Although we modified the codec shutdown callback to perform
runtime-suspend, it's still not fully effective, as this may be
resumed again at any time later.
For fixing such an unwanted resume, this patch replaces
pm_runtime_suspend() with pm_runtime_force_suspend(), and call
pm_runtime_disable() afterward. It assures to keep the device
suspended.
Also for code simplification, we apply the code unconditionally; when
it's been already suspended, nothing would happen by calls of
snd_pcm_suspend_all() and pm_runtime_force_suspend(), just proceed to
pm_runtime_disable().
As reboot_notify callback is no longer used by the codec core, let's
get rid of the unused code. Conexant codec needs a slight code change
as it used to call the reboot_notify at the codec removal, too.
So far we have a few workarounds at shutdown for each codec,
e.g. turning off the display power and setting the codec to D3.
But all those are basically a part of the suspend procedure.
Moreover, the streams are still active after that call, hence it might
hit the update on the codec that has been already put to D3.
In this patch, instead of calling each reboot_notify callback, simply
put the codec into the runtime-suspended state after the manual
suspend of all PCM streams. It makes the code and the behavior more
consistent.
The reboot_notify callback is no longer used after this patch, and
will be cleaned up later.
Conexant codecs have a workaround for the noise at shutdown to turn
off EAPD, but it wasn't applied at suspend. In the later patch, we'll
switch from reboot_notify callback to the general suspend-at-shutdown,
so let's apply the workaround to the suspend case, too.
Make sure the HDA driver's display power reference is released during
shutdown/reboot.
During the shutdown/reboot sequence the pci device core calls the
pm_runtime_resume handler for all devices before calling the driver's
shutdown callback and so the HDA driver's runtime resume callback will
acquire a display power reference (on HSW/BDW). This triggers a power
reference held WARN on HSW/BDW in the i915 driver's subsequent shutdown
handler, which expects all display power references to be released by
that time.
Since the HDA controller is stopped in the shutdown handler in any case,
let's follow here the same sequence as the one during runtime suspend.
This will also reset the HDA link and drop the display power reference,
getting rid of the above WARN.
Tested on HSW.
v2:
- Fix the build for CONFIG_PM=n (Takashi)
- s/__azx_runtime_suspend/azx_shutdown_chip/
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3618
References: https://<email address hidden>
Reported-and-tested-by: Thomas Voegtle <email address hidden>
Signed-off-by: Imre Deak <email address hidden>
Link: https://<email address hidden>
Signed-off-by: Takashi Iwai <email address hidden>
(cherry picked from commit 472e18f63c425dda97b888f40f858ea54e3efc17)
Signed-off-by: You-Sheng Yang <email address hidden>
During reboot, when the CS42L42 powers down, pops and clicks
may occur due to the codec not being shutdown gracefully.
This can be fixed by going through the suspend sequence,
which shuts down the codec cleanly inside the reboot_notify
hook, which is called on reboot.
Signed-off-by: Stefan Binding <email address hidden>
Signed-off-by: Vitaly Rodionov <email address hidden>
Signed-off-by: You-Sheng Yang <email address hidden>
bbce3d5...
by
Stefan Binding <email address hidden>
ALSA: hda/cs8409: Prevent pops and clicks during suspend
Some of the register values set for type detection cause pops during suspend,
ensure these are cleaned up after type detection completes, as well
ensuring that these are cleared when we suspend.
Signed-off-by: Stefan Binding <email address hidden>
Signed-off-by: Vitaly Rodionov <email address hidden>
Signed-off-by: You-Sheng Yang <email address hidden>