Dne 07. 02. 21 v 21:52 Hans de Goede napsal(a): > Hi Jaroslav, > > On 2/7/21 9:11 PM, Jaroslav Kysela wrote: >> [DO NOT MERGE!] >> [This is just an early proposal for comments] >> [The code is not tested / debugged] >> >> The recent laptops have usually two LEDs assigned to reflect >> the speaker and microphone mute state. This implementation >> adds a tiny layer on top of the control API which calculates >> the state for those LEDs using the driver callbacks. >> >> Two new access flags are introduced to describe the controls >> which affects the audio path settings (an easy path for drivers). >> >> The LED resource can be shared with multiple sound cards with >> this code. The user space controls may be added to the state >> chain, too. >> >> This code should replace the LED code in the HDA driver and >> add a possibility to easy extend the other drivers (ASoC >> codecs etc.). >> >> Note: The snd_ctl_notify_one() changes should be moved to >> a separate patch. I will do this, when this proposal is >> agreed. >> >> Signed-off-by: Jaroslav Kysela <perex@xxxxxxxx> >> Cc: Perry Yuan <Perry.Yuan@xxxxxxxx> >> Cc: Hans de Goede <hdegoede@xxxxxxxxxx> > > Nice I think that having this handled in a generic manner > would be a great improvement. > > I have 2 HP x2 models which are Bay Trail resp. Cherry Trail > based, which means they use Intel's LPE (Low Power Engine) > for audio which uses ASoC. > > These come with detachable USB keyboard-docks which have > a speaker mute LED. I have already added LED-class support to > the HID driver for these, but I never got around to hooking > up a trigger from the ASoC code. > > So if I understand things correctly, then if I add this > patch to a kernel and in the ASoC codec driver add > SNDRV_CTL_ELEM_ACCESS_SPK_LED to one or more of the controls, > and add .default_trigger = "audio-mute" to the led_class_dev > for the LED, then if I toggle the control on / off in > alsamixer this should then control the LED ? Yes, the ALSA control code should do just the LED trigger. The LED class driver is a separate thing. > If I got that right I will certainly give this a try. > > Any advice for which control would be the best to use? > Looking at the code, it assumes that if a control is enabled > that then there will be a route, which makes sense if there > are e.g. headphones and speaker and lineout controls. But what > about chained controls, e.g. separate volume + mute controls > or multiple volume controls chained. With ASoC this is not > unheard of. I guess the answer is to pick the control which > we will also want to use for hw volume-control if/when UCM > profiles grow hw volume-control support ? The proposed implementation just goes through all marked controls and if any of the marked controls is set to value greater than the minimal value, then the path is evaluated as "on" (thus the mute LED should be set to off). The controls tied to the PCM stream (DAC / DMA) should be marked. The physical inputs behind a multiplexer without a "global" volume / switch controls will require a bit different handling. I would just start with something simple and we can add the more complex things on demand. > > Regards, > > Hans > > > > > p.s. This would only / at least add proper support for these LEDs > at the kernel level. ATM these devices which use UCM profiles don't > actually use any kind of hw volume-control. So we would also need to > fix that (in userspace) to get these LEDs to really work when an > user hits the mute key, or lowers the volume to NIL. There's already the hw volume support in UCM and in PA (I added this to PA 14.0). Some drivers like SOF HDA / SoundWire already set the PlaybackMixerElem / CaptureMixerElem values. It's true that the hw volume controls are missing for the Intel SST hardware. The similar situation is for the SoundWire codecs (missing LED controls, missing hw volume configuration in UCM). Jaroslav -- Jaroslav Kysela <perex@xxxxxxxx> Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
