Hi Jeremy, On Wed, Oct 22, 2008 at 12:16:05AM +0100, Jeremy Morse wrote: > I'm seeing 0xdeadbeef pointers in some of those traces, this suggests > something's being freed unexpectedly... > > One possible cause of this is a behavioral change in the usb stack, > src/sys/dev/usb/usbdi.c, version 1.124, happened 10 days ago. How up > to date is your source tree?
I cvs updated before building the kernels which made crash8 - B. But I have had the problem even before 10 days ago. I first reported this to Jared McNeil on 3rd of Octobre. But I wasn't testing the uhci attachment back then, only the ehci one. So I've had this crash with and without the change, although I don't know whether it will crash with a version before the change using uhci. I will thus try this next (going back to usb/usbdi.c:1.123, boot -c and boot -1c, disable ehci, and look what I'll get). Regards, -Martin
