Re: [libvirt] [RFC PATCHv2 5/5] WIP: smartcard: turn on qemu support

On 01/14/2011 12:23 PM, Daniel P. Berrange wrote: > Alon's docs are showing the simplified syntax suitable for > end users. This doesn't guarentee a stable guest visible ABI. > Looking at the code, we need to set the 'slot' parameter on each > ccid device we have. This means we need a new address type for > smart card devices, and a corresponding <controller> instance. Thanks - that clears up a lot for me. However, I still have a question: > So in the XML we'd get (including libvirt generated aliases > and addresses): > > <devices> > <controller type='ccid' index='0'> > <alias id='ccid0'/> > </controller> Are you suggesting that the XML mandate that the user provides a <controller type=ccid'>, or should I still shoot for the idea that if the user omits <controller> but provides <smartcard> that we go ahead and auto-create a controller (assigning it to the next available slot in the process)?

> <smartcard mode='host'> > <alias id='smartcard0'/> > <address type='ccid' controller='0' slot='0'/> So I _do_ need an (optional) <address> subelement for <smartcard> after all. Will go back and get that into my series. > > In other words a hierarchy > > USB bus 0 > | > +- ccid0 > | | > | +- smartcard0 (ccid slot 0) > | +- smartcard1 (ccid slot 3) > | > +- ccid1 > | > +- smartcard2 (ccid slot 0) Very helpful.