Re: [alsa-devel] No speaker output with kernel 3.15 and later

8 Jul 2014

At Mon, 07 Jul 2014 17:05:43 +0100,
Steven Newbury wrote:
...
On Mon, 2014-07-07 at 17:41 +0200, Takashi Iwai wrote:
...
At Mon, 07 Jul 2014 16:34:10 +0100,
...
Steven Newbury wrote:
...
...
On Tue, 2014-07-01 at 13:54 +0200, Takashi Iwai wrote:
...
...
...
At Tue, 01 Jul 2014 12:49:32 +0100,
...
...
...
Steven Newbury wrote:
...
...
...
...
On Mon, 2014-06-30 at 11:24 +0200, Takashi Iwai wrote:
...
...
...
...
...
At Sun, 29 Jun 2014 15:09:56 +0100,
...
...
...
...
...
Steven Newbury wrote:
...
...
...
...
...
>    I sent
...
...
...
...
...
>   Takashi an email mentioning this regression a month ago,
> and
...
...
...
...
...
>    it's
...
...
...
...
...
>   taken me this long to send this email... :-$
...
...
...
...
...
>    Since updating to
...
...
...
...
...
>   kernel versions >=3.15 I've had no audio from my
...
...
...
...
...
>    laptop speakers.
...
...
...
...
...
>   Earlier versions worked fine.
...
...
...
...
...
>    It's a CLEVO W270EUQ laptop.  The
...
...
...
...
...
>   output mixer device is showing up,
...
...
...
...
...
>    but no sound comes out at any
...
...
...
...
...
>   volume level.
...
...
...
...
...
>    As requested I've attached compressed alsa-info.sh
...
...
...
...
...
>   output for both the
...
...
...
...
...
>    working and non-working cases.
...
...
...
...
...
>    I'm perfectly
...
...
...
...
...
>   willing to test patches etc.
...
...
...
...
...
Hm, I see no significant difference between 3.14 and 3.15
...
...
...
...
...
alsa-info.sh
...
...
...
...
...
outputs, and there's been very little change for VIA codecs
...
...
...
...
...
between
...
...
...
...
...
these versions.  So, in short, I have no idea what went
wrong.
...
...
...
...
...
Could you double-check whether it's really a kernel issue?
...
...
...
...
...
That is, switch between 3.14 and 3.15 and confirm the
problem
...
...
...
...
...
happens
...
...
...
...
...
only with 3.15.  Also, for testing the sound, try like
...
...
...
...
...
      % aplay -Dplughw -vv foo.wav

...
...
...
...
Checked again.
...
...
...
...
3.16-rc2: silence
...
...
...
...
3.14.3: works fine
...
...
...
Then I'd suggest to bisect between 3.14 and 3.15.
...
...
Somewhat surprisingly I ended up here:
...
...
dcb32ecd9a533f47ab652c5c5680bc50a7a822cd is the first bad commit
...
...
commit dcb32ecd9a533f47ab652c5c5680bc50a7a822cd
...
...
Author: Anssi Hannula anssi.hannula@iki.fi
...
...
Date:   Tue Apr 8 12:36:42 2014 +0300
...
...
 ALSA: hda - Do not assign streams in reverse order

...
...
 Currently stream numbers are assigned in reverse order.

...
...
 Unfortunately commit 7546abfb8e1f9933b5 ("ALSA: hda -

Increment
...
...
 default stream numbers for AMD HDMI controllers") assumed

this was not
...
...
 the case (specifically, it had the "old cards had single

device only"
...
...
 => "extra unused stream numbers do not matter" assumption),

causing
...
...
 non-working audio regressions for AMD Radeon HDMI users.

...
...
 Change the stream numbers to be assigned in forward order.

...
...
 The benefit is that regular audio playback will still work

even if the
...
...
 assumed stream count is too high, downside is that a too high

stream
...
...
 count may remain hidden.

...
...
 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=77002

...
...
 Reported-by: Christian GÃ¼del <cg@dmesg.ch>

...
...
 Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>

...
...
 Tested-by: Christian GÃ¼del <cg@dmesg.ch> # 3.14

...
...
 Cc: Alex Deucher <alexander.deucher@amd.com>

...
...
 Signed-off-by: Takashi Iwai <tiwai@suse.de>

...
...
:040000 040000 6ed3a982a9600333b07a42169db5173fa75b8618
b7ff28d00007f988221ba58c9d3f501fdb3a7583 M        sound
...
...
Interesting.  Does reverting this commit fix the issue?
...
Just run "git revert dcb32ecd9a533f47ab652c5c5680bc50a7a822cd", and
...
test it.
...
I just did this on top of my rebased local git tree off linus/master
   and indeed reverting does make the speakers work again.*
OK, thanks for confirmation.
So, by some reason, the controller (or codec) needs to use the last
entry (in this case it's Panther Point).  I cannot understand why this
fails, but it's so...
The patch below does revert conditionally only for Intel controller
chips.  This should work instead of git revert.  Let me know if it
really works for you.
But, we still want to understand why.  Maybe assigning the device
manually to an explicit value and figuring out which dev# doesn't work
would be helpful.
thanks,
Takashi
---

diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c
index 06f39c126ba4..8337645aa7a5 100644
--- a/sound/pci/hda/hda_controller.c
+++ b/sound/pci/hda/hda_controller.c
@@ -194,7 +194,8 @@ azx_assign_device(struct azx *chip, struct snd_pcm_substream *substream)
    			dsp_unlock(azx_dev);
    			return azx_dev;
    		}
-			if (!res)
+			if (!res ||
+			    (chip->driver_caps & AZX_DCAPS_REVERSE_ASSIGN))
    			res = azx_dev;
    	}
    	dsp_unlock(azx_dev);
diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c
index 75b52c4cd70d..dfc88831f1be 100644
--- a/sound/pci/hda/hda_intel.c
+++ b/sound/pci/hda/hda_intel.c
@@ -272,7 +272,7 @@ enum {
 /* quirks for Intel PCH */
 #define AZX_DCAPS_INTEL_PCH_NOPM \
    (AZX_DCAPS_SCH_SNOOP | AZX_DCAPS_BUFSIZE | \
-	 AZX_DCAPS_COUNT_LPIB_DELAY)
+	 AZX_DCAPS_COUNT_LPIB_DELAY | AZX_DCAPS_REVERSE_ASSIGN)
#define AZX_DCAPS_INTEL_PCH \
    (AZX_DCAPS_INTEL_PCH_NOPM | AZX_DCAPS_PM_RUNTIME)
diff --git a/sound/pci/hda/hda_priv.h b/sound/pci/hda/hda_priv.h
index 33eb5d027d38..949cd437eeb2 100644
--- a/sound/pci/hda/hda_priv.h
+++ b/sound/pci/hda/hda_priv.h
@@ -166,6 +166,7 @@ enum { SDI0, SDI1, SDI2, SDI3, SDO0, SDO1, SDO2, SDO3 };
 #define AZX_DCAPS_BUFSIZE	(1 << 21)	/* no buffer size alignment */
 #define AZX_DCAPS_ALIGN_BUFSIZE	(1 << 22)	/* buffer size alignment */
 #define AZX_DCAPS_4K_BDLE_BOUNDARY (1 << 23)	/* BDLE in 4k boundary */
+#define AZX_DCAPS_REVERSE_ASSIGN (1 << 24)	/* Assign devices in reverse order */
 #define AZX_DCAPS_COUNT_LPIB_DELAY  (1 << 25)	/* Take LPIB as delay */
 #define AZX_DCAPS_PM_RUNTIME	(1 << 26)	/* runtime PM support */
 #define AZX_DCAPS_I915_POWERWELL (1 << 27)	/* HSW i915 powerwell support */

    