[pulseaudio-discuss] Assertion 'usec <= ((pa_usec_t) 1000000ULL)*60ULL*60ULL' , failed at pulsecore/rtpoll.c:516

[Lennart Poettering]

On Mon, 10.08.09 12:35, Timo Harmonen (timo.harmonen at iki.fi) wrote:

> Hi,
> 
> I have a ltsp client with some cheap SoundBlaster card. PA fails
> to start, assertion fails, logs below.
> 
> - Timo

Your sound driver is broken. I mean, seriously, how could I ever make
this any clearer than printing this a gazillion times to syslog:

> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1443200 bytes (8181 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1443264 bytes (8181 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1377728 bytes (7810 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1312192 bytes (7438 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1246656 bytes (7067 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1181120 bytes (6695 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1115584 bytes (6324 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 1050048 bytes (5952 ms) Most
>   likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 984512 bytes (5581 ms) Most
> likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 918976 bytes (5209 ms) Most
> likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> D: alsa-util.c: snd_pcm_avail_update() returned a value that is
> exceptionally large: 853440 bytes (4838 ms) Most
> likely this is a Linux bug. Please report this issue to the ALSA
> developers.
> E: rtpoll.c: Assertion 'usec <= ((pa_usec_t) 1000000ULL)*60ULL*60ULL'
> failed at pulsecore/rtpoll.c:516, function
> pa_rtpoll_set_timer_relative(). Aborting.
> Aborted

The assert that is hit is simply the result of the broken kernel
driver you have there.

snd_pcm_avail() is broken for your driver, but we rely on this for our
time handling. Hence sooner or later our timing goes bonkers and
safety checks (i.e. asserts) are triggered.

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net
http://0pointer.net/lennart/           GnuPG 0x1A015CC4