>>> On 20.09.11 at 20:02, "Kay, Allen M" <allen.m.kay@xxxxxxxxx> wrote:
> Thanks for the reminder. This has slip through the cracks. The patch look
> good. With this patch, we shouldn't need the existing call to
> pci_enable_acs() in setup_dom0_devices(), correct?
Ah, indeed, I didn't even pay attention to this.
But you didn't say anything about the subject of the mail, which is
what keeps me from pushing the (incomplete) patch.
> I agree that the same problem exists for bus-to-bridge mapping function
> pci_scan_device(). By the way, we probably should take the opportunity to
> give it a proper name that reflects to the true purpose of this function.
It's not that far off, if you mean scan_pci_devices() (and
_scan_pci_devices() once the multi-seg patches get committed).
What alternative name were you thinking of?
Also, if the bus2bridge data got maintained dynamically (i.e. out of
pci_{add,remove}_device()), I would think these functions could go
away again (and hence renaming them would become mute).
Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|