On Wednesday 01 August 2007, Daniel Mack wrote: > on an i.MX31 ARM based system which has an EHCI controller on-chip, > kernel version 2.6.19.2 crashes occur randomly
What does "randomly" mean? And why aren't you using 2.6.23-rc1-git? > with lots of USB bulk > traffic in action, usually after around some minutes of consecutive > streaming. As the generic EHCI stack seems to be affected, I doubt > that I have any ARM specific bug here. That doesn't follow at all. Especially for "random" oopsing, it's not uncommon that the symptoms be far away from the root cause of the bug. And I suspect you're not actually using 2.6.19.2 ... at which point the first thing to suspect is one of the patches you've added on top of that. (E.g. for i.MX31 support, which I see hasn't yet merged upstream.) > Others who have similar problems report that it does not occur with > kernel version 2.6.16 and earlier. > > The Oops is attached. Did anyone else see similar behavior? Nope. Looks like someone's trashing part of your kernel data. Does it always die in mod_timer() like that? > Have such > issues been addressed in more recent version of the USB stack? I can't think of any particular changes in the USB stack which might relate to this. However, bugs are constantly getting fixed all over the kernel; some of them even prevent memory corruption. Which is why when you report any bug, one of the first questions will be whether a *current* kernel exhibits the relevant symptom. > Please Cc: me in answers, I'm not subscribed to linux-usb-devel. > > Many thanks for any pointer, > Daniel > > > [ 49.100000] Unable to handle kernel paging request at virtual > address 8065a200 > [ 49.100000] pgd = c0004000 > [ 49.100000] [8065a200] *pgd=00000000 > [ 52.100000] Internal error: Oops: 5 [#1] > [ 52.100000] Modules linked in: mxc_ide usbhid caiaq_platform > caiaq_audio_device > [ 52.100000] CPU: 0 > [ 52.100000] PC is at __mod_timer+0x60/0xf8 Looks more like something trashed the part of the heap which is used by the EHCI driver, or the timer data structures. What is that part of mod_timer() doing? Could it be that some other timer (from your platform patches) is the problem? It's most likely that EHCI is just the victim of a bug somewhere else. - Dave ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel