On Mon, 07 Aug 2023 16:26:53 +0200, Pierre-Louis Bossart wrote:
On 8/7/23 04:00, Maarten Lankhorst wrote:
Now that we can use -EPROBE_DEFER, it's no longer required to spin off the snd_hdac_i915_init into a workqueue.
Use the -EPROBE_DEFER mechanism instead, which must be returned in the probe function.
I don't think this patch is aligned with the previous discussions. What we agreed on is that snd_hdac_i915_init() would be called from and not from the workqueue.
But this patch also moves all codec initialization out of the workqueue.
I think we need two callbacks for device-specific initilization, one that is called from the probe function and one from the workqueue, otherwise we'll have a structure that differs from the snd-hda-intel - which would be rather silly in terms of support/debug.
I realize there's quite a bit of surgery involved, and most likely the SOF folks should provide this patch for you to build on.
So this patch looks like the only significant concern in the whole patch set. Can we reach to some agreement for merging to 6.6 in time?
thanks,
Takashi