kern/59882: Skipping crash dump on recursive panic

[Thomas Klausner <wiz%NetBSD.org@localhost>]

>Number:         59882
>Category:       kern
>Synopsis:       Skipping crash dump on recursive panic
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 03 07:20:02 +0000 2026
>Originator:     Thomas Klausner
>Release:        NetBSD 11.99.4
>Organization:

>Environment:
	
	
Architecture: x86_64
Machine: amd64
>Description:
I've seen a panic on 11.99.4/x86_64

(hand-corrected OCR from screenshot)

panic() at netbsd:panic+0x3c
heartbeat() at netbsd: heartbeat+0x35e
hardclock() at netbsd:hardclock+0x89
Xresume_lapic_ltimer() at netbsd:resume_lapic_Itimer+Ox1e
--- interrupt ---
_kernel_lock() at netbsd:_kernel_lock+0xec
bridge_input() at netbsd:bridge_input+0x3f6
if_percpuq_softint() at netbsd:if_percpuq_softint +0x9e
softint_dispatch() at netbsd:softint_dispatch+0x112
DDB lost frame for netbsd:Xsoft intr+0x4c, trying Oxffffa8a469e7d0f0
Xsoftintr() at netbsd:softintr+0x4c
--- interrupt ---
c7e2f560cba2f564:
cpu: End traceback...
Mutex error: mutex_vector_enter,519: assertion failed: !cpu_intr_p()

lock address : ffff94713f154c40
current cpu  :                   0
current lwp  : 0xffff94798925f400
owner field  : 0xffff946a93977000 wait/spin:                 1/0

Skipping crash dump on recursive panic
panic: lock error: Mutex: mutex_vector_enter,519: assertion failed: !cpu_intr_p(): lock Oxffff94713f154c40 cpu
 0 lwp 0xffff94798925f400

cpu0: Begin traceback...
panic() at netbsd: vpanic+0x189
panic() at netbsd:panic+0x3c
lockdebug_abort() at netbsd:lockdebug_abort+0x115
mutex_vector_enter() at netbsd:mutex_vector_enter+0x41b
suspendsched() at netbsd:suspendsched+Oxd7
cpu_reboot() at netbsd:cpu_reboot+0x18e
kern_reboot() at netbsd:kern_reboot+0x5f
vpanic() at netbsd:vpanic+0x193
panic() at netbsd:panic+0x3c
heartbeat() at netbsd:heartbeat+0x35e
hardclock() at netbsd:hardclock+0x89
Xresume_lapic_ltimer() at netbsd:Xresume_lapic_ltimer+0x1e
--- interrupt ---
_kernel_lock() at netbsd:_kernel_lock+0xec
bridge_input() at netbsd:bridge_input+0x3f6
if_percpuq_softint() at netbsd:if_percpuq_softint+0x9e
softint_dispatch() at netbsd:softint_dispatch+0x112
DDB lost frame for netbsd:Xsoftintr+0x4c, trying Oxffffa8a469e7d0f0
softintr() at netbsd:softintr+0x4c
--- interrupt ---
c7e2f560cba2f564:
cpu0: End traceback...

No particularly heavy load on either network or memory, just running digikam.

riastradh@ said that crash dumps from within an interrupt handler should work
and to file this.

>How-To-Repeat:
Be unlucky, I guess?
>Fix:
Yes please.

>Unformatted: