diff options
| author | Bjorn Helgaas <bhelgaas@google.com> | 2019-10-09 23:14:05 +0300 |
|---|---|---|
| committer | Bjorn Helgaas <bhelgaas@google.com> | 2019-10-16 00:37:10 +0300 |
| commit | fd872843ecd52820654dcacf85e680d8101a9fde (patch) | |
| tree | 5b4fb42d89aa5c2a0e8be70588a3275dd98c34b7 /include/linux | |
| parent | 54ecb8f7028c5eb3d740bb82b0f1d90f2df63c5c (diff) | |
| download | linux-fd872843ecd52820654dcacf85e680d8101a9fde.tar.xz | |
iommu/vt-d: Select PCI_PRI for INTEL_IOMMU_SVM
Previously intel-iommu.c depended on CONFIG_AMD_IOMMU in an undesirable
way. When CONFIG_INTEL_IOMMU_SVM=y, iommu_enable_dev_iotlb() calls PRI
interfaces (pci_reset_pri() and pci_enable_pri()), but those are only
implemented when CONFIG_PCI_PRI is enabled.
The INTEL_IOMMU_SVM Kconfig did nothing with PCI_PRI, but AMD_IOMMU selects
PCI_PRI. So if AMD_IOMMU was enabled, intel-iommu.c got the full PRI
interfaces, but if AMD_IOMMU was not enabled, it got the PRI stubs.
Make the iommu_enable_dev_iotlb() behavior independent of AMD_IOMMU by
having INTEL_IOMMU_SVM select PCI_PRI so iommu_enable_dev_iotlb() always
uses the full implementations of PRI interfaces.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Acked-by: Joerg Roedel <jroedel@suse.de>
Diffstat (limited to 'include/linux')
0 files changed, 0 insertions, 0 deletions