Re: [PATCH] ALSA: ymfpci: avoid changing control access under spinlock

[Runyu Xiao]

Hi,

On Mon, Sep 28, 2026 at 8:54 PM Takashi Iwai <tiwai@xxxxxxx> wrote:

> Although the intention is clear, this is just an overkill. There is no real
> race, and introducing the offloading work just for using
> snd_ctl_activate_id() makes little sense for this old code.
>
> If any, we should rather make snd_ctl_activate_id() to be callable from the
> irq context, instead. But this would need a redesign of the control-led
> layer implementation -- that's the very reason of being sleepable context.
>
> thanks,
>
> Takashi

Thanks for the review. I understand your point and will drop this
workqueue-based patch. Making snd_ctl_activate_id() callable from IRQ context
would require a separate redesign of the control layer, so I will not pursue
it in this driver change.

Regards,
Runyu