Re: KASSERT() on signal delivery

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Sat, Apr 26, 2025 at 10:23:58AM +0200, Manuel Bouyer wrote:
> Hello,
> I just got this panic on a 10.1-STABLE web server while running
> a nginx reload:
> [ 63790.8339809] panic: kernel diagnostic assertion "((uintptr_t)(__al_ptr) & (sizeof(*(__al_ptr)) - 1)) == 0" failed: file "/local/armandeche2/netbsd-10/src/sys/kern/kern_sig.c", line 1082
> [ 63790.8539802] cpu7: Begin traceback...
> [ 63790.8539802] vpanic() at netbsd:vpanic+0x183
> [ 63790.8639801] kern_assert() at netbsd:kern_assert+0x4b
> [ 63790.8639801] kpsignal() at netbsd:kpsignal+0xe2
> [ 63790.8739801] fownsignal() at netbsd:fownsignal+0xe3
> [ 63790.8739801] sowakeup() at netbsd:sowakeup+0xe6
> [ 63790.8839799] unp_rcvd() at netbsd:unp_rcvd+0x11d
> [ 63790.8839799] soreceive() at netbsd:soreceive+0x66a
> [ 63790.8939801] do_sys_recvmsg_so() at netbsd:do_sys_recvmsg_so+0x240
> [ 63790.8939801] sys_recvmsg() at netbsd:sys_recvmsg+0xa1
> [ 63790.9039799] syscall() at netbsd:syscall+0x196
> [ 63790.9039799] --- syscall (number 27) ---
> 
> Unfortunably I can't get a core dump on this host.
> 
> Does it ring a bell to someone ?

this looks reproductible (but as fsck takes more than 2 hours I'd prefer to
avoid it)

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--