On Mon, 10.08.09 12:24, Colin Guthrie (gmane at colin.guthr.ie) wrote:
> Hi,
>
> Anyone noticed this?
>
> The CPU usuage of g-s-d and g-v-c-a going through the roof (with PA
> following it).
>
> Seems to be getting stuck in some kind of loop.
>
> I have to kill both g-v-c-a and g-s-d to get things back to normal.
g-v-c-a? I onlyknow g-v-c, What's that -a about?
> Anyone noticed this?
Not sure.
> poll([{fd=5, events=POLLIN}, {fd=3, events=POLLIN}, {fd=8, events=0},
> {fd=12, events=POLLIN|POLLPRI}, {fd=14, events=POLLIN|POLLPRI}, {fd=15,
> events=POLLIN|POLLPRI}, {fd=16, events=POLLIN|POLLPRI}, {fd=19,
> events=POLLIN}, {fd=7, events=POLLIN}, {fd=20, events=POLLIN}, {fd=4,
> events=POLLIN}], 11, 0) = 1 ([{fd=4, revents=POLLIN}])
> read(3, 0x1161eb4, 4096) = -1 EAGAIN (Resource
> temporarily unavailable)
> poll([{fd=5, events=POLLIN}, {fd=3, events=POLLIN}, {fd=8, events=0},
> {fd=12, events=POLLIN|POLLPRI}, {fd=14, events=POLLIN|POLLPRI}, {fd=15,
> events=POLLIN|POLLPRI}, {fd=16, events=POLLIN|POLLPRI}, {fd=19,
> events=POLLIN}, {fd=7, events=POLLIN}, {fd=20, events=POLLIN}, {fd=4,
> events=POLLIN}], 11, 0) = 1 ([{fd=4, revents=POLLIN}])
Smells like one of those g-v-c feedback loops. i.e. user triggers
volume change, g-v-c forwards that to PA. PA informs all clients about
the vol change including g-v-c. g-v-c forwards that to gtk. gtk
informs g-v-c about the changes it just made to the UI, pavucontrol
forwards that to PA, GOTO 10.
Under no circumstances g-v-c should forward anything coming in from
PA back to PA. And under no circumstances g-v-c should forward
anything coming from gtk back to gtk. As it stands now this is however
broken in g-v-c.
A while back I sat down and fixed a couple of those issues, but I
think there are a number of issues left.
Please file bugs against g-v-c regarding this.
Lennart
--
Lennart Poettering Red Hat, Inc.
lennart [at] poettering [dot] net
http://0pointer.net/lennart/ GnuPG 0x1A015CC4
