powerpc/powernv: Fix IOMMU table for VFIO dev
authorGavin Shan <gwshan@linux.vnet.ibm.com>
Tue, 15 Jul 2014 07:00:55 +0000 (17:00 +1000)
committerBenjamin Herrenschmidt <benh@kernel.crashing.org>
Tue, 5 Aug 2014 05:41:12 +0000 (15:41 +1000)
commitdff4a39e880062b358be2f08de878a49425036c4
tree0a288cce8018c5a1d4ee682676c1530687f60663
parent9e5c6e5a3be0b2e17ff61b9b74adef4a2c9e6934
powerpc/powernv: Fix IOMMU table for VFIO dev

On PHB3, PCI devices can bypass IOMMU for DMA access. If we pass
through one PCI device, whose hose driver ever enable the bypass
mode, pdev->dev.archdata.dma_data.iommu_table_base isn't IOMMU
table. However, EEH needs access the IOMMU table when the device
is owned by guest.

The patch fixes pdev->dev.archdata.dma_data.iommu_table when
passing through the device to guest in pnv_pci_ioda2_set_bypass().

Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
arch/powerpc/platforms/powernv/pci-ioda.c