[alsa-devel] [PATCH 2/2] [RFC] ALSA: hda: hdmi - preserve non-MST PCM routing for Intel platforms

[Takashi Iwai]

On Tue, 03 Dec 2019 14:48:10 +0100,
Kai Vehmanen wrote:
> 
> Hi Takashi and Nikhil,
> 
> On Mon, 2 Dec 2019, Kai Vehmanen wrote:
> > yup, -1 for non-MST, that was just plain wrong in my Friday mail. The 
> > problem seems to be pipe id is positive in some non-MST cases (like 
> > https://github.com/thesofproject/linux/issues/1536 ), which is very 
> > suprising looking at e.g. intel_audio_codec_enable(), but that seems to be 
> > happening nevertheless.
> 
> ok, got graphics driver log from this case now, and this turns out to be a 
> system with an unusual converter setup for HDMI output. I.e. from user 
> point of view it's HDMI, but for graphics driver it looks like a DP 
> connection (with MST enabled which is not so common but apparently is the 
> case). So i915 driver is working as expected and in patch_hdmi we should 
> handle it as a normal MST case.
> 
> Takashi, considering the commit message is wrong, I think it's better
> to revert the "ALSA: hda: hdmi - preserve non-MST PCM routing for Intel 
> platforms" patch. Should I send a revert?
> 
> It also doesn't fully preseve the old routing as in the case of a single 
> DP-MST monitor (with dev_id=1), hdmi_find_pcm_slot() will not pick the 
> preferred PCM "per_pin->pin_nid_idx", but will choose "spec->num_nids + 
> (per_pin->dev_id - 1)". So seems better to just drop it.

Well, if we want to keep the old behavior for compatibility just to be
sure, how about a patch like below?


Takashi

--- a/sound/pci/hda/patch_hdmi.c
+++ b/sound/pci/hda/patch_hdmi.c
@@ -1348,21 +1348,18 @@ static int hdmi_find_pcm_slot(struct hdmi_spec *spec,
 	 * with the legacy static per_pin-pcm assignment that existed in the
 	 * days before DP-MST.
 	 *
+	 * Intel DP-MST prefers this legacy behavior for compatibility, too.
+	 *
 	 * per_pin of m!=0 prefers to get pcm=(num_nids + (m - 1)).
 	 */
 
-	if (per_pin->dev_id == 0) {
+	if (per_pin->dev_id == 0 || spec->intel_hsw_fixup) {
 		if (!test_bit(per_pin->pin_nid_idx, &spec->pcm_bitmap))
 			return per_pin->pin_nid_idx;
 	} else {
 		i = spec->num_nids + (per_pin->dev_id - 1);
 		if (i < spec->pcm_used && !(test_bit(i, &spec->pcm_bitmap)))
 			return i;
-
-		/* keep legacy assignment for dev_id>0 on Intel platforms */
-		if (spec->intel_hsw_fixup)
-			if (!test_bit(per_pin->pin_nid_idx, &spec->pcm_bitmap))
-				return per_pin->pin_nid_idx;
 	}
 
 	/* have a second try; check the area over num_nids */