[yocto] QA cycle report for 2.3.4 RC1

[Richard Purdie]

On Tue, 2018-07-10 at 06:11 +0000, Yeoh, Ee Peng wrote:
> This is the full report for 2.3.4.rc1: 
> https://wiki.yoctoproject.org/wiki/WW25_-_2018-06-22-_Full_Test_Cycle
> _-_2.3.4_rc1
>  
> ======= Summary ========
>  
> All planned tests were executed.  There were zero high milestone
> defect.  Team had found 3 new defects where yocto-bsp failed to
> create bootable image with new bsp layer [1], QEMU ended unexpectedly
> during do_testimage:core-image-lsb-sdk [2], & valgrind ptest passed
> in 2.3.3.rc1 failed at current release [3].  
>  
> ======= QA-Hints========
>  
> Found 3 non critical defects. 
>  
> ======= Bugs ========

Thanks for this Ee Peng and team!

> New Bugs
> [1] Bug 12846 - [QA 2.3.4rc1][Meta-Yocto][Case 309, Case 310]: After
> creating BSP layer using yocto-bsp, the image created not able to
> boot up as local variable 'qbsys' referenced before assignment
> https://bugzilla.yoctoproject.org/show_bug.cgi?id=12846

Has this ever worked? I agree we should have a better error message but
ultimately its due to runqemu not understanding that MACHINE name.

I don't think that is a release blocker.

> [2] Bug 12812 - [2.3.4 rc1] Qemu ended unexpectedly during
> do_testimage:core-image-lsb-sdk
> https://bugzilla.yoctoproject.org/show_bug.cgi?id=12812

We believe this is due to the infamous apic issues we've had with
certain kernel versions (IA only). Its annoying and something we should
fix but not a release blocker IMO.
 
> [3] Bug 12847 - [2.3.4 rc1] valgrind ptest failed
> https://bugzilla.yoctoproject.org/show_bug.cgi?id=12847

This is annoying but we'd probably focus on master for ptest work.

Looking at the various bugs filed against the release in the QA report,
I think there are some annoying issues but nothing to warrant an rc2
given our current resource situation so I'd suggest releasing rc1.

Cheers,

Richard