[Intel-gfx] [PATCH] Add Dmesg Triage Feature: further triage i-g-t kmsg log to reduce result noise resulted from piglit dmesg defect

[Gao, Ethan]

Actually, I sent to piglit ML before intel-gfx ML.
Why It's here is because many internal guys don't subscribe to piglit ML.

Best Regards
Ethan, Gao

-----Original Message-----
From: Lespiau, Damien 
Sent: Friday, March 20, 2015 1:16 AM
To: Gao, Ethan
Cc: intel-gfx at lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH] Add Dmesg Triage Feature: further triage i-g-t kmsg log to reduce result noise resulted from piglit dmesg defect

On Thu, Mar 19, 2015 at 10:36:51AM +0800, ethan gao wrote:
> tests/igt.py: add igt env to enable or disable dmesg triage
> framework/test/base.py: trigger dmesg triage depending on dmesg log 
> occurrence
> framework/dmesg.py: employ dmesg triage simply for Linux dmesg
> dmesg_triage/*: deal with kmsg log with pre-defined dmesg oops pattern
> 
> In general, if dmesg triage is enabled and there is new dmesg along 
> with i-g-t testcases running, the new dmesg will be captured and 
> worked out a tag and head to rollback the detail when necessary, in 
> addition, on the basis of different dmesg tag or head, the final result noise of a testcase can be reduced with defined rules or strategy.
> 
> Signed-off-by: ethan gao <ethan.gao at intel.com>

That's a patch to piglit, shouldn't be sent to the piglit ml as well to get their opinion?

--
Damien