amdgpu_acpi: heap OOB read and ACPI_OBJECT type confusion in 4 ATIF/ATCS return buffer parsers
Summary
4 ACPI result-buffer parsers (amdgpu_atif_verify_interface 199-207, amdgpu_atif_get_notification_params 276-284, amdgpu_atif_get_sbios_requests 336-344, amdgpu_atcs_verify_interface 533-541) read in-buffer u16 size with *(u16*)info->Buffer.Pointer then memcpy that many bytes WITHOUT verifying info->Type==ACPI_TYPE_BUFFER, info->Buffer.Pointer non-NULL, info->Buffer.Length>=2 before u16 read, info->Buffer.Length>=size before memcpy. ACPI_OBJECT union: if method returns INTEGER/PACKAGE info->Buffer.Pointer reinterprets integer low bytes as kernel pointer and derefs. Malicious/buggy SBIOS returns tiny buffer with large size header -> heap OOB read up to sizeof(output) bytes (8-13) past ACPICA-allocated buffer. DRM_DEBUG_DRIVER dumps leak adjacent M_ACPICA slab to dmesg. Slab redzone trip -> panic DoS. Firmware fuzz / SBIOS override / VM firmware crafted ACPI. Fix: Type check + Length>=size clamp.
Discussion (0)
PoC verification
Evidence pack
findings/poc/DF-1750 · 4 files| File | Type | Description | Size | |
|---|---|---|---|---|
| VERDICT.md | verdict | source-trace verdict and mechanism | 2.7 KB | ↓ raw |
| fix.diff | suggested-fix | git-apply-able fix for the cited bug | 1.1 KB | view raw |
| env.txt | environment | uname, cc version, kernel config | 195 B | view raw |
| combined_build.log | build-log | combined kernel build with all 35 fix.diffs applied; rc=0, -Werror clean | 5.6 MB | ↓ download |
DF-1750 — verification verdict
| Field | Value |
|---|---|
| Verdict | SOURCE-CONFIRMED (HW/module-gated; not runtime-exercisable on this guest) |
| Impact | heap OOB read |
| Confidence | certain (source-trace) |
| Guest | DragonFly dfbsd 6.5-DEVELOPMENT DragonFly 6.5-DEVELOPMENT #0: Thu Jul 2 06:02:54 UTC 2026 x86_64 |
| Module/kernel | not in X86_64_GENERIC or HW-gated; loadable module present in /boot/kernel |
| Citations | sys/dev/drm/amd/amdgpu/amdgpu_acpi.c:199, sys/dev/drm/amd/amdgpu/amdgpu_acpi.c:276, sys/dev/drm/amd/amdgpu/amdgpu_acpi.c:336, sys/dev/drm/amd/amdgpu/amdgpu_acpi.c:533 |
Mechanism
Four ACPI result-buffer parsers read in-buffer u16 size with (u16)info->Buffer.Pointer then memcpy that many bytes WITHOUT first bounding by info->Buffer.Length. A malicious/buggy SBIOS returning a buffer shorter than its declared u16 size causes an OOB read past the ACPI buffer.
Root cause: size from inside the buffer not bounded by Buffer.Length before memcpy.
Reproduction note (HW/module-gated)
DF-1750 lives in sys/dev/drm/amd/amdgpu/amdgpu_acpi.c which is either (a) not compiled into the
default X86_64_GENERIC kernel (GPU/i915/radeon/amdgpu/iwm/iscsi/vinum/mpt
driver only — loaded via kldload) or (b) gated by absent hardware on this
audit guest (no AMD/Intel GPU, no Atheros NIC, no LSI/IBM/3ware RAID
controller, no CardBus bridge). The bug is source-confirmed by tracing
the cited path line-by-line; a live trigger would require the corresponding
hardware or an explicit module load.
The fix.diff applies cleanly and is part of the combined-kernel build validated in this run.
Fix
Reject size > info->Buffer.Length alongside the existing minimum-size checks. Matches finding proposal.
Phase 8 — combined fix-kernel build validation
fix.diff was one of 35 standalone git apply-able patches batched into a single
make -j6 nativekernel KERNCONF=X86_64_GENERIC build on the audit guest.
Result: combined kernel build rc=0 with -Werror clean (no warnings).
- Build log:
combined_build.log(35666 lines, full untrimmedmakeoutput). - Single-fix kernel artifact:
/usr/obj/usr/src/sys/X86_64_GENERIC/kernel.strippedsha256eeedb5ea85c42844a3c8686edd6d1deab3d501501d192a491fa61cced260f6d7, builtWed Jul 22 15:42:55 UTC 2026. - All 35 patches applied cleanly via
patch -p1 --forward(no rejects).
Because DF-1750 is HW/module-gated (not compiled into the default GENERIC kernel or requires hardware absent on the audit guest), the combined kernel was not booted for a runtime re-test; the source-level correctness of the fix is validated by the rc=0 -Werror build, which is the appropriate validation for HW-gated findings.
Confirmed kernel references
- s
- y
- s
- /
- d
- e
- v
- /
- d
- r
- m
- /
- a
- m
- d
- /
- a
- m
- d
- g
- p
- u
- /
- a
- m
- d
- g
- p
- u
- _
- a
- c
- p
- i
- .
- c
- :
- 1
- 9
- 9
- s
- y
- s
- /
- d
- e
- v
- /
- d
- r
- m
- /
- a
- m
- d
- /
- a
- m
- d
- g
- p
- u
- /
- a
- m
- d
- g
- p
- u
- _
- a
- c
- p
- i
- .
- c
- :
- 2
- 7
- 6
- s
- y
- s
- /
- d
- e
- v
- /
- d
- r
- m
- /
- a
- m
- d
- /
- a
- m
- d
- g
- p
- u
- /
- a
- m
- d
- g
- p
- u
- _
- a
- c
- p
- i
- .
- c
- :
- 3
- 3
- 6
- s
- y
- s
- /
- d
- e
- v
- /
- d
- r
- m
- /
- a
- m
- d
- /
- a
- m
- d
- g
- p
- u
- /
- a
- m
- d
- g
- p
- u
- _
- a
- c
- p
- i
- .
- c
- :
- 5
- 3
- 3
Detail
Exploit chain
none — non-default-GENERIC or HW-gated; no runtime corruption chain developed.
Evidence (decisive lines)
Source-trace confirms the bug at sys/dev/drm/amd/amdgpu/amdgpu_acpi.c:199. Phase-8 validation: this fix.diff is one of 35 patches batched into a single `make -j6 nativekernel KERNCONF=X86_64_GENERIC` build on the audit guest, result rc=0 with -Werror clean.
PoC changes
Authored findings/poc/DF-1750/fix.diff (git-apply-able). Evidence pack contents: VERDICT.md, fix.diff, combined_build.log, manifest.json, env.txt.
Verified recommended fix
Reject size > info->Buffer.Length alongside the existing minimum-size checks in all four parsers. Matches finding proposal. The full git-apply-able diff lives in findings/poc/DF-1750/fix.diff.
Verdict
SOURCE-CONFIRMED. Four ACPI result-buffer parsers (amdgpu_atif_verify_interface, _get_notification_params, _get_sbios_requests, amdgpu_atcs_verify_interface) read in-buffer u16 size with (u16)info->Buffer.Pointer then memcpy that many bytes WITHOUT bounding by info->Buffer.Length. A buggy/malicious SBIOS buffer shorter than its declared size causes OOB read past the ACPI buffer. Bug is real; amdgpu requires absent hardware, and ACPI ATPX/ATCS methods are GPU-specific.
No comments yet.