Re: some kernel exceptions

[Jan Kiszka <jan.kiszka@siemens.com>]

On 10.10.22 17:30, Yxian.Yao wrote:
> 
> Hi Jan,
> 
>> The warnings start with perf, but I might have been too quick to shoot
>> at it. The real issue is definitely the page fault. You don't have a
>> stack backtrace for the first one?
> 
> No, these are the last messages that the kernel outputs. 
> The second exception is the same as the first one？
> The logs are printed with "BUG: unable to handle kernel painting request at 00000000000a22e8"
> 

The issue is that you generally cannot trust succeeding BUG messages
after the first one. They may point into the wrong direction, not to the
root cause.

> 
>> Also note that your kernel is not just old /wrt to its base version
>> (4.19), it's also not using the latest I-pipe patch for that
>> (ipipe-core-4.19.252-cip78-x86-23).
> 
> What is the difference between cip and no cip on ipipe core?
> 

At some point, we moved from 4.19-LTS to 4.19-CIP as baseline for the
I-pipe patches as the latter kernel will live longer (10 instead of 6
years). No further development was done on the 4.19-LTS branch you are
sitting on, which is also missing some bug fixes to I-pipe since then.

> 
>> The key switches are warned about. But that does not exclude that some
>> sleeping bug is only surfaced by turning on or off some feature that
>> others usually do not touch.
> 
> 
> How to further determine what caused it?

Hard to say yet. Please re-try with the latest 4.19-ipipe release if
possible and also check if something changes positively when you use the
config from xenomai-images.

Jan

-- 
Siemens AG, Technology
Competence Center Embedded Linux