> On 28 Apr 2023, at 15:57, Vitaliy Makkoveev <m...@openbsd.org> wrote:
>
> On Fri, Apr 28, 2023 at 02:13:15PM +0200, Alexander Bluhm wrote:
>> After running stress test successfully with this diff, next day
>> machine crashed while compiling a new kernel. It is unclear whether
>> it is related to the diff. The softdep in ps is problably processing
>> make output via ssh. Looks like recursive kernel stack overflow.
>>
>> [-- MARK -- Fri Apr 28 13:25:00 2023]
>> kernel: protection fault trap, code=0
>> Stopped at rt_if_linkstate_change+0x21: movl 0x58(%rdi),%eax
>> ddb{3}>
>>
>> 8a0) at art_table_walk+0x26c
>> art_table_walk(ffff800000390900,fffffd8746decdc0,ffffffff813bde40,ffff8000247c9
>> 8a0) at art_table_walk+0x26c
>> art_table_walk(ffff800000390900,fffffd8746decea0,ffffffff813bde40,ffff8000247c9
>> 8a0) at art_table_walk+0x26c
>> art_walk(ffff800000390900,ffffffff813bde40,ffff8000247c98a0) at art_walk+0xd1
>> rtable_walk(0,18,ffff8000247c9938,ffffffff813c2b70,ffff800000784050) at
>> rtable_
>> walk+0xa4
>> art_walk(ffff800000390900,ffffffff813bde40,ffff8000247c98a0) at art_walk+0xd1
>> rtable_walk(0,18,ffff8000247c9938,ffffffff813c2b70,ffff800000784050) at
>> rtable_
>
> This is not rtfree() related. I was surptized, but reference counters
> used by art tables are not MP safe. I propose to use refcnt API.
Please drop this.
