On Mon, 5 May 2008, Marti Raudsepp wrote:

> On Mon, May 5, 2008 at 3:13 AM, Matthew Dharm
> <mdharm-usb@one-eyed-alien.net> wrote:
> > If there is no actual activity, then this is likely not a usb-storage
> > problem. It may very well be an HCD problem, however.
> >
> > You may want to turn on CONFIG_USB_STORAGE_DEBUG and verify that there is,
> > in fact, no activity to the drive during these events.
>
> Indeed, I enabled USB_DEBUG and USB_STORAGE_DEBUG; there are no kernel
> messages at the time of the interrupt flood.
>
> Any ideas how to proceed to get this problem resolved? I added David
> Brownell (the EHCI maintainer) to recipients.

One thing you can do is to mount a debugfs filesystem, and under it go
to the EHCI directory corresponding to the controller your drive is
plugged into. The contents of the "registers" and "async" files would
be interesting -- particularly if you get them while the disk is idle.

You might even try looking at "registers" several times in a row, since
its contents may change over time.

Alan Stern

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel