> Your patch was a very successful shot. > > I prepared to running through different scenarious but the system booted > instantaneously, so i ran a little out of steam, and i'm wondering how to > proceed best. Got the system up and running with nearly a dozen usb devices > attached though i didn't tried them all.
I'm surprised -- but pleased! -- that it was that effective. I guess the usb midi driver you're using is another driver relying on probe()s that will turn down opportunities to bind to interfaces, which is why that patch matters. Greg, I hope you and Patrick will merge some version of that patch. It's essentially the same as one Andries Brouwer just posted, modulo commenting the return convention to match the for_each_dev() usage, and seems to fix a lot of strangeness. > Perhaps most importantly, the usbdevfs still fails mounting during boot time, > though it is placed in fstab. It is mountable flawlessly after that and does > not show anything suspicious on the first glance. > > I recognized an unusal long time to read from it. 3 seconds or so to complete > the list. And in recent kernels it also doesn't show the root hub strings. Likely these are not OHCI issues at all. > Find the kern.log of the session attached. Thanks, I may have comments later. I was amused to see those endpoint scheduling messages! - Dave ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
