On Thu, Dec 19, 2019 at 05:30:20PM +0100, Jean-Philippe Brucker wrote: > Add support for Substream ID and PASIDs to the SMMUv3 driver. Since v3 > [1], I added review and tested tags where appropriate and applied the > suggested changes, shown in the diff below. Thanks all! > > I'm testing using the zip accelerator on the Hisilicon KunPeng920 and > Zhangfei's uacce module [2]. The full SVA support, which I'll send out > early next year, is available on my branch sva/zip-devel at > https://jpbrucker.net/git/linux/ > > [1] https://lore.kernel.org/linux-iommu/20191209180514.272727-1-jean-philippe@xxxxxxxxxx/ > [2] https://lore.kernel.org/linux-iommu/1576465697-27946-1-git-send-email-zhangfei.gao@xxxxxxxxxx/ > > Jean-Philippe Brucker (13): > iommu/arm-smmu-v3: Drop __GFP_ZERO flag from DMA allocation > dt-bindings: document PASID property for IOMMU masters > iommu/arm-smmu-v3: Parse PASID devicetree property of platform devices > ACPI/IORT: Parse SSID property of named component node > iommu/arm-smmu-v3: Prepare arm_smmu_s1_cfg for SSID support > iommu/arm-smmu-v3: Add context descriptor tables allocators > iommu/arm-smmu-v3: Add support for Substream IDs > iommu/arm-smmu-v3: Propagate ssid_bits > iommu/arm-smmu-v3: Prepare for handling arm_smmu_write_ctx_desc() > failure > iommu/arm-smmu-v3: Add second level of context descriptor table > iommu/arm-smmu-v3: Improve add_device() error handling > PCI/ATS: Add PASID stubs > iommu/arm-smmu-v3: Add support for PCI PASID > > .../devicetree/bindings/iommu/iommu.txt | 6 + > drivers/acpi/arm64/iort.c | 18 + > drivers/iommu/arm-smmu-v3.c | 467 +++++++++++++++--- > drivers/iommu/of_iommu.c | 6 +- > include/linux/iommu.h | 2 + > include/linux/pci-ats.h | 3 + > 6 files changed, 442 insertions(+), 60 deletions(-) This is close, and I've replied to all of the patches I have comments on. To summarise: 1-5 I could queue these now 6 I can make the small change we discussed 7 I can make the changes if you agree (but I'd prefer you to change to batch submission since I can't test this) 8 Good to go once above is solved 9 Need your opinion 10 Some refactoring needed (sorry) 11 Needs Robin's input 12 Good to go once above is solved 13 Need clarification on PCIe behaviour from you In other words, I could probably take the first 8 or 9 patches for 5.6 if you can resolve those issues asap. Cheers, Will
