Intel HDA Driver Issue With PCM RW Interleaved

Intel HDA Driver Issue With PCM RW Interleaved

[re.mcclue]

I'm trying to play PCM buffer single threaded without buffer underrun.
Program source: https://gist.github.com/takaswie/d8a9553c230971d974a3dfc759b2ff67
(struct snd_pcm_sw_params).start_threshold is altered so as to allow specifying the start of audio frame transmission and prevent underrun issues.
In my application, on ioctl(fd, SNDRV_PCM_IOCTL_STATUS, &status), I receive: File descriptor in bad state
Output of $(dmesg) is: [ 69.013084] snd_hda_intel 0000:00:1f.3: Too big adjustment 128
What is causing this issue and how can I solve it?
Thanks

Sent with [ProtonMail](https://protonmail.com) Secure Email.

Re: Intel HDA Driver Issue With PCM RW Interleaved

[Takashi Iwai]

On Wed, 19 May 2021 07:48:52 +0200,
re.mcclue wrote:
> 
> I'm trying to play PCM buffer single threaded without buffer underrun.
> Program source: https://gist.github.com/takaswie/d8a9553c230971d974a3dfc759b2ff67
> (struct snd_pcm_sw_params).start_threshold is altered so as to allow specifying the start of audio frame transmission and prevent underrun issues.
> In my application, on ioctl(fd, SNDRV_PCM_IOCTL_STATUS, &status), I receive: File descriptor in bad state
> Output of $(dmesg) is: [ 69.013084] snd_hda_intel 0000:00:1f.3: Too big adjustment 128
> What is causing this issue and how can I solve it?

The warning message is likely a red herring and it's irrelevant from
your bug itself.  The driver has some small number of periods to be
put at the beginning of the buffer for correcting the reported
position.  Judging from the message, you seem to have set a quite high
rate and a very shot number, and that's possibly too tight to operate
properly.


Takashi

Intel HDA Driver Issue With PCM RW Interleaved

[Ryan McClue]

Hi,
I'm trying to play a simple PCM buffer single threaded without buffer underrun.
Program source:https://gist.github.com/takaswie/d8a9553c230971d974a3dfc759b2ff67
(struct snd_pcm_sw_params).start_thresholdis altered so as to allow specifying the start of audio frame transmission and prevent underrun issues.
In my application, onioctl(fd, SNDRV_PCM_IOCTL_STATUS, &status), I receive:File descriptor in bad state
Output of$(dmesg)is:[ 69.013084] snd_hda_intel 0000:00:1f.3: Too big adjustment 128

From a previous response on thealsa-develmailing list:

> The driver has some small number of periods to be
> put at the beginning of the buffer for correcting the reported
> position. Judging from the message, you seem to have set a quite high
> rate and a very shot number, and that's possibly too tight to operate
> properly.

However, I'm not sure what to change in this code to address this issue and get it to work. I tried alteringMY_PLAYBACK_DURATIONandparams.start_thresholdbut get the same error.

The program dumps the current hardware parameters:

Current hardware parameters:
  Changed parameters:
    access
    format
    subformat
    sample-bits
    frame-bits
    channels
    rate
    period-time
    period-size
    period-bytes
    periods
    buffer-time
    buffer-size
    buffer-bytes
    tick-time
  Runtime parameters:
    access:
      readwrite-interleaved
    format:
      s16-le
    subformat:
      std
    sample-bits:
      [16, 16], integer,
    frame-bits:
      [32, 32], integer,
    channels:
      [2, 2], integer,
    rate:
      [48000, 48000], integer,
    period-time:
      (666, 667),
    period-size:
      [32, 32], integer,
    periods:
      [32, 32], integer,
    buffer-time:
      (21333, 21334),
    buffer-size:
      [1024, 1024], integer,
    buffer-bytes:
      [4096, 4096], integer,
    tick-time:
      [0, 0], integer,
    info:
      mmap
      mmap-valid
      interleaved
      block-transfer
      pause
      sync-start
      no-period-wakeup
      has-wall-clock
      has-link-atime
      has-link-synchronized-atime
      most-significant-bits:    16
      rate_num: 48000
      rate_den: 1

Thanks