ALSA: hda/hdmi: fix failures at PCM open on Intel ICL and later #2240
Conversation
kv2019i
requested review from
bardliao,
dbaluta,
lgirdwood,
plbossart and
ranj063
as code owners
kv2019i
force-pushed
the
topic/hdmi-icljsl-pcm-open-bugfix
branch
from
f806c2d to
3aa695f
Compare
|
@kv2019i I'll let you merge this, i don't recall if this is the final version or still WIP. |
|
@kv2019i Good to catch the initialization issue. If we go through the code deeply, it seems we should not set intel_hsw_fixup flag for ICL as the latest platforms are not hsw like platform. Without the flag, the connection list can be initialized correctly. We are wrongly use this flag now. And I can also see some potential issues if we set intel_hsw_fixup flag on ICL. |
|
@libinyang wrote:
That's a good point, thanks. Let me do some tests on this. The hsw_fixup flag is used for multiple things and I don't think all have been fixed for ICL. E.g. it's used for hdmi_add_pin(). But if the snd_hda_get_raw_connections() works correctly on ICL, we could indeed remote the exception for ICL and newer. @plbossart @ranj063 Let's not merge this yet. |
|
@libinyang OK, I tried without the hsw_fixup, but it still fails. On pins with monitor connected, snd_hda_get_raw_connections() works ok and full list of cvt nids is returned, but on pins with no monitor connected at time of probe, no connections are returned. So maybe best to go with current patch still. UPDATE: @libinyang I tested with ICL and CFL systems without the fixup and the results are similar, so I'll keep the fixup. |
|
I sent this upstream now, let's see if we need to adjust more before merging to sof-dev. |
When HDMI PCM devices are opened in a specific order, with at least one HDMI/DP receiver connected, ALSA PCM open fails to -EBUSY on the connected monitor, on recent Intel platforms (ICL/JSL and newer). While this is not a typical sequence, at least Pulseaudio does this every time when it is started, to discover the available PCMs. The rootcause is an invalid assumption in hdmi_add_pin(), where the total number of converters is assumed to be known at the time the function is called. On older Intel platforms this held true, but after ICL/JSL, the order how pins and converters are in the subnode list as returned by snd_hda_get_sub_nodes(), was changed. As a result, information for some converters was not stored to per_pin->mux_nids. And this means some pins cannot be connected to all converters, and application instead gets -EBUSY instead at open. The assumption that converters are always before pins in the subnode list, is not really a valid one. Fix the problem in hdmi_parse_codec() by introducing separate loops for discovering converters and pins. BugLink: thesofproject#1978 BugLink: thesofproject#2216 BugLink: thesofproject#2217 Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200703153818.2808592-1-kai.vehmanen@linux.intel.com Signed-off-by: Takashi Iwai <tiwai@suse.de> (cherry picked from commit 5627503)
The HDMI codec driver has two debug traces printed from different functions but with identical message content: "HDMI: hinfo 000000006a6b84d9 not registered" Fix this duplication and also add a bit more context in addition to raw object pointer, to help analysis of kernel logs. Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com> Link: https://lore.kernel.org/r/20200703153818.2808592-2-kai.vehmanen@linux.intel.com Signed-off-by: Takashi Iwai <tiwai@suse.de> (cherry picked from commit 90670fd)
kv2019i
force-pushed
the
topic/hdmi-icljsl-pcm-open-bugfix
branch
from
3aa695f to
6bdaac1
Compare
|
Merged upstream. As this fixes multiple SOF issues, I'll go ahead and merge the upstream version to sof-dev right away. |
5 participants
ALSA: hda/hdmi: fix failures at PCM open on Intel ICL and later
Bugfix and one trace improvement patch.
Fixes: #1978
Fixes: #2216
Fixes: #2217