[Anuj/Steve/Ross - questions below!]

I thought I'd share a summary of the recent toolchain test
changes/work. We can see the "baseline" from the 4.2 test report:

https://downloads.yoctoproject.org/releases/yocto/milestones/yocto-4.2_M3/testresults/testresult-report.txt

showing that we run 2.9 million tests (2946881) with 174470 failures.

Once we noticed qemuarm64 toolchain test results were showing ~125,000
failures and had done since the 3.4 release, Ross was able to find and
fix it:

https://git.yoctoproject.org/poky/commit/?id=b0f1ab9810d87960d2753b0fe78039b874fd15fd
https://git.yoctoproject.org/poky/commit/?id=af39b83e58ec159a9a13d85067c2595535442566

Thanks Ross (even if you did break it too! :)

I've also merged:

https://git.yoctoproject.org/poky/commit/?id=c94fd2737e7e341188c8c41d911f1c60240088a2

which takes care of many of the qemuppc failures.

This brings the report down to:

https://autobuilder.yocto.io/pub/non-release/20230717-18/testresults/testresult-report.txt

so 3229355 tests and 15496 failures.

qemuarm has ~350 failures
qemuarm64 has ~350 failures
qemux86-64 has ~4000 (3900 in glibc)
qemux86 has ~4000 (3500 in glibc)
qemuppc has ~600 failures
qemumips64 has ~5000 failures (all over)
qemumips has ~1600 failures

We could also do with understanding the qemuarm64 ltp syscall failures.
The other ltp failures look the same for x86 and arm.

Anuj: Can Intel look into the glibc test failures on x86?
Steve: Can we backport those patches to kirkstone/mickledore?
Ross: Can ARM look into the ltp syscall issue?

Is anyone else interested/able to help in cleaning up these results a
bit further? Once have a good baseline, it will make it much easier to
look for and spot regressions.

Cheers,

Richard






-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#184498): 
https://lists.openembedded.org/g/openembedded-core/message/184498
Mute This Topic: https://lists.openembedded.org/mt/100212267/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to