Thanks
I thought this was just a test timeout, but it manually cranking up the timeout does not help, it seems to be trying to converge on something and failing to do so. >>> Test phase: media-libs/libpulse-16.99.1 * abi_mips_n64.n64: running multilib-minimal_abi_src_test meson test --print-errorlogs -C /var/tmp/portage/media-libs/libpulse-16.99.1/work/pulseaudio-16.99.1-abi_mips_n64.n64 --num-processes 16 ninja: Entering directory `/var/tmp/portage/media-libs/libpulse-16.99.1/work/pulseaudio-16.99.1-abi_mips_n64.n64' ninja: no work to do. 1/15 Validate xml file pacat.1.xml OK 0.31s 2/15 Validate xml file pactl.1.xml OK 0.31s 3/15 Validate xml file pulse-client.conf.5.xml OK 0.30s 4/15 hashmap-test OK 0.25s 5/15 proplist-test OK 0.22s 6/15 utf8-test OK 0.20s 7/15 channelmap-test OK 0.34s 8/15 core-util-test OK 0.34s 9/15 get-binary-name-test OK 0.34s 10/15 json-test OK 0.32s 11/15 mainloop-test-glib OK 0.21s 12/15 srbchannel-test OK 0.28s 13/15 mainloop-test OK 2.05s 14/15 thread-mainloop-test OK 10.06s 15/15 volume-test FAIL 120.04s exit status 1 >>> MALLOC_PERTURB_=246 UBSAN_OPTIONS=halt_on_error=1:abort_on_error=1:print_summary=1:print_stacktrace=1 LD_LIBRARY_PATH=/var/tmp/portage/media-libs/libpulse-16.99.1/work/pulse audio-16.99.1-abi_mips_n64.n64/src:/var/tmp/portage/media-libs/libpulse-16.99.1/work/pulseaudio-16.99.1-abi_mips_n64.n64/src/pulse MAKE_CHECK=1 ASAN_OPTIONS=halt_on_error=1:abor t_on_error=1:print_summary=1 /var/tmp/portage/media-libs/libpulse-16.99.1/work/pulseaudio-16.99.1-abi_mips_n64.n64/src/tests/volume-test ――――――――――――――――――――――――――――――――――――― ✀ ――――――――――――――――――――――――――――――――――――― stdout: Running suite(s): Volume 0%: Checks: 1, Failures: 0, Errors: 1 ../pulseaudio-16.99.1/src/tests/volume-test.c:107:E:volume:volume_test:0: (after this point) Test timeout expired stderr: Attenuation of sample 1 against 32767: -90.3087 dB Smallest possible attenuation > 0 applied to 32767: 0 ―――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――― Summary of Failures: 15/15 volume-test FAIL 120.04s exit status 1 Ok: 14 Expected Fail: 0 Fail: 1 Unexpected Pass: 0 Skipped: 0 Timeout: 0
IIRC there was similar issue in Debian buildd where one of arch hosts was unexpectedly slow. Could you please try this patch to increase timeout in volume-test to see if that helps it passing? diff --git a/src/tests/volume-test.c b/src/tests/volume-test.c index 55486f642..a67f30e34 100644 --- a/src/tests/volume-test.c +++ b/src/tests/volume-test.c @@ -161,7 +161,7 @@ int main(int argc, char *argv[]) { s = suite_create("Volume"); tc = tcase_create("volume"); tcase_add_test(tc, volume_test); - tcase_set_timeout(tc, 120); + tcase_set_timeout(tc, 300); suite_add_tcase(s, tc); sr = srunner_create(s);
(In reply to Igor V. Kovalenko from comment #2) > IIRC there was similar issue in Debian buildd where one of arch hosts was > unexpectedly slow. > > Could you please try this patch to increase timeout in volume-test to see if > that helps it passing? > > diff --git a/src/tests/volume-test.c b/src/tests/volume-test.c > index 55486f642..a67f30e34 100644 > --- a/src/tests/volume-test.c > +++ b/src/tests/volume-test.c > @@ -161,7 +161,7 @@ int main(int argc, char *argv[]) { > s = suite_create("Volume"); > tc = tcase_create("volume"); > tcase_add_test(tc, volume_test); > - tcase_set_timeout(tc, 120); > + tcase_set_timeout(tc, 300); > suite_add_tcase(s, tc); > > sr = srunner_create(s); This was my first thought as well, I manually cranked it up to some crazy value (like 99999) but even after letting it run overnight it just hung.
Here is debian mips64el build log with volume-test passed - maybe there is a clue. https://buildd.debian.org/status/fetch.php?pkg=pulseaudio&arch=mips64el&ver=16.1%2Bdfsg1-2%2Bb1&stamp=1665924652&raw=0
Created attachment 880303 [details] build.log and emerge --info
We investigated this timeout failure and it turned out that runtime slowdown occurs due to running this test on MIPS hardware without FPU without -msoft-float which takes about 90% in kernel fpu emulation code. Running test host with softfloat configuration resulted in significant speedup to this volume-test. Suggest either patching out volume-test for this specific mips configuration completely or reconfiguring system to use softfloat and bumping volume-test timeout to some 300 seconds.
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=684a97aa82b0b228b7faf21dd1c57dcf48757954 commit 684a97aa82b0b228b7faf21dd1c57dcf48757954 Author: Matoro Mahri <matoro_gentoo@matoro.tk> AuthorDate: 2024-03-14 20:25:02 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2024-03-21 02:26:54 +0000 media-libs/libpulse: add patch removing catch2 internal test timeouts No revbump since this is test-only fix. See upstream for justification. See: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/merge_requests/807 Bug: https://bugs.gentoo.org/918447 Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk> Closes: https://github.com/gentoo/gentoo/pull/35755 Signed-off-by: Sam James <sam@gentoo.org> .../files/pulseaudio-17.0-backport-pr807.patch | 298 +++++++++++++++++++++ media-libs/libpulse/libpulse-17.0.ebuild | 1 + 2 files changed, 299 insertions(+)
mips done all arches done
