Re: [Qemu-devel] [PATCH] usb: queue can have async packets too

[Erik Rull]

On March 5, 2012 at 11:22 AM Erik Rull <address@hidden> wrote:

>
>
>
> On March 2, 2012 at 3:15 PM Gerd Hoffmann <address@hidden> wrote:
>
> > > But: This + the latest GIT master causes again my problems with the
> >
> > With or without this one:
> >
> > commit 5ca2358ac895139e624881c5b3bf3095d3cc4515
> > Date:   Wed Feb 29 09:11:00 2012 -0600
> >
> >     Merge remote-tracking branch 'kraxel/usb.39' into staging
> >
> > ?
> >
> > cheers,
> >   Gerd
> >
> >
>
>
> Hi Gerd,
>
> I'm really sorry, but after rechecking my versions that worked (doing git
> checkouts there) I don't get it working again :-(
> I have no idea what's wrong now...
>
> The startup issue is existing now again on all versions I tested - maybe
> this was just a side effect during the bisect?
>
> So finally - with the current master + your patch the assertion failed is
> fixed but the startup issue seems to remain.
>
> Any ideas how I could help you debugging this issue?
> I tested it on Windows XP and a plain vanilla kernel 2.6.29.12 linux
system
> - both show the same behavior.
>
> Attached the usb-filtered-dmesg from my guest linux booting up.
> If you need further information, please let me know!
> The usb 1-1 is my USB msd that gets not detected properly...
>
> Thanks a lot!
>
> Best regards,
>
> Erik

It's getting pretty ugly. Today I wanted to test the executable with
enabled tracing - and it worked. I reconfigured to "nop-tracing" and it
still works - but only on Linux guest side, the Windows guest still has
some issues - with or without tracing enabled.

Would the usb-tracing output help you to find a possible source of the
problem? The logfiles are pretty large (0.5 and 1.1 MB gzip! / uncompressed
~ 70 / 140 MB).

Best regards,

Erik