Re: [Qemu-devel] [RFC/PATCH 1/1] USB code fenced for s390

[Christian Borntraeger <borntraeger@de.ibm.com>]

On 07/08/12 22:22, Blue Swirl wrote:
> On Tue, Aug 7, 2012 at 12:26 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
>> On 7 August 2012 13:19, Christian Borntraeger <borntraeger@de.ibm.com> wrote:
>>> +#if defined(TARGET_HAS_USB) && (TARGET_HAS_USB == 1)
>>>      /* init USB devices */
>>>      if (usb_enabled) {
>>>          if (foreach_device_config(DEV_USB, usb_parse) < 0)
>>>              exit(1);
>>>      }
>>> +#endif
>>
>> Whether there is USB or not is a property of the machine model,
>> not the target CPU architecture, so a TARGET_HAS_USB define
>> is definitely the wrong approach.
> 
> Yes. I'd add CONFIG_USB=y to files in default-configs/ except for
> s390*, Sparc32 and user emulators. Unless it's enough for USB to have
> a dependency on PCI?

Yes, that seems to be the right place - thanks.  A dependency on pci would also
do for us. 

On the other hand, if there are really plans to add a virtio-usb then we should
not disable usb support at all - especially since disabling it does cause some
trouble with libvirt and the current state does not cause problems (we assumed that
libvirt will detect usb capability like it does for other parameters, but it doesnt).
In other words this patch was mostly cosmetics. So lets just drop this patch, go ahead
with Li's patch set and later on we can see if we should do anything regarding s390.

Christian