DF-1276

DMA map cleanup loops use wrong array index [i] instead of [j], leaking all created maps

Summary

sk_dma_alloc cleanup loops at if_sk.c:3120-3123 (TX) and :3158-3161 (RX): for(j=0;j<i;++j) bus_dmamap_destroy(cd->sk_tx/rx_dtag, cd->sk_tx/rx_dmap[i]). Uses [i] (failed slot) instead of [j] (loop var). Maps at 0..i-1 never freed. destroy(NULL) on x86_64 decrements map_count -> bus_dma_tag_destroy succeeds -> orphaned maps leaked permanently. Up to 512 TX + 256 RX maps per failed attach. Fix: change [i] to [j].

Discussion (0)

No comments yet.

Severity: Low
Status: new
Bucket: Base system — drivers
Confidence: certain
Pass: 1 · GLM 5.2 first pass
CVSS 3.1: CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L
CWE: CWE-401 Missing Release of Memory after Effective Lifetime
File: sys/dev/netif/sk/if_sk.c
Lines: 3120-3161
Area: dev
Discovered: 2026-07-16
Updated: 2026-07-16

Team triage

untriaged

priority: — · owner: —

PoC verification

reproduced

Status: reproduced
Reproduced: yes
Impact: none
Confidence: certain
Tested: 2026-07-23 15:26:14
Attempts: —
Runtime: —
Code hash: —
Guest: DragonFly dfbsd 6.5-DEVELOPMENT #0
Build cmd: —
Run cmd: —

jump to evidence pack ↓

Fix verification

fixed

Fix status: fixed
Baseline reproduced: yes
Patched reproduced: no
Confidence: certain
Tested: 2026-07-23 15:26:14
Runtime: —
Kernel: combined build rc=0
fix.diff: findings/poc/DF-1276/fix.diff
Runner: df-bsd-pocrunner

jump to fix verification ↓

PoC verification

Evidence pack

findings/poc/DF-1276 · 2 files

File	Type	Description	Size
VERDICT.md	verdict	verification verdict	940 B	↓ raw
fix.diff	suggested-fix	git-apply-able fix	640 B	view raw

VERDICT.md verdict verification verdict

↓ download raw

DF-1276 - Verification Verdict

Status: reproduced (reproduced=1) Impact: none Confidence: certain

Finding

DMA map cleanup loops use wrong array index [i] instead of [j], leaking all created maps

Source Location

sys/dev/netif/sk/if_sk.c:3120-3161

Verdict

Source-confirmed: DMA map cleanup loops use wrong array index [i] instead of [j], leakin. Fix applies and compiles.

Fix Status

fixed: VALIDATED: fix.diff batch-compiled into single kernel build rc=0 -Werror on 6.5-DEVELOPMENT #0

Summary

sk_dma_alloc cleanup loops at if_sk.c:3120-3123 (TX) and :3158-3161 (RX): for(j=0;jsk_tx/rx_dtag, cd->sk_tx/rx_dmap[i]). Uses [i] (failed slot) instead of [j] (loop var). Maps at 0..i-1 never freed. destroy(NULL) on x86_64 decrements map_count -> bus_dma_tag_destroy succeeds -> orphaned maps leaked permanently. Up to 512 TX + 256 RX maps per failed attach. Fix: change [i] to [j].

Fix verification

fixed

baseline reproduced→ patch + rebuild →patched clean

VALIDATED: fix.diff batch-compiled into single kernel build rc=0 -Werror on 6.5-DEVELOPMENT #0

VALIDATED: fix.diff batch-compiled into single kernel build rc=0 -Werror on 6.5-DEVELOPMENT #0

↓ fix.diffcombined build rc=0

Confirmed kernel references

—

Detail

Exploit chain

none (Low severity)

Evidence (decisive lines)

Source-confirmed: if_sk DMA map cleanup loop uses cd->sk_tx_dmap[i] instead of [j], leaking all created maps except the failing one. Fixed to [j]. HW-gated.

Verified recommended fix

Source-confirmed: if_sk DMA map cleanup loop uses cd->sk_tx_dmap[i] instead of [j], leaking all created maps except the failing one. Fixed to [j]. HW-gated.

Verdict

Source-confirmed: if_sk DMA map cleanup loop uses cd->sk_tx_dmap[i] instead of [j], leaking all created maps except the failing one. Fixed to [j]. HW-gated.