On 12/09/17 11:58, Roger Pau Monné wrote:
On Tue, Sep 12, 2017 at 11:42:38AM +0100, Julien Grall wrote:
Hi Roger,
On 14/08/17 15:28, Roger Pau Monne wrote:
This functionality is going to reside in vpci.c (and the corresponding
vpci.h header), and should be arch-agnostic. The handlers introduced
in this patch setup the basic functionality required in order to trap
accesses to the PCI config space, and allow decoding the address and
finding the corresponding handler that should handle the access
(although no handlers are implemented).
If I understand correctly this patch, the virtual BDF will always correspond
to the physical BDF. Is that right? If so, would you mind to explain why
such restriction?
Yes, this is not a limitation of this patch, but of the implementation
that follows. Likely this will be expanded when support for DomU is
added, but for Dom0 at least on x86 adding such a translation layer is
not needed, since I see no reason to present a different PCI topology
to Dom0.
I think this will be the same for ARM. Dom0 will always have pBDF ==
vBDF. For guest will likely want pBDF != vBDF.
This answer to my other question regarding the plan to support vBDF !=
pBDF :). Thanks.
Thanks, Roger.
--
Julien Grall
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
