[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] Re: [PATCH] xend: fix options for assigned pci



On Mon, Jan 11, 2010 at 03:24:59PM +0800, Qing He wrote:
> On Mon, 2010-01-11 at 15:16 +0800, Simon Horman wrote:
> > > @@ -328,7 +332,7 @@
> > >      template['domain'] = "0x%04x" % domain
> > >      template['bus']    = "0x%02x" % int(pci_dev_info['bus'], 16)
> > >      template['slot']   = "0x%02x" % int(pci_dev_info['slot'], 16)
> > > -    template['key']    = pci_dev_str
> > > +    template['key']    = pci_dev_str.split(',')[0]
> > 
> > Splitting on a comma won't work as desired here, as the function
> > portion of multi-function devices may include comas.
> > 
> > Leaving the key as pci_dev_str, is probably ok. Otherwise I suggest
> > constructing the key from template['domain'], template['bus'],
> > template['slot'], pci_dev_infom['func'] and conditionally
> > pci_dev_infom['vdevfn'].
> 
> Thank you for commenting
> In fact, changing the key here doesn't explictly impact the option part
> I'm trying to fix, because ioemu doesn't seem to use backend when creating
> devices. I just changed it for consistency since I found that options
> were also attached in key using comma, but I failed to notice its
> usage as a separator of multi-function devices.
> 
> But if comma is also used for multi-function device, can you show
> me how it looks like? Is there any conflicts between it and the
> pci options?

I don't think there are any conflicts, because the function
can't include '[a-z]' and all options start with '[a-z]'

I documented the syntax on http://wiki.xensource.com/xenwiki/BDFNotation

Basically where the function used to go, it may optionally be
a comma-delimited list of:

* Function unit
* A range of function numbers, denoted by:
  + The first function unit
  + A hyphen (-)
  + The last function unit
* An asterisk (*), used to denote all physical functions that are present in
  the device

Where a function unit comprises of:

* Function number and optionally;
* An equals sign and;
* A virtual function number to use

> > IIRC, it doesn't really matter what the key is, as long as it
> > is unique for a device. Which is why I think that leaving it
> > as pci_dev_str should be ok.
> 
> Changing it back doesn't affect the pci options functionally, AFAIK.

Excellent. Lets do that.


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.