Re: ✗ Xe.CI.Full: failure for drm/xe: Eliminate usage of TIMESTAMP_OVERRIDE
From: Lucas De Marchi
Date: Thu Feb 27 2025 - 00:27:50 EST
+Cc some people for kernel/power/process.c
On Wed, Feb 26, 2025 at 01:46:31PM -0800, Matt Roper wrote:
#### Warnings ####
* igt@kms_ccs@crc-primary-suspend-4-tiled-dg2-rc-ccs:
- shard-dg2-set2: [DMESG-WARN][2] ([Intel XE#4330]) -> [DMESG-WARN][3] +1 other test dmesg-warn
[2]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-2711-c7988865b658d56bcfaac76b7d72cc59435565eb/shard-dg2-432/igt@kms_ccs@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
[3]: https://intel-gfx-ci.01.org/tree/intel-xe/xe-pw-145437v1/shard-dg2-435/igt@kms_ccs@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
<4> [319.697183] done.
This seems to be something that happens sporadically when general driver
debug messages sneak in between these two calls in
kernel/power/process.c during system resume:
pr_info("Restarting tasks ... ");
...
pr_cont("done.\n");
Since those two prints were supposed to be on the same line, the
continuation seems to be coming out with "warn" level (at least at the
point CI reads it from syslog).
that pr_cont() seems wrong to me. pr_cont() is adding a 'c' flag to mark
continuation in kmsg. That doesn't really work except on early boot.
https://www.kernel.org/doc/Documentation/ABI/testing/dev-kmsg
The flags field carries '-' by default. A 'c' indicates a
fragment of a line. Note, that these hints about continuation
lines are not necessarily correct, and the stream could be
interleaved with unrelated messages, but merging the lines in
the output usually produces better human readable results. A
similar logic is used internally when messages are printed to
the console, /proc/kmsg or the syslog() syscall.
And:
https://lore.kernel.org/lkml/CA+55aFwmwdY_mMqdEyFPpRhCKRyeqj=+aCqe5nN108v8ELFvPw@xxxxxxxxxxxxxx/
I don't thiink it should be used in thaw_processes(). Maybe it'd be
better to define a pr_fmt and convert those to non-continuation prints?
Lucas De Marchi