[pulseaudio-discuss] alsa module fails after rebooting (solved)
Pete Nesbitt
pete at linux1.ca
Sun Dec 28 16:31:35 PST 2008
On Sun, 28 Dec 2008 23:58:02 +0000
Colin Guthrie <gmane at colin.guthr.ie> wrote:
> 'Twas brillig, and Pete Nesbitt at 28/12/08 23:51 did gyre and gimble:
> > On Fri, 26 Dec 2008 14:15:27 -0800
> > Pete Nesbitt <pete at linux1.ca> wrote:
> >
> >> Hi,
> >> I have a system serving music locally and over the network. It was
> >> working properly (once I took it off wireless), however after shutting
> >> it down (to move it) the ALSA modules are failing to load.
> >>
> >> Maybe this is not a PA issue, it may be an ALSA problem but thought I
> >> would post anyway.
> >>
> >> I am auto loading the alsa modules.
> >>
> >> Here is the log information for when it was working, followed by the
> >> same info when it failed. I don't know why it would quit working after
> >> a reboot.
> >>
<snip>
> >
> > The problem was that udev overwrites permissions to the sound device on
> > reboot. I had given everyone rw (it's an appliance so I'm not worried
> > about local user security).
> >
> > I'm running Fedora 9, and had to modify /etc/udev/rules.d/40-alsa.rules
> > I changed this:
> > KERNEL=="controlC[0-9]*", NAME="snd/%k"
> > KERNEL=="hwC[D0-9]*", NAME="snd/%k"
> > KERNEL=="pcmC[D0-9cp]*", NAME="snd/%k"
> > KERNEL=="midiC[D0-9]*", NAME="snd/%k"
> > KERNEL=="timer", NAME="snd/%k"
> > KERNEL=="seq", NAME="snd/%k"
> >
> > To this:
> > KERNEL=="controlC[0-9]*", NAME="snd/%k", MODE="0666"
> > KERNEL=="hwC[D0-9]*", NAME="snd/%k", MODE="0666"
> > KERNEL=="pcmC[D0-9cp]*", NAME="snd/%k", MODE="0666"
> > KERNEL=="midiC[D0-9]*", NAME="snd/%k", MODE="0666"
> > KERNEL=="timer", NAME="snd/%k", MODE="0666"
> > KERNEL=="seq", NAME="snd/%k", MODE="0666"
>
> Please note that while I'm not certain about F9, the job of assigning
> permissions to the audio device have moved on from simple modes.
> Nowadays console-kit will assign an acl to the device nodes when a given
> user is considered "active" e.g. has the active tty login or the active
> X11 login etc. This acl cannot be seen when doing a simple ls -l
> /dev/snd. You need to use "getfacl /dev/snd/*" Also see ck-list-sessions
> and check that one of them is ACTIVE.
>
>
> > It now works great locally and over the LAN via RTP. Would still like to
> > use wireless but that seems to be too much traffic for the network.
> >
> > FYI, wireless floods my logs with 20 lines per second!
> > Dec 21 13:05:13 artigo pulseaudio[2126]: module-rtp-send.c: Failed to push chunk into memblockq.
> >
> > Oh well :(
>
> Wireless works OK for me (not via RTP, just plain pulse protocol,
> althouhg it does suffer from occasional pops etc. Tunnels are worse than
> just setting PULSE_SERVER.
>
> Col
>
I thought those ACl's were tied into SELinux (which I'm not running),
but this is for pulse user run as system daemon, and has no login
so I don't think the console-kit will have any impact.
I was thinking that when I use regular TCP (which does work on
wireless) that the processing was done at the destination but with RTP
it done it on the serving side. I wanted to have as little impact on
receiving boxes as possible. ...but I am not sure really.
Thanks.
--
Pete Nesbitt
http://www.linux1.ca
More information about the pulseaudio-discuss
mailing list