GHSA-P5PC-67G7-QCV2
Vulnerability from github – Published: 2026-03-25 12:30 – Updated: 2026-03-25 12:30In the Linux kernel, the following vulnerability has been resolved:
af_unix: Give up GC if MSG_PEEK intervened.
Igor Ushakov reported that GC purged the receive queue of an alive socket due to a race with MSG_PEEK with a nice repro.
This is the exact same issue previously fixed by commit cbcf01128d0a ("af_unix: fix garbage collect vs MSG_PEEK").
After GC was replaced with the current algorithm, the cited commit removed the locking dance in unix_peek_fds() and reintroduced the same issue.
The problem is that MSG_PEEK bumps a file refcount without interacting with GC.
Consider an SCC containing sk-A and sk-B, where sk-A is close()d but can be recv()ed via sk-B.
The bad thing happens if sk-A is recv()ed with MSG_PEEK from sk-B and sk-B is close()d while GC is checking unix_vertex_dead() for sk-A and sk-B.
GC thread User thread --------- ----------- unix_vertex_dead(sk-A) -> true <------. \ `------ recv(sk-B, MSG_PEEK) invalidate !! -> sk-A's file refcount : 1 -> 2
close(sk-B)
-> sk-B's file refcount : 2 -> 1
unix_vertex_dead(sk-B) -> true
Initially, sk-A's file refcount is 1 by the inflight fd in sk-B recvq. GC thinks sk-A is dead because the file refcount is the same as the number of its inflight fds.
However, sk-A's file refcount is bumped silently by MSG_PEEK, which invalidates the previous evaluation.
At this moment, sk-B's file refcount is 2; one by the open fd, and one by the inflight fd in sk-A. The subsequent close() releases one refcount by the former.
Finally, GC incorrectly concludes that both sk-A and sk-B are dead.
One option is to restore the locking dance in unix_peek_fds(), but we can resolve this more elegantly thanks to the new algorithm.
The point is that the issue does not occur without the subsequent close() and we actually do not need to synchronise MSG_PEEK with the dead SCC detection.
When the issue occurs, close() and GC touch the same file refcount. If GC sees the refcount being decremented by close(), it can just give up garbage-collecting the SCC.
Therefore, we only need to signal the race during MSG_PEEK with a proper memory barrier to make it visible to the GC.
Let's use seqcount_t to notify GC when MSG_PEEK occurs and let it defer the SCC to the next run.
This way no locking is needed on the MSG_PEEK side, and we can avoid imposing a penalty on every MSG_PEEK unnecessarily.
Note that we can retry within unix_scc_dead() if MSG_PEEK is detected, but we do not do so to avoid hung task splat from abusive MSG_PEEK calls.
{
"affected": [],
"aliases": [
"CVE-2026-23394"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-03-25T11:16:40Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\naf_unix: Give up GC if MSG_PEEK intervened.\n\nIgor Ushakov reported that GC purged the receive queue of\nan alive socket due to a race with MSG_PEEK with a nice repro.\n\nThis is the exact same issue previously fixed by commit\ncbcf01128d0a (\"af_unix: fix garbage collect vs MSG_PEEK\").\n\nAfter GC was replaced with the current algorithm, the cited\ncommit removed the locking dance in unix_peek_fds() and\nreintroduced the same issue.\n\nThe problem is that MSG_PEEK bumps a file refcount without\ninteracting with GC.\n\nConsider an SCC containing sk-A and sk-B, where sk-A is\nclose()d but can be recv()ed via sk-B.\n\nThe bad thing happens if sk-A is recv()ed with MSG_PEEK from\nsk-B and sk-B is close()d while GC is checking unix_vertex_dead()\nfor sk-A and sk-B.\n\n GC thread User thread\n --------- -----------\n unix_vertex_dead(sk-A)\n -\u003e true \u003c------.\n \\\n `------ recv(sk-B, MSG_PEEK)\n invalidate !! -\u003e sk-A\u0027s file refcount : 1 -\u003e 2\n\n close(sk-B)\n -\u003e sk-B\u0027s file refcount : 2 -\u003e 1\n unix_vertex_dead(sk-B)\n -\u003e true\n\nInitially, sk-A\u0027s file refcount is 1 by the inflight fd in sk-B\nrecvq. GC thinks sk-A is dead because the file refcount is the\nsame as the number of its inflight fds.\n\nHowever, sk-A\u0027s file refcount is bumped silently by MSG_PEEK,\nwhich invalidates the previous evaluation.\n\nAt this moment, sk-B\u0027s file refcount is 2; one by the open fd,\nand one by the inflight fd in sk-A. The subsequent close()\nreleases one refcount by the former.\n\nFinally, GC incorrectly concludes that both sk-A and sk-B are dead.\n\nOne option is to restore the locking dance in unix_peek_fds(),\nbut we can resolve this more elegantly thanks to the new algorithm.\n\nThe point is that the issue does not occur without the subsequent\nclose() and we actually do not need to synchronise MSG_PEEK with\nthe dead SCC detection.\n\nWhen the issue occurs, close() and GC touch the same file refcount.\nIf GC sees the refcount being decremented by close(), it can just\ngive up garbage-collecting the SCC.\n\nTherefore, we only need to signal the race during MSG_PEEK with\na proper memory barrier to make it visible to the GC.\n\nLet\u0027s use seqcount_t to notify GC when MSG_PEEK occurs and let\nit defer the SCC to the next run.\n\nThis way no locking is needed on the MSG_PEEK side, and we can\navoid imposing a penalty on every MSG_PEEK unnecessarily.\n\nNote that we can retry within unix_scc_dead() if MSG_PEEK is\ndetected, but we do not do so to avoid hung task splat from\nabusive MSG_PEEK calls.",
"id": "GHSA-p5pc-67g7-qcv2",
"modified": "2026-03-25T12:30:24Z",
"published": "2026-03-25T12:30:24Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-23394"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/37dd7ab332396eb8dd80b2dc7ea4b61abf767436"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/e5b31d988a41549037b8d8721a3c3cae893d8670"
}
],
"schema_version": "1.4.0",
"severity": []
}
Sightings
| Author | Source | Type | Date |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.