summaryrefslogtreecommitdiff
path: root/include/linux/dmar.h
diff options
context:
space:
mode:
authorDavid Woodhouse <David.Woodhouse@intel.com>2009-04-04 01:45:37 +0100
committerDavid Woodhouse <David.Woodhouse@intel.com>2009-04-04 10:43:31 +0100
commit276dbf997043cbf38f0087624e0f9c51742c8885 (patch)
treeeface2519a6ad4c25c2864ee1ee69361ea3f594c /include/linux/dmar.h
parent924b6231edfaf1e764ffb4f97ea382bf4facff58 (diff)
downloadlwn-276dbf997043cbf38f0087624e0f9c51742c8885.tar.gz
lwn-276dbf997043cbf38f0087624e0f9c51742c8885.zip
intel-iommu: Handle PCI domains appropriately.
We were comparing {bus,devfn} and assuming that a match meant it was the same device. It doesn't -- the same {bus,devfn} can exist in multiple PCI domains. Include domain number in device identification (and call it 'segment' in most places, because there's already a lot of references to 'domain' which means something else, and this code is infected with ACPI thinking already). Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Diffstat (limited to 'include/linux/dmar.h')
-rw-r--r--include/linux/dmar.h1
1 files changed, 1 insertions, 0 deletions
diff --git a/include/linux/dmar.h b/include/linux/dmar.h
index 4a0ce6f27e1b..e397dc342cda 100644
--- a/include/linux/dmar.h
+++ b/include/linux/dmar.h
@@ -34,6 +34,7 @@ struct dmar_drhd_unit {
u64 reg_base_addr; /* register base address*/
struct pci_dev **devices; /* target device array */
int devices_cnt; /* target device count */
+ u16 segment; /* PCI domain */
u8 ignored:1; /* ignore drhd */
u8 include_all:1;
struct intel_iommu *iommu;