USB2+umass: timing related bug (machine check abort)

[Hans Petter Selasky]

On Tuesday 17 February 2009, Marcel Moolenaar wrote:
> Context: MACHINE=ia64, CPU=Montecito
>
> I'm running into a timing related MCA. In short:
> 	...
> umass0: <HEWLETT PACKARD INTEGRITY SERVER, class 0/0, rev 2.00/0.a1,
> addr 2> on usbus2
> umass0:  SCSI over Bulk-Only; quirks = 0x0000
> umass0:2:0:-1: Attached to scbus2
> 	*** machine check abort ***
> ***********************************************************
> * ROM Version : 01.05
> * ROM Date    : 11/06/2006
> * BMC Version :  05.06
> ***********************************************************
> 	...
>
> When I enable EHCI debugging (level 99) this does not happen
> and between the debug output, I see:
>

It looks like there is a timing issue there. I suspect that someone is trying 
to setup a USB transfer before usb2_transfer_setup() has returned.

Could you get into the kernel debugger and dump "xfer->xroot" 
and "xfer->xroot->xfer_mtx".

--HPS