[ATrpms-users] new alsa, no mythtv sound now

[john bray]

On Wed, 2006-09-06 at 00:18 -0700, David Rees wrote:
> On 9/4/06, john bray <john at johnbray.name> wrote:
> > i seem to have no problems with gaim, xmms, xine, mplayer making sounds.
> > but, mythtv remains silent.  i've mucked around with the setup settings
> > for mixer and mixer device.  no joy yet.
> >
> > i'm sort of looking for ideas.  anyone have any directions at which to
> > look?  anyone been through this already and know the answer?
> 
> Have you checked the permissions of /dev/dsp and /dev/mixer? Mine seem
> to have changed to being only owned/read/writable by root recently,
> I'm still trying to figure out what changed.

hey dave -

i ran into a similar problem with permissions in fc4 some time ago.  i'd
needed to have multiple users able to use certain devices.  but, i'd put
up this and the problem was solved.  dunno if it fixes your problem, but
might be worth a try to put up something similar...

[root at junior ~]# cat /etc/security/console.perms.d/local.perms
#
# jmb notes:
#       - video and sound stuff is mostly for solving the mythbackend
#       access when console user owns the device.
#
#       - sound - this will likely have to be the same as video since
some
#       of the cards are wintv go cards what have to be audio'ed through
#       line in.
#
#       - scanner stuff is for adding sg0 to the perms to pick up the
#       linotype which, at least now, is the only sg device.
#       theoretically, it doesn't need different perms since it won't
get
#       shared.
#
#       note that my current understanding is that, based on the man
page
#       for pam_console_apply, that the naming convention is the reverse
#       from that of the udev .rules files.  apparently the console
perms
#       files are read in lexical order like the udev files.  BUT, later
#       defs *>replace<* prior ones.  stupid convention.  might make
sense
#       since this is likely parsed once at boot, and the other is
probably
#       parsed at device creation time.  pah.  don't believe that
either,
#       since it's likely parsed only at udev start time as well.  just
#       sucks as a convention.  but udevstart will force a rescan of the
#       udev stuffies.  mebbe this console.perms stuff happens at login?
#

# device classes -- these are shell-style globs
<v4l>=/dev/video* /dev/radio* /dev/winradio* /dev/vtx* /dev/vbi* \
      /dev/video/* /dev/dvb/adapter*/*

<sound>=/dev/dsp* /dev/audio* /dev/midi* \
        /dev/mixer* /dev/sequencer \
        /dev/sound/* /dev/beep \
        /dev/snd/*

<scanner>=/dev/scanner* /dev/usb/scanner* /dev/sg*

# permission definitions
<console>  0666 <v4l>        0666 root
<console>  0666 <sound>      0666 root
<console>  0666 <scanner>    0666 root
[root at junior ~]#


hth.

john

-- 
cleverly disguised as a responsible adult.