Race Condition Affecting kernel-modules-extra package, versions <0:4.18.0-553.22.1.el8_10

Note: Versions mentioned in the description apply only to the upstream kernel-modules-extra package and not the kernel-modules-extra package as distributed by Centos. See How to fix? for Centos:8 relevant fixed versions and status.

In the Linux kernel, the following vulnerability has been resolved:

af_unix: Fix garbage collector racing against connect()

Garbage collector does not take into account the risk of embryo getting enqueued during the garbage collection. If such embryo has a peer that carries SCM_RIGHTS, two consecutive passes of scan_children() may see a different set of children. Leading to an incorrectly elevated inflight count, and then a dangling pointer within the gc_inflight_list.

sockets are AF_UNIX/SOCK_STREAM S is an unconnected socket L is a listening in-flight socket bound to addr, not in fdtable V's fd will be passed via sendmsg(), gets inflight count bumped

connect(S, addr) sendmsg(S, [V]); close(V) __unix_gc()

NS = unix_create1() skb1 = sock_wmalloc(NS) L = unix_find_other(addr) unix_state_lock(L) unix_peer(S) = NS // V count=1 inflight=0

         NS = unix_peer(S)
         skb2 = sock_alloc()
        skb_queue_tail(NS, skb2[V])

    // V became in-flight
    // V count=2 inflight=1

    close(V)

    // V count=1 inflight=1
    // GC candidate condition met

                for u in gc_inflight_list:
                  if (total_refs == inflight_refs)
                    add u to gc_candidates

                // gc_candidates={L, V}

                for u in gc_candidates:
                  scan_children(u, dec_inflight)

                // embryo (skb1) was not
                // reachable from L yet, so V's
                // inflight remains unchanged

__skb_queue_tail(L, skb1) unix_state_unlock(L) for u in gc_candidates: if (u.inflight) scan_children(u, inc_inflight_move_tail)

                    // V count=1 inflight=2 (!)

If there is a GC-candidate listening socket, lock/unlock its state. This makes GC wait until the end of any ongoing connect() to that socket. After flipping the lock, a possibly SCM-laden embryo is already enqueued. And if there is another embryo coming, it can not possibly carry SCM_RIGHTS. At this point, unix_inflight() can not happen because unix_gc_lock is already taken. Inflight graph remains unaffected.

• https://access.redhat.com/security/cve/CVE-2024-26923
• https://git.kernel.org/stable/c/2e2a03787f4f0abc0072350654ab0ef3324d9db3
• https://git.kernel.org/stable/c/343c5372d5e17b306db5f8f3c895539b06e3177f
• https://git.kernel.org/stable/c/47d8ac011fe1c9251070e1bd64cb10b48193ec51
• https://git.kernel.org/stable/c/507cc232ffe53a352847893f8177d276c3b532a9
• https://git.kernel.org/stable/c/a36ae0ec2353015f0f6762e59f4c2dbc0c906423
• https://git.kernel.org/stable/c/b75722be422c276b699200de90527d01c602ea7c
• https://git.kernel.org/stable/c/dbdf7bec5c920200077d693193f989cb1513f009
• https://git.kernel.org/stable/c/e76c2678228f6aec74b305ae30c9374cc2f28a51
• https://lists.debian.org/debian-lts-announce/2024/06/msg00017.html
• https://lists.debian.org/debian-lts-announce/2024/06/msg00020.html