Skip to contentSkip to conversation

Strange err=25 result.

DISCUSSION James Courtier-Dutton 2 participants 3 messages Jan 14, 2004 – Jan 15, 2004

lore
Share & tools
lore
#0 lore raw

Some debug output from calling snd_pcm_wait() when a USB device is unplugged.

(pcm_hw.c from /alsa-lib/src/pcm/pcm_hw.c) file=pcm_hw.c:406 function=snd_pcm_hw_status err=25 pcm_hw.c: snd_pcm_hw_status() SNDRV_PCM_IOCTL_STATUS failed.

According to errno.h,
#define ENOTTY 25 /* Not a typewriter */

I would think that the following would be a better err result.
#define ENODEV 19 /* No such device */
or even
#define EPIPE 32 /* Broken pipe */

I have looked through alsa-lib and alsa-kernel and cannot find any place in the code that sets errno to 25!

Can anybody help?

My aim is to let the application detect when a usb audio device is unplugged, so that the application can release all alsa resources.
The application will then poll the alsa-lib trying to open the device again, and continually get errors until the audio usb device is plugged in again. For my application to achieve this, I would need a specific errno for the "usb device unplugged" that will not get confused with over/underruns or other errors that might not be "usb unplugged" related.

Cheers
James

-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html

#1 lore raw
Quoted 15 lines · James Courtier-Dutton

Some debug output from calling snd_pcm_wait() when a USB device is unplugged.

(pcm_hw.c from /alsa-lib/src/pcm/pcm_hw.c) file=pcm_hw.c:406 function=snd_pcm_hw_status err=25 pcm_hw.c: snd_pcm_hw_status() SNDRV_PCM_IOCTL_STATUS failed.

According to errno.h,
#define ENOTTY 25 /* Not a typewriter */

I would think that the following would be a better err result.
#define ENODEV 19 /* No such device */
or even
#define EPIPE 32 /* Broken pipe */

Quoted 4 lines

I have looked through alsa-lib and alsa-kernel and cannot find any place in the code that sets errno to 25!

Can anybody help?

Consult the toplevel filesystem routines. In case when the device is disconnected, all i/o (filesystem operation) calls are redirected to empty structure, so the Linux kernel filesystem code probably returns ENOTTY errors when no ioctl/read/write callbacks are associated to the device.

Jaroslav

-----
Jaroslav Kysela <[email protected]>
Linux Kernel Sound Maintainer
ALSA Project, SuSE Labs

-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html

#2 lore raw
Quoted 17 lines · Jaroslav Kysela
Quoted 14 lines · James Courtier-Dutton

Some debug output from calling snd_pcm_wait() when a USB device is unplugged.

(pcm_hw.c from /alsa-lib/src/pcm/pcm_hw.c) file=pcm_hw.c:406 function=snd_pcm_hw_status err=25 pcm_hw.c: snd_pcm_hw_status() SNDRV_PCM_IOCTL_STATUS failed.

According to errno.h,
#define ENOTTY 25 /* Not a typewriter */

I would think that the following would be a better err result.
#define ENODEV 19 /* No such device */
or even
#define EPIPE 32 /* Broken pipe */

Quoted 8 lines
Quoted 3 lines

I have looked through alsa-lib and alsa-kernel and cannot find any place in the code that sets errno to 25!

Can anybody help?

Consult the toplevel filesystem routines. In case when the device is disconnected, all i/o (filesystem operation) calls are redirected to empty structure, so the Linux kernel filesystem code probably returns ENOTTY errors when no ioctl/read/write callbacks are associated to the device.

Jaroslav

Thank you, that makes sense. :-)

Also, what should
snd_pcm_wait()
and
snd_pcm_state()
return if the USB device is unplugged.

At the moment, snd_pcm_wait() just hangs around until the timeout and returns errno -5 and snd_pcm_state() still returns state=3 which is SND_PCM_STATE_RUNNING

I would hardly say the the card is still in "RUNNING" state if you physically remove the card from the system!

I would prefer snd_pcm_wait() to imeadiately exit with an errno, and snd_pcm_state() to return a new state of "SND_PCM_STATE_DISCONNECTED" or similar.

Can I do what I want with snd_pcm_wait or not?

Cheers
James

-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html