[pulseaudio-tickets] [PulseAudio] #53: TeamSpeak problems and patches

[PulseAudio]

#53: TeamSpeak problems and patches
------------------------------------+---------------------------------------
 Reporter:  ZlatkO                  |       Owner:  lennart
     Type:  defect                  |      Status:  new    
 Priority:  normal                  |   Milestone:         
Component:  clients                 |    Severity:  normal 
 Keywords:  teamspeak padsp duplex  |  
------------------------------------+---------------------------------------
 I tried to run the Linux !TeamSpeak client through padsp, which failed.
 The problem was that padsp only works for /dev/dsp (ALSA card !#0), but my
 USB headset is /dev/dsp2 (ALSA card !#2, my ALSA card !#1 is a virtual
 UART midi device).

 {{{
 [zlatko at disclosure]:~$ cat /proc/asound/cards
  0 [ICH5           ]: ICH4 - Intel ICH5
                       Intel ICH5 with AD1985 at 0xf7fff800, irq 20
  1 [UART           ]: MPU-401 UART - MPU-401 UART
                       MPU-401 UART at 0x300, irq 10
  2 [Headset        ]: USB-Audio - Logitech USB Headset
                       Logitech Logitech USB Headset at usb-0000:00:1d.3-1,
 full speed
 }}}

 After fixing padsp to intercept all /dev/dsp* devices, !TeamSpeak still
 didn't work properly. It wants to switch the sound card to full duplex
 mode, but padsp doesn't support SNDCTL_DSP_SETDUPLEX ("unknown ioctl
 0x00005016"). As far as I understand, this should actually be a no-op, as
 !PulseAudio (or at least padsp) is always operating in full duplex mode
 internally (padsp answers SNDCTL_DSP_GETCAPS with DSP_CAP_DUPLEX by
 default).

 After adding a SNDCTL_DSP_SETDUPLEX no-op to padsp, it starts up fine
 without complaints - and !TeamSpeak actually works! :-)

 Patches attached.

-- 
Ticket URL: <http://pulseaudio.org/ticket/53>
PulseAudio <http://pulseaudio.org/>
The PulseAudio Sound Server