PT-2026-27759 · Linux · Linux

Published

2026-03-25

·

Updated

2026-03-25

·

CVE-2026-23394

None

No severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
In 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.

Related Identifiers

CVE-2026-23394

Affected Products

Linux