Re: [PATCH v3 2/4] ASoC: SDCA: Populate IRQ data earlier

[Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>]

On 7/20/26 16:37, Charles Keepax wrote:
> On Fri, Jul 17, 2026 at 10:42:46PM +0200, Pierre-Louis Bossart wrote:
>> On 7/16/26 16:26, Charles Keepax wrote:
>>> Currently, the IRQ data (attached Entity/Control/etc) is populated
>>> as the IRQ is requested. However, this can cause issues as
>>> occasionally the setup process wants to access specifics of
>>> an IRQ before the IRQ is actually enabled. To facilitate this
>>> cache all the IRQ data during sdca_irq_populate_early() and make
>>> sdca_irq_populate() simply request the outstanding IRQs. This
>>> also has the advantage that sdca_irq_populate() can now just
>>> iterate through the IRQ array which is much smaller/faster than
>>> going through every Entity in the Function for Controls.
>>>
>>> Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
>>> ---
>>
>> I may be splitting hair but shouldn't this patch be moved
>> earlier?patch1/3-4 seem to be completely tied with the introduction and
>> use of the fixup_controls() callback, I am not sure I see the link with
>> the IRQ stuff. Or conversely move this last if this makes more sense.
> 
> It can be moved to the start of the series, probably not worth a
> spin on its own but if we end up with changes from Mark's
> comments I will do so.
> 
> This is slightly related though as sdca_jack_set_jack is called
> before all the controls are registered and we populate the jack
> masks in there, so moving the IRQs to later means we need that
> information ready at that time which this accomplishes. Although
> it does also just generally make things a bit neater.
> 
>> Also while I am at it, I wouldn't hurt to have an explanation of the
>> races conditions earlier, it's only in patch3 that the race condition is
>> described in details.
> 
> Yeah I should have probably included that in the cover letter
> more. Likewise, I will update that if we end up with another
> spin. If either you or Mark feel strongly enough on these points
> to warrant another spin just for them just let me know.

I do have a new question triggered by the last patch from Shuming: what
happens to the existing users of sdca_irq_populate(), do they also need
to use the new sdca_irq_populate_early()? See my comment in the other
patch "ASoC: rt766: add RT766/RT767 SDCA driver"

If there is no impact no need for a new spin IMHO.