ACK: [SRU J][PATCH 0/1] Fix 5.15 af_netlink regression

Magali Lemes magali.lemes at canonical.com
Tue Sep 30 16:54:43 UTC 2025


On 9/30/25 12:02, Stefan Bader wrote:
> BugLink: https://bugs.launchpad.net/bugs/2126434
> 
> [SRU Justification]
> 
> == Impact ==
> Upstream stable v5.15.189 / CVE-2025-38465 added
>    ae8f160e7eb2 ("netlink: Fix wraparounds of sk->sk_rmem_alloc.")
> which is reported to cause soft-lockups / hangs on boot when loading auditd rules.
> 
> Message from syslogd at usw745 at Sep 26 12:45:22 ...
>   kernel:[22208.861925] watchdog: BUG: soft lockup - CPU#65 stuck for 26s! [kauditd:593]
> [22208.862003] FS: 0000000000000000(0000) GS:ffff8e7f84c40000(0000) knlGS:0000000000000000
> [22208.862004] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [22208.862005] CR2: 00007f6c313b1110 CR3: 0000006fa5c30005 CR4: 00000000007706e0
> [22208.862009] PKRU: 55555554
> [22208.862010] Call Trace:
> [22208.862012] <TASK>
> [22208.862017] ? wake_up_q+0x90/0x90
> [22208.862022] netlink_unicast+0xb0/0x290
> [22208.862024] kauditd_send_queue+0xb2/0x180
> 
> == Fix ==
> This was fixed upstream with the following commit which is part of pending v5.15.190
>    759dfc7d04ba netlink: avoid infinite retry looping in netlink_unicast()
> 
> == Testcase ==
> Creating and loading auditd rules on boot (details TBD)
> 
> == Regression Potential ==
> Anything related to netlink sockets.
> 
> Fedor Pchelkin (1):
>    netlink: avoid infinite retry looping in netlink_unicast()
> 
>   net/netlink/af_netlink.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 

Acked-by: Magali Lemes <magali.lemes at canonical.com>




More information about the kernel-team mailing list