS=3 on 4k EEPROM struct with calPierData2G[1][3] -> OOB type-confusion via txMask
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ath/ath_hal/ar9002/ar9285_reset.c:607 |
none |
inconclusive |
| DF-1875 |
Medium |
6.0 |
Unvalidated PSP firmware header fields cause heap OOB write in fw_pri_buf (sysdrv + sos load) — variant of DF-1838 for Vega20/Navi10
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/psp_v11_0.c:122 |
none |
inconclusive |
| DF-1895 |
Medium |
6.0 |
parse_csr_fw never validates fw->datasize before indexing fw->data: multiple OOB-read paths
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_csr.c:297 |
none |
inconclusive |
| DF-1964 |
Medium |
6.0 |
Orphaned worker thread / use-after-free when write-side _init fails in dm_target_delay_init
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/delay/dm_target_delay.c:104 |
none |
inconclusive |
| DF-0302 |
Medium |
5.9 |
Failover state machine in input path runs without synchronization: concurrent state corruption
Base system — network stack · sys/netinet/ip_carp.c:1108 |
dos |
inconclusive |
| DF-0391 |
Medium |
5.9 |
pf_fragcache: m_dup NULL dereferenced before NULL check in m_adj argument: remote DoS under memory pressure with fragcrop
Base system — network stack · sys/net/pf/pf_norm.c:661 |
panic |
inconclusive |
| DF-0423 |
Medium |
5.9 |
Stale reg_mif_num after MRT6_DEL_MIF: remote NULL-deref panic via PIM REGISTER to freed register mif
Kernel memory corruption (UAF / OOB write) · sys/netinet6/ip6_mroute.c:642 |
panic |
inconclusive |
| DF-1155 |
Medium |
5.9 |
Heap overflow in isp_handle_platform_atio CDB copy (parallel SCSI target mode)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/isp/isp_freebsd.c:1926 |
none |
not_reproduced |
| DF-0044 |
Medium |
5.8 |
mount_get_by_nc returns struct mount without a hold -> use-after-free via cache_fullpath racing dounmount
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_mount.c:1235 |
none on this non-INVARIANTS kernel (race too tight); theoretical local DoS / info-leak if the race is won (unprivileged if vfs.usermount=1, else races a privileged unmount) |
not_reproduced |
| DF-0565 |
Medium |
5.8 |
Callout lifetime: ng_uncallout non-synchronizing + nglmi_shutdown omits ng_uncallout — ng7 twin of DF-0557
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/lmi/ng_lmi.c:1053 |
panic |
not_reproduced |
| DF-1212 |
Medium |
5.8 |
OOB array access in FBSS saved-commands reissue loop (hardcoded 32 vs sc_ncmds)
Base system — drivers · sys/dev/disk/ahci/ahci.c:3972 |
none |
inconclusive |
| DF-1383 |
Medium |
5.8 |
Out-of-bounds write in tl_encap when TX mbuf chain has exactly TL_MAXFRAGS non-empty mbufs totaling <64 bytes
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/tl/if_tl.c:1794 |
none |
inconclusive |
| DF-2417 |
Medium |
5.8 |
ADW_TARGET_GROUP macro off-by-factor-4 causes heap OOB write/read on user_sdtr[4] for all wide-SCSI targets (tid >= 4)
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/advansys/adwlib.c:754 |
none |
not_reproduced |
| DF-2572 |
High |
5.7 |
Use-after-free / NULL-deref: connection list mutated and traversed with no synchronization
Kernel memory corruption (UAF / OOB write) · sys/netgraph/ng_device.c:309 |
none |
not_reproduced |
| DF-0965 |
Medium |
5.7 |
twofish_set_key performs no key-length validation, allowing OOB stack writes and ctx corruption if reached with unsupported keylen
Kernel memory corruption (UAF / OOB write) · sys/crypto/twofish/twofish.c:425 |
none |
not_reproduced |
| DF-0988 |
Medium |
5.7 |
cdsize accepts blksize=0 from READ_CAPACITY, cdstart divide-by-zero panics on read
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_cd.c:3058 |
none |
not_reproduced |
| DF-0989 |
Medium |
5.7 |
cdreaddvdstructure integer underflow in length math leaks kernel heap to userspace
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/scsi/scsi_cd.c:4288 |
none |
not_reproduced |
| DF-1072 |
Medium |
5.7 |
pnp_create_devices ANSI tag trim loop reads/writes stack below buf[0] with no lower bound
Kernel memory corruption (UAF / OOB write) · sys/bus/isa/pnp.c:398 |
none |
not_reproduced |
| DF-1090 |
Medium |
5.7 |
pnp_parse_desc accesses fixed descriptor offsets without validating payload length, causing heap OOB read and OOB byte-write
Kernel memory corruption (UAF / OOB write) · sys/bus/isa/pnpparse.c:90 |
none |
not_reproduced fixed |
| DF-1214 |
Medium |
5.7 |
Unchecked rxcd->len vs cluster size yields mbuf OOB read (guest kernel heap info leak) on RX
Kernel memory leaks / uninitialized disclosure · sys/dev/virtual/vmware/vmxnet3/if_vmx.c:2216 |
none |
inconclusive |
| DF-1396 |
Medium |
5.7 |
Firmware-controlled desc_count/desc_len in oce_get_profile_config/oce_get_func_config walks OOB past DMA buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/oce/oce_mbox.c:2013 |
none |
inconclusive |
| DF-1464 |
Medium |
5.7 |
Unbounded scatter/gather list copy overflows per-CCB sg_list when CAM_SCATTER_VALID is set
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/dpt/dpt_scsi.c:730 |
none |
inconclusive |
| DF-1709 |
Medium |
5.7 |
iwm_scan: integer overflow in iwm_scan_size()/iwm_config_umac_scan via unbounded firmware n_scan_channels -> heap OOB write
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/iwm/if_iwm_scan.c:486 |
none |
inconclusive |
| DF-1713 |
Medium |
5.7 |
evdev: use-after-free race on ev_absinfo/ev_mt vs concurrent device detach in unlocked ioctls
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/evdev/cdev.c:351 |
none |
inconclusive |
| DF-1784 |
Medium |
5.7 |
radeon_bios: OOB read in radeon_get_bios - attacker-controlled BIOS offsets index into potentially tiny BIOS buffer
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_bios.c:752 |
none |
inconclusive |
| DF-2000 |
Medium |
5.7 |
UAF race in icioctl SIOCSIFMTU vs concurrent icoutput (TX) and icintr (RX)
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/ic/if_ic.c:205 |
none |
inconclusive |
| DF-2336 |
Medium |
5.7 |
Unbounded MMIO read/write in xHCI extended-capability list traversal
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/controller/xhci_pci.c:398 |
none |
not_reproduced |
| DF-1399 |
Medium |
5.6 |
vol->plex[] set to -1 by kw_prefer logic bug causing PLEX[-1] OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/vinum/vinumconfig.c:1537 |
none |
inconclusive |
| DF-1734 |
Medium |
5.6 |
vinumio: remove_drive indexes drive[] with no bounds check; caller off-by-one > instead of >=
Base system — drivers · sys/dev/raid/vinum/vinumio.c:237 |
none |
inconclusive |
| DF-1739 |
Medium |
5.6 |
vinumioctl: validdrive/validsd/validplex/validvol/resetstats accept negative indices -> OOB heap read/write via DRIVE[idx]/SD[idx]/PLEX[idx]/VOL[idx]
Base system — drivers · sys/dev/raid/vinum/vinumioctl.c:381 |
none |
inconclusive |
| DF-1884 |
Medium |
5.6 |
Missing bounds check on plexno in parityops allows OOB read/write on PLEX array
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/vinum/vinumrevive.c:259 |
none |
inconclusive |
| DF-1327 |
High |
5.5 |
Kernel heap info leak in mpr_user_pass_thru: copyout uses user-controlled ReplySize with no upper bound
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_user.c:857 |
none |
inconclusive |
| DF-1328 |
High |
5.5 |
Kernel heap info leak in mpr_user_event_report: copyout of user-supplied Size from fixed-size recorded_events[40000]
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_user.c:2069 |
none |
inconclusive |
| DF-1329 |
High |
5.5 |
Integer overflow in mpr_diag_read_buffer bounds check allows OOB read of diag DMA buffer
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/mpr/mpr_user.c:1772 |
none |
inconclusive |
| DF-1360 |
High |
5.5 |
Kernel heap memory disclosure via mps_user_event_report copyout with attacker-controlled length
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mps/mps_user.c:1856 |
none |
inconclusive |
| DF-1976 |
High |
5.5 |
flush_workqueue hangs forever in uninterruptible tsleep (permanent kernel thread DoS)
Base system — drivers · sys/dev/drm/linux_workqueue.c:309 |
none |
inconclusive |
| DF-2163 |
High |
5.5 |
tsleep while holding queue_lock deadlocks interrupt handler permanently hanging IPS controller
Base system — drivers · sys/dev/raid/ips/ips_ioctl.c:125 |
none |
not_reproduced |
| DF-2164 |
High |
5.5 |
get_user_pages() KKASSERT inverts td_proc polarity, panicking on every user-context call
Base system — drivers · sys/dev/drm/linux_shmem.c:137 |
none |
not_reproduced |
| DF-2222 |
High |
5.5 |
wait_for_completion_interruptible spins forever on signal (no break on ERESTART/EINTR) unkillable CPU-bound local DoS
Base system — drivers · sys/dev/drm/linux_completion.c:38 |
none |
not_reproduced |
| DF-2281 |
High |
5.5 |
sgwrite: size_t->int truncation of uio_resid into buf_len plus unchecked kmalloc/M_ZERO memset-on-NULL yields local kernel panic
Kernel memory corruption (UAF / OOB write) · sys/bus/cam/scsi/scsi_sg.c:648 |
none |
not_reproduced |
| DF-0705 |
Medium |
5.5 |
Local DoS / panic via out-of-order ioctls, kthread_create failure, and negative count in show_conf
Base system — network stack · sys/net/ipfw3_basic/ip_fw3_sync.c:239 |
none |
not_reproduced |
| DF-0808 |
Medium |
5.5 |
dirfs_nrename NULL deref panic on over-length or unlinked rename paths
Base system — filesystems · sys/vfs/dirfs/dirfs_vnops.c:977 |
none |
not_reproduced |
| DF-0940 |
Medium |
5.5 |
Integer truncation in vm_map_growstack grow_amount enables unbounded kernel-memory-growth local DoS
Base system — virtual memory · sys/vm/vm_map.c:4110 |
none |
not_reproduced |
| DF-0952 |
Medium |
5.5 |
sys_vmspace_destroy leaks ve->refs on EBUSY, causing kernel panic in rb_vmspace_delete at proc exit
Base system — virtual memory · sys/vm/vm_vmspace.c:222 |
none |
not_reproduced |
| DF-0996 |
Medium |
5.5 |
pmap_unwire dereferences NULL pte when page table page does not exist -> vkernel crash
Base system — bus / platform / cpu · sys/platform/vkernel64/platform/pmap.c:2637 |
none |
not_reproduced |
| DF-1028 |
Medium |
5.5 |
OOB heap read in READ ELEMENT STATUS response parsing (device-controlled desclen/avail/flags)
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/scsi/scsi_ch.c:1118 |
none |
not_reproduced |
| DF-1029 |
Medium |
5.5 |
OOB heap read via unchecked blk_desc_len in MODE SENSE parsing (chdone/chgetparams)
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/scsi/scsi_ch.c:494 |
none |
not_reproduced |
| DF-1059 |
Medium |
5.5 |
FW_SBINDADDR / FW_CBINDADDR dereference ir (= d->ir) without NULL check — local kernel NULL-deref DoS
Base system — bus / platform / cpu · sys/bus/firewire/fwdev.c:600 |
none |
not_reproduced |
| DF-1113 |
Medium |
5.5 |
Unbounded ACPI table Length in sdt_sdth_map enables deterministic boot-time panic DoS
Base system — bus / platform / cpu · sys/platform/pc64/acpica/acpi_sdt.c:137 |
none |
not_reproduced |
| DF-1143 |
Medium |
5.5 |
Divide-by-zero in ci_thermal_setup_fan_table from VBIOS-controlled fan temperature deltas
Base system — drivers · DF-1129 · sys/dev/drm/amd/amdgpu/ci_dpm.c:1094 |
none |
not_reproduced |
| DF-1152 |
Medium |
5.5 |
BUG_ON kernel panic in atomic_check via MST connector without dc_sink
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm.c:5290 |
none |
not_reproduced |
| DF-1183 |
Medium |
5.5 |
Heap OOB read in vega10_get_soc_index_for_max_uclk: hardcoded entries[3] without count check
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega10_hwmgr.c:3444 |
none |
not_reproduced |
| DF-1193 |
Medium |
5.5 |
CISS_BIG_MAP_BUS divide-by-zero: drives_per_scsi_bus==0 from controller ID_CTLR panics the kernel
Base system — drivers · sys/dev/raid/ciss/ciss.c:4026 |
none |
not_reproduced |
| DF-1245 |
Medium |
5.5 |
Divide-by-zero DoS in evergreen_cs_track_validate_cb via user-controlled CB_COLOR_PITCH
Base system — drivers · sys/dev/drm/radeon/evergreen_cs.c:1396 |
none |
inconclusive |
| DF-1260 |
Medium |
5.5 |
AIOGCAP ioctl dereferences d->mixer_dev without NULL check, causing kernel panic
Base system — drivers · sys/dev/sound/pcm/dsp.c:1285 |
none |
inconclusive |
| DF-1274 |
Medium |
5.5 |
Divide-by-zero in SCLK/MCLK spread-spectrum math from VBIOS-controlled zero rates
Base system — drivers · sys/dev/drm/amd/powerplay/smumgr/tonga_smumgr.c:591 |
none |
inconclusive |
| DF-1393 |
Medium |
5.5 |
Off-by-one in seq_fetch_mid yields NULL kobj dereference (local kernel panic via 0666 device)
Base system — drivers · sys/dev/sound/midi/sequencer.c:724 |
none |
inconclusive |
| DF-1403 |
Medium |
5.5 |
Unprivileged OOB read / kernel memory leak in cik_get_register_value via attacker-controlled SE/SH index
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/cik.c:1026 |
none |
inconclusive |
| DF-1431 |
Medium |
5.5 |
Divide-by-zero in radeon_compute_pll_avivo via sub-10 kHz mode clock
Base system — drivers · sys/dev/drm/radeon/radeon_display.c:959 |
none |
inconclusive |
| DF-1433 |
Medium |
5.5 |
NULL pointer dereference in radeon_wb_init on GPU resume/reset (kernel panic)
Base system — drivers · sys/dev/drm/radeon/radeon_device.c:454 |
none |
inconclusive |
| DF-1434 |
Medium |
5.5 |
CDRIOCSETBLOCKSIZE with arg=0 causes kernel divide-by-zero panic via rounddown()
Privilege escalation (general use) · sys/dev/disk/nata/atapi-cd.c:723 |
none |
inconclusive |
| DF-1453 |
Medium |
5.5 |
NULL pointer dereference in radeon_legacy_tv_dac_mode_set via unguarded enc_priv
Base system — drivers · sys/dev/drm/radeon/radeon_legacy_encoders.c:1159 |
none |
inconclusive |
| DF-1467 |
Medium |
5.5 |
vi_get_register_value OOB read of gfx.config.rb_config[se][sh] via AMDGPU_INFO_READ_MMR_REG (DF-1403 twin)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/vi.c:550 |
none |
inconclusive |
| DF-1471 |
Medium |
5.5 |
NULL deref DoS in init_overdrive_limits and cac_dtp_table allocation path
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/processpptables.c:1062 |
none |
inconclusive |
| DF-1494 |
Medium |
5.5 |
midistat_read OOB kernel heap read via unchecked negative uio_offset
Kernel memory leaks / uninitialized disclosure · sys/dev/sound/midi/midi.c:1038 |
none |
inconclusive |
| DF-1495 |
Medium |
5.5 |
qlock leak in midi_read/midi_write/midisynth_writeraw blocks device permanently (local DoS)
Base system — drivers · sys/dev/sound/midi/midi.c:762 |
none |
inconclusive |
| DF-1509 |
Medium |
5.5 |
Unbounded multicast CAM index in bfe_set_rx_mode enables local NIC hang DoS
Base system — drivers · sys/dev/netif/bfe/if_bfe.c:844 |
none |
inconclusive |
| DF-1520 |
Medium |
5.5 |
Heap OOB read in ath_hal_getregdump via HAL_DIAG_REGS — loop bound on output space, not on input array size
Privilege escalation (general use) · sys/dev/netif/ath/ath_hal/ah.c:848 |
none |
inconclusive |
| DF-1521 |
Medium |
5.5 |
NULL-deref kernel panic via HAL_DIAG_REGS / HAL_DIAG_EEREAD when caller passes result=NULL or args=NULL
Privilege escalation (general use) · sys/dev/netif/ath/ath_hal/ah.c:880 |
none |
inconclusive |
| DF-1537 |
Medium |
5.5 |
Unbounded WS operand index and NULL dereference when ws=0
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/atom.c:233 |
none |
inconclusive |
| DF-1555 |
Medium |
5.5 |
Memory leak of drm_pending_vblank_event on out-fence error path in complete_signaling()
Base system — drivers · sys/dev/drm/drm_atomic_uapi.c:1083 |
none |
inconclusive |
| DF-1559 |
Medium |
5.5 |
Signedness bug in copy_batch min_t causes kernel heap overflow with large batch_len
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/i915_cmd_parser.c:1104 |
none |
inconclusive |
| DF-1575 |
Medium |
5.5 |
NULL pointer dereference in get_pcie_table/get_gpio_table when vdd_dep_on_sclk is NULL
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/process_pptables_v1_0.c:522 |
none |
inconclusive |
| DF-1587 |
Medium |
5.5 |
NULL pointer dereference in get_pcie_table when usGfxclkDependencyTableOffset is zero
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/vega10_processpptables.c:765 |
none |
inconclusive |
| DF-1593 |
Medium |
5.5 |
rv515_debugfs_ga_info issues destructive GPU reset on every read of world-readable debugfs file
Base system — drivers · sys/dev/drm/radeon/rv515.c:259 |
none |
inconclusive |
| DF-1595 |
Medium |
5.5 |
Divide-by-zero in UVD decode message parser crashes the kernel
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_uvd.c:524 |
none |
inconclusive |
| DF-1708 |
Medium |
5.5 |
radeon_clocks: divide-by-zero in calc_eng_mem_clock via zeroed reference_freq from crafted VBIOS
Base system — drivers · sys/dev/drm/radeon/radeon_clocks.c:371 |
none |
inconclusive |
| DF-1717 |
Medium |
5.5 |
amdgpu_display: GEM object reference leak on imported dma-buf framebuffer creation path (local memory-exhaustion DoS)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_display.c:531 |
none |
inconclusive |
| DF-1724 |
Medium |
5.5 |
radeon_cs: unchecked IB chunk length_dw in non-VM path triggers BUG_ON kernel panic
Base system — drivers · sys/dev/drm/radeon/radeon_cs.c:608 |
none |
inconclusive |
| DF-1726 |
Medium |
5.5 |
radeon_cs: NULL parser->ib.fence dereference in radeon_cs_parser_fini when CS has relocs but no IB chunk
Base system — drivers · sys/dev/drm/radeon/radeon_cs.c:414 |
none |
inconclusive |
| DF-1774 |
Medium |
5.5 |
mpt_user: uninitialized kernel-heap memory leak via MPTIO_READ_CFG_PAGE / READ_EXT_CFG_PAGE
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt_user.c:186 |
none |
inconclusive |
| DF-1775 |
Medium |
5.5 |
mpt_user: kernel panic via len==0 kmalloc(0) returns ZERO_LENGTH_PTR -> page fault in copyin/hdr deref
Base system — drivers · sys/dev/disk/mpt/mpt_user.c:186 |
none |
inconclusive |
| DF-1789 |
Medium |
5.5 |
ar2316: missing numChannels>0 guard in getGainBoundariesAndPdadcsForPowers -> kernel OOB read (mirrors DF-1766)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5212/ar2316.c:384 |
none |
inconclusive |
| DF-1855 |
Medium |
5.5 |
Unbounded busy-wait in psp_cmd_submit_buf hangs kernel thread on PSP submission failure or non-response
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_psp.c:130 |
none |
inconclusive |
| DF-1907 |
Medium |
5.5 |
acp_poweron dereferences powerplay.pp_funcs without NULL guard: kernel panic
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_acp.c:145 |
none |
inconclusive |
| DF-1921 |
Medium |
5.5 |
Early-return error paths leak mfi cmds and DMA resources: cmd-pool exhaustion DoS
Base system — drivers · sys/dev/raid/mrsas/mrsas_ioctl.c:193 |
none |
inconclusive |
| DF-1932 |
Medium |
5.5 |
kfree(NULL) kernel panic on any PCI attach failure path
Base system — drivers · sys/dev/raid/aac/aac_pci.c:358 |
none |
inconclusive |
| DF-1944 |
Medium |
5.5 |
afd_reinit frees softc without destroying disk cdev: orphan /dev/afd* NULL-deref/UAF on next I/O
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/nata/atapi-fd.c:168 |
none |
inconclusive |
| DF-1951 |
Medium |
5.5 |
Kernel heap info leak: uninitialized BIOS buffer exposed to unprivileged userspace via AMDGPU_INFO_VBIOS_IMAGE ioctl on ATRM path
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_bios.c:338 |
none |
inconclusive |
| DF-1954 |
Medium |
5.5 |
OOB kernel heap info leak in sndstat_read via signed/unsigned truncation in min() length math
Kernel memory leaks / uninitialized disclosure · sys/dev/sound/pcm/sndstat.c:206 |
none |
inconclusive |
| DF-1957 |
Medium |
5.5 |
Divide-by-zero in clip_scaled when destination rectangle has zero width or height
Base system — drivers · sys/dev/drm/drm_rect.c:53 |
none |
inconclusive |
| DF-1981 |
Medium |
5.5 |
Missing feature-arg-count validation in _init_features allows kernel panic via crafted dmsetup table string
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/flakey/dm_target_flakey.c:131 |
none |
inconclusive |
| DF-1987 |
Medium |
5.5 |
BUG_ON panic on oversize SA allocation request converts user-triggerable error into kernel panic (Local DoS)
Base system — drivers · sys/dev/drm/radeon/radeon_sa.c:321 |
none |
inconclusive |
| DF-2003 |
Medium |
5.5 |
Unbounded busy-wait polling in SMIC allows permanent kernel thread hang / CPU DoS
Base system — drivers · sys/dev/misc/ipmi/ipmi_smic.c:53 |
none |
inconclusive |
| DF-2215 |
Medium |
5.5 |
Unconditional memset on FREEBLKS bios writes to unmapped KVA causing kernel panic
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/dm/dm_target_zero.c:43 |
none |
not_reproduced |
| DF-2245 |
Medium |
5.5 |
Unbounded recursion in kqsort causes kernel stack overflow on adversarial input
Kernel memory corruption (UAF / OOB write) · sys/libkern/qsort.c:173 |
none |
not_reproduced |
| DF-2463 |
Medium |
5.5 |
NULL-deref panic in so_input: pdu_alloc retry uses M_NOWAIT instead of M_WAITOK then dereferences pq unconditionally
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/isc_soc.c:540 |
none |
inconclusive |
| DF-2472 |
Medium |
5.5 |
ISP_FC_GETDINFO ioctl missing channel bounds check -> kernel heap OOB read / info leak
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/isp/isp_freebsd.c:471 |
none |
not_reproduced |
| DF-2479 |
Medium |
5.5 |
Divide-by-zero in mmcsd_delete via attacker-controlled CSD erase-sector size
Base system — drivers · sys/dev/disk/mmcsd/mmcsd.c:407 |
none |
not_reproduced |
| DF-2486 |
Medium |
5.5 |
Kernel heap info leak via uninitialized DMA buffer in MPTIO_READ_CFG_PAGE / MPTIO_READ_EXT_CFG_PAGE
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt_user.c:203 |
none |
not_reproduced |
| DF-2503 |
Medium |
5.5 |
TOCTOU info leak in CDIOCREADSUBCHANNEL via double-fetch of args->data_len
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/nata/atapi-cd.c:502 |
none |
not_reproduced |
| DF-2512 |
Medium |
5.5 |
Admin request leak on failed Identify NS exhausts pool and panics kernel
Base system — drivers · sys/dev/disk/nvme/nvme_admin.c:624 |
none |
not_reproduced |
| DF-2514 |
Medium |
5.5 |
Divide-by-zero kernel panic in XPT_CALC_GEOMETRY from block_size > 1MB
Base system — drivers · sys/dev/disk/sbp/sbp.c:2384 |
none |
not_reproduced |
| DF-2535 |
Medium |
5.5 |
OOB read / deterministic panic in mpt_disable_lun via wildcard LUN (loop indexes trt[lun] not trt[i])
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/mpt/mpt_cam.c:4334 |
none |
not_reproduced |
| DF-2561 |
Medium |
5.5 |
Accept-filter NULL-deref/UAF race in soisconnected vs do_setopt_accept_filter
Kernel memory corruption (UAF / OOB write) · sys/kern/uipc_socket2.c:252 |
none |
not_reproduced |
| DF-2568 |
Medium |
5.5 |
NULL-deref panic in hammer2_flush retry loop when chain loses parent during flush
Hammer2 · sys/vfs/hammer2/hammer2_flush.c:404 |
panic |
not_reproduced |
| DF-2390 |
Low |
5.5 |
Boot-time kernel panic from inconsistent DPF handling between resource counter and link_add_crs
Base system — drivers · sys/dev/acpica/acpi_pci_link.c:180 |
none |
not_reproduced |
| DF-0282 |
Medium |
5.4 |
Signed tx_cred overflow and OOB credit byte read in UIH reception
Base system — network stack · sys/netgraph7/bluetooth/socket/ng_btsocket_rfcomm.c:2409 |
none |
inconclusive |
| DF-0431 |
Medium |
5.4 |
Dead expire-scaling code: imported state expiry is always raw attacker value, enabling infinite-lifetime states
Base system — network stack · sys/net/pf/if_pfsync.c:402 |
none |
inconclusive |
| DF-0526 |
Medium |
5.4 |
ng_fec_choose_port dereferences ether/IP headers without mbuf length validation: OOB read
Kernel memory leaks / uninitialized disclosure · sys/netgraph/fec/ng_fec.c:897 |
none |
not_reproduced |
| DF-1239 |
Medium |
5.4 |
iwi_frame_intr frame->len bound too loose: up to 28-byte OOB read past mbuf cluster
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwi/if_iwi.c:1248 |
none |
not_reproduced |
| DF-2578 |
Medium |
5.4 |
OOB heap write via unchecked tdma_slot in ieee80211_parse_tdma
Kernel memory corruption (UAF / OOB write) · sys/netproto/802_11/wlan/ieee80211_tdma.c:657 |
none |
not_reproduced |
| DF-1145 |
High |
5.3 |
Unbounded firmware-supplied MPDU length in RX paths causes kernel heap OOB read
Kernel memory leaks / uninitialized disclosure · DF-1123 · sys/dev/netif/iwm/if_iwm.c:3221 |
none |
inconclusive |
| DF-0162 |
Medium |
5.3 |
Global modules TAILQ mutated without mod_token: unpriv readers race with privileged kldload/unload
Kernel memory corruption (UAF / OOB write) · sys/kern/kern_module.c:141 |
none |
not_reproduced fixed |
| DF-0176 |
Medium |
5.3 |
cttyioctl forwards ioctls to ttyvp without vnode reference (UAF race)
Kernel memory corruption (UAF / OOB write) · sys/kern/tty_tty.c:238 |
none |
not_reproduced fixed |
| DF-0271 |
Medium |
5.3 |
NULL deref in bridge_input: unchecked bridge_lookup_member_if result (race with member deletion)
Base system — network stack · sys/net/bridge/if_bridge.c:2738 |
panic |
not_reproduced |
| DF-0406 |
Medium |
5.3 |
in_delayed_cksum: unchecked m_pullup return leads to NULL pointer write panic
Base system — network stack · sys/netinet/ip_output.c:940 |
panic |
inconclusive |
| DF-0522 |
Medium |
5.3 |
thread0 root-credential fallback for socket operations (DF-0510 v1 twin)
Privilege escalation (general use) · sys/netgraph/ksocket/ng_ksocket.c:559 |
none |
not_reproduced |
| DF-0566 |
Medium |
5.3 |
ACL reassembly want is uint16_t: overshoot wraps -> L2CAP RX stall + unbounded memory growth
Base system — network stack · sys/netbt/hci_link.c:421 |
none |
not_reproduced |
| DF-0928 |
Medium |
5.3 |
UFS inode-hash has no lock; concurrent ffs_vget() inserts orphan an inode and panic on reclaim (dual-vnode on production)
Base system — filesystems · sys/vfs/ufs/ufs_ihash.c:78 |
none |
not_reproduced |
| DF-1070 |
Medium |
5.3 |
Teardown race in scsi_targ_bh: fixed hz/2 sleep with no wakeup allows CTIO completion after softc kfree (UAF)
Kernel memory corruption (UAF / OOB write) · DF-1045 · sys/bus/cam/scsi/scsi_targ_bh.c:342 |
none |
not_reproduced |
| DF-1231 |
Medium |
5.3 |
Use-after-free race in aac_getnext_aif: AIF context list traversed and dereferenced without aac_aifq_lock
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/aac/aac.c:3519 |
none |
not_reproduced |
| DF-1254 |
Medium |
5.3 |
OD8 capability/settings arrays indexed by constants but allocated to VBIOS-supplied length -> heap OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/vega20_hwmgr.c:894 |
none |
inconclusive |
| DF-1371 |
Medium |
5.3 |
oce_rx/oce_discard_rx_comp loop past empty RQ ring on firmware-misreported fragment count, NULL-deref panic
Base system — drivers · sys/dev/netif/oce/oce_if.c:1274 |
none |
inconclusive |
| DF-1484 |
Medium |
5.3 |
Missing size validation for AMDGPU_CHUNK_ID_IB allows OOB read of drm_amdgpu_cs_chunk_ib fields
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_cs.c:166 |
none |
inconclusive |
| DF-1492 |
Medium |
5.3 |
Missing minimum COUNT validation in r300_packet3_check allows OOB read/write past IB buffer
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/r300.c:1176 |
none |
inconclusive |
| DF-1761 |
Medium |
5.3 |
isc_sm: remote memory leak / pdu-pool exhaustion via TASK_RSP/default cases never freeing received pduq
Legacy / ancient (PPP, iSCSI, NTFS, SMB…) — deferred · sys/dev/disk/iscsi/initiator/isc_sm.c:451 |
none |
inconclusive |
| DF-1909 |
Medium |
5.3 |
acp_hw_fini frees genpd without pm_genpd_remove: dangling UAF in global genpd list
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/amdgpu_acp.c:217 |
none |
inconclusive |
| DF-1953 |
Medium |
5.3 |
Unmasked ioread32(vaddr+head+4) reads 4 bytes past ring-buffer mapping in semaphore_waits_for
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_hangcheck.c:103 |
none |
inconclusive |
| DF-2271 |
Medium |
5.3 |
Device-controlled avail in chgetelemstatus walks desc past data buffer -> kernel heap info leak via copyout
Kernel memory leaks / uninitialized disclosure · sys/bus/cam/scsi/scsi_ch.c:1159 |
none |
not_reproduced |
| DF-2589 |
Medium |
5.3 |
Unchecked AcpiEvaluateObject return in acpi_panasonic_sinf causes NULL-deref panic and heap OOB read
Kernel memory leaks / uninitialized disclosure · sys/dev/acpica/acpi_panasonic/acpi_panasonic.c:277 |
none |
not_reproduced |
| DF-0365 |
Medium |
5.2 |
ieee80211_media_setup unbounded rate-set merge overflows stack buffer rs_rates[15]: 17 unique rates across 11a/b/g/HALF/QUARTER
Kernel memory corruption (UAF / OOB write) · disputed · 👤 orchestrator · sys/netproto/802_11/wlan/ieee80211.c:1564 |
none |
not_reproduced |
| DF-0633 |
Medium |
5.2 |
ip_fw3_ctl_state_get reads every other CPU RB tree without synchronization
Kernel memory corruption (UAF / OOB write) · sys/net/ipfw3_basic/ip_fw3_state.c:422 |
none |
not_reproduced |
| DF-0683 |
Medium |
5.2 |
hci_disable modifies queues and link list without hci_devlock, racing with hci_intr (panic / mbuf UAF)
Kernel memory corruption (UAF / OOB write) · sys/netbt/hci_unit.c:250 |
none |
not_reproduced |
| DF-0689 |
Medium |
5.2 |
Unsynchronized encaptab list yields use-after-free in encap4_input/encap6_input dispatch
Kernel memory corruption (UAF / OOB write) · sys/netinet/ip_encap.c:174 |
none |
not_reproduced |
| DF-1217 |
Medium |
5.2 |
Unprivileged users get raw I2C-bus, EEPROM-write, and GPIO control over the capture card
Privilege escalation (general use) · sys/dev/video/bktr/bktr_core.c:2058 |
none |
inconclusive |
| DF-1415 |
Medium |
5.2 |
Unchecked gpio pin count from u16 structuresize indexes fixed gpio_pin[8] array
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/bios/bios_parser2.c:353 |
none |
inconclusive |
| DF-1498 |
Medium |
5.2 |
Unvalidated VBIOS USHORT offsets and UCHAR counts in atomctrl_get_leakage_vddc_base_on_leakage allow OOB read of BIOS image
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:1498 |
none |
inconclusive |
| DF-1804 |
Medium |
5.2 |
snp: UAF / infinite-loop race between snpclose and concurrent snplwrite/snp_in on snooped tty
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/snp/snp.c:484 |
none |
inconclusive |
| DF-1874 |
Medium |
5.2 |
Signed int32 (daddr_t) overflow in block/offset math corrupts data.data pointer for >1 TiB SDXC cards
Base system — drivers · sys/dev/disk/mmcsd/mmcsd.c:338 |
none |
inconclusive |
| DF-2219 |
Medium |
5.2 |
Unbounded array index in encode_pcie_lane_width() allows OOB .rodata read via crafted GPU VBIOS
Privilege escalation (general use) · sys/dev/drm/amd/powerplay/hwmgr/pppcielanes.c:56 |
none |
not_reproduced |
| DF-1252 |
Medium |
5.1 |
Unvalidated 16-bit BIOS offsets dereferenced in LVDS mode-table and GPIO voltage-table parsing (OOB read up to 64KB)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/radeon_combios.c:1246 |
none |
inconclusive |
| DF-1386 |
Medium |
5.1 |
Divide-by-zero panic in cypress_calculate_burst_time when memory_clock<=1
Base system — drivers · sys/dev/drm/radeon/cypress_dpm.c:907 |
none |
inconclusive |
| DF-1530 |
Medium |
5.1 |
hpt_get_channel_info uses caller-controlled bus with no bounds check (OOB read + kernel pointer leak / panic)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/hptmv/gui_lib.c:446 |
none |
inconclusive |
| DF-1540 |
Medium |
5.1 |
Use-after-free read of dangling sch->hook in NGM_TEXT_STATUS after hook disconnect
Kernel memory corruption (UAF / OOB write) · sys/dev/netif/mn/if_mn.c:474 |
none |
inconclusive |
| DF-1802 |
Medium |
5.1 |
snp: SNPSTTY with NOUDEV drops snp_token before snp_down/snp_detach -> ASSERT_LWKT_TOKEN_HELD panic or unlocked race
Base system — drivers · sys/dev/misc/snp/snp.c:540 |
none |
inconclusive |
| DF-1978 |
Medium |
5.1 |
queue_delayed_work silently ignores caller workqueue always runs on system_wq breaking ordered/highpri semantics
Base system — drivers · sys/dev/drm/linux_workqueue.c:151 |
none |
inconclusive |
| DF-1989 |
Medium |
5.1 |
No BAR bounds validation on TPM-reported buffer offsets and sizes
Kernel memory corruption (UAF / OOB write) · sys/dev/crypto/tpm/tpm_crb.c:187 |
none |
inconclusive |
| DF-2027 |
Medium |
5.1 |
All VBIOS table parsers dereference attacker-controlled offsets without bounds validation -- OOB heap read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_atomfirmware.c:42 |
none |
inconclusive |
| DF-2410 |
Medium |
5.1 |
TPM-reported CRB buffer offsets/sizes used as BAR offsets without any BAR bounds check (and off-by-one in offset normalization)
Kernel memory corruption (UAF / OOB write) · sys/dev/crypto/tpm/tpm_crb.c:188 |
none |
not_reproduced |
| DF-0008 |
Low |
5.1 |
vfs_setpublicfs() use-after-vput of root vnode + refcount leak on VFS_VPTOFH error
Kernel memory corruption (UAF / OOB write) · sys/kern/vfs_subr.c:2255 |
none |
not_reproduced |
| DF-0610 |
Low |
5.1 |
Heap OOB read in NGM_NAT_PROXY_RULE via non-NUL-terminated user string
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/ng_nat.c:643 |
none |
not_reproduced |
| DF-0726 |
Low |
5.1 |
if_cloners list and if_cloners_count accessed without synchronization — unprivileged SIOCIFGCLONERS races module load/unload UAF
Kernel memory corruption (UAF / OOB write) · sys/net/if_clone.c:43 |
none |
not_reproduced fixed |
| DF-2303 |
Low |
5.1 |
Latent negative-length bcopy in fwohci_arcv split-header reassembly can cause kernel panic
Kernel memory corruption (UAF / OOB write) · sys/bus/firewire/fwohci.c:2729 |
corruption |
not_reproduced fixed |
| DF-0106 |
Medium |
5.0 |
dkcksum32 OOB read via crafted disklabel in writedisklabel path
Kernel memory corruption (UAF / OOB write) · sys/kern/subr_disklabel32.c:358 |
no observable runtime panic/leak on the writedisklabel path (bug real in source); sibling DF-0107 demonstrates the shared root-cause panic |
not_reproduced |
| DF-0484 |
Medium |
4.8 |
SYN-cookie crypto state global unsynchronized across netisr CPUs: racy MD5_CTX + tcp_secret[] defeats SYN-flood mitigation
Base system — network stack · sys/netinet/tcp_syncache.c:1351 |
none |
not_reproduced |
| DF-2165 |
High |
4.7 |
shem_read_mapping_page() leaks VM_OBJECT_LOCK and hold_count on every error path
Base system — drivers · sys/dev/drm/linux_shmem.c:52 |
none |
not_reproduced |
| DF-2166 |
High |
4.7 |
pagecache_write_begin() stores shmem ERR_PTR in *pagep and returns success, causing kmap panic
Base system — drivers · sys/dev/drm/linux_shmem.c:102 |
none |
not_reproduced |
| DF-0014 |
Medium |
4.7 |
enterpgrp() lwkt_reltoken on an un-acquired token -> race-triggered kernel panic
Base system — kernel core (kern) · sys/kern/kern_proc.c:763 |
none |
not_reproduced |
| DF-0039 |
Medium |
4.7 |
ptsopen check-then-use TOCTOU on dev->si_drv1 -> NULL-deref kernel panic (local DoS)
Base system — kernel core (kern) · sys/kern/tty_pty.c:313 |
none |
not_reproduced |
| DF-0047 |
Medium |
4.7 |
mtx_wait_link lock-leak race: chain can grant lock during mtx_delete_link window, caller returns error despite holding the lock (permanent deadlock)
Base system — kernel core (kern) · sys/kern/kern_mutex.c:1002 |
none |
not_reproduced |
| DF-0325 |
Medium |
4.7 |
Deadlock: callout_stop under pcb_lock while timeout callback requires pcb_lock
Base system — network stack · sys/netgraph7/bluetooth/socket/ng_btsocket_l2cap.c:2765 |
none |
inconclusive |
| DF-0366 |
Medium |
4.7 |
Lockless sc_count pre-check in lagg_start races with port destroy causing divide-by-zero panic
Base system — network stack · sys/net/lagg/if_lagg.c:1758 |
none |
inconclusive |
| DF-0696 |
Medium |
4.7 |
Missing return after NG_FREE_DATA in ng_etf_rcvdata turns discard path into NULL-deref panic
Kernel memory corruption (UAF / OOB write) · sys/netgraph/etf/ng_etf.c:381 |
none |
not_reproduced |
| DF-0813 |
Medium |
4.7 |
NULL deref in hammer2_flush() retry loop when chain->parent becomes NULL during concurrent flush — missing NULL guards present in setup/teardown
Hammer2 · sys/vfs/hammer2/hammer2_flush.c:397 |
none |
not_reproduced |
| DF-0847 |
Medium |
4.7 |
Global dquot hash/free-list manipulated with no lock - concurrent ufs_dqget cache-miss corrupts lists and panics
Base system — filesystems · sys/vfs/ufs/ufs_quota.c:769 |
none |
not_reproduced |
| DF-0955 |
Medium |
4.7 |
Race-induced kernel panic in vm_contig_pg_alloc when first iteration of alloc loop fails (vm_contig_pg_free called with size=0)
Base system — virtual memory · sys/vm/vm_contig.c:398 |
none |
not_reproduced |
| DF-1035 |
Medium |
4.7 |
bus_dmamap_load_uio uses stale absolute segment index after advancing segs pointer -> KKASSERT panic or OOB write
Kernel memory corruption (UAF / OOB write) · sys/platform/pc64/x86_64/busdma_machdep.c:961 |
none |
not_reproduced |
| DF-1151 |
Medium |
4.7 |
Divide-by-zero DoS in ci_thermal_setup_fan_table from unvalidated VBIOS fan temperatures
Base system — drivers · DF-1129 · sys/dev/drm/radeon/ci_dpm.c:1018 |
none |
not_reproduced |
| DF-1391 |
Medium |
4.7 |
VBIOS power-table parsing dereferences attacker-controlled offsets with no bounds checks
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/radeon/rv6xx_dpm.c:1889 |
none |
inconclusive |
| DF-1424 |
Medium |
4.7 |
TOCTOU race: interrupt completion re-reads unchecked user buffer sizes from ioctl ucmd
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/iir/iir.c:1600 |
none |
inconclusive |
| DF-1527 |
Medium |
4.7 |
i915_request_skip reads uninitialized rq->postfix, causing ring-buffer memset with stale slab offset
Base system — drivers · sys/dev/drm/i915/i915_request.c:1019 |
none |
inconclusive |
| DF-1581 |
Medium |
4.7 |
Off-by-one SMID bounds check in mfi_tbolt_complete_cmd allows OOB read of mfi_cmd_pool_tbolt[]
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mfi/mfi_tbolt.c:666 |
none |
inconclusive |
| DF-1796 |
Medium |
4.7 |
ar2413: missing numChannels>0 guard in getGainBoundariesAndPdadcsForPowers -> deterministic kernel OOB read panic (mirrors DF-1766/DF-1789)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5212/ar2413.c:353 |
none |
inconclusive |
| DF-1806 |
Medium |
4.7 |
ar2317: missing numChannels>0 guard in getGainBoundariesAndPdadcsForPowers -> OOB read panic (mirrors DF-1766/1789/1796)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5212/ar2317.c:363 |
none |
inconclusive |
| DF-1815 |
Medium |
4.7 |
ppatomfwctrl: voltage-object VBIOS walk uses attacker-controlled structuresize/object_size -> OOB read + infinite-loop DoS
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/ppatomfwctrl.c:33 |
none |
inconclusive |
| DF-1834 |
Medium |
4.7 |
NULL pointer deref in drm_sched_job_timedout first loop (missing parent check)
Base system — drivers · sys/dev/drm/scheduler/sched_main.c:259 |
none |
inconclusive |
| DF-1949 |
Medium |
4.7 |
ASSERT-only divide-by-zero guard in fixed-point division panics kernel in production builds
Base system — drivers · sys/dev/drm/amd/display/dc/basics/fixpt31_32.c:43 |
none |
inconclusive |
| DF-1974 |
Medium |
4.7 |
NULL pointer dereference in radeon_dig_monitor_is_duallink when encoder has no linked connector
Base system — drivers · sys/dev/drm/radeon/radeon_encoders.c:368 |
none |
inconclusive |
| DF-1985 |
Medium |
4.7 |
u32 to uint16 truncation of pitch into fb_info.stride enables syscons mmap SIZE_MAX bound-check bypass (kernel memory read primitive)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_fb.c:261 |
none |
inconclusive |
| DF-2025 |
Medium |
4.7 |
Missing upper-bound on BMC-supplied response byte count causes kernel stack OOB read into userspace reply
Kernel memory leaks / uninitialized disclosure · sys/dev/misc/ipmi/ipmi_ssif.c:183 |
none |
inconclusive |
| DF-2026 |
Medium |
4.7 |
Multi-part SSIF response read loop has no iteration cap -- infinite-loop DoS via malicious BMC
Kernel memory corruption (UAF / OOB write) · sys/dev/misc/ipmi/ipmi_ssif.c:255 |
none |
inconclusive |
| DF-2509 |
Medium |
4.7 |
ast_reinit frees softc/NULLs ivars; ast_detach then NULL-derefs (same class as DF-2507 never fixed here)
Base system — drivers · sys/dev/disk/nata/atapi-tape.c:202 |
none |
not_reproduced |
| DF-0800 |
Low |
4.7 |
Missing NULL check on hammer2_inode_chain in truncation reset path — panic + %s OOB read on crafted image filename
Hammer2 · sys/vfs/hammer2/hammer2_xops.c:1600 |
none |
not_reproduced fixed |
| DF-0818 |
Low |
4.7 |
Integer divide-by-zero in bulkfree_pass progress computation for total_size<10000 (crafted image volu_size=0)
Hammer2 · sys/vfs/hammer2/hammer2_bulkfree.c:727 |
none |
not_reproduced fixed |
| DF-0935 |
Low |
4.7 |
Stack buffer overflow via unbounded ksprintf into psbuf[512] in /proc/<pid>/rlimit
Kernel memory corruption (UAF / OOB write) · sys/vfs/procfs/procfs_rlimit.c:64 |
none |
not_reproduced |
| DF-2299 |
Low |
4.7 |
Missing NULL check on kmalloc return in fwdma_malloc_multiseg
Base system — bus / platform / cpu · sys/bus/firewire/fwdma.c:162 |
none |
not_reproduced |
| DF-2396 |
Low |
4.7 |
UAF: sysctl sensor read races kfree of sensor arrays in detach
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/aibs/atk0110.c:268 |
corruption |
not_reproduced |
| DF-2208 |
Info |
4.7 |
Missing NULL checks on kzalloc/kasprintf returns in drm_sysfs_connector_add cause NULL-deref panic under OOM
Base system — drivers · sys/dev/drm/drm_sysfs.c:39 |
none |
not_reproduced fixed |
| DF-1017 |
High |
4.6 |
Untrusted READ_CAPACITY block_len=0 causes kernel divide-by-zero panic (secsize used as divisor without validation)
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_da.c:2280 |
none |
inconclusive |
| DF-0621 |
Medium |
4.6 |
Privilege bypass: rip6_output uses cr_uid==0 instead of caps_priv_check for per-packet IPv6 options
Privilege escalation (general use) · sys/netinet6/raw_ip6.c:297 |
none |
not_reproduced |
| DF-0990 |
Medium |
4.6 |
PCI capability-list walk has no cycle detection, malicious device hangs kernel forever
Base system — bus / platform / cpu · sys/bus/pci/pci.c:847 |
none |
not_reproduced |
| DF-1023 |
Medium |
4.6 |
Kernel panic on unrecognized SD CSD version from malicious SD card (csd_structure=2 or 3 triggers panic not graceful error)
Base system — bus / platform / cpu · sys/bus/mmc/mmc.c:1056 |
none |
not_reproduced |
| DF-1038 |
Medium |
4.6 |
NULL pointer dereference in wsp_intr_callback when ntouch==0 (malicious USB / TYPE1 regression)
Base system — bus / platform / cpu · sys/bus/u4b/input/wsp.c:845 |
none |
not_reproduced |
| DF-1040 |
Medium |
4.6 |
Unvalidated 32-bit CIS longlink/MFC target address causes wild bus_space_read_1 and kernel panic
Kernel memory corruption (UAF / OOB write) · sys/bus/pccard/pccard_cis.c:386 |
panic |
inconclusive |
| DF-1093 |
Medium |
4.6 |
NULL pointer dereference in ppb_pnp_detect via crafted IEEE 1284 PnP string
Base system — bus / platform / cpu · sys/bus/ppbus/ppbconf.c:242 |
none |
not_reproduced fixed |
| DF-1106 |
Medium |
4.6 |
Unbounded XHCI extended-capability pointer walk allows OOB MMIO read/write and kernel DoS from malicious PCI controller
Kernel memory leaks / uninitialized disclosure · DF-1092 · sys/bus/u4b/controller/xhci_pci.c:398 |
none |
not_reproduced |
| DF-1240 |
Medium |
4.6 |
iwi_checkforqos unbounded notif->len allows up to 292-byte OOB read past RX mbuf on crafted ASSOC RESP
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/iwi/if_iwi.c:1368 |
none |
not_reproduced |
| DF-1269 |
Medium |
4.6 |
Missing bounds checks on VBIOS-supplied indices in kv_parse_power_table (OOB reads)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/kv_dpm.c:2745 |
none |
inconclusive |
| DF-1278 |
Medium |
4.6 |
Infinite recursion in mpt_configure_ioc due to tn++ post-increment bypasses retry limit -> stack overflow
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt.c:2630 |
none |
inconclusive |
| DF-1279 |
Medium |
4.6 |
Divide-by-zero panic if malicious IOC reports RequestFrameSize == 0
Base system — drivers · sys/dev/disk/mpt/mpt.c:2654 |
none |
inconclusive |
| DF-1414 |
Medium |
4.6 |
Unchecked VBIOS number_of_path (u8) indexes fixed display_path[8] at 6 call sites
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/dc/bios/bios_parser2.c:162 |
none |
inconclusive |
| DF-1452 |
Medium |
4.6 |
ae_rxeof lacks upper-bound check on NIC-controlled rxd->len allowing OOB heap read via m_devget
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ae/if_ae.c:609 |
none |
inconclusive |
| DF-1463 |
Medium |
4.6 |
Divide-by-zero panic in XPT_CALC_GEOMETRY from block_size==0
Base system — drivers · sys/dev/raid/dpt/dpt_scsi.c:1025 |
none |
inconclusive |
| DF-1809 |
Medium |
4.6 |
ar2425: missing numChannels>0 guard in getGainBoundariesAndPdadcsForPowers -> OOB read panic (mirrors DF-1766/1789/1796/1806)
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ath/ath_hal/ar5212/ar2425.c:264 |
none |
inconclusive |
| DF-1923 |
Medium |
4.6 |
Detach loop indexes mem_res[] by slot position rather than slot->num: NULL-deref panic + resource leak after partial slot-init failure
Base system — drivers · sys/dev/disk/sdhci/sdhci_pci.c:350 |
none |
inconclusive |
| DF-1940 |
Medium |
4.6 |
Divide-by-zero in ata_intel_31244_tf_write via crafted IDENTIFY CHS geometry (heads=0 or sectors=0)
Base system — drivers · sys/dev/disk/nata/chipsets/ata-intel.c:471 |
none |
inconclusive |
| DF-1966 |
Medium |
4.6 |
NULL pointer dereference in dm_dp_mst_get_modes via unchecked dc_link_add_remote_sink return
Base system — drivers · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c:237 |
none |
inconclusive |
| DF-2338 |
Medium |
4.6 |
NULL-deref in wsp_intr_callback when first HID report yields ntouch=0 (always for TYPE1 devices)
Base system — bus / platform / cpu · sys/bus/u4b/input/wsp.c:884 |
none |
not_reproduced |
| DF-2415 |
Medium |
4.6 |
Divide-by-zero panic in XPT_CALC_GEOMETRY from device-controlled block_size
Base system — drivers · sys/dev/disk/advansys/adwcam.c:734 |
none |
not_reproduced |
| DF-0510 |
Medium |
4.5 |
Credential bypass via thread0 fallback in socket operations: root creds for all ksocket ops
Privilege escalation (general use) · sys/netgraph7/ksocket/ng_ksocket.c:546 |
none |
not_reproduced |
| DF-2272 |
Info |
4.5 |
TOCTOU on user-supplied chio.h ioctl structs enables re-validation bypass in move/exchange/position/setvoltag
Base system — bus / platform / cpu · sys/bus/cam/scsi/scsi_ch.c:805 |
none |
not_reproduced |
| DF-2571 |
High |
4.4 |
Kernel stack overflow via user-controlled VLA in ngdread() and ngdwrite()
Kernel memory corruption (UAF / OOB write) · sys/netgraph/ng_device.c:509 |
none |
not_reproduced |
| DF-0234 |
Medium |
4.4 |
Signed-integer overflow in callout timer calc (min_period*hz/2) -> self-perpetuating CPU-burn or watchdog defeat
Base system — kernel core (kern) · sys/kern/kern_wdog.c:94 |
dos |
not_reproduced |
| DF-0236 |
Medium |
4.4 |
Driver callbacks (wdog_fn) invoked under global spinlock with interrupts disabled
Base system — kernel core (kern) · sys/kern/kern_wdog.c:84 |
dos |
not_reproduced |
| DF-0411 |
Medium |
4.4 |
Infinite loop in ng_parse_skip_value on unclosed quoted string inside brackets: kernel thread hang DoS
Base system — network stack · sys/netgraph7/netgraph/ng_parse.c:1651 |
none |
not_reproduced |
| DF-0436 |
Medium |
4.4 |
Heap OOB read + security-filter bypass via OGF=0/event=0: negative bitstr index into ng_btsocket_hci_raw_sec_filter
Privilege escalation (general use) · sys/netgraph7/bluetooth/socket/ng_btsocket_hci_raw.c:669 |
none |
not_reproduced |
| DF-0495 |
Medium |
4.4 |
rn_delete integer underflow: klen-head_off wraps to huge size_t when key sa_len < tree offset -> kernel panic
Kernel memory leaks / uninitialized disclosure · sys/net/radix.c:884 |
none |
not_reproduced |
| DF-0502 |
Medium |
4.4 |
ng_fec_shutdown infinite loop when member interface destroyed: dangling pointer + unkillable loop
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ng_fec.c:1335 |
dos |
inconclusive |
| DF-0527 |
Medium |
4.4 |
ng_fec_choose_port NULL dereference when computed port index removed: reachable panic after port delete
Base system — network stack · sys/netgraph/fec/ng_fec.c:896 |
none |
not_reproduced |
| DF-0528 |
Medium |
4.4 |
ng_fec_rmnode infinite loop when member interface vanished: dangling ptr + unkillable loop
Kernel memory corruption (UAF / OOB write) · sys/netgraph/fec/ng_fec.c:1224 |
none |
not_reproduced |
| DF-0535 |
Medium |
4.4 |
Integer underflow in ngc_send path-length math: sg_len<2 -> bcopy with SIZE_MAX -> kernel heap smash
Base system — network stack · sys/netgraph7/socket/ng_socket.c:245 |
none |
not_reproduced |
| DF-0622 |
Medium |
4.4 |
NULL-pointer dereference panic: ng_car_disconnect does not cancel the pending SHAPE callout, ng_car_q_event dereferences purged queue
Base system — network stack · sys/netgraph7/ng_car.c:553 |
none |
not_reproduced |
| DF-0666 |
Medium |
4.4 |
NULL mbuf deref/panic on m_devget failure in ngdwrite
Base system — network stack · sys/netgraph/ng_device.c:589 |
none |
not_reproduced |
| DF-1095 |
Medium |
4.4 |
iicsmb_bwrite leaves I2C bus in STARTED state on write failure — permanent SMBus bridge lockup
Base system — bus / platform / cpu · sys/bus/iicbus/iicsmb.c:462 |
none |
not_reproduced fixed |
| DF-1204 |
Medium |
4.4 |
Divide-by-zero in spread spectrum clock calculation via ss.rate == 0
Base system — drivers · sys/dev/drm/radeon/ni_dpm.c:2044 |
none |
not_reproduced |
| DF-1234 |
Medium |
4.4 |
Kernel heap info leak via u_int16_t truncation in ASR_fillMessage on user-controlled reply size
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/asr/asr.c:486 |
none |
not_reproduced |
| DF-1364 |
Medium |
4.4 |
Unchecked fd_formb_nsecs in FD_FORMAT causes OOB heap read via DMA/bcopy
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/fd/fd.c:2272 |
none |
inconclusive |
| DF-1423 |
Medium |
4.4 |
Integer overflow in raw-service ioctl bounds check bypasses scratch-size guard
Kernel memory corruption (UAF / OOB write) · sys/dev/raid/iir/iir.c:1036 |
none |
inconclusive |
| DF-1474 |
Medium |
4.4 |
Topology/IR-config/PCIe-config change list loops trust firmware NumEntries/NumElements without bounding to allocation (heap OOB read)
Kernel memory leaks / uninitialized disclosure · sys/dev/raid/mpr/mpr_sas_lsi.c:216 |
none |
inconclusive |
| DF-1480 |
Medium |
4.4 |
Signed integer underflow in P-state sysctl readers leaks kernel stack to unprivileged users
Kernel memory leaks / uninitialized disclosure · sys/dev/acpica/acpi_cpu_pstate.c:1255 |
none |
inconclusive |
| DF-1516 |
Medium |
4.4 |
Integer overflow in buffer bounds check bypasses offset validation in check_overlay_src
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/i915/intel_overlay.c:1033 |
none |
inconclusive |
| DF-1798 |
Medium |
4.4 |
smu_helper: heap OOB read in phm_get_sclk_for_voltage_evv via unvalidated vddInd from VBIOS
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/powerplay/hwmgr/smu_helper.c:468 |
none |
inconclusive |
| DF-1801 |
Medium |
4.4 |
snp: NULL-deref panic when read/kqfilter issued on unattached /dev/snpN device
Base system — drivers · sys/dev/misc/snp/snp.c:263 |
none |
inconclusive |
| DF-1965 |
Medium |
4.4 |
Unconditional _objcache_destroy in MOD_UNLOAD leaves module obj_cache NULL while still in use
Base system — drivers · sys/dev/disk/dm/delay/dm_target_delay.c:440 |
none |
inconclusive |
| DF-2333 |
Medium |
4.4 |
UAC2.0 Feature Unit descriptor accepted too short causing OOB read of bmaControls[0]
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/audio/uaudio.c:4000 |
none |
not_reproduced |
| DF-0607 |
Low |
4.4 |
BPF write to ng_iface reads uninitialized sa_data, can KASSERT-panic the kernel
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/iface/ng_iface.c:428 |
none |
not_reproduced |
| DF-0648 |
Low |
4.4 |
NULL-deref panic in NGM_NETFLOW_INFO and NGM_NETFLOW_IFINFO response allocation
Base system — network stack · sys/netgraph7/netflow/ng_netflow.c:305 |
panic |
inconclusive |
| DF-0649 |
Low |
4.4 |
Per-node ~1.5 MB kernel-memory amplification via fixed NG_NETFLOW_MAXIFACES array
Base system — network stack · sys/netgraph7/netflow/ng_netflow.c:172 |
dos |
inconclusive |
| DF-2439 |
Low |
4.4 |
FREEBLKS bio falls through delay target switch and is dispatched to underlying pdev without offset translation
Base system — drivers · sys/dev/disk/dm/delay/dm_target_delay.c:213 |
none |
not_reproduced |
| DF-2455 |
Low |
4.4 |
FD_STYPE accepts unvalidated fd_type enabling divide-by-zero panics and shift UB
Base system — drivers · sys/dev/disk/fd/fd.c:2325 |
none |
not_reproduced |
| DF-2483 |
Low |
4.4 |
mpt_pci_attach returns success on every failure path causing NULL-deref / linked-list corruption on detach
Base system — drivers · sys/dev/disk/mpt/mpt_pci.c:576 |
none |
not_reproduced |
| DF-2487 |
Low |
4.4 |
Modifying ioctls (MPTIO_WRITE_CFG_PAGE MPTIO_RAID_ACTION) allowed via read-only file descriptor
Privilege escalation (general use) · sys/dev/disk/mpt/mpt_user.c:116 |
none |
not_reproduced |
| DF-2540 |
Low |
4.4 |
Signed integer overflow in amdsb_watchdog() timeout calc programs watchdog count=0 -> deterministic reboot
Base system — drivers · sys/dev/misc/amdsbwd/amdsbwd.c:217 |
none |
not_reproduced |
| DF-2308 |
Info |
4.4 |
I2CREAD ioctl leaks uninitialized kernel heap memory to userspace
Kernel memory leaks / uninitialized disclosure · sys/bus/iicbus/iic.c:343 |
none |
not_reproduced |
| DF-0540 |
Medium |
4.3 |
Uninitialized cmd.ident sent in Command Reject: remote 1-byte kernel stack info leak
Kernel memory leaks / uninitialized disclosure · sys/netbt/l2cap_signal.c:70 |
none |
not_reproduced |
| DF-1481 |
Medium |
4.2 |
vr_rxeof trusts 11-bit NIC-reported RX length as m_devget copy size without MCLBYTES bound: OOB heap read past RX mbuf cluster
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/vr/if_vr.c:1005 |
none |
inconclusive |
| DF-1514 |
Medium |
4.2 |
ste_rxeof trusts 13-bit NIC-reported FRAMELEN as m_len without MCLBYTES bound: large OOB heap read past RX mbuf cluster
Kernel memory leaks / uninitialized disclosure · sys/dev/netif/ste/if_ste.c:678 |
none |
inconclusive |
| DF-1968 |
Medium |
4.2 |
Integer overflow in radeon fbcon size math yields undersized GEM object with stale full pitch
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/radeon/radeon_fb.c:154 |
none |
inconclusive |
| DF-1392 |
Medium |
4.1 |
Divide-by-zero panic when VBIOS supplies zero SCLK or zero spread-spectrum rate
Base system — drivers · sys/dev/drm/radeon/rv6xx_dpm.c:411 |
none |
inconclusive |
| DF-1497 |
Medium |
4.1 |
Missing NULL check on voltage_object in atomctrl_get_svi2_info dereferences lookup result
Base system — drivers · sys/dev/drm/amd/powerplay/hwmgr/ppatomctrl.c:1444 |
none |
inconclusive |
| DF-1511 |
Medium |
4.1 |
TOCTOU NULL-pointer dereference of handler in vga_switcheroo_force_migd causes kernel panic
Base system — drivers · sys/dev/video/vga/vga_switcheroo.c:1231 |
none |
inconclusive |
| DF-1699 |
Medium |
4.1 |
vn: unsynchronized UAF race between vnstrategy I/O path and VNIOCDETACH/vnclear
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/vn/vn.c:260 |
none |
inconclusive |
| DF-1744 |
Medium |
4.1 |
mpt_pci: use-after-free - DMA request pool freed before interrupt handler torn down
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/mpt/mpt_pci.c:657 |
none |
inconclusive |
| DF-1821 |
Medium |
4.1 |
vinumraid5: missing plex->subdisks bounds check enables division-by-zero panic via config-race
Base system — drivers · sys/dev/raid/vinum/vinumraid5.c:142 |
none |
inconclusive |
| DF-2481 |
Medium |
4.1 |
Divide-by-zero kernel panic via unchecked IOCFACTS RequestFrameSize
Base system — drivers · sys/dev/disk/mpt/mpt.c:2683 |
none |
not_reproduced |
| DF-2607 |
Medium |
4.1 |
TOCTOU/UAF on bus->ifp between usbpf_xfertap (bus-locked) and usbpf_clone_destroy (unlocked)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/usb_pf.c:362 |
none |
not_reproduced |
| DF-0587 |
Low |
4.1 |
NULL vap deref in scan_curchan_task: scan state not re-validated after dropping IEEE80211_LOCK across ic_set_channel
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_scan_sw.c:741 |
panic |
not_reproduced |
| DF-0637 |
Low |
4.1 |
ngt_close frees sc while ngt_timeout callout can still reference it (UAF / deadlock)
Kernel memory corruption (UAF / OOB write) · sys/netgraph/tty/ng_tty.c:256 |
none |
not_reproduced |
| DF-0713 |
Low |
4.1 |
cac_timeout assigns dfs->newchan (may be NULL) to iv_des_chan — downstream hostap/adhoc/mesh dereference NULL via IEEE80211_IS_CHAN_RADAR
Base system — network stack · sys/netproto/802_11/wlan/ieee80211_dfs.c:154 |
none |
inconclusive |
| DF-1906 |
Low |
4.1 |
UAF on clockmod_dom in clockmod_dom_sysctl_select across sysctl_handle_string sleep
Kernel memory corruption (UAF / OOB write) · sys/dev/powermng/clockmod/clockmod.c:430 |
none |
not_reproduced fixed |
| DF-2228 |
Low |
4.1 |
ACPI notify-handler use-after-free in detach via DragonFly no-op AcpiOsWaitEventsComplete stub
Kernel memory corruption (UAF / OOB write) · sys/gnu/dev/misc/apple_gmux/apple-gmux.c:919 |
none |
not_reproduced |
| DF-2326 |
Low |
4.1 |
pci_devq traversal in PCIOCGETCONF is unprotected: pci_token is single-taker races pci_freecfg (UAF)
Kernel memory corruption (UAF / OOB write) · sys/bus/pci/pci_user.c:59 |
none |
not_reproduced |
| DF-2374 |
Low |
4.1 |
Use-after-free race between acpiioctl() hook dispatch and acpi_deregister_ioctl()
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi.c:3407 |
corruption |
not_reproduced |
| DF-2377 |
Low |
4.1 |
Missing cst_cx_count==0 guard causes OOB heap read in set_lowest_oncpu and NULL-deref panic in idle
Base system — drivers · sys/dev/acpica/acpi_cpu_cstate.c:504 |
panic |
not_reproduced |
| DF-2378 |
Low |
4.1 |
HT-sibling C-state copy in acpi_cst_cx_probe_cst writes another CPU live softc without synchronization
Base system — drivers · sys/dev/acpica/acpi_cpu_cstate.c:621 |
panic |
inconclusive |
| DF-2393 |
Low |
4.1 |
Use-after-free of softc in notify handler during detach (AcpiOsWaitEventsComplete is a no-op) DF-2381 class
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_toshiba/acpi_toshiba.c:273 |
panic |
not_reproduced |
| DF-2408 |
Low |
4.1 |
tpm_legacy_read unconditionally writes through a NULL count pointer panicking on every multi-chunk read of a legacy TPM
Base system — drivers · sys/dev/crypto/tpm/tpm.c:1092 |
none |
not_reproduced |
| DF-2528 |
Low |
4.1 |
Unlocked vn_softc state in vnstrategy/vnget races vnclear -> NULL deref or UAF
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/vn/vn.c:272 |
none |
not_reproduced |
| DF-2243 |
Info |
4.0 |
Early-boot RC4 stream keyed with uninitialized kernel stack because read_random() returns 0 before CSPRNG initialized
Base system — libkern / libprop / ddb · sys/libkern/arc4random.c:53 |
none |
not_reproduced |
| DF-2313 |
Info |
3.9 |
IIC_LOCK/IIC_UNLOCK are empty no-op macros - all shared state access is unsynchronized
Base system — bus / platform / cpu · sys/bus/iicbus/iic.c:60 |
— |
not_reproduced |
| DF-0514 |
Low |
3.7 |
REJECT_RST copies full 20-byte tcphdr after PULLUP_TO guaranteed only 14: stale mbuf buffer leak in RST
Kernel memory leaks / uninitialized disclosure · sys/net/ip6fw/ip6_fw.c:629 |
none |
not_reproduced |
| DF-0303 |
Medium |
3.6 |
HMAC precomputed context torn-read race between config and input paths
Base system — network stack · sys/netinet/ip_carp.c:504 |
dos |
inconclusive |
| DF-1960 |
Medium |
3.6 |
Uninitialized heap memory leaked to any local user via sysctl when BIOS PSB has no matching PST
Kernel memory leaks / uninitialized disclosure · sys/dev/powermng/powernow/powernow.c:366 |
none |
inconclusive |
| DF-0851 |
Low |
3.6 |
Stale dereference of rootp after pribp released — use-after-free style read on buffer cache recycle
Kernel memory corruption (UAF / OOB write) · sys/vfs/isofs/cd9660/cd9660_vfsops.c:435 |
none |
not_reproduced fixed |
| DF-1043 |
Low |
3.6 |
ufoma sysctl handlers access freed sc_modetable after detach (UAF)
Kernel memory corruption (UAF / OOB write) · sys/bus/u4b/serial/ufoma.c:453 |
none |
not_reproduced |
| DF-1765 |
Low |
3.6 |
amdgpu_dm_debugfs: uninitialized stack read of param[] in dp_link_settings_write / dp_phy_settings_write
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c:155 |
none |
inconclusive |
| DF-0266 |
Medium |
3.5 |
Uninitialized inflate window: kernel heap info leak via stale window data
Kernel memory leaks / uninitialized disclosure · sys/net/zlib.c:3706 |
leak |
not_reproduced |
| DF-1091 |
Medium |
3.4 |
Slave-controlled block-read count drives OOB MMIO reads and inflates returned byte count (amdsmb_bread)
Kernel memory leaks / uninitialized disclosure · DF-1076 · sys/bus/smbus/amdsmb/amdsmb.c:517 |
none |
not_reproduced fixed |
| DF-0722 |
Low |
3.4 |
Unit-number bitmap length tracked in unsigned char: truncates after ~2040 nodes causing duplicate unit allocation (LATENT — file does not compile)
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ng_sppp.c:103 |
dos |
inconclusive |
| DF-0185 |
Medium |
3.3 |
Uninitialized kernel stack memory leaked via acl_get_file/acl_get_fd
Kernel memory leaks / uninitialized disclosure · sys/kern/kern_acl.c:92 |
none |
not_reproduced fixed |
| DF-1078 |
Medium |
3.3 |
vkernel DDB backtrace leaks kernel addresses to world-readable msgbuf (DF-1077 class) + missing Xfast_syscall boundary stop
Kernel memory corruption (UAF / OOB write) · DF-1077 · sys/platform/vkernel64/x86_64/db_trace.c:280 |
none |
not_reproduced |
| DF-1712 |
Medium |
3.3 |
intel_fbdev: stolen-memory framebuffer exposed unzeroed on DragonFly (info leak)
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/i915/intel_fbdev.c:276 |
none |
inconclusive |
| DF-1982 |
Medium |
3.3 |
GEM object refcount leak on amdgpufb_create error path (VRAM buffer permanently leaked)
Base system — drivers · sys/dev/drm/amd/amdgpu/amdgpu_fb.c:308 |
none |
inconclusive |
| DF-2379 |
Medium |
3.3 |
Integer underflow in P-state sysctl readers leaks uninitialized kernel stack
Privilege escalation (general use) · sys/dev/acpica/acpi_cpu_pstate.c:1258 |
none |
not_reproduced |
| DF-0690 |
Low |
3.3 |
mbuf + item leak on debug-hook data path (unreachable free condition) in dormant template
Base system — network stack · sys/netgraph7/ng_sample.c:337 |
none |
inconclusive |
| DF-1764 |
Low |
3.3 |
amdgpu_dm_debugfs: memory leak on put_user failure in dp_link_settings_read / dp_phy_settings_read
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c:89 |
none |
inconclusive |
| DF-2217 |
Low |
3.3 |
Missing privilege check on NVMEIOCGETLOG allows unprivileged NVMe admin command execution
Privilege escalation (general use) · sys/dev/disk/nvme/nvme_ioctl.c:37 |
none |
not_reproduced |
| DF-2434 |
Low |
3.3 |
ii_index terminator allocation leaked on every CCDIOCSET/CCDIOCCLR cycle (ileave>0)
Base system — drivers · sys/dev/disk/ccd/ccd.c:629 |
none |
not_reproduced |
| DF-2581 |
Low |
3.3 |
radix_node_head leak on table_delete and on table re-create enables kernel memory exhaustion DoS
Base system — network stack · sys/net/ipfw3_basic/ip_fw3_table.c:100 |
none |
not_reproduced |
| DF-2608 |
Low |
3.3 |
Uninitialized kernel stack memory leaked to userspace via USB FIFO in packet reassembly
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/input/uep.c:200 |
none |
not_reproduced |
| DF-0724 |
Low |
3.1 |
Unsynchronized timer-vs-caller double-close race in rfcomm_dlc_close — no lock/atomic on CLOSED transition
Base system — network stack · sys/netbt/rfcomm_dlc.c:156 |
none |
not_reproduced |
| DF-2516 |
Low |
3.1 |
Missing payload length validation in sbp_recv1 allows stale/uninitialized heap data to drive control flow and leak to userspace
Kernel memory leaks / uninitialized disclosure · sys/dev/disk/sbp/sbp.c:1572 |
none |
not_reproduced |
| DF-2542 |
Low |
3.0 |
mfc entries kmalloc without M_ZERO only [0 numvifs) of mfc_ttls/mfc_flags initialized later add_vif exposes uninitialized heap bytes via forwarding decisions
Kernel memory leaks / uninitialized disclosure · sys/net/ip_mroute/ip_mroute.c:1055 |
none |
not_reproduced |
| DF-0505 |
Low |
2.9 |
ng_fec_input defined 2 params but assigned to 4-param if_input slot: calling-convention UB
Base system — network stack · sys/netgraph7/ng_fec.c:862 |
none |
not_reproduced |
| DF-0245 |
Medium |
2.5 |
Per-cpu iowbytes counter underflow via thread migration accounting break
Privilege escalation (general use) · sys/kern/kern_iosched.c:79 |
dos |
not_reproduced |
| DF-0031 |
Low |
2.5 |
pipe->open_count underflow on pipe_create partial failure leaks kernel KVA and pipe struct
Kernel memory leaks / uninitialized disclosure · sys/kern/sys_pipe.c:433 |
none |
not_reproduced |
| DF-2363 |
Low |
2.5 |
Sensitive key material left on kernel stack across multiple paths
Kernel memory leaks / uninitialized disclosure · sys/crypto/chachapoly.c:52 |
none |
not_reproduced |
| DF-2433 |
Low |
2.5 |
TOCTOU on CCDF_INITED allows concurrent CCDIOCSET to leak memory and vnodes
Kernel memory corruption (UAF / OOB write) · sys/dev/disk/ccd/ccd.c:1310 |
none |
not_reproduced |
| DF-0601 |
Info |
2.5 |
ng_source_rcvdata races unsynchronized on snd_queue (latent: file is orphaned and non-compiling)
Kernel memory corruption (UAF / OOB write) · sys/netgraph7/ng_source.c:270 |
none |
not_reproduced |
| DF-2211 |
Info |
2.5 |
Uninitialized res.clock field returned by value from amdgpu_afmt_acr()
Kernel memory leaks / uninitialized disclosure · sys/dev/drm/amd/amdgpu/amdgpu_afmt.c:90 |
— |
not_reproduced |
| DF-2218 |
Info |
2.5 |
Missing NULL check on nvme_get_admin_request return potential kernel NULL-deref panic
Base system — drivers · sys/dev/disk/nvme/nvme_ioctl.c:49 |
— |
not_reproduced |
| DF-2375 |
Info |
2.5 |
Off-by-one NUL heap write in acpi_battery_clean_str past the _BIF oeminfo field
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_battery.c:321 |
none |
not_reproduced |
| DF-2402 |
Info |
2.5 |
Partial MMIO key read returns success without zeroing unfilled buffer tail enabling uninitialized-stack leak through sysctl
Kernel memory leaks / uninitialized disclosure · sys/dev/apple/smc/smc_mmio.c:75 |
leak |
not_reproduced |
| DF-1790 |
Medium |
2.4 |
amdgpu_atombios_dp: OOB kernel-stack read in DP link training via malicious sink MAX_LANE_COUNT > 4
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/amd/amdgpu/atombios_dp.c:264 |
none |
inconclusive |
| DF-2588 |
Low |
2.4 |
Signed/unsigned comparison in RX aggregation loop permits heap OOB read of RX descriptor
Kernel memory leaks / uninitialized disclosure · sys/bus/u4b/wlan/if_urtwn.c:984 |
none |
not_reproduced |
| DF-0723 |
Low |
2.3 |
Error-path kfree uses wrong malloc type M_NETGRAPH_SPPP for M_IFNET allocation and bypasses if_free (LATENT — file does not compile)
Base system — network stack · sys/netgraph7/ng_sppp.c:261 |
dos |
inconclusive |
| DF-2400 |
Info |
2.3 |
agp_mmap off-by-one allows mapping one physical page past the AGP aperture BAR
Base system — drivers · sys/dev/agp/agp.c:860 |
— |
not_reproduced |
| DF-0530 |
Low |
1.9 |
ng_fec_free_unit increments usage counter instead of decrementing + global bitmap mutated without locks
Base system — network stack · sys/netgraph/fec/ng_fec.c:290 |
none |
not_reproduced |
| DF-2395 |
Low |
1.9 |
Heap OOB read via zero-element ACPI SIF package - no count check before o[0]
Kernel memory leaks / uninitialized disclosure · sys/dev/acpica/aibs/atk0110.c:169 |
corruption |
inconclusive |
| DF-2257 |
Info |
1.6 |
Unbounded while(1) loop in ar9300_rf_gain_cap_apply can walk past rx_gain_table[256] (defense-in-depth; not exploitable with current static data)
Vendored third-party (contrib) · sys/contrib/dev/ath/ath_hal/ar9300/ar9300_attach.c:3397 |
— |
not_reproduced |
| DF-0602 |
Info |
1.1 |
NGM_SOURCE_GET_COUNTER dereferences msg->data without arglen check (latent: file is orphaned and non-compiling)
Kernel memory leaks / uninitialized disclosure · sys/netgraph7/ng_source.c:479 |
none |
not_reproduced |
| DF-1858 |
Low |
0.8 |
Negative framebuffer offset in bmp_SetPix via RLE delta-move: OOB write below va_window
Kernel memory corruption (UAF / OOB write) · sys/dev/video/fb/bmp/splash_bmp.c:258 |
none |
not_reproduced |
| DF-2388 |
Info |
0.5 |
Detach does not drain in-flight ACPI notifies (AcpiOsWaitEventsComplete no-op on DragonFly) - DF-2381 class
Kernel memory corruption (UAF / OOB write) · sys/dev/acpica/acpi_panasonic/acpi_panasonic.c:210 |
— |
not_reproduced |
| DF-1977 |
High |
— |
destroy_workqueue is #if 0 no-op; drain_workqueue does not wait for in-flight works; driver teardown UAF + permanent memory/thread leak
Kernel memory corruption (UAF / OOB write) · sys/dev/drm/linux_workqueue.c:243 |
none |
inconclusive |