Re: [Qemu-devel] EHCI USB regression in 1.2.0 - ehci_state_fetchqtd() asserting

[Hans de Goede <hdegoede@redhat.com>]

Hi,

On 09/21/2012 08:46 PM, Shawn Starr wrote:
> On Friday, September 21, 2012 07:35:42 PM Hans de Goede wrote:
>> Hi,
>>
>> On 09/21/2012 05:39 PM, Shawn Starr wrote:
>>> On Friday, September 21, 2012 02:19:08 PM Hans de Goede wrote:
>>>> Hi,
>>>
>>> <snip>
>>>
>>>> Ok, so your using host redirection here, see the device usb-host
>>>>
>>>>> In addition, if i disable usb-redirection device the cam works as it did
>>>>> before (although like before, the video image is upside down in some
>>>>> applications).
>>>>
>>>> When you say disable the usb-redirection device, do you mean removing
>>>> the host redirection, and selecting the cam for redirection in a
>>>> spice client (ie remote-viewer) instead, or... ?
>>>>
>>>> Can you please try, disabling host-redirection, then connecting to
>>>> the vm with remote-viewer and then selecting the cam from
>>>> remote-viewers UI?
>>>>
>>>> Thanks,
>>>>
>>>> Hans
>>>
>>> Hi Hans,
>>>
>>> I am using usb-host and it works. If I keep usb-host and add
>>> usb-redirection it shows those errors.
>>
>> Hmm, this sounds like you are using spice usb-redir too and the 2 are
>> fighting each other. Which spice-client are you using? And are you plugging
>> in the device after starting up the vm ?
>>
>>> Your patch however did fix using host-usb alone.
>>
>> Good.
>>
>>> I don't have a setup that can let me use usb-redirection unfortunately.
>>
>> Why not? Looking at your qemu cmdline you are using spice. Hmm, I guesss you
>> may want to keep the device redirected when no client is connected?
>>
>
> Using spicec, it doesn't give me an option to connection devices.

spicec is deprecated you really should be using remote-viewer instead. That
will give you an option to select devices.

 > I assume
> usb-redirection for devices *not* on the same machine as the running VM? If
> so, that use case I wasn't testing originally (although useful)

spice's usbredirection works fine locally too, and it gives you a nice
UI to redirect usb devices, without needing to need usb-ids, etc.


>
>> Anyway, if you could remove the host-redirection temporarily and try with
>> spice usb-redir and collect logs that would be great. If you don't have
>> time for that, that is ok too, then I'll consider this issue resolved.
>
> I consider the issue resolved solely on usb-host issue vs any regressions with
> usb-redirection itself which I never used/tried.
>
> However, if someone can point out how you can test this, I can do this also.

See above :)

Regards,

Hans