Re: [Xen-users] Q: how to solve "shared RMRR" / "gfn already mapped" / GSI ?

[George Dunlap <george.dunlap@xxxxxxxxxx>]

On 14/10/15 16:40, HÃkon Alstadheim wrote:
> Den 14. okt. 2015 15:05, skrev George Dunlap:
>> Ah -- yeah, that makes a bisection pretty much impossible. :-( Out of
>> curiosity, do you actually need to pass through the entire USB
>> controller? Could you pass through just the device itself?
> 
> My main reason for firing up windows is to run a high-res usb attached
> photnegative-scanner -- proprietary driver :-( ; I don't know whether it
> will work, any tips/caveats before I start ? Will
> 'usb=1;usbdevice=["host:mumble:mumble"]'  work ?

Something like that *should* work, yes.  (Read the man page for the
exact syntax, of course.)  I would think the only potential issue you
might have would be with performance.

Then again, passing through the USB controller *should* work as well;
there are obviously occasional quirks.  It's often hardware which
doesn't quite follow the spec but actually works on real hardware.  When
you switch it to a virtual environment, suddenly all the things which
happened to work on real hardware don't anymore.

Come to think of it -- one more thing to check: Does the USB controller
you're passing through have more than one virtual function, and if so,
are you passing in both (/all)?

As I understand it, it's part of the official spec that USB3 controllers
have two functions (basically two different logical devices in the same
PCI slot): one a full-speed USB3 controller, and one a legacy USB1
controller.  If you were only passing in one, it's possible that the
Linux driver works but the Windows driver is confused.

Checking "lspci" should tell you if you've got multiple functions in
your USB controller.  Passing them all in (and making sure they all had
the same virtual device) might be something worth trying.

 -George


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users