Re: kern/57644: ixl(4) may panic or hang

[Manuel Bouyer <manuel.bouyer%lip6.fr@localhost>]

The following reply was made to PR kern/57644; it has been noted by GNATS.

From: Manuel Bouyer <manuel.bouyer%lip6.fr@localhost>
To: Taylor R Campbell <riastradh%NetBSD.org@localhost>
Cc: gnats-bugs%NetBSD.org@localhost, knakahara%NetBSD.org@localhost, yamaguchi%NetBSD.org@localhost,
        rin%NetBSD.org@localhost
Subject: Re: kern/57644: ixl(4) may panic or hang
Date: Fri, 6 Oct 2023 14:21:39 +0200

 On Fri, Oct 06, 2023 at 09:43:27AM +0000, Taylor R Campbell wrote:
 > bouyer: Can you try a current kernel and reproduce the hang?  This
 > should trigger a heartbeat panic in current, which should give us more
 > diagnostic information.
 
 With -current, I got:
 [  33.2819124] kern.module.path=/stand/amd64/10.99.9/modules
 [  34.2019101] Mutex error: mutex_vector_enter,517: assertion failed: !cpu_intr_
 p()
 
 [  34.2019101] lock address : ffff85001f2ce090
 [  34.2019101] current cpu  :                  0
 [  34.2019101] current lwp  : 0xffff82d575583040
 [  34.2019101] owner field  : 0xffff82d21876ab80 wait/spin:                1/0
 
 [  34.2019101] panic: lock error: Mutex: mutex_vector_enter,517: assertion faile
 d: !cpu_intr_p(): lock 0xffff85001f2ce090 cpu 0 lwp 0xffff82d575583040
 [  34.2019101] cpu0: Begin traceback...
 [  34.2019101] vpanic() at netbsd:vpanic+0x173
 [  34.2019101] panic() at netbsd:panic+0x3c
 [  34.2019101] lockdebug_abort() at netbsd:lockdebug_abort+0x114
 [  34.2119104] mutex_vector_enter() at netbsd:mutex_vector_enter+0x3ef
 [  34.2119104] ixl_link_state_update() at netbsd:ixl_link_state_update+0x24
 [  34.2119104] ixl_get_link_status_done() at netbsd:ixl_get_link_status_done+0x4
 9
 [  34.2119104] ixl_atq_done_locked() at netbsd:ixl_atq_done_locked+0x74
 [  34.2119104] ixl_other_intr() at netbsd:ixl_other_intr+0x2cc
 [  34.2119104] intr_kdtrace_wrapper() at netbsd:intr_kdtrace_wrapper+0x26
 [  34.2119104] Xhandle_ioapic_edge18() at netbsd:Xhandle_ioapic_edge18+0x75
 [  34.2119104] --- interrupt ---
 [  34.2119104] x86_mwait() at netbsd:x86_mwait+0xd
 [  34.2119104] acpicpu_cstate_idle() at netbsd:acpicpu_cstate_idle+0x19a
 [  34.2119104] idle_loop() at netbsd:idle_loop+0x114
 [  34.2119104] cpu0: End traceback...
 
 [  34.2119104] dumping to dev 168,2 (offset=8, size=4122263):
 [  34.2119104] dump WARNING: lwp 0 (system swapper) flags 0x20020080: timecounte
 r went backwards from (34 + 0x7494c0960ce053a8/2^64) sec to (34 + 0x7494788dfc2b
 6cf0/2^64) sec in netbsd:mi_switch+0x112
 crash> 
 
 but no hang
 
 -- 
 Manuel Bouyer, LIP6, Sorbonne Université.           Manuel.Bouyer%lip6.fr@localhost
      NetBSD: 26 ans d'experience feront toujours la difference
 --