Re: help debugging failures on riscv64
Hi,
On Thu, Apr 17, 2025 at 10:55 PM Paul Gevers <elbrus@debian.org> wrote:
>
> Hi,
>
> On 17-04-2025 16:47, Michael Stone wrote:
> > What's reasonable? I don't really want to just keep uploading packages
> > with longer and longer timeouts to see what works, that's why I was
> > asking if it was possible to run manually on one of the machines having
> > issues to see if there's something wrong here other than the timeout.
>
>
> My question was also mostly towards vimer.
>
I did try running the autopkgtest manually on Unmatched using various
large timeout values including here[0], but the test always fails like
this:
```
Processing triggers for libc-bin (2.41-7) ...
autopkgtest [03:30:49]: test savegames: [-----------------------
timeout: sending signal TERM to command ‘xvfb-run’
/tmp/autopkgtest-lxc.8b92fbd4/downtmp/wrapper.sh: Killing leaked
background processes: 1950 PID TTY STAT TIME COMMAND 1950 ? Zl 103:25
[uqm] <defunct>
autopkgtest [04:30:51]: test savegames: -----------------------]
autopkgtest [04:30:51]: test savegames: - - - - - - - - - - results -
- - - - - - - - - savegames FAIL non-zero exit status 124
autopkgtest [04:30:51]: test savegames: - - - - - - - - - - stderr - -
- - - - - - - - timeout: sending signal TERM to command ‘xvfb-run’
autopkgtest [04:30:51]: - - - - - - - - - - running shell - - - - - - - - - -
root@autopkgtest-lxc-iuojpu:/tmp/autopkgtest-lxc.8b92fbd4/downtmp/build.Jg9/real-tree#
root@autopkgtest-lxc-iuojpu:/tmp/autopkgtest-lxc.8b92fbd4/downtmp/build.Jg9/real-tree#
```
However, on the SiFive P550, it consistently passes without issue. I
have compared these two log files[1][2] but not have idea about this.
Interestingly, running `uqm` directly in a full desktop environment on
Unmatched works -- the game is playable there, so this doesn't seem
like a fundamental issue with the platform or the binary itself.
I suspect the problem is with the combination of xvfb-run and uqm then
one timeout issue[0] causes the entire autopkgtest to exit with a
timeout signal. Or in other words, this does trigger a bug, but I
don't have enough evidence to prove it.I also wan to run xvfb-run and
uqm manually to reproduce this issue but no lucky until now.
If possible, I would like to open reportbug to track the issue under
debian-riscv@, similar to #1102540 did if we add this package to
reject_list. In fact, we have observed several similar debci reports
caused by different riscv64 workers. It's frustrating but we will
continue to investigate if there is a real issue here.
BR,
Bo
[0]: https://salsa.debian.org/games-team/uqm/-/blob/master/debian/tests/savegames_x?ref_type=heads
[1]: https://gist.github.com/yuzibo/64eec987da753b672e8f6500e9b7274e
[2]: https://gist.github.com/yuzibo/88420de1a73a3a7cb015662e87ed517e
[3]: https://bugs.debian.org/1102540
> Paul
>
Reply to: