Thanks. Interestingly, the lsusb output after resume shows some errors. Could you give the dmesg output after resume (that is indicating the bug)? It might be irrelevant, but just to be sure. Also, I'm building a test kernel with a patch that might be related with the USB audio driver like yours. It's still built in OBS home:tiwai:bsc1206766 repo. Once after the build finishes, could you give it a try?