Created attachment 211151 [details] alsa-info 4.6-rc1 With 4.6-rc1 I can no longer use my Scarlett 2i4, it is not listed as an input or output device on pavucontrol and on alsamixer it says "This sound device does not have any controls." It is working on mainline 4.5 kernel. I had the same BUG described here: https://lkml.org/lkml/2016/3/15/238 I applied the attached patch and gets rid of the BUG but I still cannot use my 2i4 My 2i4 recieves power, leds turn on and lsusb detects it.
Created attachment 211161 [details] lsusb 4.6-rc1
Created attachment 211171 [details] dmesg 4.6-rc1
Created attachment 211191 [details] config 4.6-rc1
There aren't so many commits in usb-audio driver, so could you try to revert the commit one by one? Start with c0fd9cdf940b5a2b5ffd118831bddf88afe922e4 On the git tree, run "git revert c0fd9cdf940b5a2b5ffd118831bddf88afe922e4" This alone won't change anything, so revert e2703363316278cd4a8880671d38f783f7de18ba together. These will revert the addition of media controller in 4.6-rc1. If these reverts don't help, try to revert e2703363316278cd4a8880671d38f783f7de18ba If this doesn't, try 79289e24194a9d099bf18f200894832c5760cd83 If this still doesn't help, it might be something else. You can copy the whole sound/usb/* files from 4.5, and see whether it works.
Hi Takashi, thanks for the quick reply. I actually had a similar thought that it was related to media controller. Instead of reverting the commits I unset SND_USB_AUDIO_USE_MEDIA_CONTROLLER My 2i4 is now properly detected and audio works fine now. Attached alsa-info. I'll try to further isolate the issue tomorrow.
Created attachment 211211 [details] alsa-info SND_USB_AUDIO_USE_MEDIA_CONTROLLER not set
Thanks, that's already helpful. I found one problem at least. Could you check the patch below? --- a/sound/usb/media.c +++ b/sound/usb/media.c @@ -57,7 +57,7 @@ int media_snd_stream_init(struct snd_usb_substream *subs, struct snd_pcm *pcm, mdev = subs->stream->chip->media_dev; if (!mdev) - return -ENODEV; + return 0; if (subs->media_ctl) return 0;
There are other issues in the recent MC code. Below are the three fix patches (including the previous one). Let me know if these work for you.
Created attachment 211221 [details] Fix patch #1
Created attachment 211231 [details] Fix patch #2
Created attachment 211241 [details] Fix patch #3
Patch #1 by itself fixes the issue, thanks! Patch #2 and #3 apply fine too.