[alsa-devel] [PATCH] ALSA: hda - Autosuspend controller after probe even if codecs are already suspended
From: Ville Syrjälä ville.syrjala@linux.intel.com
azx_probe_continue() uses pm_runtime_put_noidle() to drop the rpm usage_count, which means that if it's the last reference the autosuspend of the controller won't actually happen. So if the codecs autosuspend before the azx_probe_continue() drops the last reference we'll fail to autosuspend the controller. This does happen in practice, but not every time. As can be seen in [1] the controller autosuspend attempt fails due to the usage_count when suspending the codecs. A bit later we see the the contoller usage_count dropping to zero without further attempts at autosuspend.
Fix the problem by using pm_runtime_put_autosuspend() instead, which will kick off the autosuspend of the controller even if the codecs are already asleep. As can be seen in [2] the controller autosuspend still fails while suspending the codecs, but later on we see another autosuspend attempt after dropping the usage_count to 0.
I was also a bit worried that there might still be a race between the controller autosuspend and the rest of the code in azx_probe_continue(). So I also tried replacing the the put_noidle() with put_sync_suspend(). No explosions occurred, so I'm somewhat satisfied that there are no serious problems in this area.
[1] kworker/1:2-122 [001] .... 63.661310: __pm_runtime_suspend: hdaudioC0D0 usage_count 0 kworker/1:2-122 [001] d..2 63.661316: rpm_suspend: hdaudioC0D0 flags-d cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/1:2-122 [001] d..1 63.661317: rpm_check_suspend_allowed: hdaudioC0D0 retval 0 kworker/1:2-122 [001] d..2 63.661332: rpm_return_int: rpm_suspend+0x406/0x5e8:hdaudioC0D0 ret=0 kworker/1:1-72 [001] d..2 63.661543: rpm_suspend: hdaudioC0D0 flags-a cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/1:1-72 [001] d..1 63.661544: rpm_check_suspend_allowed: hdaudioC0D0 retval 0 kworker/1:1-72 [001] .... 63.661545: hda_codec_runtime_suspend: hdaudioC0D0 suspend kworker/1:1-72 [001] d..2 63.661614: rpm_idle: 0000:00:03.0 flags-1 cnt-1 dep-0 auto-1 p-0 irq-0 child-0 kworker/1:1-72 [001] d..1 63.661615: rpm_check_suspend_allowed: 0000:00:03.0 usage_count 1 kworker/1:1-72 [001] d..1 63.661615: rpm_check_suspend_allowed: 0000:00:03.0 retval -11 kworker/1:1-72 [001] d..2 63.661616: rpm_return_int: rpm_idle+0x249/0x487:0000:00:03.0 ret=-11 kworker/1:1-72 [001] d..2 63.661616: rpm_return_int: rpm_suspend+0x406/0x5e8:hdaudioC0D0 ret=0 kworker/1:2-122 [001] d..2 63.664834: rpm_idle: hdaudioC0D0 flags-8 cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/1:2-122 [001] d..1 63.664835: rpm_check_suspend_allowed: hdaudioC0D0 retval 1 kworker/1:2-122 [001] d..2 63.664836: rpm_return_int: rpm_idle+0x249/0x487:hdaudioC0D0 ret=-11 kworker/1:2-122 [001] d..2 63.664841: rpm_idle: hdaudioC0D0 flags-8 cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/1:2-122 [001] d..1 63.664841: rpm_check_suspend_allowed: hdaudioC0D0 retval 1 kworker/1:2-122 [001] d..2 63.664841: rpm_return_int: rpm_idle+0x249/0x487:hdaudioC0D0 ret=-11 kworker/1:2-122 [001] .... 63.664842: azx_probe_continue: 0000:00:03.0 usage_count=0
[2] kworker/0:0-4 [000] .... 50.354567: __pm_runtime_suspend: hdaudioC0D0 usage_count 0 kworker/0:0-4 [000] d..2 50.354574: rpm_suspend: hdaudioC0D0 flags-d cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:0-4 [000] d..1 50.354575: rpm_check_suspend_allowed: hdaudioC0D0 retval 0 kworker/0:0-4 [000] d..2 50.354589: rpm_return_int: rpm_suspend+0x406/0x5e8:hdaudioC0D0 ret=0 kworker/0:2-135 [000] d..2 50.354809: rpm_suspend: hdaudioC0D0 flags-a cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:2-135 [000] d..1 50.354810: rpm_check_suspend_allowed: hdaudioC0D0 retval 0 kworker/0:2-135 [000] .... 50.354816: hda_codec_runtime_suspend: hdaudioC0D0 suspend kworker/0:2-135 [000] d..2 50.354908: rpm_idle: 0000:00:03.0 flags-1 cnt-1 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:2-135 [000] d..1 50.354909: rpm_check_suspend_allowed: 0000:00:03.0 usage_count 1 kworker/0:2-135 [000] d..1 50.354909: rpm_check_suspend_allowed: 0000:00:03.0 retval -11 kworker/0:2-135 [000] d..2 50.354909: rpm_return_int: rpm_idle+0x249/0x487:0000:00:03.0 ret=-11 kworker/0:2-135 [000] d..2 50.354910: rpm_return_int: rpm_suspend+0x406/0x5e8:hdaudioC0D0 ret=0 kworker/0:0-4 [000] d..2 50.373791: rpm_idle: hdaudioC0D0 flags-8 cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:0-4 [000] d..1 50.373792: rpm_check_suspend_allowed: hdaudioC0D0 retval 1 kworker/0:0-4 [000] d..2 50.373793: rpm_return_int: rpm_idle+0x249/0x487:hdaudioC0D0 ret=-11 kworker/0:0-4 [000] d..2 50.373797: rpm_idle: hdaudioC0D0 flags-8 cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:0-4 [000] d..1 50.373798: rpm_check_suspend_allowed: hdaudioC0D0 retval 1 kworker/0:0-4 [000] d..2 50.373798: rpm_return_int: rpm_idle+0x249/0x487:hdaudioC0D0 ret=-11 kworker/0:0-4 [000] .... 50.373799: __pm_runtime_suspend: 0000:00:03.0 usage_count 0 kworker/0:0-4 [000] d..2 50.373800: rpm_suspend: 0000:00:03.0 flags-d cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:0-4 [000] d..1 50.373800: rpm_check_suspend_allowed: 0000:00:03.0 retval 0 kworker/0:0-4 [000] d..2 50.373803: rpm_return_int: rpm_suspend+0x406/0x5e8:0000:00:03.0 ret=0 kworker/0:0-4 [000] d..2 50.385164: rpm_suspend: 0000:00:03.0 flags-a cnt-0 dep-0 auto-1 p-0 irq-0 child-0 kworker/0:0-4 [000] d..1 50.385165: rpm_check_suspend_allowed: 0000:00:03.0 retval 0 kworker/0:0-4 [000] .... 50.385174: azx_runtime_suspend: 0000:00:03.0 azx suspend releaseing power well kworker/0:0-4 [000] .... 50.385179: azx_runtime_suspend: 0000:00:03.0 azx suspend kworker/0:0-4 [000] d..2 50.386872: rpm_return_int: rpm_suspend+0x406/0x5e8:0000:00:03.0 ret=0
Cc: Takashi Iwai tiwai@suse.de Signed-off-by: Ville Syrjälä ville.syrjala@linux.intel.com --- sound/pci/hda/hda_intel.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c index e5240cb3749f..2624cfe98884 100644 --- a/sound/pci/hda/hda_intel.c +++ b/sound/pci/hda/hda_intel.c @@ -2145,7 +2145,7 @@ static int azx_probe_continue(struct azx *chip) azx_add_card_list(chip); snd_hda_set_power_save(&chip->bus, power_save * 1000); if (azx_has_pm_runtime(chip) || hda->use_vga_switcheroo) - pm_runtime_put_noidle(&pci->dev); + pm_runtime_put_autosuspend(&pci->dev);
out_free: if (chip->driver_caps & AZX_DCAPS_I915_POWERWELL
On Fri, 26 Feb 2016 18:39:57 +0100, ville.syrjala@linux.intel.com wrote:
From: Ville Syrjälä ville.syrjala@linux.intel.com
azx_probe_continue() uses pm_runtime_put_noidle() to drop the rpm usage_count, which means that if it's the last reference the autosuspend of the controller won't actually happen. So if the codecs autosuspend before the azx_probe_continue() drops the last reference we'll fail to autosuspend the controller. This does happen in practice, but not every time. As can be seen in [1] the controller autosuspend attempt fails due to the usage_count when suspending the codecs. A bit later we see the the contoller usage_count dropping to zero without further attempts at autosuspend.
Fix the problem by using pm_runtime_put_autosuspend() instead, which will kick off the autosuspend of the controller even if the codecs are already asleep. As can be seen in [2] the controller autosuspend still fails while suspending the codecs, but later on we see another autosuspend attempt after dropping the usage_count to 0.
I was also a bit worried that there might still be a race between the controller autosuspend and the rest of the code in azx_probe_continue(). So I also tried replacing the the put_noidle() with put_sync_suspend(). No explosions occurred, so I'm somewhat satisfied that there are no serious problems in this area.
A good catch!
The change should be safe, as this is the point where already all initialization has been done.
I queued it to for-next branch, since this is no super-urgent issue to be fixed in 4.5.
thanks,
Takashi
participants (2)
-
Takashi Iwai
-
ville.syrjala@linux.intel.com