At Wed, 8 Apr 2015 23:24:52 -0700, Matthew Garrett wrote:

On Wed, Apr 8, 2015 at 10:20 PM, Takashi Iwai tiwai@suse.de wrote:

...

At Wed, 8 Apr 2015 18:53:48 -0700, Matthew Garrett wrote:

...

Modern hardware will often have multiple HDA devices, and the desired power saving configuration may vary depending on the codecs attached to each of them. Push the power_save value down to the individual codec structures, keeping the module parameter as a global control mechanism for compatibility purposes.

This patch doesn't apply any longer at all to the recent code, since the power saving stuff has been already translated to the standard runtime PM. Take a look at linux-next tree.

Ah! Wonderful!

...

Also we can't break the current behavior controlling the power save via option. Many programs rely on this mechanism, thus changing this would break its behavior.

Yeah, I kept that behaviour - anything that touches /sys/module/snd_hda_intel/parameters/power_save just changes the settings for all codecs.

...

In the latest code, the common control via power_save module option is still there for compatibility, but you can also adjust the value for each codec in runtime attribute of each codec sysfs in addition.

That's ideal - I'll pull down next and rewrite my udev rules. Thanks!

Let me know if you find anything broken with the new code. It's pretty new, so any intensive testing is appreciated.

thanks,

Takashi