https://bugs.freedesktop.org/show_bug.cgi?id=100064
GitLab Migration User changed:
What|Removed |Added
Status|NEW |RESOLVED
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #42 from Martin Ejdestig ---
Indeed, commenting out module-x11-publish and module-x11-xsmp
(module-x11-cork-request was already commented out... on Arch) makes everything
behave as expected (HDMI source is
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #41 from Tanu Kaskinen ---
Now I remember why I didn't earlier suggest removing module-x11-xsmp: it
doesn't only keep pulseaudio running, it's also required for telling the other
X11 modules to unload themselves when
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #40 from Tanu Kaskinen ---
I see it's been more than a month without a reply. I'm sorry, I forgot about
this bug...
ice_io_cb() could have a codition that if flags contains PA_IO_EVENT_HANGUP
like it does in this
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #37 from Martin Ejdestig ---
When testing this I also noticed another thing. I managed to get output
switched back to "Line Out" for user B by just VT-switching. I did not capture
logs for that though. But...
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #36 from Martin Ejdestig ---
(In reply to Tanu Kaskinen from comment #35)
> The exit seems to be happening in pa__done() of
> src/modules/x11/module-x11-xsmp.c. Does removing the
> IceRemoveConnectionWatch() call
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #35 from Tanu Kaskinen ---
(In reply to Tanu Kaskinen from comment #34)
> PulseAudio shouldn't call exit(1) by itself. I suspect some of the X libraries
> does that.
I forgot that the journal actually tells that quite
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #34 from Tanu Kaskinen ---
Ok, now we're getting a bit forward. These messages give a clue:
( 0.513| 0.000) D: [pulseaudio] module-card-restore.c: foobar -
alsa_card.pci-_01_00.1: e->preferred_output_port =
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #32 from Martin Ejdestig ---
Btw, why are there two stores? One with null and one with hdmi-output-2.
( 21.405| 0.000) I: [pulseaudio] module-card-restore.c: Storing profile
hdmi-stereo-extra2 for port
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #31 from Martin Ejdestig ---
(In reply to Tanu Kaskinen from comment #28)
> I suspect that the "extra" pulseaudio instance overwrites the database.
> Things might happen like this:
>
> Instance 1 (the "real"
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #30 from Tanu Kaskinen ---
Hmm, no, it doesn't explain the problem. Please ignore the previous comment
(comment #29). That message is printed after the card is already created. The
database is read at an earlier stage,
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #29 from Tanu Kaskinen ---
(In reply to Tanu Kaskinen from comment #28)
> Is that message from an "extra" pulseaudio instance? When you sent the
> package with 8 logs, every other of them were from those extra
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #28 from Tanu Kaskinen ---
Is that message from an "extra" pulseaudio instance? When you sent the package
with 8 logs, every other of them were from those extra pulseaudio instances
that shut down immediately after
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #27 from Martin Ejdestig ---
> Was not able get a log of when it works after trying ~10 times.
I have only ever seen it working once though (the one time mentioned above) so
would probably have to try _a lot_
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #26 from Martin Ejdestig ---
Created attachment 130130
--> https://bugs.freedesktop.org/attachment.cgi?id=130130=edit
Log with 0001-more-logging.patch applied
Attached log with 0001-more-logging.patch applied.
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #25 from Tanu Kaskinen ---
Created attachment 130114
--> https://bugs.freedesktop.org/attachment.cgi?id=130114=edit
0001-more-logging.patch
When preferred_output_port is set, it should be remembered. The logs don't
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #24 from Martin Ejdestig ---
Created attachment 130092
--> https://bugs.freedesktop.org/attachment.cgi?id=130092=edit
Debug patch and logs
Patch that adds pa_log_debug() and logs. pulseverbose.log.2 is log for
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #23 from Martin Ejdestig ---
I tested a bit with pa_log_debug():s sprinkled everywhere and it seems like
port is NULL in is_preferred_port_active_and_available() so it returns false.
Once out of my 5 test runs it
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #21 from Tanu Kaskinen ---
(In reply to Martin Ejdestig from comment #19)
> The patch does not seem to work. Will attach new log soon.
The new log looks fine, the HDMI card profile doesn't any more get changed from
https://bugs.freedesktop.org/show_bug.cgi?id=100064
Tanu Kaskinen changed:
What|Removed |Added
Attachment #130091|text/x-log |text/plain
mime
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #20 from Martin Ejdestig ---
Created attachment 130091
--> https://bugs.freedesktop.org/attachment.cgi?id=130091=edit
pulseverbose-patch1.log
--
You are receiving this mail because:
You are the assignee for
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #19 from Martin Ejdestig ---
The patch does not seem to work. Will attach new log soon.
> I hope this is helpful:
I was mainly looking for how to not have to do "make install" as root (once
bitten...).
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #18 from Tanu Kaskinen ---
I hope this is helpful:
https://www.freedesktop.org/wiki/Software/PulseAudio/Documentation/Developer/PulseAudioFromGit/
--
You are receiving this mail because:
You are the QA Contact for
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #17 from Martin Ejdestig ---
Thank you!
I have applied it on a local branch but will have to build and test tomorrow (I
am being forcibly removed from the computer ;) .
I looked quickly on the home page and in
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #16 from Tanu Kaskinen ---
Created attachment 130075
--> https://bugs.freedesktop.org/attachment.cgi?id=130075=edit
0001-switch-on-port-available-avoid-switching-away-from-p.patch
Here's an untested patch. Are you
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #15 from Martin Ejdestig ---
OK. :)
(Also, probably not relevant any longer then since you seem to understand the
problem, but the time stamps in the journalctl --user output above is close to
when the user logs
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #14 from Tanu Kaskinen ---
Yes, but when switching the card profile, the HDMI 3 sink becomes unavailable.
The HDMI 3 sink was previously selected as the default sink by you, but now
there's no sink that's explicitly
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #13 from Martin Ejdestig ---
(In reply to Tanu Kaskinen from comment #11)
> This should be pretty easy to fix: just don't switch to HDMI 2 if you've
> previously selected HDMI 3 and it's still available.
Hum, but
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #12 from Martin Ejdestig ---
If I run journalctl --user I see this for first login of user B:
mar 05 15:31:41 gromit pulseaudio[31127]: ICE default IO error handler doing an
exit(), pid = 31127, errno = 32
mar 05
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #11 from Tanu Kaskinen ---
At one point something makes HDMI 2 change state from unavailable to available,
and module-switch-on-available will then switch the HDMI card to the HDMI 2
profile. You want to keep using
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #10 from Tanu Kaskinen ---
That's weird indeed.
The second instance removes the pid file:
W: [pulseaudio] pid.c: Stale PID file, overwriting.
If the pid file was created by the first instance, the second instance
https://bugs.freedesktop.org/show_bug.cgi?id=100064
Tanu Kaskinen changed:
What|Removed |Added
Attachment #130071|text/x-log |text/plain
mime
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #9 from Martin Ejdestig ---
Do not know if this is related to my problems, but why are log.1 and log.3
created? .log and .log.2 seems to be the "real" logs. Something tries to spawn
an extra pulseaudio instance
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #8 from Martin Ejdestig ---
As user A (my main desktop user) I su:ed to user B and created the config file.
Verified that no user B process was running as user A, VT switched to GDM and
logged in as user B.
When
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #7 from Martin Ejdestig ---
Created attachment 130074
--> https://bugs.freedesktop.org/attachment.cgi?id=130074=edit
pulseverbose.log.3
--
You are receiving this mail because:
You are the assignee for the bug.
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #5 from Martin Ejdestig ---
Created attachment 130072
--> https://bugs.freedesktop.org/attachment.cgi?id=130072=edit
pulseverbose.log.1
--
You are receiving this mail because:
You are the assignee for the bug.
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #4 from Martin Ejdestig ---
Created attachment 130071
--> https://bugs.freedesktop.org/attachment.cgi?id=130071=edit
pulseverbose.log
--
You are receiving this mail because:
You are the QA Contact for the bug.
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #3 from Tanu Kaskinen ---
Ok. Let's try to figure out what happens during the logout-login cycle. Add
these lines to ~/.config/pulse/daemon.conf (for the B user):
log-target = newfile:/tmp/pulseverbose.log
log-level =
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #2 from Martin Ejdestig ---
Nope. Killing the "pulseaudio" process while logged in as B and letting it
respawn does not seem to reset the output device to "Line Out".
--
You are receiving this mail because:
You
https://bugs.freedesktop.org/show_bug.cgi?id=100064
--- Comment #1 from Tanu Kaskinen ---
Does the output device get reset also if you just restart pulseaudio while
logged in as B?
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee
40 matches
Mail list logo