Re: panic in drm or vt or deadlock on mutex or ...

From: Hans Petter Selasky <hps_at_selasky.org>
Date: Tue, 9 Mar 2021 09:45:06 +0100
On 3/9/21 9:31 AM, Alastair Hogge wrote:
> On 2021-02-08 21:01, Hans Petter Selasky wrote:
>> On 2/8/21 1:53 PM, Alastair Hogge wrote:
>>> Boot to multi-user; login (getty):
>>> $ doas kldload /boot/modules/amdgpu.ko
>>> $ sysctl
>>> [panic]
>>>
>>> ..is a guaranteed way to panic my system.
>>
>> Hi,
>>
>> Maybe you could do a hack and edit the sysctl source code:
>>
>> 1) print the sysctl before it is queried.
>> 2) sleep 1 second between print and query.
>>
>> Should be easy to nail this down!
> 
> Always panics at:
> $ /tmp/sysctl-with-delay -a
> [...]
> p1003_1b.mapped_files: 200112
> p1003_1b.memlock: 0
> p1003_1b.memlock_range: 0
> p1003_1b.memory_protection: 0
> p1003_1b.message_passing: 0
> p1003_1b.prioritized_io: 0
> p1003_1b.priority_scheduling: 200112
> p1003_1b.realtime_signals: 200112
> p1003_1b.semaphores: 0
> p1003_1b.fsync: 200112
> p1003_1b.shared_memory_objects: 200112
> p1003_1b.synchronized_io: 0
> p1003_1b.timers: 200112
> p1003_1b.aio_listio_max: 256
> p1003_1b.aio_max: 1024
> p1003_1b.aio_prio_delta_max: 0
> p1003_1b.delaytimer_max: 2147483647
> p1003_1b.mq_open_max: 0
> p1003_1b.pagesize: 4096
> p1003_1b.rtsig_max: 62
> p1003_1b.sem_nsems_max: 0
> p1003_1b.sem_value_max: 0
> p1003_1b.sigqueue_max: 128
> p1003_1b.timer_max: 32
> 

Hi,

On my system:

p1003_1b.timer_max: 32
sys.class.drm.card0-HDMI-A-1.modes: 1680x1050
1600x1200
1280x1024
1440x900
1280x960
1024x768
800x600
640x480
720x400

So that means some modes sysctl is causing the panic.

Emmanuel: Do you want to debug this?

--HPS
Received on Tue Mar 09 2021 - 07:45:32 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:27 UTC