iommufd follows the same design as KVM and uses memory cgroups to limit the amount of kernel memory a iommufd file descriptor can pin down. The various internal data structures already use GFP_KERNEL_ACCOUNT to charge its own memory. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain and these allocations are not tracked. This series is the first step in fixing it. The iommu driver contract already includes a 'gfp' argument to the map_pages op, allowing iommufd to specify GFP_KERNEL_ACCOUNT and then having the driver allocate the IOPTE tables with that flag will capture a significant amount of the allocations. Update the iommu_map() API to pass in the GFP argument, and fix all call sites. Replace iommu_map_atomic(). Audit the "enterprise" iommu drivers to make sure they do the right thing. Intel and S390 ignore the GFP argument and always use GFP_ATOMIC. This is problematic for iommufd anyhow, so fix it. AMD and ARM SMMUv2/3 are already correct. A follow up series will be needed to capture the allocations made when the iommu_domain itself is allocated, which will complete the job. v3: - Leave a GFP_ATOMIC in "Add a gfp parameter to iommu_map_sg()" and move the conversion to gfp argument to "Use the gfp parameter in __iommu_dma_alloc_noncontiguous()" - Mask off the zone/policy flags from gfp before doing internal allocations and add a comment about Robin's note that this is to keep the buffer and internal seperate. v2: https://lore.kernel.org/r/0-v2-ce66f632bd0d+484-iommu_map_gfp_jgg@xxxxxxxxxx - Prohibit bad GFP flags in the iommu wrappers - Split out the new GFP_KERNEL usages into dedicated patches so it is easier to check. No code change after the full series v1: https://lore.kernel.org/r/0-v1-6e8b3997c46d+89e-iommu_map_gfp_jgg@xxxxxxxxxx Jason Gunthorpe (10): iommu: Add a gfp parameter to iommu_map() iommu: Remove iommu_map_atomic() iommu: Add a gfp parameter to iommu_map_sg() iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() iommu/intel: Add a gfp parameter to alloc_pgtable_page() iommu/intel: Support the gfp argument to the map_pages op iommu/intel: Use GFP_KERNEL in sleepable contexts iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s iommu/s390: Use GFP_KERNEL in sleepable contexts arch/arm/mm/dma-mapping.c | 11 ++-- arch/s390/include/asm/pci_dma.h | 5 +- arch/s390/pci/pci_dma.c | 31 ++++++----- .../drm/nouveau/nvkm/subdev/instmem/gk20a.c | 3 +- drivers/gpu/drm/tegra/drm.c | 2 +- drivers/gpu/host1x/cdma.c | 2 +- drivers/infiniband/hw/usnic/usnic_uiom.c | 4 +- drivers/iommu/dma-iommu.c | 18 +++++-- drivers/iommu/intel/iommu.c | 36 +++++++------ drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 53 +++++++------------ drivers/iommu/iommufd/pages.c | 6 ++- drivers/iommu/s390-iommu.c | 15 +++--- drivers/media/platform/qcom/venus/firmware.c | 2 +- drivers/net/ipa/ipa_mem.c | 6 ++- drivers/net/wireless/ath/ath10k/snoc.c | 2 +- drivers/net/wireless/ath/ath11k/ahb.c | 4 +- drivers/remoteproc/remoteproc_core.c | 5 +- drivers/vfio/vfio_iommu_type1.c | 9 ++-- drivers/vhost/vdpa.c | 2 +- include/linux/iommu.h | 31 +++-------- 22 files changed, 126 insertions(+), 125 deletions(-) base-commit: 5dc4c995db9eb45f6373a956eb1f69460e69e6d4 -- 2.39.0
