panic: malloc(M_WAITOK) with sleeping prohibited with main-n245383-15565e0a2177

[Peter Holm]

On Thu, Mar 11, 2021 at 12:56:02PM -0500, Mark Johnston wrote:
> On Thu, Mar 11, 2021 at 06:32:13PM +0100, Peter Holm wrote:
> > I just got this panic:
> > 
> > panic: malloc(M_WAITOK) with sleeping prohibited
> > cpuid = 0
> > time = 1615472733
> > KDB: stack backtrace:
> > db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe00e49748b0
> > vpanic() at vpanic+0x181/frame 0xfffffe00e4974900
> > panic() at panic+0x43/frame 0xfffffe00e4974960
> > malloc_dbg() at malloc_dbg+0xd4/frame 0xfffffe00e4974980
> > malloc() at malloc+0x34/frame 0xfffffe00e49749e0
> > g_mirror_event_send() at g_mirror_event_send+0x30/frame 0xfffffe00e4974a30
> > softclock_call_cc() at softclock_call_cc+0x15d/frame 0xfffffe00e4974b00
> > softclock() at softclock+0x66/frame 0xfffffe00e4974b20
> > ithread_loop() at ithread_loop+0x279/frame 0xfffffe00e4974bb0
> > fork_exit() at fork_exit+0x80/frame 0xfffffe00e4974bf0
> > fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe00e4974bf0
> > 
> > https://people.freebsd.org/~pho/stress/log/log0078.txt
> 
> Hi Peter,
> 
> Could you try the patch here? https://reviews.freebsd.org/D29223

This fixed the problem for me. I ran the problem test for an hour and
then the rest of the g_mirror tests. No problems seen.

- Peter