return_bounce_pages wakes wrong map (typo: passes map instead of wait_map to add_map_callback)
Summary
return_bounce_pages at busdma_machdep.c:1311 add_map_callback(map) should be add_map_callback(wait_map). wait_map=get_map_waiting(dmat) at :1306 dequeues waiter from bounce_map_waitinglist + reserve_bounce_pages(commit=1). Passing map (the just-unloaded map) instead of wait_map (the dequeued waiter): (1) waiter I/O stalls forever (callback never scheduled, issuing thread blocks); (2) unloaded map gets spurious busdma_swi callback invoking bus_dmamap_load on stale map->buf/buflen/callback/arg. free_bounce_page :1399-1404 handles correctly (passes dequeued map). Requires bounce zone over-subscription (BUS_DMA_COULD_BOUNCE + exhausted bpages). Fix: add_map_callback(wait_map).
Discussion (0)
PoC verification
Evidence pack
findings/poc/DF-1036 · 3 files| File | Type | Description | Size | |
|---|---|---|---|---|
| fix.diff | suggested-fix | git-apply-able fix for the cited path | 341 B | view raw |
| VERDICT.md | verdict | source-confirmation narrative | 933 B | ↓ raw |
| env.txt | environment | guest uname + toolchain | 247 B | view raw |
DF-1036 source-confirmation
Verdict: REPRODUCED (source-confirmed) Impact: none Confidence: likely
Kernel ref: sys/platform/pc64/x86_64/busdma_machdep.c:1311
Mechanism
return_bounce_pages wakes wrong map (typo): add_map_callback(map) should be add_map_callback(wait_map); waiter I/O stalls forever, stale map gets spurious callback. confirmed.
Confirmation method
source-only Low-severity; confirmation by code inspection. Runtime PoC not exercised for this Low-severity item; confirmation is by code inspection against sys/.
Recommended fix
See fix.diff in this folder (git-apply-able unified diff).
Phase 8 (combined build)
This fix is part of the batched 70-finding combined patch
(../_batch70/combined_70.patch) applied to in-guest /usr/src. A single
make -j6 nativekernel KERNCONF=X86_64_GENERIC build is validated rc=0 with 0
errors under -Werror (../_batch70/fix_build.log).
Fix verification
fixedVALIDATED via combined build: fix in combined_70.patch; single make -j6 nativekernel built rc=0, 0 errors under -Werror (../_batch70/fix_build.log). Cited line corrected. Source-only -> validation = clean -Werror compile.
'>>> Kernel build for X86_64_GENERIC completed' + 'NK_DONE rc=0'; grep -cE 'error:|undefined reference' fix_build.log = 0
Confirmed kernel references
- s
- y
- s
- /
- p
- l
- a
- t
- f
- o
- r
- m
- /
- p
- c
- 6
- 4
- /
- x
- 8
- 6
- _
- 6
- 4
- /
- b
- u
- s
- d
- m
- a
- _
- m
- a
- c
- h
- d
- e
- p
- .
- c
- :
- 1
- 3
- 1
- 1
Detail
Exploit chain
none (source-only Low finding, not memory-corruption driven to runtime; no escalation chain)
Evidence (decisive lines)
baseline (with-src #0): bug at sys/platform/pc64/x86_64/busdma_machdep.c:1311. combined-70 fix kernel: NK_DONE rc=0 (0 errors, -Werror).
PoC changes
authored/validated fix.diff (findings/poc/DF-1036/fix.diff); part of combined_70 kernel build.
Verified recommended fix
See findings/poc/DF-1036/fix.diff (git-apply-able). Matches finding proposal.
Verdict
REAL: return_bounce_pages passes map instead of dequeued wait_map to add_map_callback -> waiter stalls forever + stale-map callback. confirmed.
No comments yet.