Hi Pierre,
On 8/30/2024 2:52 AM, Pierre-Louis Bossart wrote:
>> +/* Stream disable request timeout during USB device disconnect */
>> +#define DEV_RELEASE_WAIT_TIMEOUT 10000 /* in ms */
> 10s really? That seems rather large for a stream disable timeout...
Hmm, yes that is overkill, will adjust it accordingly.
>> +static struct snd_usb_platform_ops offload_ops = {
>> + .connect_cb = qc_usb_audio_offload_probe,
>> + .disconnect_cb = qc_usb_audio_offload_disconnect,
>> + .suspend_cb = qc_usb_audio_offload_suspend,
>> +};
> You probably want to explain why there's no .resume_cb?
>
> The comments mention also that the suspend_cb has to stop playback, but
> then who resumes playback :-)
>
I can add a comment. Ideally, the suspend_cb is only used for the case of PM suspend/system suspend. If usb autosuspend is enabled, then the QC offload driver will handle the voting based on the audio stream being active or not. Is there a use case where the ASoC layer re-opens any previously active audio streams so that userspace doesn't have to? Currently, I was under the assumption that the audio stream would have to be re-opened by the application.
Thanks
Wesley Cheng
[Index of Archives]
[Pulseaudio]
[Linux Audio Users]
[ALSA Devel]
[Fedora Desktop]
[Fedora SELinux]
[Big List of Linux Books]
[Yosemite News]
[KDE Users]