Alan, I've applied your patch to 3.1.5 and it seems that some troubles fixed. HCI doesn't die now even with too small audio latencies (died without this patch). Will continue testing. Current dmesg log is in attachment. 2011/12/20 Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>: > Michael and Anton: > > Here is a preliminary patch meant to address the problem you have both > encountered: ATI's EHCI controller doesn't turn off the periodic > schedule as quickly as it should. > > I wrote this patch against 3.2-rc5, but I think it will apply (perhaps > with minor changes) to any recent kernel. This isn't a complete > solution by any means; that will require a significantly larger change. > But I'd appreciate it if you can test this and see how well it does. > > The patch adds a few ehci_info() calls, which will print some debugging > information in the system log. It will be interesting to see what > shows up. > > Alan Stern
Attachment:
FastTrackUltra_Linux-3.1.5_dmesg.log
Description: Binary data
