On Sat, 2010-04-03 at 18:58 -0400, Alan Stern wrote: > Come to think of it, let's start by testing whether the problem goes > away when sitd's are never deallocated. Let's just leak them. > > For this, all you have to do is change the line > > list_move(&sitd->sitd_list, &stream->free_list); > > near the end of sitd_complete() to > > list_del(&sitd->sitd_list); > > If you still get fatal errors after making that change, then the > problem has a different cause. Nice! Leaking the sitds eliminated the problem for both my custom board setup with an ISP1568A controller, and also the VT6212 controller setup. Sounds like your original idea of delaying their re-use is what needs to be implemented. If you would like hardware to replicate this problem, let me know and I could probably get some shipped out to you. On the usbaudio side, I wonder if something still needs to be done about the sync URBs for this case. If I re-enable them (SYNC_URBS = 2) after making the change to leak sitds, my audio loopback tests fail. My playback/record timing seemed to be screwed up and I eventually get XRUN on ALSA writes. It could be a problem with my own code since aplay/arecord seem to work properly. Thanks again. - Nate Case <ncase@xxxxxxxxxxx> -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
