On Fri, 7 Feb 2014, Devin Heitmueller wrote: > Hi Alan, > > Thanks for taking the time to respond. > > On Thu, Feb 6, 2014 at 5:08 PM, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: > > Is it possible to get any debugging information from the machines in > > the field? > > At this point all I know is the usb_submit_urb() is returning -EBIGF. > In the 3.7 ehci_hcd driver that can come from one of three places and > I'm in the process of deploying some additional logging so at least I > know which of those three is causing it. I do know that when it > occurs, it's not necessarily at the start of streaming (i.e. it's not > some sort of initial bandwidth reservation issue). I see it happening > when the URBs are being actively serviced by the completion handler > and resubmitted. That sounds like it may be the sort of problem that was addressed by the snd-usb-audio changes. > >> I know this is an atypical thing to consider since most devs work on > >> the trunk, but can anybody give me a feel for how much breakage is > >> likely to occur If I move the tree between those two revs? Have there > >> been a ton of changes to common frameworks in the kernel that > >> drivers/usb relies on? > >> > >> Suggestions welcome. > > > > That's a pretty hard question to answer. Probably the best way to find > > out is try it and see what breaks. It may be easier to back-port just > > the EHCI driver family and not the whole stack. > > Yeah, I'll take a crack at it. Figured I would ask to see if anybody > would come out and say "don't even try it - we know it to be a lost > cause." I just remembered that some serious changes in the way ehci-hcd hands back URBs were merged about half a year ago. This means it won't be possible to back-port ehci-hcd without the rest of the USB core. > > I can point out that in the last one or two kernel releases there were > > significant changes to the snd-usb-audio driver that affect the way it > > interacts with ehci-hcd. Just to make your life more difficult... > > Supposedly they never had this issue in 2.6.33, so presumably it's > some sort of regression in either the EHCI driver or snd-usb-audio > between 2.6.33 and 3.7. That said, I'm hoping it's something already > fixed in 3.13 and I just need to find the right change to backport. > > I know the ISO_ASAP behavior changed, but aside from that do you see > any reason that I wouldn't be able to backport the entire > snd-usb-audio driver from 3.13 (leaving the ehci driver at the 3.7 > level)? You might. I don't know what else has been happening to that driver. Alan Stern -- 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
