-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 W dniu 2015-10-05 o 06:21, Mikhail Morfikov pisze: > W dniu 2015-10-05 o 01:27, Tanu Kaskinen pisze: >> On Sun, 2015-10-04 at 20:28 +0200, Mikhail Morfikov wrote: >>> In the case of snd_hda_intel kernel module, there's two options called >>> power_save and power_save_controller . I've set them to the following >>> values: >>> >>> # systool -v -m snd_hda_intel | grep power >>> power_save = "10" >>> power_save_controller= "Y" >>> >>> According to archlinux wiki (https://wiki.archlinux.org/index.php/Power_management#Audio), >>> the power_save parameter is measured in seconds, and it should idle the >>> audio card after 10s. >>> >>> The question is how could I know whether this works or not? Is there >>> any way to be sure that the powersave mode was activated? I've heard >>> that there's could be some "clicking" when the state of the sound >>> card changes, but I don't hear anything. Some people reported that the >>> volume behaves really strange, but it seems that it works just fine on >>> my machine. >>> >>> Does the powersave mode in the kernel module is somehow related to the >>> following output? >>> >>> $ pacmd list-sinks | egrep -i "state|suspend" >>> state: SUSPENDED >>> suspend cause: IDLE >>> >>> I noticed that when some app plays audio stream, the state becomes >>> RUNNING. When the app stops playing, the state changes to IDLE, and >>> about 5s later, it changes to the state that you can see above. This >>> behavior is the same no matter what I set in the power_save option >>> in the kernel module. I've tested 0, 1 and 10, and actually I didn't >>> notice any difference. >>> >>> So, is the powersave mode enabled or not? > >> I'm not familiar with the kenel module, so I can't answer most of your >> questions. However, I can tell you that the PulseAudio sink state (IDLE >> or SUSPENDED) is not the same thing as the kernel driver's power save >> state. Based on your description, the driver enters the power save mode >> 10 seconds after the PulseAudio sink state has changed to SUSPENDED >> (suspending means that PulseAudio closes the alsa device). > > > So it should work. Though I have no idea how to check the state of the > soundcard. > It looks like the powersave feature doesn't work for my device. According to the following log: $ cat /proc/asound/card0/codec\#0 | grep -i power Power states: D0 D1 D2 D3 CLKSTOP EPSS Power: setting=D0, actual=D0 $ cat /proc/asound/card0/codec\#3 | grep -i power Power states: D0 D3 Power: setting=D0, actual=D0 Power states: D0 D3 Power: setting=D0, actual=D0 Power states: D0 D3 Power: setting=D0, actual=D0 Power states: D0 D3 Power: setting=D0, actual=D0 Power states: D0 D3 Power: setting=D0, actual=D0 Power states: D0 D3 Power: setting=D0, actual=D0 there are power switches, but none of them switch after the period set in /sys/module/snd_hda_intel/parameters/power_save . There's info at https://www.kernel.org/doc/Documentation/power/pci.txt that "The implementation of the PCI PM Spec is optional for conventional PCI devices", so apparently it's not supported here and that's why nothing happens after 10s. -----BEGIN PGP SIGNATURE----- iQIcBAEBCgAGBQJWErv2AAoJEM0EaBB3G2Ug+QYP/jM1E6xrWSke0X85jKnfsNKG 8Ob7n7qkNIX5FthNzAKkoX+4p8G6U3EAu/HahW7OzdoYZnF7YhJLe9+IhMPhHCdR ldpA0mPtS2j7CeqQfND8wtJb7peXND3P54RliOKyrC7Ip9gmiJNuDTDrG+x7Cnw1 xxCbXG3Fenz54gczDew86sZDdgb+HsGrOUO/IHLvyvP+P/i2/Xl2nVsU1/6mIvBB H/ggr5sWfOpnodEfY2rKLIO35Hn/sM3sKgGGc1QEh8r8y6OSBZdeaUqX4Kvzy+YT uYA1Qwxah+wHY/b+pF5gP8ldPTKmwHLeJ4QOcPX+fU4qJvMSd5hblhrlS4czs5rk q03QsgqI+sXn+sHiPlMCLwTrs1sxQpKuRvV+/apE1k4pti7H8mD+qy8V2Lcu62Gw P61GW9cg8ZRP+c6p3INlHKaOjZT7WpPqN5bwwgH2RssiZGIUMuuQrG0c7KMk1Szf 70szXpAdmd24+UEsWsi34+HzZUTalyw+Yh8eENN9yymI/Mg3Uh8fUEmDE2DDmhlc G0VKTBks1KX1JbdWajm8yegVlwxRPRtfizfFStGh4BPtO7eJrSE0wjKZ2z/pHbKX Q1SFcKHLcSGzar+F6N+LrJiTJEGde+ARZUsWb+VBDh2ryU/Ak0m9jbUJuvj/86lI MREffMkqUVkpqeCuIrzy =sdH2 -----END PGP SIGNATURE-----
