On 27.10.10 at 07:43, Weidong Han <weidong.han@xxxxxxxxx> wrote:
Jan Beulich wrote:
The question now is whether some similar check should be restored,
or whether pdev->domain should get updated earlier. This may
I prefer to add the check.
Like this (not tested yet, simplifying the code a little at once):
--- a/xen/drivers/passthrough/vtd/iommu.c
+++ b/xen/drivers/passthrough/vtd/iommu.c
@@ -1371,23 +1371,16 @@ static int domain_context_mapping(struct
if ( find_upstream_bridge(&bus, &devfn, &secbus) < 1 )
break;
- /* PCIe to PCI/PCIx bridge */
- if ( pdev_type(bus, devfn) == DEV_TYPE_PCIe2PCI_BRIDGE )
- {
- ret = domain_context_mapping_one(domain, drhd->iommu, bus, devfn);
- if ( ret )
- return ret;
+ ret = domain_context_mapping_one(domain, drhd->iommu, bus, devfn);
- /*
- * Devices behind PCIe-to-PCI/PCIx bridge may generate
- * different requester-id. It may originate from devfn=0
- * on the secondary bus behind the bridge. Map that id
- * as well.
- */
+ /*
+ * Devices behind PCIe-to-PCI/PCIx bridge may generate different
+ * requester-id. It may originate from devfn=0 on the secondary bus
+ * behind the bridge. Map that id as well if we didn't already.
+ */
+ if ( !ret && pdev_type(bus, devfn) == DEV_TYPE_PCIe2PCI_BRIDGE &&
+ (secbus != pdev->bus || pdev->devfn != 0) )
ret = domain_context_mapping_one(domain, drhd->iommu, secbus, 0);
- }
- else /* Legacy PCI bridge */
- ret = domain_context_mapping_one(domain, drhd->iommu, bus, devfn);
break;
