https://bugzilla.kernel.org/show_bug.cgi?id=212955 Bug ID: 212955 Summary: Possible kernel regression USB2 (not USB3) port EDIROL UA-101 (in USB 1.1 mode, not USB2) error -110 Product: Drivers Version: 2.5 Kernel Version: 5.12.0-rc8-next-20210422 Hardware: All OS: Linux Tree: Mainline Status: NEW Severity: normal Priority: P1 Component: USB Assignee: drivers_usb@xxxxxxxxxxxxxxxxxxxxxx Reporter: jaffa225man@xxxxxxxxx Regression: No While working helping to figure out universal EDIROL/Roland/BOSS snd-usb-audio support, I noticed a new issue with the last kernel. Prior to using kernel 5.12.0-rc8-next-20210422, I'd been using 5.9.0-1-rt-amd64 from Debian's package, and it worked on it. I have a recollection of it working on kernel 5.11 too, but I'm away from the testing computer and don't have any notes on it at the moment. The EDIROL UA-101 has a physical switch which sets it to either "Full speed" (USB 1.1) mode, or "High speed" (USB 2) mode. High Speed mode works on all kernels, but its Full speed mode only works either with prior kernels (both USB2 & USB3 port versions) or only USB3 ports on the current kernel. These results are unvaried and quite reproducible on my hardware. Whatever causes it, breaks both capture and playback, but here's just a capture example: arecord -D plughw:UA101 -f S24_3LE -r 48000 -c 2 ./file.wav arecord: main:830: audio open error: Connection timed out (dmesg:) [ 188.956481] usb 4-1.5: cannot initialize interface; error -110: unknown error [ 194.076450] usb 4-1.5: cannot initialize interface; error -110: unknown error [ 199.196407] usb 4-1.5: cannot initialize interface; error -110: unknown error I have yet to narrow down the cause in the code, but plan to try eventually. Takashi Iwai on the alsa-devel list, had this suggestion for me: The error -110 is ETIMEDOUT, and the symptom implies that the problem is rather in USB core side. You'd better ask USB devs. It might be worth to run git bisect if it's a kernel regression and reliably reproducible. It seems I need the git repository to use bisect, and the sheer bandwidth required would take me days or perhaps a week downloading, so if you can suggest a method of downloading just the past 5.11.0 to 5.12.0 history I'd be happy to do that. I'll be glad to test whatever you wish, thanks! -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
