* [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. Jason Gunthorpe (8): 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/s390: Push the gfp parameter to the kmem_cache_alloc()'s 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 | 11 ++--- drivers/iommu/intel/iommu.c | 36 +++++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 43 +++++-------------- 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, 109 insertions(+), 125 deletions(-) base-commit: 88603b6dc419445847923fcb7fe5080067a30f98 -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media 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. Jason Gunthorpe (8): 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/s390: Push the gfp parameter to the kmem_cache_alloc()'s 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 | 11 ++--- drivers/iommu/intel/iommu.c | 36 +++++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 43 +++++-------------- 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, 109 insertions(+), 125 deletions(-) base-commit: 88603b6dc419445847923fcb7fe5080067a30f98 -- 2.39.0 ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. Jason Gunthorpe (8): 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/s390: Push the gfp parameter to the kmem_cache_alloc()'s 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 | 11 ++--- drivers/iommu/intel/iommu.c | 36 +++++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 43 +++++-------------- 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, 109 insertions(+), 125 deletions(-) base-commit: 88603b6dc419445847923fcb7fe5080067a30f98 -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media 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. Jason Gunthorpe (8): 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/s390: Push the gfp parameter to the kmem_cache_alloc()'s 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 | 11 ++--- drivers/iommu/intel/iommu.c | 36 +++++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 43 +++++-------------- 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, 109 insertions(+), 125 deletions(-) base-commit: 88603b6dc419445847923fcb7fe5080067a30f98 -- 2.39.0 ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. Jason Gunthorpe (8): 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/s390: Push the gfp parameter to the kmem_cache_alloc()'s 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 | 11 ++--- drivers/iommu/intel/iommu.c | 36 +++++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 43 +++++-------------- 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, 109 insertions(+), 125 deletions(-) base-commit: 88603b6dc419445847923fcb7fe5080067a30f98 -- 2.39.0 ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. Jason Gunthorpe (8): 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/s390: Push the gfp parameter to the kmem_cache_alloc()'s 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 | 11 ++--- drivers/iommu/intel/iommu.c | 36 +++++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/iommu.c | 43 +++++-------------- 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, 109 insertions(+), 125 deletions(-) base-commit: 88603b6dc419445847923fcb7fe5080067a30f98 -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization The internal mechanisms support this, but instead of exposting the gfp to the caller it wrappers it into iommu_map() and iommu_map_atomic() Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/arm/mm/dma-mapping.c | 11 +++++++---- .../gpu/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 | 2 +- drivers/iommu/iommu.c | 17 ++++++----------- drivers/iommu/iommufd/pages.c | 6 ++++-- 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 | 4 ++-- 16 files changed, 43 insertions(+), 38 deletions(-) diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index c135f6e37a00ca..8bc01071474ab7 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, len = (j - i) << PAGE_SHIFT; ret = iommu_map(mapping->domain, iova, phys, len, - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), + GFP_KERNEL); if (ret < 0) goto fail; iova += len; @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, iova, phys, len, prot); + ret = iommu_map(mapping->domain, iova, phys, len, prot, + GFP_KERNEL); if (ret < 0) goto fail; count += len >> PAGE_SHIFT; @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, + prot, GFP_KERNEL); if (ret < 0) goto fail; @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); if (ret < 0) goto fail; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, u32 offset = (r->offset + i) << imem->iommu_pgshift; ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret < 0) { nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 7bd2e65c2a16c5..6ca9f396e55be4 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) *dma = iova_dma_addr(&tegra->carveout.domain, alloc); err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), - size, IOMMU_READ | IOMMU_WRITE); + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (err < 0) goto free_iova; diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 103fda055394ab..4ddfcd2138c95b 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) pb->dma = iova_dma_addr(&host1x->iova, alloc); err = iommu_map(host1x->domain, pb->dma, pb->phys, size, - IOMMU_READ); + IOMMU_READ, GFP_KERNEL); if (err) goto iommu_free_iova; } else { diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index c301b3be9f303d..aeeaca65ace96a 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index f798c44e090337..8bdb65e7686ff9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, if (!iova) goto out_free_page; - if (iommu_map(domain, iova, msi_addr, size, prot)) + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) goto out_free_iova; INIT_LIST_HEAD(&msi_page->list); diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index de91dd88705bd3..fe29fc2140b132 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, if (map_size) { ret = iommu_map(domain, addr - map_size, addr - map_size, map_size, - entry->prot); + entry->prot, GFP_KERNEL); if (ret) goto out; map_size = 0; @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, return ret; } -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) +int iommu_map(struct iommu_domain *domain, unsigned long iova, + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); return ret; } - -int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - might_sleep(); - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map); int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot) { - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); } EXPORT_SYMBOL_GPL(iommu_map_atomic); diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 1e1d3509efae5e..22cc3bb0c6c55a 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, size % PAGE_SIZE); while (size) { - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, + GFP_KERNEL); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, else rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, - next_iova - iova, area->iommu_prot); + next_iova - iova, area->iommu_prot, + GFP_KERNEL); if (rc) goto err_unmap; iova = next_iova; diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c index 142d4c74017c04..07d4dceb5e72c7 100644 --- a/drivers/media/platform/qcom/venus/firmware.c +++ b/drivers/media/platform/qcom/venus/firmware.c @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, core->fw.mapped_mem_size = mem_size; ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); if (ret) { dev_err(dev, "could not map video firmware region\n"); return ret; diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c index 9ec5af323f731d..991a7d39f06661 100644 --- a/drivers/net/ipa/ipa_mem.c +++ b/drivers/net/ipa/ipa_mem.c @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index cfcb759a87deac..9a82f0336d9537 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, ar->msa.paddr, ar->msa.mem_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath10k_err(ar, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c index d34a4d6325b2b4..df8fdc7067f99c 100644 --- a/drivers/net/wireless/ath/ath11k/ahb.c +++ b/drivers/net/wireless/ath/ath11k/ahb.c @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); goto err_iommu_unmap; diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 1cd4815a6dd197..80072b6b628358 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, if (!mapping) return -ENOMEM; - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, + GFP_KERNEL); if (ret) { dev_err(dev, "failed to map devmem: %d\n", ret); goto out; @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, } ret = iommu_map(rproc->domain, mem->da, dma, mem->len, - mem->flags); + mem->flags, GFP_KERNEL); if (ret) { dev_err(dev, "iommu_map failed: %d\n", ret); goto free_mapping; diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 23c24fe98c00d4..e14f86a8ef5258 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, list_for_each_entry(d, &iommu->domain_list, next) { ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, - npage << PAGE_SHIFT, prot | IOMMU_CACHE); + npage << PAGE_SHIFT, prot | IOMMU_CACHE, + GFP_KERNEL); if (ret) goto unwind; @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, size = npage << PAGE_SHIFT; } - ret = iommu_map(domain->domain, iova, phys, - size, dma->prot | IOMMU_CACHE); + ret = iommu_map(domain->domain, iova, phys, size, + dma->prot | IOMMU_CACHE, GFP_KERNEL); if (ret) { if (!dma->iommu_mapped) { vfio_unpin_pages_remote(dma, iova, @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) return; ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); if (!ret) { size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index 166044642fd5cc..e555c3bd1c030b 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, r = ops->set_map(vdpa, asid, iotlb); } else { r = iommu_map(v->domain, iova, pa, size, - perm_to_iommu_flags(perm)); + perm_to_iommu_flags(perm), GFP_KERNEL); } if (r) { vhost_iotlb_del_range(iotlb, iova, iova + size - 1); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 46e1347bfa2286..d2020994f292db 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) } static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { return -ENODEV; } -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media The internal mechanisms support this, but instead of exposting the gfp to the caller it wrappers it into iommu_map() and iommu_map_atomic() Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/arm/mm/dma-mapping.c | 11 +++++++---- .../gpu/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 | 2 +- drivers/iommu/iommu.c | 17 ++++++----------- drivers/iommu/iommufd/pages.c | 6 ++++-- 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 | 4 ++-- 16 files changed, 43 insertions(+), 38 deletions(-) diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index c135f6e37a00ca..8bc01071474ab7 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, len = (j - i) << PAGE_SHIFT; ret = iommu_map(mapping->domain, iova, phys, len, - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), + GFP_KERNEL); if (ret < 0) goto fail; iova += len; @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, iova, phys, len, prot); + ret = iommu_map(mapping->domain, iova, phys, len, prot, + GFP_KERNEL); if (ret < 0) goto fail; count += len >> PAGE_SHIFT; @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, + prot, GFP_KERNEL); if (ret < 0) goto fail; @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); if (ret < 0) goto fail; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, u32 offset = (r->offset + i) << imem->iommu_pgshift; ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret < 0) { nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 7bd2e65c2a16c5..6ca9f396e55be4 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) *dma = iova_dma_addr(&tegra->carveout.domain, alloc); err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), - size, IOMMU_READ | IOMMU_WRITE); + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (err < 0) goto free_iova; diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 103fda055394ab..4ddfcd2138c95b 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) pb->dma = iova_dma_addr(&host1x->iova, alloc); err = iommu_map(host1x->domain, pb->dma, pb->phys, size, - IOMMU_READ); + IOMMU_READ, GFP_KERNEL); if (err) goto iommu_free_iova; } else { diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index c301b3be9f303d..aeeaca65ace96a 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index f798c44e090337..8bdb65e7686ff9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, if (!iova) goto out_free_page; - if (iommu_map(domain, iova, msi_addr, size, prot)) + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) goto out_free_iova; INIT_LIST_HEAD(&msi_page->list); diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index de91dd88705bd3..fe29fc2140b132 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, if (map_size) { ret = iommu_map(domain, addr - map_size, addr - map_size, map_size, - entry->prot); + entry->prot, GFP_KERNEL); if (ret) goto out; map_size = 0; @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, return ret; } -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) +int iommu_map(struct iommu_domain *domain, unsigned long iova, + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); return ret; } - -int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - might_sleep(); - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map); int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot) { - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); } EXPORT_SYMBOL_GPL(iommu_map_atomic); diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 1e1d3509efae5e..22cc3bb0c6c55a 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, size % PAGE_SIZE); while (size) { - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, + GFP_KERNEL); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, else rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, - next_iova - iova, area->iommu_prot); + next_iova - iova, area->iommu_prot, + GFP_KERNEL); if (rc) goto err_unmap; iova = next_iova; diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c index 142d4c74017c04..07d4dceb5e72c7 100644 --- a/drivers/media/platform/qcom/venus/firmware.c +++ b/drivers/media/platform/qcom/venus/firmware.c @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, core->fw.mapped_mem_size = mem_size; ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); if (ret) { dev_err(dev, "could not map video firmware region\n"); return ret; diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c index 9ec5af323f731d..991a7d39f06661 100644 --- a/drivers/net/ipa/ipa_mem.c +++ b/drivers/net/ipa/ipa_mem.c @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index cfcb759a87deac..9a82f0336d9537 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, ar->msa.paddr, ar->msa.mem_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath10k_err(ar, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c index d34a4d6325b2b4..df8fdc7067f99c 100644 --- a/drivers/net/wireless/ath/ath11k/ahb.c +++ b/drivers/net/wireless/ath/ath11k/ahb.c @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); goto err_iommu_unmap; diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 1cd4815a6dd197..80072b6b628358 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, if (!mapping) return -ENOMEM; - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, + GFP_KERNEL); if (ret) { dev_err(dev, "failed to map devmem: %d\n", ret); goto out; @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, } ret = iommu_map(rproc->domain, mem->da, dma, mem->len, - mem->flags); + mem->flags, GFP_KERNEL); if (ret) { dev_err(dev, "iommu_map failed: %d\n", ret); goto free_mapping; diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 23c24fe98c00d4..e14f86a8ef5258 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, list_for_each_entry(d, &iommu->domain_list, next) { ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, - npage << PAGE_SHIFT, prot | IOMMU_CACHE); + npage << PAGE_SHIFT, prot | IOMMU_CACHE, + GFP_KERNEL); if (ret) goto unwind; @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, size = npage << PAGE_SHIFT; } - ret = iommu_map(domain->domain, iova, phys, - size, dma->prot | IOMMU_CACHE); + ret = iommu_map(domain->domain, iova, phys, size, + dma->prot | IOMMU_CACHE, GFP_KERNEL); if (ret) { if (!dma->iommu_mapped) { vfio_unpin_pages_remote(dma, iova, @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) return; ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); if (!ret) { size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index 166044642fd5cc..e555c3bd1c030b 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, r = ops->set_map(vdpa, asid, iotlb); } else { r = iommu_map(v->domain, iova, pa, size, - perm_to_iommu_flags(perm)); + perm_to_iommu_flags(perm), GFP_KERNEL); } if (r) { vhost_iotlb_del_range(iotlb, iova, iova + size - 1); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 46e1347bfa2286..d2020994f292db 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) } static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { return -ENODEV; } -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization The internal mechanisms support this, but instead of exposting the gfp to the caller it wrappers it into iommu_map() and iommu_map_atomic() Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/arm/mm/dma-mapping.c | 11 +++++++---- .../gpu/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 | 2 +- drivers/iommu/iommu.c | 17 ++++++----------- drivers/iommu/iommufd/pages.c | 6 ++++-- 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 | 4 ++-- 16 files changed, 43 insertions(+), 38 deletions(-) diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index c135f6e37a00ca..8bc01071474ab7 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, len = (j - i) << PAGE_SHIFT; ret = iommu_map(mapping->domain, iova, phys, len, - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), + GFP_KERNEL); if (ret < 0) goto fail; iova += len; @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, iova, phys, len, prot); + ret = iommu_map(mapping->domain, iova, phys, len, prot, + GFP_KERNEL); if (ret < 0) goto fail; count += len >> PAGE_SHIFT; @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, + prot, GFP_KERNEL); if (ret < 0) goto fail; @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); if (ret < 0) goto fail; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, u32 offset = (r->offset + i) << imem->iommu_pgshift; ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret < 0) { nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 7bd2e65c2a16c5..6ca9f396e55be4 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) *dma = iova_dma_addr(&tegra->carveout.domain, alloc); err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), - size, IOMMU_READ | IOMMU_WRITE); + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (err < 0) goto free_iova; diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 103fda055394ab..4ddfcd2138c95b 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) pb->dma = iova_dma_addr(&host1x->iova, alloc); err = iommu_map(host1x->domain, pb->dma, pb->phys, size, - IOMMU_READ); + IOMMU_READ, GFP_KERNEL); if (err) goto iommu_free_iova; } else { diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index c301b3be9f303d..aeeaca65ace96a 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index f798c44e090337..8bdb65e7686ff9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, if (!iova) goto out_free_page; - if (iommu_map(domain, iova, msi_addr, size, prot)) + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) goto out_free_iova; INIT_LIST_HEAD(&msi_page->list); diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index de91dd88705bd3..fe29fc2140b132 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, if (map_size) { ret = iommu_map(domain, addr - map_size, addr - map_size, map_size, - entry->prot); + entry->prot, GFP_KERNEL); if (ret) goto out; map_size = 0; @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, return ret; } -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) +int iommu_map(struct iommu_domain *domain, unsigned long iova, + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); return ret; } - -int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - might_sleep(); - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map); int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot) { - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); } EXPORT_SYMBOL_GPL(iommu_map_atomic); diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 1e1d3509efae5e..22cc3bb0c6c55a 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, size % PAGE_SIZE); while (size) { - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, + GFP_KERNEL); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, else rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, - next_iova - iova, area->iommu_prot); + next_iova - iova, area->iommu_prot, + GFP_KERNEL); if (rc) goto err_unmap; iova = next_iova; diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c index 142d4c74017c04..07d4dceb5e72c7 100644 --- a/drivers/media/platform/qcom/venus/firmware.c +++ b/drivers/media/platform/qcom/venus/firmware.c @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, core->fw.mapped_mem_size = mem_size; ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); if (ret) { dev_err(dev, "could not map video firmware region\n"); return ret; diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c index 9ec5af323f731d..991a7d39f06661 100644 --- a/drivers/net/ipa/ipa_mem.c +++ b/drivers/net/ipa/ipa_mem.c @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index cfcb759a87deac..9a82f0336d9537 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, ar->msa.paddr, ar->msa.mem_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath10k_err(ar, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c index d34a4d6325b2b4..df8fdc7067f99c 100644 --- a/drivers/net/wireless/ath/ath11k/ahb.c +++ b/drivers/net/wireless/ath/ath11k/ahb.c @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); goto err_iommu_unmap; diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 1cd4815a6dd197..80072b6b628358 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, if (!mapping) return -ENOMEM; - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, + GFP_KERNEL); if (ret) { dev_err(dev, "failed to map devmem: %d\n", ret); goto out; @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, } ret = iommu_map(rproc->domain, mem->da, dma, mem->len, - mem->flags); + mem->flags, GFP_KERNEL); if (ret) { dev_err(dev, "iommu_map failed: %d\n", ret); goto free_mapping; diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 23c24fe98c00d4..e14f86a8ef5258 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, list_for_each_entry(d, &iommu->domain_list, next) { ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, - npage << PAGE_SHIFT, prot | IOMMU_CACHE); + npage << PAGE_SHIFT, prot | IOMMU_CACHE, + GFP_KERNEL); if (ret) goto unwind; @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, size = npage << PAGE_SHIFT; } - ret = iommu_map(domain->domain, iova, phys, - size, dma->prot | IOMMU_CACHE); + ret = iommu_map(domain->domain, iova, phys, size, + dma->prot | IOMMU_CACHE, GFP_KERNEL); if (ret) { if (!dma->iommu_mapped) { vfio_unpin_pages_remote(dma, iova, @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) return; ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); if (!ret) { size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index 166044642fd5cc..e555c3bd1c030b 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, r = ops->set_map(vdpa, asid, iotlb); } else { r = iommu_map(v->domain, iova, pa, size, - perm_to_iommu_flags(perm)); + perm_to_iommu_flags(perm), GFP_KERNEL); } if (r) { vhost_iotlb_del_range(iotlb, iova, iova + size - 1); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 46e1347bfa2286..d2020994f292db 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) } static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { return -ENODEV; } -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media The internal mechanisms support this, but instead of exposting the gfp to the caller it wrappers it into iommu_map() and iommu_map_atomic() Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/arm/mm/dma-mapping.c | 11 +++++++---- .../gpu/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 | 2 +- drivers/iommu/iommu.c | 17 ++++++----------- drivers/iommu/iommufd/pages.c | 6 ++++-- 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 | 4 ++-- 16 files changed, 43 insertions(+), 38 deletions(-) diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index c135f6e37a00ca..8bc01071474ab7 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, len = (j - i) << PAGE_SHIFT; ret = iommu_map(mapping->domain, iova, phys, len, - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), + GFP_KERNEL); if (ret < 0) goto fail; iova += len; @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, iova, phys, len, prot); + ret = iommu_map(mapping->domain, iova, phys, len, prot, + GFP_KERNEL); if (ret < 0) goto fail; count += len >> PAGE_SHIFT; @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, + prot, GFP_KERNEL); if (ret < 0) goto fail; @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); if (ret < 0) goto fail; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, u32 offset = (r->offset + i) << imem->iommu_pgshift; ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret < 0) { nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 7bd2e65c2a16c5..6ca9f396e55be4 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) *dma = iova_dma_addr(&tegra->carveout.domain, alloc); err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), - size, IOMMU_READ | IOMMU_WRITE); + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (err < 0) goto free_iova; diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 103fda055394ab..4ddfcd2138c95b 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) pb->dma = iova_dma_addr(&host1x->iova, alloc); err = iommu_map(host1x->domain, pb->dma, pb->phys, size, - IOMMU_READ); + IOMMU_READ, GFP_KERNEL); if (err) goto iommu_free_iova; } else { diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index c301b3be9f303d..aeeaca65ace96a 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index f798c44e090337..8bdb65e7686ff9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, if (!iova) goto out_free_page; - if (iommu_map(domain, iova, msi_addr, size, prot)) + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) goto out_free_iova; INIT_LIST_HEAD(&msi_page->list); diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index de91dd88705bd3..fe29fc2140b132 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, if (map_size) { ret = iommu_map(domain, addr - map_size, addr - map_size, map_size, - entry->prot); + entry->prot, GFP_KERNEL); if (ret) goto out; map_size = 0; @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, return ret; } -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) +int iommu_map(struct iommu_domain *domain, unsigned long iova, + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); return ret; } - -int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - might_sleep(); - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map); int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot) { - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); } EXPORT_SYMBOL_GPL(iommu_map_atomic); diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 1e1d3509efae5e..22cc3bb0c6c55a 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, size % PAGE_SIZE); while (size) { - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, + GFP_KERNEL); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, else rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, - next_iova - iova, area->iommu_prot); + next_iova - iova, area->iommu_prot, + GFP_KERNEL); if (rc) goto err_unmap; iova = next_iova; diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c index 142d4c74017c04..07d4dceb5e72c7 100644 --- a/drivers/media/platform/qcom/venus/firmware.c +++ b/drivers/media/platform/qcom/venus/firmware.c @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, core->fw.mapped_mem_size = mem_size; ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); if (ret) { dev_err(dev, "could not map video firmware region\n"); return ret; diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c index 9ec5af323f731d..991a7d39f06661 100644 --- a/drivers/net/ipa/ipa_mem.c +++ b/drivers/net/ipa/ipa_mem.c @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index cfcb759a87deac..9a82f0336d9537 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, ar->msa.paddr, ar->msa.mem_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath10k_err(ar, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c index d34a4d6325b2b4..df8fdc7067f99c 100644 --- a/drivers/net/wireless/ath/ath11k/ahb.c +++ b/drivers/net/wireless/ath/ath11k/ahb.c @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); goto err_iommu_unmap; diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 1cd4815a6dd197..80072b6b628358 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, if (!mapping) return -ENOMEM; - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, + GFP_KERNEL); if (ret) { dev_err(dev, "failed to map devmem: %d\n", ret); goto out; @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, } ret = iommu_map(rproc->domain, mem->da, dma, mem->len, - mem->flags); + mem->flags, GFP_KERNEL); if (ret) { dev_err(dev, "iommu_map failed: %d\n", ret); goto free_mapping; diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 23c24fe98c00d4..e14f86a8ef5258 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, list_for_each_entry(d, &iommu->domain_list, next) { ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, - npage << PAGE_SHIFT, prot | IOMMU_CACHE); + npage << PAGE_SHIFT, prot | IOMMU_CACHE, + GFP_KERNEL); if (ret) goto unwind; @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, size = npage << PAGE_SHIFT; } - ret = iommu_map(domain->domain, iova, phys, - size, dma->prot | IOMMU_CACHE); + ret = iommu_map(domain->domain, iova, phys, size, + dma->prot | IOMMU_CACHE, GFP_KERNEL); if (ret) { if (!dma->iommu_mapped) { vfio_unpin_pages_remote(dma, iova, @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) return; ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); if (!ret) { size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index 166044642fd5cc..e555c3bd1c030b 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, r = ops->set_map(vdpa, asid, iotlb); } else { r = iommu_map(v->domain, iova, pa, size, - perm_to_iommu_flags(perm)); + perm_to_iommu_flags(perm), GFP_KERNEL); } if (r) { vhost_iotlb_del_range(iotlb, iova, iova + size - 1); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 46e1347bfa2286..d2020994f292db 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) } static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { return -ENODEV; } -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization The internal mechanisms support this, but instead of exposting the gfp to the caller it wrappers it into iommu_map() and iommu_map_atomic() Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/arm/mm/dma-mapping.c | 11 +++++++---- .../gpu/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 | 2 +- drivers/iommu/iommu.c | 17 ++++++----------- drivers/iommu/iommufd/pages.c | 6 ++++-- 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 | 4 ++-- 16 files changed, 43 insertions(+), 38 deletions(-) diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index c135f6e37a00ca..8bc01071474ab7 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, len = (j - i) << PAGE_SHIFT; ret = iommu_map(mapping->domain, iova, phys, len, - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), + GFP_KERNEL); if (ret < 0) goto fail; iova += len; @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, iova, phys, len, prot); + ret = iommu_map(mapping->domain, iova, phys, len, prot, + GFP_KERNEL); if (ret < 0) goto fail; count += len >> PAGE_SHIFT; @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, + prot, GFP_KERNEL); if (ret < 0) goto fail; @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); if (ret < 0) goto fail; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, u32 offset = (r->offset + i) << imem->iommu_pgshift; ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret < 0) { nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 7bd2e65c2a16c5..6ca9f396e55be4 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) *dma = iova_dma_addr(&tegra->carveout.domain, alloc); err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), - size, IOMMU_READ | IOMMU_WRITE); + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (err < 0) goto free_iova; diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 103fda055394ab..4ddfcd2138c95b 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) pb->dma = iova_dma_addr(&host1x->iova, alloc); err = iommu_map(host1x->domain, pb->dma, pb->phys, size, - IOMMU_READ); + IOMMU_READ, GFP_KERNEL); if (err) goto iommu_free_iova; } else { diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index c301b3be9f303d..aeeaca65ace96a 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index f798c44e090337..8bdb65e7686ff9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, if (!iova) goto out_free_page; - if (iommu_map(domain, iova, msi_addr, size, prot)) + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) goto out_free_iova; INIT_LIST_HEAD(&msi_page->list); diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index de91dd88705bd3..fe29fc2140b132 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, if (map_size) { ret = iommu_map(domain, addr - map_size, addr - map_size, map_size, - entry->prot); + entry->prot, GFP_KERNEL); if (ret) goto out; map_size = 0; @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, return ret; } -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) +int iommu_map(struct iommu_domain *domain, unsigned long iova, + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); return ret; } - -int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - might_sleep(); - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map); int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot) { - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); } EXPORT_SYMBOL_GPL(iommu_map_atomic); diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 1e1d3509efae5e..22cc3bb0c6c55a 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, size % PAGE_SIZE); while (size) { - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, + GFP_KERNEL); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, else rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, - next_iova - iova, area->iommu_prot); + next_iova - iova, area->iommu_prot, + GFP_KERNEL); if (rc) goto err_unmap; iova = next_iova; diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c index 142d4c74017c04..07d4dceb5e72c7 100644 --- a/drivers/media/platform/qcom/venus/firmware.c +++ b/drivers/media/platform/qcom/venus/firmware.c @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, core->fw.mapped_mem_size = mem_size; ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); if (ret) { dev_err(dev, "could not map video firmware region\n"); return ret; diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c index 9ec5af323f731d..991a7d39f06661 100644 --- a/drivers/net/ipa/ipa_mem.c +++ b/drivers/net/ipa/ipa_mem.c @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index cfcb759a87deac..9a82f0336d9537 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, ar->msa.paddr, ar->msa.mem_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath10k_err(ar, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c index d34a4d6325b2b4..df8fdc7067f99c 100644 --- a/drivers/net/wireless/ath/ath11k/ahb.c +++ b/drivers/net/wireless/ath/ath11k/ahb.c @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); goto err_iommu_unmap; diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 1cd4815a6dd197..80072b6b628358 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, if (!mapping) return -ENOMEM; - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, + GFP_KERNEL); if (ret) { dev_err(dev, "failed to map devmem: %d\n", ret); goto out; @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, } ret = iommu_map(rproc->domain, mem->da, dma, mem->len, - mem->flags); + mem->flags, GFP_KERNEL); if (ret) { dev_err(dev, "iommu_map failed: %d\n", ret); goto free_mapping; diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 23c24fe98c00d4..e14f86a8ef5258 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, list_for_each_entry(d, &iommu->domain_list, next) { ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, - npage << PAGE_SHIFT, prot | IOMMU_CACHE); + npage << PAGE_SHIFT, prot | IOMMU_CACHE, + GFP_KERNEL); if (ret) goto unwind; @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, size = npage << PAGE_SHIFT; } - ret = iommu_map(domain->domain, iova, phys, - size, dma->prot | IOMMU_CACHE); + ret = iommu_map(domain->domain, iova, phys, size, + dma->prot | IOMMU_CACHE, GFP_KERNEL); if (ret) { if (!dma->iommu_mapped) { vfio_unpin_pages_remote(dma, iova, @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) return; ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); if (!ret) { size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index 166044642fd5cc..e555c3bd1c030b 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, r = ops->set_map(vdpa, asid, iotlb); } else { r = iommu_map(v->domain, iova, pa, size, - perm_to_iommu_flags(perm)); + perm_to_iommu_flags(perm), GFP_KERNEL); } if (r) { vhost_iotlb_del_range(iotlb, iova, iova + size - 1); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 46e1347bfa2286..d2020994f292db 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) } static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { return -ENODEV; } -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization The internal mechanisms support this, but instead of exposting the gfp to the caller it wrappers it into iommu_map() and iommu_map_atomic() Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/arm/mm/dma-mapping.c | 11 +++++++---- .../gpu/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 | 2 +- drivers/iommu/iommu.c | 17 ++++++----------- drivers/iommu/iommufd/pages.c | 6 ++++-- 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 | 4 ++-- 16 files changed, 43 insertions(+), 38 deletions(-) diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c index c135f6e37a00ca..8bc01071474ab7 100644 --- a/arch/arm/mm/dma-mapping.c +++ b/arch/arm/mm/dma-mapping.c @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, len = (j - i) << PAGE_SHIFT; ret = iommu_map(mapping->domain, iova, phys, len, - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), + GFP_KERNEL); if (ret < 0) goto fail; iova += len; @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, iova, phys, len, prot); + ret = iommu_map(mapping->domain, iova, phys, len, prot, + GFP_KERNEL); if (ret < 0) goto fail; count += len >> PAGE_SHIFT; @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, prot = __dma_info_to_prot(dir, attrs); - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, + prot, GFP_KERNEL); if (ret < 0) goto fail; @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); if (ret < 0) goto fail; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, u32 offset = (r->offset + i) << imem->iommu_pgshift; ret = iommu_map(imem->domain, offset, node->dma_addrs[i], - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret < 0) { nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index 7bd2e65c2a16c5..6ca9f396e55be4 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) *dma = iova_dma_addr(&tegra->carveout.domain, alloc); err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), - size, IOMMU_READ | IOMMU_WRITE); + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (err < 0) goto free_iova; diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index 103fda055394ab..4ddfcd2138c95b 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) pb->dma = iova_dma_addr(&host1x->iova, alloc); err = iommu_map(host1x->domain, pb->dma, pb->phys, size, - IOMMU_READ); + IOMMU_READ, GFP_KERNEL); if (err) goto iommu_free_iova; } else { diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c index c301b3be9f303d..aeeaca65ace96a 100644 --- a/drivers/infiniband/hw/usnic/usnic_uiom.c +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", va_start, &pa_start, size, flags); err = iommu_map(pd->domain, va_start, pa_start, - size, flags); + size, flags, GFP_KERNEL); if (err) { usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", va_start, &pa_start, size, err); diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index f798c44e090337..8bdb65e7686ff9 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, if (!iova) goto out_free_page; - if (iommu_map(domain, iova, msi_addr, size, prot)) + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) goto out_free_iova; INIT_LIST_HEAD(&msi_page->list); diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index de91dd88705bd3..fe29fc2140b132 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, if (map_size) { ret = iommu_map(domain, addr - map_size, addr - map_size, map_size, - entry->prot); + entry->prot, GFP_KERNEL); if (ret) goto out; map_size = 0; @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, return ret; } -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) +int iommu_map(struct iommu_domain *domain, unsigned long iova, + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); return ret; } - -int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - might_sleep(); - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map); int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot) { - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); } EXPORT_SYMBOL_GPL(iommu_map_atomic); diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 1e1d3509efae5e..22cc3bb0c6c55a 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, size % PAGE_SIZE); while (size) { - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, + GFP_KERNEL); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, else rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, - next_iova - iova, area->iommu_prot); + next_iova - iova, area->iommu_prot, + GFP_KERNEL); if (rc) goto err_unmap; iova = next_iova; diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c index 142d4c74017c04..07d4dceb5e72c7 100644 --- a/drivers/media/platform/qcom/venus/firmware.c +++ b/drivers/media/platform/qcom/venus/firmware.c @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, core->fw.mapped_mem_size = mem_size; ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); if (ret) { dev_err(dev, "could not map video firmware region\n"); return ret; diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c index 9ec5af323f731d..991a7d39f06661 100644 --- a/drivers/net/ipa/ipa_mem.c +++ b/drivers/net/ipa/ipa_mem.c @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) size = PAGE_ALIGN(size + addr - phys); iova = phys; /* We just want a direct mapping */ - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, + GFP_KERNEL); if (ret) return ret; diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c index cfcb759a87deac..9a82f0336d9537 100644 --- a/drivers/net/wireless/ath/ath10k/snoc.c +++ b/drivers/net/wireless/ath/ath10k/snoc.c @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, ar->msa.paddr, ar->msa.mem_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath10k_err(ar, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c index d34a4d6325b2b4..df8fdc7067f99c 100644 --- a/drivers/net/wireless/ath/ath11k/ahb.c +++ b/drivers/net/wireless/ath/ath11k/ahb.c @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware region: %d\n", ret); goto err_iommu_detach; @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, - IOMMU_READ | IOMMU_WRITE); + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); if (ret) { ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); goto err_iommu_unmap; diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c index 1cd4815a6dd197..80072b6b628358 100644 --- a/drivers/remoteproc/remoteproc_core.c +++ b/drivers/remoteproc/remoteproc_core.c @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, if (!mapping) return -ENOMEM; - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, + GFP_KERNEL); if (ret) { dev_err(dev, "failed to map devmem: %d\n", ret); goto out; @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, } ret = iommu_map(rproc->domain, mem->da, dma, mem->len, - mem->flags); + mem->flags, GFP_KERNEL); if (ret) { dev_err(dev, "iommu_map failed: %d\n", ret); goto free_mapping; diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 23c24fe98c00d4..e14f86a8ef5258 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, list_for_each_entry(d, &iommu->domain_list, next) { ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, - npage << PAGE_SHIFT, prot | IOMMU_CACHE); + npage << PAGE_SHIFT, prot | IOMMU_CACHE, + GFP_KERNEL); if (ret) goto unwind; @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, size = npage << PAGE_SHIFT; } - ret = iommu_map(domain->domain, iova, phys, - size, dma->prot | IOMMU_CACHE); + ret = iommu_map(domain->domain, iova, phys, size, + dma->prot | IOMMU_CACHE, GFP_KERNEL); if (ret) { if (!dma->iommu_mapped) { vfio_unpin_pages_remote(dma, iova, @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) return; ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); if (!ret) { size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c index 166044642fd5cc..e555c3bd1c030b 100644 --- a/drivers/vhost/vdpa.c +++ b/drivers/vhost/vdpa.c @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, r = ops->set_map(vdpa, asid, iotlb); } else { r = iommu_map(v->domain, iova, pa, size, - perm_to_iommu_flags(perm)); + perm_to_iommu_flags(perm), GFP_KERNEL); } if (r) { vhost_iotlb_del_range(iotlb, iova, iova + size - 1); diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 46e1347bfa2286..d2020994f292db 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) } static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) { return -ENODEV; } -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-06 16:42 ` Jason Gunthorpe ` (4 preceding siblings ...) (?) @ 2023-01-06 17:15 ` Robin Murphy -1 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:15 ` Robin Murphy 0 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:15 ` Robin Murphy 0 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:15 ` Robin Murphy 0 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, netdev, ath10k, linux-wireless, linux-tegra, virtualization, ath11k, linux-arm-kernel, linux-media On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:15 ` Robin Murphy 0 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:15 ` Robin Murphy 0 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:15 ` Robin Murphy 0 siblings, 0 replies; 144+ messages in thread From: Robin Murphy @ 2023-01-06 17:15 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On 2023-01-06 16:42, Jason Gunthorpe wrote: > The internal mechanisms support this, but instead of exposting the gfp to > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. FWIW, since we *do* have two variants already, I think I'd have a mild preference for leaving the regular map calls as-is (i.e. implicit GFP_KERNEL), and just generalising the _atomic versions for the special cases. However, echoing the recent activity over on the DMA API side of things, I think it's still worth proactively constraining the set of permissible flags, lest we end up with more weird problems if stuff that doesn't really make sense, like GFP_COMP or zone flags, manages to leak through (that may have been part of the reason for having the current wrappers rather than a bare gfp argument in the first place, I forget now). Thanks, Robin. > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/arm/mm/dma-mapping.c | 11 +++++++---- > .../gpu/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 | 2 +- > drivers/iommu/iommu.c | 17 ++++++----------- > drivers/iommu/iommufd/pages.c | 6 ++++-- > 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 | 4 ++-- > 16 files changed, 43 insertions(+), 38 deletions(-) > > diff --git a/arch/arm/mm/dma-mapping.c b/arch/arm/mm/dma-mapping.c > index c135f6e37a00ca..8bc01071474ab7 100644 > --- a/arch/arm/mm/dma-mapping.c > +++ b/arch/arm/mm/dma-mapping.c > @@ -984,7 +984,8 @@ __iommu_create_mapping(struct device *dev, struct page **pages, size_t size, > > len = (j - i) << PAGE_SHIFT; > ret = iommu_map(mapping->domain, iova, phys, len, > - __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs)); > + __dma_info_to_prot(DMA_BIDIRECTIONAL, attrs), > + GFP_KERNEL); > if (ret < 0) > goto fail; > iova += len; > @@ -1207,7 +1208,8 @@ static int __map_sg_chunk(struct device *dev, struct scatterlist *sg, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, iova, phys, len, prot); > + ret = iommu_map(mapping->domain, iova, phys, len, prot, > + GFP_KERNEL); > if (ret < 0) > goto fail; > count += len >> PAGE_SHIFT; > @@ -1379,7 +1381,8 @@ static dma_addr_t arm_iommu_map_page(struct device *dev, struct page *page, > > prot = __dma_info_to_prot(dir, attrs); > > - ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, prot); > + ret = iommu_map(mapping->domain, dma_addr, page_to_phys(page), len, > + prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > @@ -1443,7 +1446,7 @@ static dma_addr_t arm_iommu_map_resource(struct device *dev, > > prot = __dma_info_to_prot(dir, attrs) | IOMMU_MMIO; > > - ret = iommu_map(mapping->domain, dma_addr, addr, len, prot); > + ret = iommu_map(mapping->domain, dma_addr, addr, len, prot, GFP_KERNEL); > if (ret < 0) > goto fail; > > diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > index 648ecf5a8fbc2a..a4ac94a2ab57fc 100644 > --- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/gk20a.c > @@ -475,7 +475,8 @@ gk20a_instobj_ctor_iommu(struct gk20a_instmem *imem, u32 npages, u32 align, > u32 offset = (r->offset + i) << imem->iommu_pgshift; > > ret = iommu_map(imem->domain, offset, node->dma_addrs[i], > - PAGE_SIZE, IOMMU_READ | IOMMU_WRITE); > + PAGE_SIZE, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret < 0) { > nvkm_error(subdev, "IOMMU mapping failure: %d\n", ret); > > diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c > index 7bd2e65c2a16c5..6ca9f396e55be4 100644 > --- a/drivers/gpu/drm/tegra/drm.c > +++ b/drivers/gpu/drm/tegra/drm.c > @@ -1057,7 +1057,7 @@ void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma) > > *dma = iova_dma_addr(&tegra->carveout.domain, alloc); > err = iommu_map(tegra->domain, *dma, virt_to_phys(virt), > - size, IOMMU_READ | IOMMU_WRITE); > + size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (err < 0) > goto free_iova; > > diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c > index 103fda055394ab..4ddfcd2138c95b 100644 > --- a/drivers/gpu/host1x/cdma.c > +++ b/drivers/gpu/host1x/cdma.c > @@ -105,7 +105,7 @@ static int host1x_pushbuffer_init(struct push_buffer *pb) > > pb->dma = iova_dma_addr(&host1x->iova, alloc); > err = iommu_map(host1x->domain, pb->dma, pb->phys, size, > - IOMMU_READ); > + IOMMU_READ, GFP_KERNEL); > if (err) > goto iommu_free_iova; > } else { > diff --git a/drivers/infiniband/hw/usnic/usnic_uiom.c b/drivers/infiniband/hw/usnic/usnic_uiom.c > index c301b3be9f303d..aeeaca65ace96a 100644 > --- a/drivers/infiniband/hw/usnic/usnic_uiom.c > +++ b/drivers/infiniband/hw/usnic/usnic_uiom.c > @@ -277,7 +277,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > @@ -294,7 +294,7 @@ static int usnic_uiom_map_sorted_intervals(struct list_head *intervals, > usnic_dbg("va 0x%lx pa %pa size 0x%zx flags 0x%x\n", > va_start, &pa_start, size, flags); > err = iommu_map(pd->domain, va_start, pa_start, > - size, flags); > + size, flags, GFP_KERNEL); > if (err) { > usnic_err("Failed to map va 0x%lx pa %pa size 0x%zx with err %d\n", > va_start, &pa_start, size, err); > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c > index f798c44e090337..8bdb65e7686ff9 100644 > --- a/drivers/iommu/dma-iommu.c > +++ b/drivers/iommu/dma-iommu.c > @@ -1615,7 +1615,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, > if (!iova) > goto out_free_page; > > - if (iommu_map(domain, iova, msi_addr, size, prot)) > + if (iommu_map(domain, iova, msi_addr, size, prot, GFP_KERNEL)) > goto out_free_iova; > > INIT_LIST_HEAD(&msi_page->list); > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c > index de91dd88705bd3..fe29fc2140b132 100644 > --- a/drivers/iommu/iommu.c > +++ b/drivers/iommu/iommu.c > @@ -930,7 +930,7 @@ static int iommu_create_device_direct_mappings(struct iommu_group *group, > if (map_size) { > ret = iommu_map(domain, addr - map_size, > addr - map_size, map_size, > - entry->prot); > + entry->prot, GFP_KERNEL); > if (ret) > goto out; > map_size = 0; > @@ -2360,31 +2360,26 @@ static int __iommu_map(struct iommu_domain *domain, unsigned long iova, > return ret; > } > > -static int _iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > +int iommu_map(struct iommu_domain *domain, unsigned long iova, > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > const struct iommu_domain_ops *ops = domain->ops; > int ret; > > + might_sleep_if(gfpflags_allow_blocking(gfp)); > + > ret = __iommu_map(domain, iova, paddr, size, prot, gfp); > if (ret == 0 && ops->iotlb_sync_map) > ops->iotlb_sync_map(domain, iova, size); > > return ret; > } > - > -int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > -{ > - might_sleep(); > - return _iommu_map(domain, iova, paddr, size, prot, GFP_KERNEL); > -} > EXPORT_SYMBOL_GPL(iommu_map); > > int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot) > { > - return _iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > + return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); > } > EXPORT_SYMBOL_GPL(iommu_map_atomic); > > diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c > index 1e1d3509efae5e..22cc3bb0c6c55a 100644 > --- a/drivers/iommu/iommufd/pages.c > +++ b/drivers/iommu/iommufd/pages.c > @@ -456,7 +456,8 @@ static int batch_iommu_map_small(struct iommu_domain *domain, > size % PAGE_SIZE); > > while (size) { > - rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot); > + rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova += PAGE_SIZE; > @@ -500,7 +501,8 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, > else > rc = iommu_map(domain, iova, > PFN_PHYS(batch->pfns[cur]) + page_offset, > - next_iova - iova, area->iommu_prot); > + next_iova - iova, area->iommu_prot, > + GFP_KERNEL); > if (rc) > goto err_unmap; > iova = next_iova; > diff --git a/drivers/media/platform/qcom/venus/firmware.c b/drivers/media/platform/qcom/venus/firmware.c > index 142d4c74017c04..07d4dceb5e72c7 100644 > --- a/drivers/media/platform/qcom/venus/firmware.c > +++ b/drivers/media/platform/qcom/venus/firmware.c > @@ -158,7 +158,7 @@ static int venus_boot_no_tz(struct venus_core *core, phys_addr_t mem_phys, > core->fw.mapped_mem_size = mem_size; > > ret = iommu_map(iommu, VENUS_FW_START_ADDR, mem_phys, mem_size, > - IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV); > + IOMMU_READ | IOMMU_WRITE | IOMMU_PRIV, GFP_KERNEL); > if (ret) { > dev_err(dev, "could not map video firmware region\n"); > return ret; > diff --git a/drivers/net/ipa/ipa_mem.c b/drivers/net/ipa/ipa_mem.c > index 9ec5af323f731d..991a7d39f06661 100644 > --- a/drivers/net/ipa/ipa_mem.c > +++ b/drivers/net/ipa/ipa_mem.c > @@ -466,7 +466,8 @@ static int ipa_imem_init(struct ipa *ipa, unsigned long addr, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > @@ -574,7 +575,8 @@ static int ipa_smem_init(struct ipa *ipa, u32 item, size_t size) > size = PAGE_ALIGN(size + addr - phys); > iova = phys; /* We just want a direct mapping */ > > - ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE); > + ret = iommu_map(domain, iova, phys, size, IOMMU_READ | IOMMU_WRITE, > + GFP_KERNEL); > if (ret) > return ret; > > diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c > index cfcb759a87deac..9a82f0336d9537 100644 > --- a/drivers/net/wireless/ath/ath10k/snoc.c > +++ b/drivers/net/wireless/ath/ath10k/snoc.c > @@ -1639,7 +1639,7 @@ static int ath10k_fw_init(struct ath10k *ar) > > ret = iommu_map(iommu_dom, ar_snoc->fw.fw_start_addr, > ar->msa.paddr, ar->msa.mem_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath10k_err(ar, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > diff --git a/drivers/net/wireless/ath/ath11k/ahb.c b/drivers/net/wireless/ath/ath11k/ahb.c > index d34a4d6325b2b4..df8fdc7067f99c 100644 > --- a/drivers/net/wireless/ath/ath11k/ahb.c > +++ b/drivers/net/wireless/ath/ath11k/ahb.c > @@ -1021,7 +1021,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.msa_paddr, > ab_ahb->fw.msa_paddr, ab_ahb->fw.msa_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware region: %d\n", ret); > goto err_iommu_detach; > @@ -1029,7 +1029,7 @@ static int ath11k_ahb_fw_resources_init(struct ath11k_base *ab) > > ret = iommu_map(iommu_dom, ab_ahb->fw.ce_paddr, > ab_ahb->fw.ce_paddr, ab_ahb->fw.ce_size, > - IOMMU_READ | IOMMU_WRITE); > + IOMMU_READ | IOMMU_WRITE, GFP_KERNEL); > if (ret) { > ath11k_err(ab, "failed to map firmware CE region: %d\n", ret); > goto err_iommu_unmap; > diff --git a/drivers/remoteproc/remoteproc_core.c b/drivers/remoteproc/remoteproc_core.c > index 1cd4815a6dd197..80072b6b628358 100644 > --- a/drivers/remoteproc/remoteproc_core.c > +++ b/drivers/remoteproc/remoteproc_core.c > @@ -643,7 +643,8 @@ static int rproc_handle_devmem(struct rproc *rproc, void *ptr, > if (!mapping) > return -ENOMEM; > > - ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags); > + ret = iommu_map(rproc->domain, rsc->da, rsc->pa, rsc->len, rsc->flags, > + GFP_KERNEL); > if (ret) { > dev_err(dev, "failed to map devmem: %d\n", ret); > goto out; > @@ -737,7 +738,7 @@ static int rproc_alloc_carveout(struct rproc *rproc, > } > > ret = iommu_map(rproc->domain, mem->da, dma, mem->len, > - mem->flags); > + mem->flags, GFP_KERNEL); > if (ret) { > dev_err(dev, "iommu_map failed: %d\n", ret); > goto free_mapping; > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 23c24fe98c00d4..e14f86a8ef5258 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -1480,7 +1480,8 @@ static int vfio_iommu_map(struct vfio_iommu *iommu, dma_addr_t iova, > > list_for_each_entry(d, &iommu->domain_list, next) { > ret = iommu_map(d->domain, iova, (phys_addr_t)pfn << PAGE_SHIFT, > - npage << PAGE_SHIFT, prot | IOMMU_CACHE); > + npage << PAGE_SHIFT, prot | IOMMU_CACHE, > + GFP_KERNEL); > if (ret) > goto unwind; > > @@ -1777,8 +1778,8 @@ static int vfio_iommu_replay(struct vfio_iommu *iommu, > size = npage << PAGE_SHIFT; > } > > - ret = iommu_map(domain->domain, iova, phys, > - size, dma->prot | IOMMU_CACHE); > + ret = iommu_map(domain->domain, iova, phys, size, > + dma->prot | IOMMU_CACHE, GFP_KERNEL); > if (ret) { > if (!dma->iommu_mapped) { > vfio_unpin_pages_remote(dma, iova, > @@ -1866,7 +1867,7 @@ static void vfio_test_domain_fgsp(struct vfio_domain *domain) > return; > > ret = iommu_map(domain->domain, 0, page_to_phys(pages), PAGE_SIZE * 2, > - IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE); > + IOMMU_READ | IOMMU_WRITE | IOMMU_CACHE, GFP_KERNEL); > if (!ret) { > size_t unmapped = iommu_unmap(domain->domain, 0, PAGE_SIZE); > > diff --git a/drivers/vhost/vdpa.c b/drivers/vhost/vdpa.c > index 166044642fd5cc..e555c3bd1c030b 100644 > --- a/drivers/vhost/vdpa.c > +++ b/drivers/vhost/vdpa.c > @@ -777,7 +777,7 @@ static int vhost_vdpa_map(struct vhost_vdpa *v, struct vhost_iotlb *iotlb, > r = ops->set_map(vdpa, asid, iotlb); > } else { > r = iommu_map(v->domain, iova, pa, size, > - perm_to_iommu_flags(perm)); > + perm_to_iommu_flags(perm), GFP_KERNEL); > } > if (r) { > vhost_iotlb_del_range(iotlb, iova, iova + size - 1); > diff --git a/include/linux/iommu.h b/include/linux/iommu.h > index 46e1347bfa2286..d2020994f292db 100644 > --- a/include/linux/iommu.h > +++ b/include/linux/iommu.h > @@ -467,7 +467,7 @@ extern int iommu_sva_unbind_gpasid(struct iommu_domain *domain, > extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); > extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); > extern int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot); > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp); > extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, > phys_addr_t paddr, size_t size, int prot); > extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, > @@ -773,7 +773,7 @@ static inline struct iommu_domain *iommu_get_domain_for_dev(struct device *dev) > } > > static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, > - phys_addr_t paddr, size_t size, int prot) > + phys_addr_t paddr, size_t size, int prot, gfp_t gfp) > { > return -ENODEV; > } -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-06 17:15 ` Robin Murphy ` (3 preceding siblings ...) (?) @ 2023-01-06 17:24 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 17:24 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > On 2023-01-06 16:42, Jason Gunthorpe wrote: > > The internal mechanisms support this, but instead of exposting the gfp to > > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. > > FWIW, since we *do* have two variants already, I think I'd have a mild > preference for leaving the regular map calls as-is (i.e. implicit > GFP_KERNEL), and just generalising the _atomic versions for the special > cases. I think it is just better to follow kernel convention and have allocation functions include the GFP because it is a clear signal to the user that there is an allocation hidden inside the API. The whole point of gfp is not to have multitudes of every function for every allocation mode. There are not so many callers that it seems worth worrying about removing the extra GFP_KERNEL argument. > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). Yeah, that can be done Thanks, Jason _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:24 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 17:24 UTC (permalink / raw) To: Robin Murphy Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, Joerg Roedel, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Lu Baolu On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > On 2023-01-06 16:42, Jason Gunthorpe wrote: > > The internal mechanisms support this, but instead of exposting the gfp to > > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. > > FWIW, since we *do* have two variants already, I think I'd have a mild > preference for leaving the regular map calls as-is (i.e. implicit > GFP_KERNEL), and just generalising the _atomic versions for the special > cases. I think it is just better to follow kernel convention and have allocation functions include the GFP because it is a clear signal to the user that there is an allocation hidden inside the API. The whole point of gfp is not to have multitudes of every function for every allocation mode. There are not so many callers that it seems worth worrying about removing the extra GFP_KERNEL argument. > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). Yeah, that can be done Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:24 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 17:24 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > On 2023-01-06 16:42, Jason Gunthorpe wrote: > > The internal mechanisms support this, but instead of exposting the gfp to > > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. > > FWIW, since we *do* have two variants already, I think I'd have a mild > preference for leaving the regular map calls as-is (i.e. implicit > GFP_KERNEL), and just generalising the _atomic versions for the special > cases. I think it is just better to follow kernel convention and have allocation functions include the GFP because it is a clear signal to the user that there is an allocation hidden inside the API. The whole point of gfp is not to have multitudes of every function for every allocation mode. There are not so many callers that it seems worth worrying about removing the extra GFP_KERNEL argument. > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). Yeah, that can be done Thanks, Jason _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:24 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 17:24 UTC (permalink / raw) To: Robin Murphy Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, Joerg Roedel, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Lu Baolu On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > On 2023-01-06 16:42, Jason Gunthorpe wrote: > > The internal mechanisms support this, but instead of exposting the gfp to > > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. > > FWIW, since we *do* have two variants already, I think I'd have a mild > preference for leaving the regular map calls as-is (i.e. implicit > GFP_KERNEL), and just generalising the _atomic versions for the special > cases. I think it is just better to follow kernel convention and have allocation functions include the GFP because it is a clear signal to the user that there is an allocation hidden inside the API. The whole point of gfp is not to have multitudes of every function for every allocation mode. There are not so many callers that it seems worth worrying about removing the extra GFP_KERNEL argument. > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). Yeah, that can be done Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:24 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 17:24 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > On 2023-01-06 16:42, Jason Gunthorpe wrote: > > The internal mechanisms support this, but instead of exposting the gfp to > > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. > > FWIW, since we *do* have two variants already, I think I'd have a mild > preference for leaving the regular map calls as-is (i.e. implicit > GFP_KERNEL), and just generalising the _atomic versions for the special > cases. I think it is just better to follow kernel convention and have allocation functions include the GFP because it is a clear signal to the user that there is an allocation hidden inside the API. The whole point of gfp is not to have multitudes of every function for every allocation mode. There are not so many callers that it seems worth worrying about removing the extra GFP_KERNEL argument. > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). Yeah, that can be done Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-06 17:24 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 17:24 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > On 2023-01-06 16:42, Jason Gunthorpe wrote: > > The internal mechanisms support this, but instead of exposting the gfp to > > the caller it wrappers it into iommu_map() and iommu_map_atomic() > > > > Fix this instead of adding more variants for GFP_KERNEL_ACCOUNT. > > FWIW, since we *do* have two variants already, I think I'd have a mild > preference for leaving the regular map calls as-is (i.e. implicit > GFP_KERNEL), and just generalising the _atomic versions for the special > cases. I think it is just better to follow kernel convention and have allocation functions include the GFP because it is a clear signal to the user that there is an allocation hidden inside the API. The whole point of gfp is not to have multitudes of every function for every allocation mode. There are not so many callers that it seems worth worrying about removing the extra GFP_KERNEL argument. > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). Yeah, that can be done Thanks, Jason -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-06 17:24 ` Jason Gunthorpe ` (4 preceding siblings ...) (?) @ 2023-01-20 9:24 ` Joerg Roedel -1 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 9:24 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 9:24 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 9:24 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, ath11k, linux-media, Niklas Schnelle, linux-arm-msm, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 9:24 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 9:24 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 9:24 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-20 9:24 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > I think it is just better to follow kernel convention and have > allocation functions include the GFP because it is a clear signal to > the user that there is an allocation hidden inside the API. The whole > point of gfp is not to have multitudes of every function for every > allocation mode. Well, having GFP parameters is not a strict kernel convention. There are places doing it differently and have sleeping and atomic variants of APIs. I have to say I like the latter more. But given that this leads to an invasion of API functions here which all do the same under the hood, I agree it is better to go with a GFP parameter here. Regards, Joerg -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-20 9:24 ` Joerg Roedel ` (3 preceding siblings ...) (?) @ 2023-01-20 17:53 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-20 17:53 UTC (permalink / raw) To: Joerg Roedel Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 10:24:55AM +0100, Joerg Roedel wrote: > On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > > I think it is just better to follow kernel convention and have > > allocation functions include the GFP because it is a clear signal to > > the user that there is an allocation hidden inside the API. The whole > > point of gfp is not to have multitudes of every function for every > > allocation mode. > > Well, having GFP parameters is not a strict kernel convention. There are > places doing it differently and have sleeping and atomic variants of > APIs. I have to say I like the latter more. But given that this leads to > an invasion of API functions here which all do the same under the hood, I > agree it is better to go with a GFP parameter here. Ok, I think we are done with this series, I'll stick it in linux-next for a bit and send you a PR so the trees stay in sync Thanks, Jason _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 17:53 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-20 17:53 UTC (permalink / raw) To: Joerg Roedel Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 20, 2023 at 10:24:55AM +0100, Joerg Roedel wrote: > On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > > I think it is just better to follow kernel convention and have > > allocation functions include the GFP because it is a clear signal to > > the user that there is an allocation hidden inside the API. The whole > > point of gfp is not to have multitudes of every function for every > > allocation mode. > > Well, having GFP parameters is not a strict kernel convention. There are > places doing it differently and have sleeping and atomic variants of > APIs. I have to say I like the latter more. But given that this leads to > an invasion of API functions here which all do the same under the hood, I > agree it is better to go with a GFP parameter here. Ok, I think we are done with this series, I'll stick it in linux-next for a bit and send you a PR so the trees stay in sync Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 17:53 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-20 17:53 UTC (permalink / raw) To: Joerg Roedel Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 10:24:55AM +0100, Joerg Roedel wrote: > On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > > I think it is just better to follow kernel convention and have > > allocation functions include the GFP because it is a clear signal to > > the user that there is an allocation hidden inside the API. The whole > > point of gfp is not to have multitudes of every function for every > > allocation mode. > > Well, having GFP parameters is not a strict kernel convention. There are > places doing it differently and have sleeping and atomic variants of > APIs. I have to say I like the latter more. But given that this leads to > an invasion of API functions here which all do the same under the hood, I > agree it is better to go with a GFP parameter here. Ok, I think we are done with this series, I'll stick it in linux-next for a bit and send you a PR so the trees stay in sync Thanks, Jason _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 17:53 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-20 17:53 UTC (permalink / raw) To: Joerg Roedel Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 20, 2023 at 10:24:55AM +0100, Joerg Roedel wrote: > On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > > I think it is just better to follow kernel convention and have > > allocation functions include the GFP because it is a clear signal to > > the user that there is an allocation hidden inside the API. The whole > > point of gfp is not to have multitudes of every function for every > > allocation mode. > > Well, having GFP parameters is not a strict kernel convention. There are > places doing it differently and have sleeping and atomic variants of > APIs. I have to say I like the latter more. But given that this leads to > an invasion of API functions here which all do the same under the hood, I > agree it is better to go with a GFP parameter here. Ok, I think we are done with this series, I'll stick it in linux-next for a bit and send you a PR so the trees stay in sync Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 17:53 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-20 17:53 UTC (permalink / raw) To: Joerg Roedel Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 10:24:55AM +0100, Joerg Roedel wrote: > On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > > I think it is just better to follow kernel convention and have > > allocation functions include the GFP because it is a clear signal to > > the user that there is an allocation hidden inside the API. The whole > > point of gfp is not to have multitudes of every function for every > > allocation mode. > > Well, having GFP parameters is not a strict kernel convention. There are > places doing it differently and have sleeping and atomic variants of > APIs. I have to say I like the latter more. But given that this leads to > an invasion of API functions here which all do the same under the hood, I > agree it is better to go with a GFP parameter here. Ok, I think we are done with this series, I'll stick it in linux-next for a bit and send you a PR so the trees stay in sync Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-20 17:53 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-20 17:53 UTC (permalink / raw) To: Joerg Roedel Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 10:24:55AM +0100, Joerg Roedel wrote: > On Fri, Jan 06, 2023 at 01:24:11PM -0400, Jason Gunthorpe wrote: > > I think it is just better to follow kernel convention and have > > allocation functions include the GFP because it is a clear signal to > > the user that there is an allocation hidden inside the API. The whole > > point of gfp is not to have multitudes of every function for every > > allocation mode. > > Well, having GFP parameters is not a strict kernel convention. There are > places doing it differently and have sleeping and atomic variants of > APIs. I have to say I like the latter more. But given that this leads to > an invasion of API functions here which all do the same under the hood, I > agree it is better to go with a GFP parameter here. Ok, I think we are done with this series, I'll stick it in linux-next for a bit and send you a PR so the trees stay in sync Thanks, Jason -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-20 17:53 ` Jason Gunthorpe ` (4 preceding siblings ...) (?) @ 2023-01-23 9:59 ` Joerg Roedel -1 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-23 9:59 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-23 9:59 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-23 9:59 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, ath11k, linux-media, Niklas Schnelle, linux-arm-msm, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-23 9:59 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Robin Murphy, Lu Baolu On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-23 9:59 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-23 9:59 ` Joerg Roedel 0 siblings, 0 replies; 144+ messages in thread From: Joerg Roedel @ 2023-01-23 9:59 UTC (permalink / raw) To: Jason Gunthorpe Cc: Robin Murphy, Lu Baolu, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 20, 2023 at 01:53:40PM -0400, Jason Gunthorpe wrote: > > Well, having GFP parameters is not a strict kernel convention. There are > > places doing it differently and have sleeping and atomic variants of > > APIs. I have to say I like the latter more. But given that this leads to > > an invasion of API functions here which all do the same under the hood, I > > agree it is better to go with a GFP parameter here. > > Ok, I think we are done with this series, I'll stick it in linux-next > for a bit and send you a PR so the trees stay in sync This series mostly touches parts outside of IOMMUFD, so we should follow the process here and let this reach linux-next via the IOMMU tree. Please send me a new version and I will put it into a separate branch where you can pull it from. Regards, Joerg -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() 2023-01-06 17:15 ` Robin Murphy ` (3 preceding siblings ...) (?) @ 2023-01-16 18:12 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-16 18:12 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). I did it like this: --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2368,6 +2368,11 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); @@ -2477,6 +2482,11 @@ ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); Will post a v2 when the driver people take a look Thanks, Jason _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-16 18:12 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-16 18:12 UTC (permalink / raw) To: Robin Murphy Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, Joerg Roedel, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Lu Baolu On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). I did it like this: --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2368,6 +2368,11 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); @@ -2477,6 +2482,11 @@ ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); Will post a v2 when the driver people take a look Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-16 18:12 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-16 18:12 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). I did it like this: --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2368,6 +2368,11 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); @@ -2477,6 +2482,11 @@ ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); Will post a v2 when the driver people take a look Thanks, Jason _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-16 18:12 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-16 18:12 UTC (permalink / raw) To: Robin Murphy Cc: kvm, nouveau, linux-remoteproc, dri-devel, virtualization, linux-stm32, linux-s390, Matthew Rosato, linux-rdma, Joerg Roedel, ath10k, iommu, Christian Borntraeger, ath11k, linux-media, Kevin Tian, Niklas Schnelle, linux-arm-msm, Alex Williamson, linux-tegra, linux-arm-kernel, netdev, linux-wireless, Lu Baolu On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). I did it like this: --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2368,6 +2368,11 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); @@ -2477,6 +2482,11 @@ ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); Will post a v2 when the driver people take a look Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-16 18:12 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-16 18:12 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). I did it like this: --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2368,6 +2368,11 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); @@ -2477,6 +2482,11 @@ ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); Will post a v2 when the driver people take a look Thanks, Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() @ 2023-01-16 18:12 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-16 18:12 UTC (permalink / raw) To: Robin Murphy Cc: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization On Fri, Jan 06, 2023 at 05:15:28PM +0000, Robin Murphy wrote: > However, echoing the recent activity over on the DMA API side of things, I > think it's still worth proactively constraining the set of permissible > flags, lest we end up with more weird problems if stuff that doesn't really > make sense, like GFP_COMP or zone flags, manages to leak through (that may > have been part of the reason for having the current wrappers rather than a > bare gfp argument in the first place, I forget now). I did it like this: --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2368,6 +2368,11 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + ret = __iommu_map(domain, iova, paddr, size, prot, gfp); if (ret == 0 && ops->iotlb_sync_map) ops->iotlb_sync_map(domain, iova, size); @@ -2477,6 +2482,11 @@ ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, might_sleep_if(gfpflags_allow_blocking(gfp)); + /* Discourage passing strange GFP flags */ + if (WARN_ON_ONCE(gfp & (__GFP_COMP | __GFP_DMA | __GFP_DMA32 | + __GFP_HIGHMEM))) + return -EINVAL; + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); Will post a v2 when the driver people take a look Thanks, Jason -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 2/8] iommu: Remove iommu_map_atomic() 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization There is only one call site and it can now just pass the GFP_ATOMIC to the normal iommu_map(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- drivers/iommu/iommu.c | 7 ------- include/linux/iommu.h | 9 --------- 3 files changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8bdb65e7686ff9..7016db569f81fc 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -713,7 +713,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys, if (!iova) return DMA_MAPPING_ERROR; - if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) { + if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) { iommu_dma_free_iova(cookie, iova, size, NULL); return DMA_MAPPING_ERROR; } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fe29fc2140b132..fee37bb246f3ea 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2376,13 +2376,6 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, } EXPORT_SYMBOL_GPL(iommu_map); -int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); -} -EXPORT_SYMBOL_GPL(iommu_map_atomic); - static size_t __iommu_unmap_pages(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d2020994f292db..521cd79700f4d8 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -468,8 +468,6 @@ extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot, gfp_t gfp); -extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size); extern size_t iommu_unmap_fast(struct iommu_domain *domain, @@ -778,13 +776,6 @@ static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, return -ENODEV; } -static inline int iommu_map_atomic(struct iommu_domain *domain, - unsigned long iova, phys_addr_t paddr, - size_t size, int prot) -{ - return -ENODEV; -} - static inline size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size) { -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 2/8] iommu: Remove iommu_map_atomic() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media There is only one call site and it can now just pass the GFP_ATOMIC to the normal iommu_map(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- drivers/iommu/iommu.c | 7 ------- include/linux/iommu.h | 9 --------- 3 files changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8bdb65e7686ff9..7016db569f81fc 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -713,7 +713,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys, if (!iova) return DMA_MAPPING_ERROR; - if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) { + if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) { iommu_dma_free_iova(cookie, iova, size, NULL); return DMA_MAPPING_ERROR; } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fe29fc2140b132..fee37bb246f3ea 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2376,13 +2376,6 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, } EXPORT_SYMBOL_GPL(iommu_map); -int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); -} -EXPORT_SYMBOL_GPL(iommu_map_atomic); - static size_t __iommu_unmap_pages(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d2020994f292db..521cd79700f4d8 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -468,8 +468,6 @@ extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot, gfp_t gfp); -extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size); extern size_t iommu_unmap_fast(struct iommu_domain *domain, @@ -778,13 +776,6 @@ static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, return -ENODEV; } -static inline int iommu_map_atomic(struct iommu_domain *domain, - unsigned long iova, phys_addr_t paddr, - size_t size, int prot) -{ - return -ENODEV; -} - static inline size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size) { -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 2/8] iommu: Remove iommu_map_atomic() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization There is only one call site and it can now just pass the GFP_ATOMIC to the normal iommu_map(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- drivers/iommu/iommu.c | 7 ------- include/linux/iommu.h | 9 --------- 3 files changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8bdb65e7686ff9..7016db569f81fc 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -713,7 +713,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys, if (!iova) return DMA_MAPPING_ERROR; - if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) { + if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) { iommu_dma_free_iova(cookie, iova, size, NULL); return DMA_MAPPING_ERROR; } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fe29fc2140b132..fee37bb246f3ea 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2376,13 +2376,6 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, } EXPORT_SYMBOL_GPL(iommu_map); -int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); -} -EXPORT_SYMBOL_GPL(iommu_map_atomic); - static size_t __iommu_unmap_pages(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d2020994f292db..521cd79700f4d8 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -468,8 +468,6 @@ extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot, gfp_t gfp); -extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size); extern size_t iommu_unmap_fast(struct iommu_domain *domain, @@ -778,13 +776,6 @@ static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, return -ENODEV; } -static inline int iommu_map_atomic(struct iommu_domain *domain, - unsigned long iova, phys_addr_t paddr, - size_t size, int prot) -{ - return -ENODEV; -} - static inline size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size) { -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 2/8] iommu: Remove iommu_map_atomic() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media There is only one call site and it can now just pass the GFP_ATOMIC to the normal iommu_map(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- drivers/iommu/iommu.c | 7 ------- include/linux/iommu.h | 9 --------- 3 files changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8bdb65e7686ff9..7016db569f81fc 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -713,7 +713,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys, if (!iova) return DMA_MAPPING_ERROR; - if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) { + if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) { iommu_dma_free_iova(cookie, iova, size, NULL); return DMA_MAPPING_ERROR; } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fe29fc2140b132..fee37bb246f3ea 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2376,13 +2376,6 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, } EXPORT_SYMBOL_GPL(iommu_map); -int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); -} -EXPORT_SYMBOL_GPL(iommu_map_atomic); - static size_t __iommu_unmap_pages(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d2020994f292db..521cd79700f4d8 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -468,8 +468,6 @@ extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot, gfp_t gfp); -extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size); extern size_t iommu_unmap_fast(struct iommu_domain *domain, @@ -778,13 +776,6 @@ static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, return -ENODEV; } -static inline int iommu_map_atomic(struct iommu_domain *domain, - unsigned long iova, phys_addr_t paddr, - size_t size, int prot) -{ - return -ENODEV; -} - static inline size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size) { -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 2/8] iommu: Remove iommu_map_atomic() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization There is only one call site and it can now just pass the GFP_ATOMIC to the normal iommu_map(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- drivers/iommu/iommu.c | 7 ------- include/linux/iommu.h | 9 --------- 3 files changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8bdb65e7686ff9..7016db569f81fc 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -713,7 +713,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys, if (!iova) return DMA_MAPPING_ERROR; - if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) { + if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) { iommu_dma_free_iova(cookie, iova, size, NULL); return DMA_MAPPING_ERROR; } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fe29fc2140b132..fee37bb246f3ea 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2376,13 +2376,6 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, } EXPORT_SYMBOL_GPL(iommu_map); -int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); -} -EXPORT_SYMBOL_GPL(iommu_map_atomic); - static size_t __iommu_unmap_pages(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d2020994f292db..521cd79700f4d8 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -468,8 +468,6 @@ extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot, gfp_t gfp); -extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size); extern size_t iommu_unmap_fast(struct iommu_domain *domain, @@ -778,13 +776,6 @@ static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, return -ENODEV; } -static inline int iommu_map_atomic(struct iommu_domain *domain, - unsigned long iova, phys_addr_t paddr, - size_t size, int prot) -{ - return -ENODEV; -} - static inline size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size) { -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 2/8] iommu: Remove iommu_map_atomic() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization There is only one call site and it can now just pass the GFP_ATOMIC to the normal iommu_map(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- drivers/iommu/iommu.c | 7 ------- include/linux/iommu.h | 9 --------- 3 files changed, 1 insertion(+), 17 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8bdb65e7686ff9..7016db569f81fc 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -713,7 +713,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys, if (!iova) return DMA_MAPPING_ERROR; - if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) { + if (iommu_map(domain, iova, phys - iova_off, size, prot, GFP_ATOMIC)) { iommu_dma_free_iova(cookie, iova, size, NULL); return DMA_MAPPING_ERROR; } diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fe29fc2140b132..fee37bb246f3ea 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2376,13 +2376,6 @@ int iommu_map(struct iommu_domain *domain, unsigned long iova, } EXPORT_SYMBOL_GPL(iommu_map); -int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot) -{ - return iommu_map(domain, iova, paddr, size, prot, GFP_ATOMIC); -} -EXPORT_SYMBOL_GPL(iommu_map_atomic); - static size_t __iommu_unmap_pages(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather) diff --git a/include/linux/iommu.h b/include/linux/iommu.h index d2020994f292db..521cd79700f4d8 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -468,8 +468,6 @@ extern struct iommu_domain *iommu_get_domain_for_dev(struct device *dev); extern struct iommu_domain *iommu_get_dma_domain(struct device *dev); extern int iommu_map(struct iommu_domain *domain, unsigned long iova, phys_addr_t paddr, size_t size, int prot, gfp_t gfp); -extern int iommu_map_atomic(struct iommu_domain *domain, unsigned long iova, - phys_addr_t paddr, size_t size, int prot); extern size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size); extern size_t iommu_unmap_fast(struct iommu_domain *domain, @@ -778,13 +776,6 @@ static inline int iommu_map(struct iommu_domain *domain, unsigned long iova, return -ENODEV; } -static inline int iommu_map_atomic(struct iommu_domain *domain, - unsigned long iova, phys_addr_t paddr, - size_t size, int prot) -{ - return -ENODEV; -} - static inline size_t iommu_unmap(struct iommu_domain *domain, unsigned long iova, size_t size) { -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Follow the pattern for iommu_map() and remove iommu_map_sg_atomic(). This allows __iommu_dma_alloc_noncontiguous() to use a GFP_KERNEL allocation here, based on the provided gfp flags. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 5 +++-- drivers/iommu/iommu.c | 21 +++++---------------- include/linux/iommu.h | 18 +++++------------- 3 files changed, 13 insertions(+), 31 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7016db569f81fc..8c2788633c1766 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -833,7 +833,8 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, arch_dma_prep_coherent(sg_page(sg), sg->length); } - ret = iommu_map_sg_atomic(domain, iova, sgt->sgl, sgt->orig_nents, ioprot); + ret = iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, ioprot, + gfp); if (ret < 0 || ret < size) goto out_free_sg; @@ -1281,7 +1282,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, * We'll leave any physical concatenation to the IOMMU driver's * implementation - it knows better than we do. */ - ret = iommu_map_sg_atomic(domain, iova, sg, nents, prot); + ret = iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); if (ret < 0 || ret < iova_len) goto out_free_iova; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fee37bb246f3ea..11fb3981e25642 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2465,9 +2465,9 @@ size_t iommu_unmap_fast(struct iommu_domain *domain, } EXPORT_SYMBOL_GPL(iommu_unmap_fast); -static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot, - gfp_t gfp) +ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, + struct scatterlist *sg, unsigned int nents, int prot, + gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; size_t len = 0, mapped = 0; @@ -2475,6 +2475,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, unsigned int i = 0; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); @@ -2514,21 +2516,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, return ret; } - -ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - might_sleep(); - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map_sg); -ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); -} - /** * report_iommu_fault() - report about an IOMMU fault to the IOMMU framework * @domain: the iommu domain where the fault has happened diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 521cd79700f4d8..d5c16dc33c87de 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -474,10 +474,8 @@ extern size_t iommu_unmap_fast(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather); extern ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot); -extern ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot); + struct scatterlist *sg, unsigned int nents, + int prot, gfp_t gfp); extern phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain, dma_addr_t iova); extern void iommu_set_fault_handler(struct iommu_domain *domain, iommu_fault_handler_t handler, void *token); @@ -791,14 +789,7 @@ static inline size_t iommu_unmap_fast(struct iommu_domain *domain, static inline ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) -{ - return -ENODEV; -} - -static inline ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) + unsigned int nents, int prot, gfp_t gfp) { return -ENODEV; } @@ -1109,7 +1100,8 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, static inline size_t iommu_map_sgtable(struct iommu_domain *domain, unsigned long iova, struct sg_table *sgt, int prot) { - return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot); + return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot, + GFP_KERNEL); } #ifdef CONFIG_IOMMU_DEBUGFS -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media Follow the pattern for iommu_map() and remove iommu_map_sg_atomic(). This allows __iommu_dma_alloc_noncontiguous() to use a GFP_KERNEL allocation here, based on the provided gfp flags. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 5 +++-- drivers/iommu/iommu.c | 21 +++++---------------- include/linux/iommu.h | 18 +++++------------- 3 files changed, 13 insertions(+), 31 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7016db569f81fc..8c2788633c1766 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -833,7 +833,8 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, arch_dma_prep_coherent(sg_page(sg), sg->length); } - ret = iommu_map_sg_atomic(domain, iova, sgt->sgl, sgt->orig_nents, ioprot); + ret = iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, ioprot, + gfp); if (ret < 0 || ret < size) goto out_free_sg; @@ -1281,7 +1282,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, * We'll leave any physical concatenation to the IOMMU driver's * implementation - it knows better than we do. */ - ret = iommu_map_sg_atomic(domain, iova, sg, nents, prot); + ret = iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); if (ret < 0 || ret < iova_len) goto out_free_iova; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fee37bb246f3ea..11fb3981e25642 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2465,9 +2465,9 @@ size_t iommu_unmap_fast(struct iommu_domain *domain, } EXPORT_SYMBOL_GPL(iommu_unmap_fast); -static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot, - gfp_t gfp) +ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, + struct scatterlist *sg, unsigned int nents, int prot, + gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; size_t len = 0, mapped = 0; @@ -2475,6 +2475,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, unsigned int i = 0; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); @@ -2514,21 +2516,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, return ret; } - -ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - might_sleep(); - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map_sg); -ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); -} - /** * report_iommu_fault() - report about an IOMMU fault to the IOMMU framework * @domain: the iommu domain where the fault has happened diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 521cd79700f4d8..d5c16dc33c87de 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -474,10 +474,8 @@ extern size_t iommu_unmap_fast(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather); extern ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot); -extern ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot); + struct scatterlist *sg, unsigned int nents, + int prot, gfp_t gfp); extern phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain, dma_addr_t iova); extern void iommu_set_fault_handler(struct iommu_domain *domain, iommu_fault_handler_t handler, void *token); @@ -791,14 +789,7 @@ static inline size_t iommu_unmap_fast(struct iommu_domain *domain, static inline ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) -{ - return -ENODEV; -} - -static inline ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) + unsigned int nents, int prot, gfp_t gfp) { return -ENODEV; } @@ -1109,7 +1100,8 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, static inline size_t iommu_map_sgtable(struct iommu_domain *domain, unsigned long iova, struct sg_table *sgt, int prot) { - return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot); + return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot, + GFP_KERNEL); } #ifdef CONFIG_IOMMU_DEBUGFS -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Follow the pattern for iommu_map() and remove iommu_map_sg_atomic(). This allows __iommu_dma_alloc_noncontiguous() to use a GFP_KERNEL allocation here, based on the provided gfp flags. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 5 +++-- drivers/iommu/iommu.c | 21 +++++---------------- include/linux/iommu.h | 18 +++++------------- 3 files changed, 13 insertions(+), 31 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7016db569f81fc..8c2788633c1766 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -833,7 +833,8 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, arch_dma_prep_coherent(sg_page(sg), sg->length); } - ret = iommu_map_sg_atomic(domain, iova, sgt->sgl, sgt->orig_nents, ioprot); + ret = iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, ioprot, + gfp); if (ret < 0 || ret < size) goto out_free_sg; @@ -1281,7 +1282,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, * We'll leave any physical concatenation to the IOMMU driver's * implementation - it knows better than we do. */ - ret = iommu_map_sg_atomic(domain, iova, sg, nents, prot); + ret = iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); if (ret < 0 || ret < iova_len) goto out_free_iova; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fee37bb246f3ea..11fb3981e25642 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2465,9 +2465,9 @@ size_t iommu_unmap_fast(struct iommu_domain *domain, } EXPORT_SYMBOL_GPL(iommu_unmap_fast); -static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot, - gfp_t gfp) +ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, + struct scatterlist *sg, unsigned int nents, int prot, + gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; size_t len = 0, mapped = 0; @@ -2475,6 +2475,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, unsigned int i = 0; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); @@ -2514,21 +2516,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, return ret; } - -ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - might_sleep(); - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map_sg); -ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); -} - /** * report_iommu_fault() - report about an IOMMU fault to the IOMMU framework * @domain: the iommu domain where the fault has happened diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 521cd79700f4d8..d5c16dc33c87de 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -474,10 +474,8 @@ extern size_t iommu_unmap_fast(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather); extern ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot); -extern ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot); + struct scatterlist *sg, unsigned int nents, + int prot, gfp_t gfp); extern phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain, dma_addr_t iova); extern void iommu_set_fault_handler(struct iommu_domain *domain, iommu_fault_handler_t handler, void *token); @@ -791,14 +789,7 @@ static inline size_t iommu_unmap_fast(struct iommu_domain *domain, static inline ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) -{ - return -ENODEV; -} - -static inline ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) + unsigned int nents, int prot, gfp_t gfp) { return -ENODEV; } @@ -1109,7 +1100,8 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, static inline size_t iommu_map_sgtable(struct iommu_domain *domain, unsigned long iova, struct sg_table *sgt, int prot) { - return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot); + return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot, + GFP_KERNEL); } #ifdef CONFIG_IOMMU_DEBUGFS -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media Follow the pattern for iommu_map() and remove iommu_map_sg_atomic(). This allows __iommu_dma_alloc_noncontiguous() to use a GFP_KERNEL allocation here, based on the provided gfp flags. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 5 +++-- drivers/iommu/iommu.c | 21 +++++---------------- include/linux/iommu.h | 18 +++++------------- 3 files changed, 13 insertions(+), 31 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7016db569f81fc..8c2788633c1766 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -833,7 +833,8 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, arch_dma_prep_coherent(sg_page(sg), sg->length); } - ret = iommu_map_sg_atomic(domain, iova, sgt->sgl, sgt->orig_nents, ioprot); + ret = iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, ioprot, + gfp); if (ret < 0 || ret < size) goto out_free_sg; @@ -1281,7 +1282,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, * We'll leave any physical concatenation to the IOMMU driver's * implementation - it knows better than we do. */ - ret = iommu_map_sg_atomic(domain, iova, sg, nents, prot); + ret = iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); if (ret < 0 || ret < iova_len) goto out_free_iova; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fee37bb246f3ea..11fb3981e25642 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2465,9 +2465,9 @@ size_t iommu_unmap_fast(struct iommu_domain *domain, } EXPORT_SYMBOL_GPL(iommu_unmap_fast); -static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot, - gfp_t gfp) +ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, + struct scatterlist *sg, unsigned int nents, int prot, + gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; size_t len = 0, mapped = 0; @@ -2475,6 +2475,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, unsigned int i = 0; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); @@ -2514,21 +2516,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, return ret; } - -ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - might_sleep(); - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map_sg); -ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); -} - /** * report_iommu_fault() - report about an IOMMU fault to the IOMMU framework * @domain: the iommu domain where the fault has happened diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 521cd79700f4d8..d5c16dc33c87de 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -474,10 +474,8 @@ extern size_t iommu_unmap_fast(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather); extern ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot); -extern ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot); + struct scatterlist *sg, unsigned int nents, + int prot, gfp_t gfp); extern phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain, dma_addr_t iova); extern void iommu_set_fault_handler(struct iommu_domain *domain, iommu_fault_handler_t handler, void *token); @@ -791,14 +789,7 @@ static inline size_t iommu_unmap_fast(struct iommu_domain *domain, static inline ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) -{ - return -ENODEV; -} - -static inline ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) + unsigned int nents, int prot, gfp_t gfp) { return -ENODEV; } @@ -1109,7 +1100,8 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, static inline size_t iommu_map_sgtable(struct iommu_domain *domain, unsigned long iova, struct sg_table *sgt, int prot) { - return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot); + return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot, + GFP_KERNEL); } #ifdef CONFIG_IOMMU_DEBUGFS -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Follow the pattern for iommu_map() and remove iommu_map_sg_atomic(). This allows __iommu_dma_alloc_noncontiguous() to use a GFP_KERNEL allocation here, based on the provided gfp flags. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 5 +++-- drivers/iommu/iommu.c | 21 +++++---------------- include/linux/iommu.h | 18 +++++------------- 3 files changed, 13 insertions(+), 31 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7016db569f81fc..8c2788633c1766 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -833,7 +833,8 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, arch_dma_prep_coherent(sg_page(sg), sg->length); } - ret = iommu_map_sg_atomic(domain, iova, sgt->sgl, sgt->orig_nents, ioprot); + ret = iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, ioprot, + gfp); if (ret < 0 || ret < size) goto out_free_sg; @@ -1281,7 +1282,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, * We'll leave any physical concatenation to the IOMMU driver's * implementation - it knows better than we do. */ - ret = iommu_map_sg_atomic(domain, iova, sg, nents, prot); + ret = iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); if (ret < 0 || ret < iova_len) goto out_free_iova; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fee37bb246f3ea..11fb3981e25642 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2465,9 +2465,9 @@ size_t iommu_unmap_fast(struct iommu_domain *domain, } EXPORT_SYMBOL_GPL(iommu_unmap_fast); -static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot, - gfp_t gfp) +ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, + struct scatterlist *sg, unsigned int nents, int prot, + gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; size_t len = 0, mapped = 0; @@ -2475,6 +2475,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, unsigned int i = 0; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); @@ -2514,21 +2516,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, return ret; } - -ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - might_sleep(); - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map_sg); -ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); -} - /** * report_iommu_fault() - report about an IOMMU fault to the IOMMU framework * @domain: the iommu domain where the fault has happened diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 521cd79700f4d8..d5c16dc33c87de 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -474,10 +474,8 @@ extern size_t iommu_unmap_fast(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather); extern ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot); -extern ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot); + struct scatterlist *sg, unsigned int nents, + int prot, gfp_t gfp); extern phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain, dma_addr_t iova); extern void iommu_set_fault_handler(struct iommu_domain *domain, iommu_fault_handler_t handler, void *token); @@ -791,14 +789,7 @@ static inline size_t iommu_unmap_fast(struct iommu_domain *domain, static inline ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) -{ - return -ENODEV; -} - -static inline ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) + unsigned int nents, int prot, gfp_t gfp) { return -ENODEV; } @@ -1109,7 +1100,8 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, static inline size_t iommu_map_sgtable(struct iommu_domain *domain, unsigned long iova, struct sg_table *sgt, int prot) { - return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot); + return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot, + GFP_KERNEL); } #ifdef CONFIG_IOMMU_DEBUGFS -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Follow the pattern for iommu_map() and remove iommu_map_sg_atomic(). This allows __iommu_dma_alloc_noncontiguous() to use a GFP_KERNEL allocation here, based on the provided gfp flags. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 5 +++-- drivers/iommu/iommu.c | 21 +++++---------------- include/linux/iommu.h | 18 +++++------------- 3 files changed, 13 insertions(+), 31 deletions(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7016db569f81fc..8c2788633c1766 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -833,7 +833,8 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, arch_dma_prep_coherent(sg_page(sg), sg->length); } - ret = iommu_map_sg_atomic(domain, iova, sgt->sgl, sgt->orig_nents, ioprot); + ret = iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, ioprot, + gfp); if (ret < 0 || ret < size) goto out_free_sg; @@ -1281,7 +1282,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, * We'll leave any physical concatenation to the IOMMU driver's * implementation - it knows better than we do. */ - ret = iommu_map_sg_atomic(domain, iova, sg, nents, prot); + ret = iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); if (ret < 0 || ret < iova_len) goto out_free_iova; diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index fee37bb246f3ea..11fb3981e25642 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2465,9 +2465,9 @@ size_t iommu_unmap_fast(struct iommu_domain *domain, } EXPORT_SYMBOL_GPL(iommu_unmap_fast); -static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot, - gfp_t gfp) +ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, + struct scatterlist *sg, unsigned int nents, int prot, + gfp_t gfp) { const struct iommu_domain_ops *ops = domain->ops; size_t len = 0, mapped = 0; @@ -2475,6 +2475,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, unsigned int i = 0; int ret; + might_sleep_if(gfpflags_allow_blocking(gfp)); + while (i <= nents) { phys_addr_t s_phys = sg_phys(sg); @@ -2514,21 +2516,8 @@ static ssize_t __iommu_map_sg(struct iommu_domain *domain, unsigned long iova, return ret; } - -ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - might_sleep(); - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_KERNEL); -} EXPORT_SYMBOL_GPL(iommu_map_sg); -ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot) -{ - return __iommu_map_sg(domain, iova, sg, nents, prot, GFP_ATOMIC); -} - /** * report_iommu_fault() - report about an IOMMU fault to the IOMMU framework * @domain: the iommu domain where the fault has happened diff --git a/include/linux/iommu.h b/include/linux/iommu.h index 521cd79700f4d8..d5c16dc33c87de 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -474,10 +474,8 @@ extern size_t iommu_unmap_fast(struct iommu_domain *domain, unsigned long iova, size_t size, struct iommu_iotlb_gather *iotlb_gather); extern ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, - struct scatterlist *sg, unsigned int nents, int prot); -extern ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot); + struct scatterlist *sg, unsigned int nents, + int prot, gfp_t gfp); extern phys_addr_t iommu_iova_to_phys(struct iommu_domain *domain, dma_addr_t iova); extern void iommu_set_fault_handler(struct iommu_domain *domain, iommu_fault_handler_t handler, void *token); @@ -791,14 +789,7 @@ static inline size_t iommu_unmap_fast(struct iommu_domain *domain, static inline ssize_t iommu_map_sg(struct iommu_domain *domain, unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) -{ - return -ENODEV; -} - -static inline ssize_t iommu_map_sg_atomic(struct iommu_domain *domain, - unsigned long iova, struct scatterlist *sg, - unsigned int nents, int prot) + unsigned int nents, int prot, gfp_t gfp) { return -ENODEV; } @@ -1109,7 +1100,8 @@ iommu_get_domain_for_dev_pasid(struct device *dev, ioasid_t pasid, static inline size_t iommu_map_sgtable(struct iommu_domain *domain, unsigned long iova, struct sg_table *sgt, int prot) { - return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot); + return iommu_map_sg(domain, iova, sgt->sgl, sgt->orig_nents, prot, + GFP_KERNEL); } #ifdef CONFIG_IOMMU_DEBUGFS -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Change the sg_alloc_table_from_pages() allocation that was hardwired to GFP_KERNEL to use the gfp parameter like the other allocations in this function. Auditing says this is never called from an atomic context, so it is safe as is, but reads wrong. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8c2788633c1766..e4bf1bb159f7c7 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -822,7 +822,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, if (!iova) goto out_free_pages; - if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, GFP_KERNEL)) + if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, gfp)) goto out_free_iova; if (!(ioprot & IOMMU_CACHE)) { -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media Change the sg_alloc_table_from_pages() allocation that was hardwired to GFP_KERNEL to use the gfp parameter like the other allocations in this function. Auditing says this is never called from an atomic context, so it is safe as is, but reads wrong. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8c2788633c1766..e4bf1bb159f7c7 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -822,7 +822,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, if (!iova) goto out_free_pages; - if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, GFP_KERNEL)) + if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, gfp)) goto out_free_iova; if (!(ioprot & IOMMU_CACHE)) { -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Change the sg_alloc_table_from_pages() allocation that was hardwired to GFP_KERNEL to use the gfp parameter like the other allocations in this function. Auditing says this is never called from an atomic context, so it is safe as is, but reads wrong. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8c2788633c1766..e4bf1bb159f7c7 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -822,7 +822,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, if (!iova) goto out_free_pages; - if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, GFP_KERNEL)) + if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, gfp)) goto out_free_iova; if (!(ioprot & IOMMU_CACHE)) { -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media Change the sg_alloc_table_from_pages() allocation that was hardwired to GFP_KERNEL to use the gfp parameter like the other allocations in this function. Auditing says this is never called from an atomic context, so it is safe as is, but reads wrong. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8c2788633c1766..e4bf1bb159f7c7 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -822,7 +822,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, if (!iova) goto out_free_pages; - if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, GFP_KERNEL)) + if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, gfp)) goto out_free_iova; if (!(ioprot & IOMMU_CACHE)) { -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Change the sg_alloc_table_from_pages() allocation that was hardwired to GFP_KERNEL to use the gfp parameter like the other allocations in this function. Auditing says this is never called from an atomic context, so it is safe as is, but reads wrong. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8c2788633c1766..e4bf1bb159f7c7 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -822,7 +822,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, if (!iova) goto out_free_pages; - if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, GFP_KERNEL)) + if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, gfp)) goto out_free_iova; if (!(ioprot & IOMMU_CACHE)) { -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Change the sg_alloc_table_from_pages() allocation that was hardwired to GFP_KERNEL to use the gfp parameter like the other allocations in this function. Auditing says this is never called from an atomic context, so it is safe as is, but reads wrong. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/dma-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 8c2788633c1766..e4bf1bb159f7c7 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -822,7 +822,7 @@ static struct page **__iommu_dma_alloc_noncontiguous(struct device *dev, if (!iova) goto out_free_pages; - if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, GFP_KERNEL)) + if (sg_alloc_table_from_pages(sgt, pages, count, 0, size, gfp)) goto out_free_iova; if (!(ioprot & IOMMU_CACHE)) { -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain. Many drivers will allocate these at iommu_map() time and will trivially do the right thing if we pass in GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/iommufd/pages.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 22cc3bb0c6c55a..f8d92c9bb65b60 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -457,7 +457,7 @@ static int batch_iommu_map_small(struct iommu_domain *domain, while (size) { rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -502,7 +502,7 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, next_iova - iova, area->iommu_prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova = next_iova; -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media 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. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain. Many drivers will allocate these at iommu_map() time and will trivially do the right thing if we pass in GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/iommufd/pages.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 22cc3bb0c6c55a..f8d92c9bb65b60 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -457,7 +457,7 @@ static int batch_iommu_map_small(struct iommu_domain *domain, while (size) { rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -502,7 +502,7 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, next_iova - iova, area->iommu_prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova = next_iova; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain. Many drivers will allocate these at iommu_map() time and will trivially do the right thing if we pass in GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/iommufd/pages.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 22cc3bb0c6c55a..f8d92c9bb65b60 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -457,7 +457,7 @@ static int batch_iommu_map_small(struct iommu_domain *domain, while (size) { rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -502,7 +502,7 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, next_iova - iova, area->iommu_prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova = next_iova; -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media 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. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain. Many drivers will allocate these at iommu_map() time and will trivially do the right thing if we pass in GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/iommufd/pages.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 22cc3bb0c6c55a..f8d92c9bb65b60 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -457,7 +457,7 @@ static int batch_iommu_map_small(struct iommu_domain *domain, while (size) { rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -502,7 +502,7 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, next_iova - iova, area->iommu_prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova = next_iova; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain. Many drivers will allocate these at iommu_map() time and will trivially do the right thing if we pass in GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/iommufd/pages.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 22cc3bb0c6c55a..f8d92c9bb65b60 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -457,7 +457,7 @@ static int batch_iommu_map_small(struct iommu_domain *domain, while (size) { rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -502,7 +502,7 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, next_iova - iova, area->iommu_prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova = next_iova; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization 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. However, one of the biggest consumers of kernel memory is the IOPTEs stored under the iommu_domain. Many drivers will allocate these at iommu_map() time and will trivially do the right thing if we pass in GFP_KERNEL_ACCOUNT. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/iommufd/pages.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/iommu/iommufd/pages.c b/drivers/iommu/iommufd/pages.c index 22cc3bb0c6c55a..f8d92c9bb65b60 100644 --- a/drivers/iommu/iommufd/pages.c +++ b/drivers/iommu/iommufd/pages.c @@ -457,7 +457,7 @@ static int batch_iommu_map_small(struct iommu_domain *domain, while (size) { rc = iommu_map(domain, iova, paddr, PAGE_SIZE, prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova += PAGE_SIZE; @@ -502,7 +502,7 @@ static int batch_to_domain(struct pfn_batch *batch, struct iommu_domain *domain, rc = iommu_map(domain, iova, PFN_PHYS(batch->pfns[cur]) + page_offset, next_iova - iova, area->iommu_prot, - GFP_KERNEL); + GFP_KERNEL_ACCOUNT); if (rc) goto err_unmap; iova = next_iova; -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization This is eventually called by iommufd through intel_iommu_map_pages() and it should not be forced to atomic. Push the GFP_ATOMIC to all callers. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 14 +++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 59df7e42fd533c..e3807776971563 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str) } __setup("intel_iommu=", intel_iommu_setup); -void *alloc_pgtable_page(int node) +void *alloc_pgtable_page(int node, gfp_t gfp) { struct page *page; void *vaddr = NULL; - page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0); + page = alloc_pages_node(node, gfp | __GFP_ZERO, 0); if (page) vaddr = page_address(page); return vaddr; @@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus, if (!alloc) return NULL; - context = alloc_pgtable_page(iommu->node); + context = alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!context) return NULL; @@ -935,7 +935,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid); + tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!tmp_page) return NULL; @@ -1186,7 +1186,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu) { struct root_entry *root; - root = (struct root_entry *)alloc_pgtable_page(iommu->node); + root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!root) { pr_err("Allocating root entry for %s failed\n", iommu->name); @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu *iommu, if (!old_ce) goto out; - new_ce = alloc_pgtable_page(iommu->node); + new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) goto out_unmap; @@ -4136,7 +4136,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width) domain->max_addr = 0; /* always allocate the top pgd */ - domain->pgd = alloc_pgtable_page(domain->nid); + domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!domain->pgd) return -ENOMEM; domain_flush_cache(domain, domain->pgd, PAGE_SIZE); diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 06e61e4748567a..ca9a035e0110af 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc, extern int dmar_ir_support(void); -void *alloc_pgtable_page(int node); +void *alloc_pgtable_page(int node, gfp_t gfp); void free_pgtable_page(void *vaddr); void iommu_flush_write_buffer(struct intel_iommu *iommu); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index fb3c7020028d07..c5bf74e9372d62 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid) retry: entries = get_pasid_table_from_pde(&dir[dir_index]); if (!entries) { - entries = alloc_pgtable_page(info->iommu->node); + entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC); if (!entries) return NULL; -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media This is eventually called by iommufd through intel_iommu_map_pages() and it should not be forced to atomic. Push the GFP_ATOMIC to all callers. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 14 +++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 59df7e42fd533c..e3807776971563 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str) } __setup("intel_iommu=", intel_iommu_setup); -void *alloc_pgtable_page(int node) +void *alloc_pgtable_page(int node, gfp_t gfp) { struct page *page; void *vaddr = NULL; - page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0); + page = alloc_pages_node(node, gfp | __GFP_ZERO, 0); if (page) vaddr = page_address(page); return vaddr; @@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus, if (!alloc) return NULL; - context = alloc_pgtable_page(iommu->node); + context = alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!context) return NULL; @@ -935,7 +935,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid); + tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!tmp_page) return NULL; @@ -1186,7 +1186,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu) { struct root_entry *root; - root = (struct root_entry *)alloc_pgtable_page(iommu->node); + root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!root) { pr_err("Allocating root entry for %s failed\n", iommu->name); @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu *iommu, if (!old_ce) goto out; - new_ce = alloc_pgtable_page(iommu->node); + new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) goto out_unmap; @@ -4136,7 +4136,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width) domain->max_addr = 0; /* always allocate the top pgd */ - domain->pgd = alloc_pgtable_page(domain->nid); + domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!domain->pgd) return -ENOMEM; domain_flush_cache(domain, domain->pgd, PAGE_SIZE); diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 06e61e4748567a..ca9a035e0110af 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc, extern int dmar_ir_support(void); -void *alloc_pgtable_page(int node); +void *alloc_pgtable_page(int node, gfp_t gfp); void free_pgtable_page(void *vaddr); void iommu_flush_write_buffer(struct intel_iommu *iommu); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index fb3c7020028d07..c5bf74e9372d62 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid) retry: entries = get_pasid_table_from_pde(&dir[dir_index]); if (!entries) { - entries = alloc_pgtable_page(info->iommu->node); + entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC); if (!entries) return NULL; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization This is eventually called by iommufd through intel_iommu_map_pages() and it should not be forced to atomic. Push the GFP_ATOMIC to all callers. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 14 +++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 59df7e42fd533c..e3807776971563 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str) } __setup("intel_iommu=", intel_iommu_setup); -void *alloc_pgtable_page(int node) +void *alloc_pgtable_page(int node, gfp_t gfp) { struct page *page; void *vaddr = NULL; - page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0); + page = alloc_pages_node(node, gfp | __GFP_ZERO, 0); if (page) vaddr = page_address(page); return vaddr; @@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus, if (!alloc) return NULL; - context = alloc_pgtable_page(iommu->node); + context = alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!context) return NULL; @@ -935,7 +935,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid); + tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!tmp_page) return NULL; @@ -1186,7 +1186,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu) { struct root_entry *root; - root = (struct root_entry *)alloc_pgtable_page(iommu->node); + root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!root) { pr_err("Allocating root entry for %s failed\n", iommu->name); @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu *iommu, if (!old_ce) goto out; - new_ce = alloc_pgtable_page(iommu->node); + new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) goto out_unmap; @@ -4136,7 +4136,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width) domain->max_addr = 0; /* always allocate the top pgd */ - domain->pgd = alloc_pgtable_page(domain->nid); + domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!domain->pgd) return -ENOMEM; domain_flush_cache(domain, domain->pgd, PAGE_SIZE); diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 06e61e4748567a..ca9a035e0110af 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc, extern int dmar_ir_support(void); -void *alloc_pgtable_page(int node); +void *alloc_pgtable_page(int node, gfp_t gfp); void free_pgtable_page(void *vaddr); void iommu_flush_write_buffer(struct intel_iommu *iommu); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index fb3c7020028d07..c5bf74e9372d62 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid) retry: entries = get_pasid_table_from_pde(&dir[dir_index]); if (!entries) { - entries = alloc_pgtable_page(info->iommu->node); + entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC); if (!entries) return NULL; -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media This is eventually called by iommufd through intel_iommu_map_pages() and it should not be forced to atomic. Push the GFP_ATOMIC to all callers. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 14 +++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 59df7e42fd533c..e3807776971563 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str) } __setup("intel_iommu=", intel_iommu_setup); -void *alloc_pgtable_page(int node) +void *alloc_pgtable_page(int node, gfp_t gfp) { struct page *page; void *vaddr = NULL; - page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0); + page = alloc_pages_node(node, gfp | __GFP_ZERO, 0); if (page) vaddr = page_address(page); return vaddr; @@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus, if (!alloc) return NULL; - context = alloc_pgtable_page(iommu->node); + context = alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!context) return NULL; @@ -935,7 +935,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid); + tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!tmp_page) return NULL; @@ -1186,7 +1186,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu) { struct root_entry *root; - root = (struct root_entry *)alloc_pgtable_page(iommu->node); + root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!root) { pr_err("Allocating root entry for %s failed\n", iommu->name); @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu *iommu, if (!old_ce) goto out; - new_ce = alloc_pgtable_page(iommu->node); + new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) goto out_unmap; @@ -4136,7 +4136,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width) domain->max_addr = 0; /* always allocate the top pgd */ - domain->pgd = alloc_pgtable_page(domain->nid); + domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!domain->pgd) return -ENOMEM; domain_flush_cache(domain, domain->pgd, PAGE_SIZE); diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 06e61e4748567a..ca9a035e0110af 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc, extern int dmar_ir_support(void); -void *alloc_pgtable_page(int node); +void *alloc_pgtable_page(int node, gfp_t gfp); void free_pgtable_page(void *vaddr); void iommu_flush_write_buffer(struct intel_iommu *iommu); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index fb3c7020028d07..c5bf74e9372d62 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid) retry: entries = get_pasid_table_from_pde(&dir[dir_index]); if (!entries) { - entries = alloc_pgtable_page(info->iommu->node); + entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC); if (!entries) return NULL; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization This is eventually called by iommufd through intel_iommu_map_pages() and it should not be forced to atomic. Push the GFP_ATOMIC to all callers. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 14 +++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 59df7e42fd533c..e3807776971563 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str) } __setup("intel_iommu=", intel_iommu_setup); -void *alloc_pgtable_page(int node) +void *alloc_pgtable_page(int node, gfp_t gfp) { struct page *page; void *vaddr = NULL; - page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0); + page = alloc_pages_node(node, gfp | __GFP_ZERO, 0); if (page) vaddr = page_address(page); return vaddr; @@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus, if (!alloc) return NULL; - context = alloc_pgtable_page(iommu->node); + context = alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!context) return NULL; @@ -935,7 +935,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid); + tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!tmp_page) return NULL; @@ -1186,7 +1186,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu) { struct root_entry *root; - root = (struct root_entry *)alloc_pgtable_page(iommu->node); + root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!root) { pr_err("Allocating root entry for %s failed\n", iommu->name); @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu *iommu, if (!old_ce) goto out; - new_ce = alloc_pgtable_page(iommu->node); + new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) goto out_unmap; @@ -4136,7 +4136,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width) domain->max_addr = 0; /* always allocate the top pgd */ - domain->pgd = alloc_pgtable_page(domain->nid); + domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!domain->pgd) return -ENOMEM; domain_flush_cache(domain, domain->pgd, PAGE_SIZE); diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 06e61e4748567a..ca9a035e0110af 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc, extern int dmar_ir_support(void); -void *alloc_pgtable_page(int node); +void *alloc_pgtable_page(int node, gfp_t gfp); void free_pgtable_page(void *vaddr); void iommu_flush_write_buffer(struct intel_iommu *iommu); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index fb3c7020028d07..c5bf74e9372d62 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid) retry: entries = get_pasid_table_from_pde(&dir[dir_index]); if (!entries) { - entries = alloc_pgtable_page(info->iommu->node); + entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC); if (!entries) return NULL; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization This is eventually called by iommufd through intel_iommu_map_pages() and it should not be forced to atomic. Push the GFP_ATOMIC to all callers. Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 14 +++++++------- drivers/iommu/intel/iommu.h | 2 +- drivers/iommu/intel/pasid.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 59df7e42fd533c..e3807776971563 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str) } __setup("intel_iommu=", intel_iommu_setup); -void *alloc_pgtable_page(int node) +void *alloc_pgtable_page(int node, gfp_t gfp) { struct page *page; void *vaddr = NULL; - page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0); + page = alloc_pages_node(node, gfp | __GFP_ZERO, 0); if (page) vaddr = page_address(page); return vaddr; @@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus, if (!alloc) return NULL; - context = alloc_pgtable_page(iommu->node); + context = alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!context) return NULL; @@ -935,7 +935,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid); + tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!tmp_page) return NULL; @@ -1186,7 +1186,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu) { struct root_entry *root; - root = (struct root_entry *)alloc_pgtable_page(iommu->node); + root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC); if (!root) { pr_err("Allocating root entry for %s failed\n", iommu->name); @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu *iommu, if (!old_ce) goto out; - new_ce = alloc_pgtable_page(iommu->node); + new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) goto out_unmap; @@ -4136,7 +4136,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width) domain->max_addr = 0; /* always allocate the top pgd */ - domain->pgd = alloc_pgtable_page(domain->nid); + domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC); if (!domain->pgd) return -ENOMEM; domain_flush_cache(domain, domain->pgd, PAGE_SIZE); diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 06e61e4748567a..ca9a035e0110af 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc, extern int dmar_ir_support(void); -void *alloc_pgtable_page(int node); +void *alloc_pgtable_page(int node, gfp_t gfp); void free_pgtable_page(void *vaddr); void iommu_flush_write_buffer(struct intel_iommu *iommu); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index fb3c7020028d07..c5bf74e9372d62 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid) retry: entries = get_pasid_table_from_pde(&dir[dir_index]); if (!entries) { - entries = alloc_pgtable_page(info->iommu->node); + entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC); if (!entries) return NULL; -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() 2023-01-06 16:42 ` Jason Gunthorpe ` (4 preceding siblings ...) (?) @ 2023-01-17 3:35 ` Tian, Kevin -1 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 3:35 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, Alex Williamson, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, Christian Borntraeger, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 3:35 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 3:35 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 3:35 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, Alex Williamson, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, Christian Borntraeger, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 3:35 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 3:35 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:35 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > *iommu, > if (!old_ce) > goto out; > > - new_ce = alloc_pgtable_page(iommu->node); > + new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); GFP_ATOMIC -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() 2023-01-17 3:35 ` Tian, Kevin ` (3 preceding siblings ...) (?) @ 2023-01-17 13:30 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:30 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > > *iommu, > > if (!old_ce) > > goto out; > > > > - new_ce = alloc_pgtable_page(iommu->node); > > + new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > GFP_ATOMIC Can't be: old_ce = memremap(old_ce_phys, PAGE_SIZE, MEMREMAP_WB); if (!old_ce) goto out; new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) memremap() is sleeping. And the only caller is: ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); if (!ctxt_tbls) goto out_unmap; for (bus = 0; bus < 256; bus++) { ret = copy_context_table(iommu, &old_rt[bus], ctxt_tbls, bus, ext); Jason _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 13:30 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:30 UTC (permalink / raw) To: Tian, Kevin Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > > *iommu, > > if (!old_ce) > > goto out; > > > > - new_ce = alloc_pgtable_page(iommu->node); > > + new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > GFP_ATOMIC Can't be: old_ce = memremap(old_ce_phys, PAGE_SIZE, MEMREMAP_WB); if (!old_ce) goto out; new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) memremap() is sleeping. And the only caller is: ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); if (!ctxt_tbls) goto out_unmap; for (bus = 0; bus < 256; bus++) { ret = copy_context_table(iommu, &old_rt[bus], ctxt_tbls, bus, ext); Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 13:30 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:30 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > > *iommu, > > if (!old_ce) > > goto out; > > > > - new_ce = alloc_pgtable_page(iommu->node); > > + new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > GFP_ATOMIC Can't be: old_ce = memremap(old_ce_phys, PAGE_SIZE, MEMREMAP_WB); if (!old_ce) goto out; new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) memremap() is sleeping. And the only caller is: ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); if (!ctxt_tbls) goto out_unmap; for (bus = 0; bus < 256; bus++) { ret = copy_context_table(iommu, &old_rt[bus], ctxt_tbls, bus, ext); Jason _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 13:30 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:30 UTC (permalink / raw) To: Tian, Kevin Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > > *iommu, > > if (!old_ce) > > goto out; > > > > - new_ce = alloc_pgtable_page(iommu->node); > > + new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > GFP_ATOMIC Can't be: old_ce = memremap(old_ce_phys, PAGE_SIZE, MEMREMAP_WB); if (!old_ce) goto out; new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) memremap() is sleeping. And the only caller is: ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); if (!ctxt_tbls) goto out_unmap; for (bus = 0; bus < 256; bus++) { ret = copy_context_table(iommu, &old_rt[bus], ctxt_tbls, bus, ext); Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 13:30 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:30 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > > *iommu, > > if (!old_ce) > > goto out; > > > > - new_ce = alloc_pgtable_page(iommu->node); > > + new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > GFP_ATOMIC Can't be: old_ce = memremap(old_ce_phys, PAGE_SIZE, MEMREMAP_WB); if (!old_ce) goto out; new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) memremap() is sleeping. And the only caller is: ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); if (!ctxt_tbls) goto out_unmap; for (bus = 0; bus < 256; bus++) { ret = copy_context_table(iommu, &old_rt[bus], ctxt_tbls, bus, ext); Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-17 13:30 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:30 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct intel_iommu > > *iommu, > > if (!old_ce) > > goto out; > > > > - new_ce = alloc_pgtable_page(iommu->node); > > + new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > GFP_ATOMIC Can't be: old_ce = memremap(old_ce_phys, PAGE_SIZE, MEMREMAP_WB); if (!old_ce) goto out; new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL); if (!new_ce) memremap() is sleeping. And the only caller is: ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); if (!ctxt_tbls) goto out_unmap; for (bus = 0; bus < 256; bus++) { ret = copy_context_table(iommu, &old_rt[bus], ctxt_tbls, bus, ext); Jason -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() 2023-01-17 13:30 ` Jason Gunthorpe ` (4 preceding siblings ...) (?) @ 2023-01-18 1:18 ` Tian, Kevin -1 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 1:18 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 1:18 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 1:18 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 1:18 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 1:18 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 1:18 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-18 1:18 UTC (permalink / raw) To: Jason Gunthorpe Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Tuesday, January 17, 2023 9:30 PM > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > intel_iommu > > > *iommu, > > > if (!old_ce) > > > goto out; > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > + new_ce = alloc_pgtable_page(iommu->node, > > > GFP_KERNEL); > > > > GFP_ATOMIC > > Can't be: > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > MEMREMAP_WB); > if (!old_ce) > goto out; > > new_ce = alloc_pgtable_page(iommu->node, > GFP_KERNEL); > if (!new_ce) > > memremap() is sleeping. > > And the only caller is: > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > if (!ctxt_tbls) > goto out_unmap; > > for (bus = 0; bus < 256; bus++) { > ret = copy_context_table(iommu, &old_rt[bus], > ctxt_tbls, bus, ext); > Yes, but the patch description says "Push the GFP_ATOMIC to all callers." implying it's purely a refactoring w/o changing those semantics. I'm fine with doing this change in this patch, but it should worth a clarification in the patch description. -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() 2023-01-18 1:18 ` Tian, Kevin ` (3 preceding siblings ...) (?) @ 2023-01-18 15:15 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-18 15:15 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Wed, Jan 18, 2023 at 01:18:18AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Tuesday, January 17, 2023 9:30 PM > > > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > > intel_iommu > > > > *iommu, > > > > if (!old_ce) > > > > goto out; > > > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > > + new_ce = alloc_pgtable_page(iommu->node, > > > > GFP_KERNEL); > > > > > > GFP_ATOMIC > > > > Can't be: > > > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > > MEMREMAP_WB); > > if (!old_ce) > > goto out; > > > > new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > if (!new_ce) > > > > memremap() is sleeping. > > > > And the only caller is: > > > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > > if (!ctxt_tbls) > > goto out_unmap; > > > > for (bus = 0; bus < 256; bus++) { > > ret = copy_context_table(iommu, &old_rt[bus], > > ctxt_tbls, bus, ext); > > > > Yes, but the patch description says "Push the GFP_ATOMIC to all > callers." implying it's purely a refactoring w/o changing those > semantics. Sure, lets split the patch, it is a good idea Jason _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 15:15 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-18 15:15 UTC (permalink / raw) To: Tian, Kevin Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu On Wed, Jan 18, 2023 at 01:18:18AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Tuesday, January 17, 2023 9:30 PM > > > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > > intel_iommu > > > > *iommu, > > > > if (!old_ce) > > > > goto out; > > > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > > + new_ce = alloc_pgtable_page(iommu->node, > > > > GFP_KERNEL); > > > > > > GFP_ATOMIC > > > > Can't be: > > > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > > MEMREMAP_WB); > > if (!old_ce) > > goto out; > > > > new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > if (!new_ce) > > > > memremap() is sleeping. > > > > And the only caller is: > > > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > > if (!ctxt_tbls) > > goto out_unmap; > > > > for (bus = 0; bus < 256; bus++) { > > ret = copy_context_table(iommu, &old_rt[bus], > > ctxt_tbls, bus, ext); > > > > Yes, but the patch description says "Push the GFP_ATOMIC to all > callers." implying it's purely a refactoring w/o changing those > semantics. Sure, lets split the patch, it is a good idea Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 15:15 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-18 15:15 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Wed, Jan 18, 2023 at 01:18:18AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Tuesday, January 17, 2023 9:30 PM > > > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > > intel_iommu > > > > *iommu, > > > > if (!old_ce) > > > > goto out; > > > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > > + new_ce = alloc_pgtable_page(iommu->node, > > > > GFP_KERNEL); > > > > > > GFP_ATOMIC > > > > Can't be: > > > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > > MEMREMAP_WB); > > if (!old_ce) > > goto out; > > > > new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > if (!new_ce) > > > > memremap() is sleeping. > > > > And the only caller is: > > > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > > if (!ctxt_tbls) > > goto out_unmap; > > > > for (bus = 0; bus < 256; bus++) { > > ret = copy_context_table(iommu, &old_rt[bus], > > ctxt_tbls, bus, ext); > > > > Yes, but the patch description says "Push the GFP_ATOMIC to all > callers." implying it's purely a refactoring w/o changing those > semantics. Sure, lets split the patch, it is a good idea Jason _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 15:15 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-18 15:15 UTC (permalink / raw) To: Tian, Kevin Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu On Wed, Jan 18, 2023 at 01:18:18AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Tuesday, January 17, 2023 9:30 PM > > > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > > intel_iommu > > > > *iommu, > > > > if (!old_ce) > > > > goto out; > > > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > > + new_ce = alloc_pgtable_page(iommu->node, > > > > GFP_KERNEL); > > > > > > GFP_ATOMIC > > > > Can't be: > > > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > > MEMREMAP_WB); > > if (!old_ce) > > goto out; > > > > new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > if (!new_ce) > > > > memremap() is sleeping. > > > > And the only caller is: > > > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > > if (!ctxt_tbls) > > goto out_unmap; > > > > for (bus = 0; bus < 256; bus++) { > > ret = copy_context_table(iommu, &old_rt[bus], > > ctxt_tbls, bus, ext); > > > > Yes, but the patch description says "Push the GFP_ATOMIC to all > callers." implying it's purely a refactoring w/o changing those > semantics. Sure, lets split the patch, it is a good idea Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 15:15 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-18 15:15 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Wed, Jan 18, 2023 at 01:18:18AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Tuesday, January 17, 2023 9:30 PM > > > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > > intel_iommu > > > > *iommu, > > > > if (!old_ce) > > > > goto out; > > > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > > + new_ce = alloc_pgtable_page(iommu->node, > > > > GFP_KERNEL); > > > > > > GFP_ATOMIC > > > > Can't be: > > > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > > MEMREMAP_WB); > > if (!old_ce) > > goto out; > > > > new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > if (!new_ce) > > > > memremap() is sleeping. > > > > And the only caller is: > > > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > > if (!ctxt_tbls) > > goto out_unmap; > > > > for (bus = 0; bus < 256; bus++) { > > ret = copy_context_table(iommu, &old_rt[bus], > > ctxt_tbls, bus, ext); > > > > Yes, but the patch description says "Push the GFP_ATOMIC to all > callers." implying it's purely a refactoring w/o changing those > semantics. Sure, lets split the patch, it is a good idea Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() @ 2023-01-18 15:15 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-18 15:15 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Wed, Jan 18, 2023 at 01:18:18AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Tuesday, January 17, 2023 9:30 PM > > > > On Tue, Jan 17, 2023 at 03:35:08AM +0000, Tian, Kevin wrote: > > > > From: Jason Gunthorpe <jgg@nvidia.com> > > > > Sent: Saturday, January 7, 2023 12:43 AM > > > > > > > > @@ -2676,7 +2676,7 @@ static int copy_context_table(struct > > intel_iommu > > > > *iommu, > > > > if (!old_ce) > > > > goto out; > > > > > > > > - new_ce = alloc_pgtable_page(iommu->node); > > > > + new_ce = alloc_pgtable_page(iommu->node, > > > > GFP_KERNEL); > > > > > > GFP_ATOMIC > > > > Can't be: > > > > old_ce = memremap(old_ce_phys, PAGE_SIZE, > > MEMREMAP_WB); > > if (!old_ce) > > goto out; > > > > new_ce = alloc_pgtable_page(iommu->node, > > GFP_KERNEL); > > if (!new_ce) > > > > memremap() is sleeping. > > > > And the only caller is: > > > > ctxt_tbls = kcalloc(ctxt_table_entries, sizeof(void *), GFP_KERNEL); > > if (!ctxt_tbls) > > goto out_unmap; > > > > for (bus = 0; bus < 256; bus++) { > > ret = copy_context_table(iommu, &old_rt[bus], > > ctxt_tbls, bus, ext); > > > > Yes, but the patch description says "Push the GFP_ATOMIC to all > callers." implying it's purely a refactoring w/o changing those > semantics. Sure, lets split the patch, it is a good idea Jason -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Flow it down to alloc_pgtable_page() via pfn_to_dma_pte() and __domain_mapping(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index e3807776971563..a1a66798e1f06c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -908,7 +908,8 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, #endif static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, - unsigned long pfn, int *target_level) + unsigned long pfn, int *target_level, + gfp_t gfp) { struct dma_pte *parent, *pte; int level = agaw_to_level(domain->agaw); @@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); + tmp_page = alloc_pgtable_page(domain->nid, gfp); if (!tmp_page) return NULL; @@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain, while (start_pfn <= end_pfn) { if (!pte) - pte = pfn_to_dma_pte(domain, start_pfn, &level); + pte = pfn_to_dma_pte(domain, start_pfn, &level, + GFP_ATOMIC); if (dma_pte_present(pte)) { dma_pte_free_pagetable(domain, start_pfn, @@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain, static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, - unsigned long phys_pfn, unsigned long nr_pages, int prot) + unsigned long phys_pfn, unsigned long nr_pages, int prot, + gfp_t gfp) { struct dma_pte *first_pte = NULL, *pte = NULL; unsigned int largepage_lvl = 0; @@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, nr_pages); - pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl); + pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl, + gfp); if (!pte) return -ENOMEM; first_pte = pte; @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain, return __domain_mapping(domain, first_vpfn, first_vpfn, last_vpfn - first_vpfn + 1, - DMA_PTE_READ|DMA_PTE_WRITE); + DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL); } static int md_domain_init(struct dmar_domain *domain, int guest_width); @@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain, the low bits of hpa would take us onto the next page */ size = aligned_nrpages(hpa, size); return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT, - hpa >> VTD_PAGE_SHIFT, size, prot); + hpa >> VTD_PAGE_SHIFT, size, prot, gfp); } static int intel_iommu_map_pages(struct iommu_domain *domain, @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain, /* Cope with horrid API which requires us to unmap more than the size argument if it happens to be a large-page mapping. */ - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level)); + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC)); if (size < VTD_PAGE_SIZE << level_to_offset_bits(level)) size = VTD_PAGE_SIZE << level_to_offset_bits(level); @@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain, int level = 0; u64 phys = 0; - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level); + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC); if (pte && dma_pte_present(pte)) phys = dma_pte_addr(pte) + (iova & (BIT_MASK(level_to_offset_bits(level) + -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media Flow it down to alloc_pgtable_page() via pfn_to_dma_pte() and __domain_mapping(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index e3807776971563..a1a66798e1f06c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -908,7 +908,8 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, #endif static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, - unsigned long pfn, int *target_level) + unsigned long pfn, int *target_level, + gfp_t gfp) { struct dma_pte *parent, *pte; int level = agaw_to_level(domain->agaw); @@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); + tmp_page = alloc_pgtable_page(domain->nid, gfp); if (!tmp_page) return NULL; @@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain, while (start_pfn <= end_pfn) { if (!pte) - pte = pfn_to_dma_pte(domain, start_pfn, &level); + pte = pfn_to_dma_pte(domain, start_pfn, &level, + GFP_ATOMIC); if (dma_pte_present(pte)) { dma_pte_free_pagetable(domain, start_pfn, @@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain, static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, - unsigned long phys_pfn, unsigned long nr_pages, int prot) + unsigned long phys_pfn, unsigned long nr_pages, int prot, + gfp_t gfp) { struct dma_pte *first_pte = NULL, *pte = NULL; unsigned int largepage_lvl = 0; @@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, nr_pages); - pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl); + pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl, + gfp); if (!pte) return -ENOMEM; first_pte = pte; @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain, return __domain_mapping(domain, first_vpfn, first_vpfn, last_vpfn - first_vpfn + 1, - DMA_PTE_READ|DMA_PTE_WRITE); + DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL); } static int md_domain_init(struct dmar_domain *domain, int guest_width); @@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain, the low bits of hpa would take us onto the next page */ size = aligned_nrpages(hpa, size); return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT, - hpa >> VTD_PAGE_SHIFT, size, prot); + hpa >> VTD_PAGE_SHIFT, size, prot, gfp); } static int intel_iommu_map_pages(struct iommu_domain *domain, @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain, /* Cope with horrid API which requires us to unmap more than the size argument if it happens to be a large-page mapping. */ - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level)); + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC)); if (size < VTD_PAGE_SIZE << level_to_offset_bits(level)) size = VTD_PAGE_SIZE << level_to_offset_bits(level); @@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain, int level = 0; u64 phys = 0; - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level); + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC); if (pte && dma_pte_present(pte)) phys = dma_pte_addr(pte) + (iova & (BIT_MASK(level_to_offset_bits(level) + -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Flow it down to alloc_pgtable_page() via pfn_to_dma_pte() and __domain_mapping(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index e3807776971563..a1a66798e1f06c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -908,7 +908,8 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, #endif static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, - unsigned long pfn, int *target_level) + unsigned long pfn, int *target_level, + gfp_t gfp) { struct dma_pte *parent, *pte; int level = agaw_to_level(domain->agaw); @@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); + tmp_page = alloc_pgtable_page(domain->nid, gfp); if (!tmp_page) return NULL; @@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain, while (start_pfn <= end_pfn) { if (!pte) - pte = pfn_to_dma_pte(domain, start_pfn, &level); + pte = pfn_to_dma_pte(domain, start_pfn, &level, + GFP_ATOMIC); if (dma_pte_present(pte)) { dma_pte_free_pagetable(domain, start_pfn, @@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain, static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, - unsigned long phys_pfn, unsigned long nr_pages, int prot) + unsigned long phys_pfn, unsigned long nr_pages, int prot, + gfp_t gfp) { struct dma_pte *first_pte = NULL, *pte = NULL; unsigned int largepage_lvl = 0; @@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, nr_pages); - pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl); + pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl, + gfp); if (!pte) return -ENOMEM; first_pte = pte; @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain, return __domain_mapping(domain, first_vpfn, first_vpfn, last_vpfn - first_vpfn + 1, - DMA_PTE_READ|DMA_PTE_WRITE); + DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL); } static int md_domain_init(struct dmar_domain *domain, int guest_width); @@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain, the low bits of hpa would take us onto the next page */ size = aligned_nrpages(hpa, size); return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT, - hpa >> VTD_PAGE_SHIFT, size, prot); + hpa >> VTD_PAGE_SHIFT, size, prot, gfp); } static int intel_iommu_map_pages(struct iommu_domain *domain, @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain, /* Cope with horrid API which requires us to unmap more than the size argument if it happens to be a large-page mapping. */ - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level)); + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC)); if (size < VTD_PAGE_SIZE << level_to_offset_bits(level)) size = VTD_PAGE_SIZE << level_to_offset_bits(level); @@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain, int level = 0; u64 phys = 0; - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level); + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC); if (pte && dma_pte_present(pte)) phys = dma_pte_addr(pte) + (iova & (BIT_MASK(level_to_offset_bits(level) + -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media Flow it down to alloc_pgtable_page() via pfn_to_dma_pte() and __domain_mapping(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index e3807776971563..a1a66798e1f06c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -908,7 +908,8 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, #endif static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, - unsigned long pfn, int *target_level) + unsigned long pfn, int *target_level, + gfp_t gfp) { struct dma_pte *parent, *pte; int level = agaw_to_level(domain->agaw); @@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); + tmp_page = alloc_pgtable_page(domain->nid, gfp); if (!tmp_page) return NULL; @@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain, while (start_pfn <= end_pfn) { if (!pte) - pte = pfn_to_dma_pte(domain, start_pfn, &level); + pte = pfn_to_dma_pte(domain, start_pfn, &level, + GFP_ATOMIC); if (dma_pte_present(pte)) { dma_pte_free_pagetable(domain, start_pfn, @@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain, static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, - unsigned long phys_pfn, unsigned long nr_pages, int prot) + unsigned long phys_pfn, unsigned long nr_pages, int prot, + gfp_t gfp) { struct dma_pte *first_pte = NULL, *pte = NULL; unsigned int largepage_lvl = 0; @@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, nr_pages); - pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl); + pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl, + gfp); if (!pte) return -ENOMEM; first_pte = pte; @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain, return __domain_mapping(domain, first_vpfn, first_vpfn, last_vpfn - first_vpfn + 1, - DMA_PTE_READ|DMA_PTE_WRITE); + DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL); } static int md_domain_init(struct dmar_domain *domain, int guest_width); @@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain, the low bits of hpa would take us onto the next page */ size = aligned_nrpages(hpa, size); return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT, - hpa >> VTD_PAGE_SHIFT, size, prot); + hpa >> VTD_PAGE_SHIFT, size, prot, gfp); } static int intel_iommu_map_pages(struct iommu_domain *domain, @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain, /* Cope with horrid API which requires us to unmap more than the size argument if it happens to be a large-page mapping. */ - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level)); + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC)); if (size < VTD_PAGE_SIZE << level_to_offset_bits(level)) size = VTD_PAGE_SIZE << level_to_offset_bits(level); @@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain, int level = 0; u64 phys = 0; - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level); + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC); if (pte && dma_pte_present(pte)) phys = dma_pte_addr(pte) + (iova & (BIT_MASK(level_to_offset_bits(level) + -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Flow it down to alloc_pgtable_page() via pfn_to_dma_pte() and __domain_mapping(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index e3807776971563..a1a66798e1f06c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -908,7 +908,8 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, #endif static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, - unsigned long pfn, int *target_level) + unsigned long pfn, int *target_level, + gfp_t gfp) { struct dma_pte *parent, *pte; int level = agaw_to_level(domain->agaw); @@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); + tmp_page = alloc_pgtable_page(domain->nid, gfp); if (!tmp_page) return NULL; @@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain, while (start_pfn <= end_pfn) { if (!pte) - pte = pfn_to_dma_pte(domain, start_pfn, &level); + pte = pfn_to_dma_pte(domain, start_pfn, &level, + GFP_ATOMIC); if (dma_pte_present(pte)) { dma_pte_free_pagetable(domain, start_pfn, @@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain, static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, - unsigned long phys_pfn, unsigned long nr_pages, int prot) + unsigned long phys_pfn, unsigned long nr_pages, int prot, + gfp_t gfp) { struct dma_pte *first_pte = NULL, *pte = NULL; unsigned int largepage_lvl = 0; @@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, nr_pages); - pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl); + pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl, + gfp); if (!pte) return -ENOMEM; first_pte = pte; @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain, return __domain_mapping(domain, first_vpfn, first_vpfn, last_vpfn - first_vpfn + 1, - DMA_PTE_READ|DMA_PTE_WRITE); + DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL); } static int md_domain_init(struct dmar_domain *domain, int guest_width); @@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain, the low bits of hpa would take us onto the next page */ size = aligned_nrpages(hpa, size); return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT, - hpa >> VTD_PAGE_SHIFT, size, prot); + hpa >> VTD_PAGE_SHIFT, size, prot, gfp); } static int intel_iommu_map_pages(struct iommu_domain *domain, @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain, /* Cope with horrid API which requires us to unmap more than the size argument if it happens to be a large-page mapping. */ - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level)); + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC)); if (size < VTD_PAGE_SIZE << level_to_offset_bits(level)) size = VTD_PAGE_SIZE << level_to_offset_bits(level); @@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain, int level = 0; u64 phys = 0; - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level); + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC); if (pte && dma_pte_present(pte)) phys = dma_pte_addr(pte) + (iova & (BIT_MASK(level_to_offset_bits(level) + -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization Flow it down to alloc_pgtable_page() via pfn_to_dma_pte() and __domain_mapping(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- drivers/iommu/intel/iommu.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index e3807776971563..a1a66798e1f06c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -908,7 +908,8 @@ void dmar_fault_dump_ptes(struct intel_iommu *iommu, u16 source_id, #endif static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, - unsigned long pfn, int *target_level) + unsigned long pfn, int *target_level, + gfp_t gfp) { struct dma_pte *parent, *pte; int level = agaw_to_level(domain->agaw); @@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain, if (!dma_pte_present(pte)) { uint64_t pteval; - tmp_page = alloc_pgtable_page(domain->nid, GFP_ATOMIC); + tmp_page = alloc_pgtable_page(domain->nid, gfp); if (!tmp_page) return NULL; @@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain, while (start_pfn <= end_pfn) { if (!pte) - pte = pfn_to_dma_pte(domain, start_pfn, &level); + pte = pfn_to_dma_pte(domain, start_pfn, &level, + GFP_ATOMIC); if (dma_pte_present(pte)) { dma_pte_free_pagetable(domain, start_pfn, @@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain, static int __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, - unsigned long phys_pfn, unsigned long nr_pages, int prot) + unsigned long phys_pfn, unsigned long nr_pages, int prot, + gfp_t gfp) { struct dma_pte *first_pte = NULL, *pte = NULL; unsigned int largepage_lvl = 0; @@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn, largepage_lvl = hardware_largepage_caps(domain, iov_pfn, phys_pfn, nr_pages); - pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl); + pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl, + gfp); if (!pte) return -ENOMEM; first_pte = pte; @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain, return __domain_mapping(domain, first_vpfn, first_vpfn, last_vpfn - first_vpfn + 1, - DMA_PTE_READ|DMA_PTE_WRITE); + DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL); } static int md_domain_init(struct dmar_domain *domain, int guest_width); @@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain, the low bits of hpa would take us onto the next page */ size = aligned_nrpages(hpa, size); return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT, - hpa >> VTD_PAGE_SHIFT, size, prot); + hpa >> VTD_PAGE_SHIFT, size, prot, gfp); } static int intel_iommu_map_pages(struct iommu_domain *domain, @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain, /* Cope with horrid API which requires us to unmap more than the size argument if it happens to be a large-page mapping. */ - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level)); + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC)); if (size < VTD_PAGE_SIZE << level_to_offset_bits(level)) size = VTD_PAGE_SIZE << level_to_offset_bits(level); @@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain, int level = 0; u64 phys = 0; - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level); + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level, + GFP_ATOMIC); if (pte && dma_pte_present(pte)) phys = dma_pte_addr(pte) + (iova & (BIT_MASK(level_to_offset_bits(level) + -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* RE: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op 2023-01-06 16:42 ` Jason Gunthorpe ` (4 preceding siblings ...) (?) @ 2023-01-17 3:38 ` Tian, Kevin -1 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 3:38 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, Alex Williamson, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, Christian Borntraeger, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 3:38 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 3:38 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto _______________________________________________ Virtualization mailing list Virtualization@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/virtualization ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 3:38 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, Alex Williamson, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, Christian Borntraeger, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 3:38 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto ^ permalink raw reply [flat|nested] 144+ messages in thread
* RE: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 3:38 ` Tian, Kevin 0 siblings, 0 replies; 144+ messages in thread From: Tian, Kevin @ 2023-01-17 3:38 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org > From: Jason Gunthorpe <jgg@nvidia.com> > Sent: Saturday, January 7, 2023 12:43 AM > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > dmar_domain *domain, > > return __domain_mapping(domain, first_vpfn, > first_vpfn, last_vpfn - first_vpfn + 1, > - DMA_PTE_READ|DMA_PTE_WRITE); > + DMA_PTE_READ|DMA_PTE_WRITE, > GFP_KERNEL); > } Baolu, can you help confirm whether switching from GFP_ATOMIC to GFP_KERNEL is OK in this path? it looks fine to me in a quick glance but want to be conservative here. > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > iommu_domain *domain, > > /* Cope with horrid API which requires us to unmap more than the > size argument if it happens to be a large-page mapping. */ > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level)); > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC)); with level==0 it implies it's only lookup w/o pgtable allocation. From this angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. > @@ -4392,7 +4397,8 @@ static phys_addr_t > intel_iommu_iova_to_phys(struct iommu_domain *domain, > int level = 0; > u64 phys = 0; > > - pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level); > + pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > &level, > + GFP_ATOMIC); ditto -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op 2023-01-17 3:38 ` Tian, Kevin ` (3 preceding siblings ...) (?) @ 2023-01-17 8:35 ` Baolu Lu -1 siblings, 0 replies; 144+ messages in thread From: Baolu Lu @ 2023-01-17 8:35 UTC (permalink / raw) To: Tian, Kevin, Jason Gunthorpe, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: baolu.lu, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On 2023/1/17 11:38, Tian, Kevin wrote: >> From: Jason Gunthorpe<jgg@nvidia.com> >> Sent: Saturday, January 7, 2023 12:43 AM >> >> @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct >> dmar_domain *domain, >> >> return __domain_mapping(domain, first_vpfn, >> first_vpfn, last_vpfn - first_vpfn + 1, >> - DMA_PTE_READ|DMA_PTE_WRITE); >> + DMA_PTE_READ|DMA_PTE_WRITE, >> GFP_KERNEL); >> } > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. This is also good for me. The memory notifier callback runs in a process context and allowed to block. Best regards, baolu _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 8:35 ` Baolu Lu 0 siblings, 0 replies; 144+ messages in thread From: Baolu Lu @ 2023-01-17 8:35 UTC (permalink / raw) To: Tian, Kevin, Jason Gunthorpe, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, linux-media@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, Alex Williamson, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, Christian Borntraeger, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, baolu.lu On 2023/1/17 11:38, Tian, Kevin wrote: >> From: Jason Gunthorpe<jgg@nvidia.com> >> Sent: Saturday, January 7, 2023 12:43 AM >> >> @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct >> dmar_domain *domain, >> >> return __domain_mapping(domain, first_vpfn, >> first_vpfn, last_vpfn - first_vpfn + 1, >> - DMA_PTE_READ|DMA_PTE_WRITE); >> + DMA_PTE_READ|DMA_PTE_WRITE, >> GFP_KERNEL); >> } > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. This is also good for me. The memory notifier callback runs in a process context and allowed to block. Best regards, baolu ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 8:35 ` Baolu Lu 0 siblings, 0 replies; 144+ messages in thread From: Baolu Lu @ 2023-01-17 8:35 UTC (permalink / raw) To: Tian, Kevin, Jason Gunthorpe, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: baolu.lu, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On 2023/1/17 11:38, Tian, Kevin wrote: >> From: Jason Gunthorpe<jgg@nvidia.com> >> Sent: Saturday, January 7, 2023 12:43 AM >> >> @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct >> dmar_domain *domain, >> >> return __domain_mapping(domain, first_vpfn, >> first_vpfn, last_vpfn - first_vpfn + 1, >> - DMA_PTE_READ|DMA_PTE_WRITE); >> + DMA_PTE_READ|DMA_PTE_WRITE, >> GFP_KERNEL); >> } > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. This is also good for me. The memory notifier callback runs in a process context and allowed to block. Best regards, baolu _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 8:35 ` Baolu Lu 0 siblings, 0 replies; 144+ messages in thread From: Baolu Lu @ 2023-01-17 8:35 UTC (permalink / raw) To: Tian, Kevin, Jason Gunthorpe, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: linux-s390@vger.kernel.org, linux-media@vger.kernel.org, kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-rdma@vger.kernel.org, linux-arm-msm@vger.kernel.org, Niklas Schnelle, linux-remoteproc@vger.kernel.org, iommu@lists.linux.dev, dri-devel@lists.freedesktop.org, linux-stm32@st-md-mailman.stormreply.com, Alex Williamson, netdev@vger.kernel.org, ath10k@lists.infradead.org, linux-wireless@vger.kernel.org, linux-tegra@vger.kernel.org, Christian Borntraeger, virtualization@lists.linux-foundation.org, ath11k@lists.infradead.org, linux-arm-kernel@lists.infradead.org, baolu.lu On 2023/1/17 11:38, Tian, Kevin wrote: >> From: Jason Gunthorpe<jgg@nvidia.com> >> Sent: Saturday, January 7, 2023 12:43 AM >> >> @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct >> dmar_domain *domain, >> >> return __domain_mapping(domain, first_vpfn, >> first_vpfn, last_vpfn - first_vpfn + 1, >> - DMA_PTE_READ|DMA_PTE_WRITE); >> + DMA_PTE_READ|DMA_PTE_WRITE, >> GFP_KERNEL); >> } > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. This is also good for me. The memory notifier callback runs in a process context and allowed to block. Best regards, baolu ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 8:35 ` Baolu Lu 0 siblings, 0 replies; 144+ messages in thread From: Baolu Lu @ 2023-01-17 8:35 UTC (permalink / raw) To: Tian, Kevin, Jason Gunthorpe, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: baolu.lu, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On 2023/1/17 11:38, Tian, Kevin wrote: >> From: Jason Gunthorpe<jgg@nvidia.com> >> Sent: Saturday, January 7, 2023 12:43 AM >> >> @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct >> dmar_domain *domain, >> >> return __domain_mapping(domain, first_vpfn, >> first_vpfn, last_vpfn - first_vpfn + 1, >> - DMA_PTE_READ|DMA_PTE_WRITE); >> + DMA_PTE_READ|DMA_PTE_WRITE, >> GFP_KERNEL); >> } > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. This is also good for me. The memory notifier callback runs in a process context and allowed to block. Best regards, baolu ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 8:35 ` Baolu Lu 0 siblings, 0 replies; 144+ messages in thread From: Baolu Lu @ 2023-01-17 8:35 UTC (permalink / raw) To: Tian, Kevin, Jason Gunthorpe, Joerg Roedel, Matthew Rosato, Robin Murphy Cc: baolu.lu, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On 2023/1/17 11:38, Tian, Kevin wrote: >> From: Jason Gunthorpe<jgg@nvidia.com> >> Sent: Saturday, January 7, 2023 12:43 AM >> >> @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct >> dmar_domain *domain, >> >> return __domain_mapping(domain, first_vpfn, >> first_vpfn, last_vpfn - first_vpfn + 1, >> - DMA_PTE_READ|DMA_PTE_WRITE); >> + DMA_PTE_READ|DMA_PTE_WRITE, >> GFP_KERNEL); >> } > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. This is also good for me. The memory notifier callback runs in a process context and allowed to block. Best regards, baolu -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op 2023-01-17 3:38 ` Tian, Kevin ` (3 preceding siblings ...) (?) @ 2023-01-17 13:28 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:28 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:38:51AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > > dmar_domain *domain, > > > > return __domain_mapping(domain, first_vpfn, > > first_vpfn, last_vpfn - first_vpfn + 1, > > - DMA_PTE_READ|DMA_PTE_WRITE); > > + DMA_PTE_READ|DMA_PTE_WRITE, > > GFP_KERNEL); > > } > > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. I checked it carefully myself as well, good to check again. > > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > > iommu_domain *domain, > > > > /* Cope with horrid API which requires us to unmap more than the > > size argument if it happens to be a large-page mapping. */ > > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level)); > > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level, > > + GFP_ATOMIC)); > > with level==0 it implies it's only lookup w/o pgtable allocation. From this > angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. We should only write GFP_KERNEL if it is actually a sleepable context because it will be mighty confusing if it isn't. I couldn't tell what the context is so I left it as ATOMIC. You are correct this is only just a lookup and so the value is never used / doesn't matter. Jason _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 13:28 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:28 UTC (permalink / raw) To: Tian, Kevin Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu On Tue, Jan 17, 2023 at 03:38:51AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > > dmar_domain *domain, > > > > return __domain_mapping(domain, first_vpfn, > > first_vpfn, last_vpfn - first_vpfn + 1, > > - DMA_PTE_READ|DMA_PTE_WRITE); > > + DMA_PTE_READ|DMA_PTE_WRITE, > > GFP_KERNEL); > > } > > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. I checked it carefully myself as well, good to check again. > > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > > iommu_domain *domain, > > > > /* Cope with horrid API which requires us to unmap more than the > > size argument if it happens to be a large-page mapping. */ > > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level)); > > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level, > > + GFP_ATOMIC)); > > with level==0 it implies it's only lookup w/o pgtable allocation. From this > angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. We should only write GFP_KERNEL if it is actually a sleepable context because it will be mighty confusing if it isn't. I couldn't tell what the context is so I left it as ATOMIC. You are correct this is only just a lookup and so the value is never used / doesn't matter. Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 13:28 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:28 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:38:51AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > > dmar_domain *domain, > > > > return __domain_mapping(domain, first_vpfn, > > first_vpfn, last_vpfn - first_vpfn + 1, > > - DMA_PTE_READ|DMA_PTE_WRITE); > > + DMA_PTE_READ|DMA_PTE_WRITE, > > GFP_KERNEL); > > } > > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. I checked it carefully myself as well, good to check again. > > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > > iommu_domain *domain, > > > > /* Cope with horrid API which requires us to unmap more than the > > size argument if it happens to be a large-page mapping. */ > > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level)); > > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level, > > + GFP_ATOMIC)); > > with level==0 it implies it's only lookup w/o pgtable allocation. From this > angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. We should only write GFP_KERNEL if it is actually a sleepable context because it will be mighty confusing if it isn't. I couldn't tell what the context is so I left it as ATOMIC. You are correct this is only just a lookup and so the value is never used / doesn't matter. Jason _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 13:28 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:28 UTC (permalink / raw) To: Tian, Kevin Cc: kvm@vger.kernel.org, nouveau@lists.freedesktop.org, linux-remoteproc@vger.kernel.org, dri-devel@lists.freedesktop.org, virtualization@lists.linux-foundation.org, linux-stm32@st-md-mailman.stormreply.com, linux-s390@vger.kernel.org, Matthew Rosato, linux-rdma@vger.kernel.org, Joerg Roedel, ath10k@lists.infradead.org, iommu@lists.linux.dev, Christian Borntraeger, ath11k@lists.infradead.org, linux-media@vger.kernel.org, Niklas Schnelle, linux-arm-msm@vger.kernel.org, Alex Williamson, linux-tegra@vger.kernel.org, linux-arm-kernel@lists.infradead.org, netdev@vger.kernel.org, linux-wireless@vger.kernel.org, Robin Murphy, Lu Baolu On Tue, Jan 17, 2023 at 03:38:51AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > > dmar_domain *domain, > > > > return __domain_mapping(domain, first_vpfn, > > first_vpfn, last_vpfn - first_vpfn + 1, > > - DMA_PTE_READ|DMA_PTE_WRITE); > > + DMA_PTE_READ|DMA_PTE_WRITE, > > GFP_KERNEL); > > } > > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. I checked it carefully myself as well, good to check again. > > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > > iommu_domain *domain, > > > > /* Cope with horrid API which requires us to unmap more than the > > size argument if it happens to be a large-page mapping. */ > > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level)); > > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level, > > + GFP_ATOMIC)); > > with level==0 it implies it's only lookup w/o pgtable allocation. From this > angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. We should only write GFP_KERNEL if it is actually a sleepable context because it will be mighty confusing if it isn't. I couldn't tell what the context is so I left it as ATOMIC. You are correct this is only just a lookup and so the value is never used / doesn't matter. Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 13:28 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:28 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:38:51AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > > dmar_domain *domain, > > > > return __domain_mapping(domain, first_vpfn, > > first_vpfn, last_vpfn - first_vpfn + 1, > > - DMA_PTE_READ|DMA_PTE_WRITE); > > + DMA_PTE_READ|DMA_PTE_WRITE, > > GFP_KERNEL); > > } > > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. I checked it carefully myself as well, good to check again. > > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > > iommu_domain *domain, > > > > /* Cope with horrid API which requires us to unmap more than the > > size argument if it happens to be a large-page mapping. */ > > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level)); > > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level, > > + GFP_ATOMIC)); > > with level==0 it implies it's only lookup w/o pgtable allocation. From this > angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. We should only write GFP_KERNEL if it is actually a sleepable context because it will be mighty confusing if it isn't. I couldn't tell what the context is so I left it as ATOMIC. You are correct this is only just a lookup and so the value is never used / doesn't matter. Jason ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op @ 2023-01-17 13:28 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-17 13:28 UTC (permalink / raw) To: Tian, Kevin Cc: Lu Baolu, Joerg Roedel, Matthew Rosato, Robin Murphy, Alex Williamson, ath10k@lists.infradead.org, ath11k@lists.infradead.org, Christian Borntraeger, dri-devel@lists.freedesktop.org, iommu@lists.linux.dev, kvm@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-arm-msm@vger.kernel.org, linux-media@vger.kernel.org, linux-rdma@vger.kernel.org, linux-remoteproc@vger.kernel.org, linux-s390@vger.kernel.org, linux-stm32@st-md-mailman.stormreply.com, linux-tegra@vger.kernel.org, linux-wireless@vger.kernel.org, netdev@vger.kernel.org, nouveau@lists.freedesktop.org, Niklas Schnelle, virtualization@lists.linux-foundation.org On Tue, Jan 17, 2023 at 03:38:51AM +0000, Tian, Kevin wrote: > > From: Jason Gunthorpe <jgg@nvidia.com> > > Sent: Saturday, January 7, 2023 12:43 AM > > > > @@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct > > dmar_domain *domain, > > > > return __domain_mapping(domain, first_vpfn, > > first_vpfn, last_vpfn - first_vpfn + 1, > > - DMA_PTE_READ|DMA_PTE_WRITE); > > + DMA_PTE_READ|DMA_PTE_WRITE, > > GFP_KERNEL); > > } > > Baolu, can you help confirm whether switching from GFP_ATOMIC to > GFP_KERNEL is OK in this path? it looks fine to me in a quick glance > but want to be conservative here. I checked it carefully myself as well, good to check again. > > @@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct > > iommu_domain *domain, > > > > /* Cope with horrid API which requires us to unmap more than the > > size argument if it happens to be a large-page mapping. */ > > - BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level)); > > + BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, > > &level, > > + GFP_ATOMIC)); > > with level==0 it implies it's only lookup w/o pgtable allocation. From this > angle it reads better to use a more relaxed gfp e.g. GFP_KERNEL here. We should only write GFP_KERNEL if it is actually a sleepable context because it will be mighty confusing if it isn't. I couldn't tell what the context is so I left it as ATOMIC. You are correct this is only just a lookup and so the value is never used / doesn't matter. Jason -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
* [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-06 16:42 ` Jason Gunthorpe -1 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by iommufd through s390_iommu_map_pages() and it should not be forced to atomic. Thread the gfp parameter through the call chain starting from s390_iommu_map_pages(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/s390/include/asm/pci_dma.h | 5 +++-- arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- drivers/iommu/s390-iommu.c | 15 +++++++++------ 3 files changed, 29 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/pci_dma.h b/arch/s390/include/asm/pci_dma.h index 91e63426bdc53f..7119c04c51c5c8 100644 --- a/arch/s390/include/asm/pci_dma.h +++ b/arch/s390/include/asm/pci_dma.h @@ -186,9 +186,10 @@ static inline unsigned long *get_st_pto(unsigned long entry) /* Prototypes */ void dma_free_seg_table(unsigned long); -unsigned long *dma_alloc_cpu_table(void); +unsigned long *dma_alloc_cpu_table(gfp_t gfp); void dma_cleanup_tables(unsigned long *); -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr); +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp); void dma_update_cpu_trans(unsigned long *entry, phys_addr_t page_addr, int flags); extern const struct dma_map_ops s390_pci_dma_ops; diff --git a/arch/s390/pci/pci_dma.c b/arch/s390/pci/pci_dma.c index ea478d11fbd132..2d9b01d7ca4c5c 100644 --- a/arch/s390/pci/pci_dma.c +++ b/arch/s390/pci/pci_dma.c @@ -27,11 +27,11 @@ static int zpci_refresh_global(struct zpci_dev *zdev) zdev->iommu_pages * PAGE_SIZE); } -unsigned long *dma_alloc_cpu_table(void) +unsigned long *dma_alloc_cpu_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_region_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_region_table_cache, gfp); if (!table) return NULL; @@ -45,11 +45,11 @@ static void dma_free_cpu_table(void *table) kmem_cache_free(dma_region_table_cache, table); } -static unsigned long *dma_alloc_page_table(void) +static unsigned long *dma_alloc_page_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_page_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_page_table_cache, gfp); if (!table) return NULL; @@ -63,7 +63,7 @@ static void dma_free_page_table(void *table) kmem_cache_free(dma_page_table_cache, table); } -static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) +static unsigned long *dma_get_seg_table_origin(unsigned long *rtep, gfp_t gfp) { unsigned long old_rte, rte; unsigned long *sto; @@ -72,7 +72,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) if (reg_entry_isvalid(rte)) { sto = get_rt_sto(rte); } else { - sto = dma_alloc_cpu_table(); + sto = dma_alloc_cpu_table(gfp); if (!sto) return NULL; @@ -90,7 +90,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) return sto; } -static unsigned long *dma_get_page_table_origin(unsigned long *step) +static unsigned long *dma_get_page_table_origin(unsigned long *step, gfp_t gfp) { unsigned long old_ste, ste; unsigned long *pto; @@ -99,7 +99,7 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) if (reg_entry_isvalid(ste)) { pto = get_st_pto(ste); } else { - pto = dma_alloc_page_table(); + pto = dma_alloc_page_table(gfp); if (!pto) return NULL; set_st_pto(&ste, virt_to_phys(pto)); @@ -116,18 +116,19 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) return pto; } -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr) +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp) { unsigned long *sto, *pto; unsigned int rtx, sx, px; rtx = calc_rtx(dma_addr); - sto = dma_get_seg_table_origin(&rto[rtx]); + sto = dma_get_seg_table_origin(&rto[rtx], gfp); if (!sto) return NULL; sx = calc_sx(dma_addr); - pto = dma_get_page_table_origin(&sto[sx]); + pto = dma_get_page_table_origin(&sto[sx], gfp); if (!pto) return NULL; @@ -170,7 +171,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, return -EINVAL; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) { rc = -ENOMEM; goto undo_cpu_trans; @@ -186,7 +188,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, while (i-- > 0) { page_addr -= PAGE_SIZE; dma_addr -= PAGE_SIZE; - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) break; dma_update_cpu_trans(entry, page_addr, flags); @@ -576,7 +579,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev) spin_lock_init(&zdev->iommu_bitmap_lock); - zdev->dma_table = dma_alloc_cpu_table(); + zdev->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!zdev->dma_table) { rc = -ENOMEM; goto out; diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c index ed33c6cce08362..7dcfffed260e6b 100644 --- a/drivers/iommu/s390-iommu.c +++ b/drivers/iommu/s390-iommu.c @@ -52,7 +52,7 @@ static struct iommu_domain *s390_domain_alloc(unsigned domain_type) if (!s390_domain) return NULL; - s390_domain->dma_table = dma_alloc_cpu_table(); + s390_domain->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!s390_domain->dma_table) { kfree(s390_domain); return NULL; @@ -260,7 +260,8 @@ static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain, static int s390_iommu_validate_trans(struct s390_domain *s390_domain, phys_addr_t pa, dma_addr_t dma_addr, - unsigned long nr_pages, int flags) + unsigned long nr_pages, int flags, + gfp_t gfp) { phys_addr_t page_addr = pa & PAGE_MASK; unsigned long *entry; @@ -268,7 +269,8 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, int rc; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + gfp); if (unlikely(!entry)) { rc = -ENOMEM; goto undo_cpu_trans; @@ -284,7 +286,7 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, while (i-- > 0) { dma_addr -= PAGE_SIZE; entry = dma_walk_cpu_trans(s390_domain->dma_table, - dma_addr); + dma_addr, gfp); if (!entry) break; dma_update_cpu_trans(entry, 0, ZPCI_PTE_INVALID); @@ -301,7 +303,8 @@ static int s390_iommu_invalidate_trans(struct s390_domain *s390_domain, int rc = 0; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + GFP_ATOMIC); if (unlikely(!entry)) { rc = -EINVAL; break; @@ -339,7 +342,7 @@ static int s390_iommu_map_pages(struct iommu_domain *domain, flags |= ZPCI_TABLE_PROTECTED; rc = s390_iommu_validate_trans(s390_domain, paddr, iova, - pgcount, flags); + pgcount, flags, gfp); if (!rc) *mapped = size; -- 2.39.0 _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by iommufd through s390_iommu_map_pages() and it should not be forced to atomic. Thread the gfp parameter through the call chain starting from s390_iommu_map_pages(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/s390/include/asm/pci_dma.h | 5 +++-- arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- drivers/iommu/s390-iommu.c | 15 +++++++++------ 3 files changed, 29 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/pci_dma.h b/arch/s390/include/asm/pci_dma.h index 91e63426bdc53f..7119c04c51c5c8 100644 --- a/arch/s390/include/asm/pci_dma.h +++ b/arch/s390/include/asm/pci_dma.h @@ -186,9 +186,10 @@ static inline unsigned long *get_st_pto(unsigned long entry) /* Prototypes */ void dma_free_seg_table(unsigned long); -unsigned long *dma_alloc_cpu_table(void); +unsigned long *dma_alloc_cpu_table(gfp_t gfp); void dma_cleanup_tables(unsigned long *); -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr); +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp); void dma_update_cpu_trans(unsigned long *entry, phys_addr_t page_addr, int flags); extern const struct dma_map_ops s390_pci_dma_ops; diff --git a/arch/s390/pci/pci_dma.c b/arch/s390/pci/pci_dma.c index ea478d11fbd132..2d9b01d7ca4c5c 100644 --- a/arch/s390/pci/pci_dma.c +++ b/arch/s390/pci/pci_dma.c @@ -27,11 +27,11 @@ static int zpci_refresh_global(struct zpci_dev *zdev) zdev->iommu_pages * PAGE_SIZE); } -unsigned long *dma_alloc_cpu_table(void) +unsigned long *dma_alloc_cpu_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_region_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_region_table_cache, gfp); if (!table) return NULL; @@ -45,11 +45,11 @@ static void dma_free_cpu_table(void *table) kmem_cache_free(dma_region_table_cache, table); } -static unsigned long *dma_alloc_page_table(void) +static unsigned long *dma_alloc_page_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_page_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_page_table_cache, gfp); if (!table) return NULL; @@ -63,7 +63,7 @@ static void dma_free_page_table(void *table) kmem_cache_free(dma_page_table_cache, table); } -static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) +static unsigned long *dma_get_seg_table_origin(unsigned long *rtep, gfp_t gfp) { unsigned long old_rte, rte; unsigned long *sto; @@ -72,7 +72,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) if (reg_entry_isvalid(rte)) { sto = get_rt_sto(rte); } else { - sto = dma_alloc_cpu_table(); + sto = dma_alloc_cpu_table(gfp); if (!sto) return NULL; @@ -90,7 +90,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) return sto; } -static unsigned long *dma_get_page_table_origin(unsigned long *step) +static unsigned long *dma_get_page_table_origin(unsigned long *step, gfp_t gfp) { unsigned long old_ste, ste; unsigned long *pto; @@ -99,7 +99,7 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) if (reg_entry_isvalid(ste)) { pto = get_st_pto(ste); } else { - pto = dma_alloc_page_table(); + pto = dma_alloc_page_table(gfp); if (!pto) return NULL; set_st_pto(&ste, virt_to_phys(pto)); @@ -116,18 +116,19 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) return pto; } -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr) +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp) { unsigned long *sto, *pto; unsigned int rtx, sx, px; rtx = calc_rtx(dma_addr); - sto = dma_get_seg_table_origin(&rto[rtx]); + sto = dma_get_seg_table_origin(&rto[rtx], gfp); if (!sto) return NULL; sx = calc_sx(dma_addr); - pto = dma_get_page_table_origin(&sto[sx]); + pto = dma_get_page_table_origin(&sto[sx], gfp); if (!pto) return NULL; @@ -170,7 +171,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, return -EINVAL; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) { rc = -ENOMEM; goto undo_cpu_trans; @@ -186,7 +188,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, while (i-- > 0) { page_addr -= PAGE_SIZE; dma_addr -= PAGE_SIZE; - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) break; dma_update_cpu_trans(entry, page_addr, flags); @@ -576,7 +579,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev) spin_lock_init(&zdev->iommu_bitmap_lock); - zdev->dma_table = dma_alloc_cpu_table(); + zdev->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!zdev->dma_table) { rc = -ENOMEM; goto out; diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c index ed33c6cce08362..7dcfffed260e6b 100644 --- a/drivers/iommu/s390-iommu.c +++ b/drivers/iommu/s390-iommu.c @@ -52,7 +52,7 @@ static struct iommu_domain *s390_domain_alloc(unsigned domain_type) if (!s390_domain) return NULL; - s390_domain->dma_table = dma_alloc_cpu_table(); + s390_domain->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!s390_domain->dma_table) { kfree(s390_domain); return NULL; @@ -260,7 +260,8 @@ static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain, static int s390_iommu_validate_trans(struct s390_domain *s390_domain, phys_addr_t pa, dma_addr_t dma_addr, - unsigned long nr_pages, int flags) + unsigned long nr_pages, int flags, + gfp_t gfp) { phys_addr_t page_addr = pa & PAGE_MASK; unsigned long *entry; @@ -268,7 +269,8 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, int rc; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + gfp); if (unlikely(!entry)) { rc = -ENOMEM; goto undo_cpu_trans; @@ -284,7 +286,7 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, while (i-- > 0) { dma_addr -= PAGE_SIZE; entry = dma_walk_cpu_trans(s390_domain->dma_table, - dma_addr); + dma_addr, gfp); if (!entry) break; dma_update_cpu_trans(entry, 0, ZPCI_PTE_INVALID); @@ -301,7 +303,8 @@ static int s390_iommu_invalidate_trans(struct s390_domain *s390_domain, int rc = 0; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + GFP_ATOMIC); if (unlikely(!entry)) { rc = -EINVAL; break; @@ -339,7 +342,7 @@ static int s390_iommu_map_pages(struct iommu_domain *domain, flags |= ZPCI_TABLE_PROTECTED; rc = s390_iommu_validate_trans(s390_domain, paddr, iova, - pgcount, flags); + pgcount, flags, gfp); if (!rc) *mapped = size; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by iommufd through s390_iommu_map_pages() and it should not be forced to atomic. Thread the gfp parameter through the call chain starting from s390_iommu_map_pages(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/s390/include/asm/pci_dma.h | 5 +++-- arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- drivers/iommu/s390-iommu.c | 15 +++++++++------ 3 files changed, 29 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/pci_dma.h b/arch/s390/include/asm/pci_dma.h index 91e63426bdc53f..7119c04c51c5c8 100644 --- a/arch/s390/include/asm/pci_dma.h +++ b/arch/s390/include/asm/pci_dma.h @@ -186,9 +186,10 @@ static inline unsigned long *get_st_pto(unsigned long entry) /* Prototypes */ void dma_free_seg_table(unsigned long); -unsigned long *dma_alloc_cpu_table(void); +unsigned long *dma_alloc_cpu_table(gfp_t gfp); void dma_cleanup_tables(unsigned long *); -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr); +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp); void dma_update_cpu_trans(unsigned long *entry, phys_addr_t page_addr, int flags); extern const struct dma_map_ops s390_pci_dma_ops; diff --git a/arch/s390/pci/pci_dma.c b/arch/s390/pci/pci_dma.c index ea478d11fbd132..2d9b01d7ca4c5c 100644 --- a/arch/s390/pci/pci_dma.c +++ b/arch/s390/pci/pci_dma.c @@ -27,11 +27,11 @@ static int zpci_refresh_global(struct zpci_dev *zdev) zdev->iommu_pages * PAGE_SIZE); } -unsigned long *dma_alloc_cpu_table(void) +unsigned long *dma_alloc_cpu_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_region_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_region_table_cache, gfp); if (!table) return NULL; @@ -45,11 +45,11 @@ static void dma_free_cpu_table(void *table) kmem_cache_free(dma_region_table_cache, table); } -static unsigned long *dma_alloc_page_table(void) +static unsigned long *dma_alloc_page_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_page_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_page_table_cache, gfp); if (!table) return NULL; @@ -63,7 +63,7 @@ static void dma_free_page_table(void *table) kmem_cache_free(dma_page_table_cache, table); } -static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) +static unsigned long *dma_get_seg_table_origin(unsigned long *rtep, gfp_t gfp) { unsigned long old_rte, rte; unsigned long *sto; @@ -72,7 +72,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) if (reg_entry_isvalid(rte)) { sto = get_rt_sto(rte); } else { - sto = dma_alloc_cpu_table(); + sto = dma_alloc_cpu_table(gfp); if (!sto) return NULL; @@ -90,7 +90,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) return sto; } -static unsigned long *dma_get_page_table_origin(unsigned long *step) +static unsigned long *dma_get_page_table_origin(unsigned long *step, gfp_t gfp) { unsigned long old_ste, ste; unsigned long *pto; @@ -99,7 +99,7 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) if (reg_entry_isvalid(ste)) { pto = get_st_pto(ste); } else { - pto = dma_alloc_page_table(); + pto = dma_alloc_page_table(gfp); if (!pto) return NULL; set_st_pto(&ste, virt_to_phys(pto)); @@ -116,18 +116,19 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) return pto; } -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr) +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp) { unsigned long *sto, *pto; unsigned int rtx, sx, px; rtx = calc_rtx(dma_addr); - sto = dma_get_seg_table_origin(&rto[rtx]); + sto = dma_get_seg_table_origin(&rto[rtx], gfp); if (!sto) return NULL; sx = calc_sx(dma_addr); - pto = dma_get_page_table_origin(&sto[sx]); + pto = dma_get_page_table_origin(&sto[sx], gfp); if (!pto) return NULL; @@ -170,7 +171,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, return -EINVAL; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) { rc = -ENOMEM; goto undo_cpu_trans; @@ -186,7 +188,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, while (i-- > 0) { page_addr -= PAGE_SIZE; dma_addr -= PAGE_SIZE; - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) break; dma_update_cpu_trans(entry, page_addr, flags); @@ -576,7 +579,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev) spin_lock_init(&zdev->iommu_bitmap_lock); - zdev->dma_table = dma_alloc_cpu_table(); + zdev->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!zdev->dma_table) { rc = -ENOMEM; goto out; diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c index ed33c6cce08362..7dcfffed260e6b 100644 --- a/drivers/iommu/s390-iommu.c +++ b/drivers/iommu/s390-iommu.c @@ -52,7 +52,7 @@ static struct iommu_domain *s390_domain_alloc(unsigned domain_type) if (!s390_domain) return NULL; - s390_domain->dma_table = dma_alloc_cpu_table(); + s390_domain->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!s390_domain->dma_table) { kfree(s390_domain); return NULL; @@ -260,7 +260,8 @@ static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain, static int s390_iommu_validate_trans(struct s390_domain *s390_domain, phys_addr_t pa, dma_addr_t dma_addr, - unsigned long nr_pages, int flags) + unsigned long nr_pages, int flags, + gfp_t gfp) { phys_addr_t page_addr = pa & PAGE_MASK; unsigned long *entry; @@ -268,7 +269,8 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, int rc; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + gfp); if (unlikely(!entry)) { rc = -ENOMEM; goto undo_cpu_trans; @@ -284,7 +286,7 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, while (i-- > 0) { dma_addr -= PAGE_SIZE; entry = dma_walk_cpu_trans(s390_domain->dma_table, - dma_addr); + dma_addr, gfp); if (!entry) break; dma_update_cpu_trans(entry, 0, ZPCI_PTE_INVALID); @@ -301,7 +303,8 @@ static int s390_iommu_invalidate_trans(struct s390_domain *s390_domain, int rc = 0; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + GFP_ATOMIC); if (unlikely(!entry)) { rc = -EINVAL; break; @@ -339,7 +342,7 @@ static int s390_iommu_map_pages(struct iommu_domain *domain, flags |= ZPCI_TABLE_PROTECTED; rc = s390_iommu_validate_trans(s390_domain, paddr, iova, - pgcount, flags); + pgcount, flags, gfp); if (!rc) *mapped = size; -- 2.39.0 _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [Nouveau] [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, Niklas Schnelle, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by iommufd through s390_iommu_map_pages() and it should not be forced to atomic. Thread the gfp parameter through the call chain starting from s390_iommu_map_pages(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/s390/include/asm/pci_dma.h | 5 +++-- arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- drivers/iommu/s390-iommu.c | 15 +++++++++------ 3 files changed, 29 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/pci_dma.h b/arch/s390/include/asm/pci_dma.h index 91e63426bdc53f..7119c04c51c5c8 100644 --- a/arch/s390/include/asm/pci_dma.h +++ b/arch/s390/include/asm/pci_dma.h @@ -186,9 +186,10 @@ static inline unsigned long *get_st_pto(unsigned long entry) /* Prototypes */ void dma_free_seg_table(unsigned long); -unsigned long *dma_alloc_cpu_table(void); +unsigned long *dma_alloc_cpu_table(gfp_t gfp); void dma_cleanup_tables(unsigned long *); -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr); +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp); void dma_update_cpu_trans(unsigned long *entry, phys_addr_t page_addr, int flags); extern const struct dma_map_ops s390_pci_dma_ops; diff --git a/arch/s390/pci/pci_dma.c b/arch/s390/pci/pci_dma.c index ea478d11fbd132..2d9b01d7ca4c5c 100644 --- a/arch/s390/pci/pci_dma.c +++ b/arch/s390/pci/pci_dma.c @@ -27,11 +27,11 @@ static int zpci_refresh_global(struct zpci_dev *zdev) zdev->iommu_pages * PAGE_SIZE); } -unsigned long *dma_alloc_cpu_table(void) +unsigned long *dma_alloc_cpu_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_region_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_region_table_cache, gfp); if (!table) return NULL; @@ -45,11 +45,11 @@ static void dma_free_cpu_table(void *table) kmem_cache_free(dma_region_table_cache, table); } -static unsigned long *dma_alloc_page_table(void) +static unsigned long *dma_alloc_page_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_page_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_page_table_cache, gfp); if (!table) return NULL; @@ -63,7 +63,7 @@ static void dma_free_page_table(void *table) kmem_cache_free(dma_page_table_cache, table); } -static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) +static unsigned long *dma_get_seg_table_origin(unsigned long *rtep, gfp_t gfp) { unsigned long old_rte, rte; unsigned long *sto; @@ -72,7 +72,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) if (reg_entry_isvalid(rte)) { sto = get_rt_sto(rte); } else { - sto = dma_alloc_cpu_table(); + sto = dma_alloc_cpu_table(gfp); if (!sto) return NULL; @@ -90,7 +90,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) return sto; } -static unsigned long *dma_get_page_table_origin(unsigned long *step) +static unsigned long *dma_get_page_table_origin(unsigned long *step, gfp_t gfp) { unsigned long old_ste, ste; unsigned long *pto; @@ -99,7 +99,7 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) if (reg_entry_isvalid(ste)) { pto = get_st_pto(ste); } else { - pto = dma_alloc_page_table(); + pto = dma_alloc_page_table(gfp); if (!pto) return NULL; set_st_pto(&ste, virt_to_phys(pto)); @@ -116,18 +116,19 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) return pto; } -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr) +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp) { unsigned long *sto, *pto; unsigned int rtx, sx, px; rtx = calc_rtx(dma_addr); - sto = dma_get_seg_table_origin(&rto[rtx]); + sto = dma_get_seg_table_origin(&rto[rtx], gfp); if (!sto) return NULL; sx = calc_sx(dma_addr); - pto = dma_get_page_table_origin(&sto[sx]); + pto = dma_get_page_table_origin(&sto[sx], gfp); if (!pto) return NULL; @@ -170,7 +171,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, return -EINVAL; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) { rc = -ENOMEM; goto undo_cpu_trans; @@ -186,7 +188,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, while (i-- > 0) { page_addr -= PAGE_SIZE; dma_addr -= PAGE_SIZE; - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) break; dma_update_cpu_trans(entry, page_addr, flags); @@ -576,7 +579,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev) spin_lock_init(&zdev->iommu_bitmap_lock); - zdev->dma_table = dma_alloc_cpu_table(); + zdev->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!zdev->dma_table) { rc = -ENOMEM; goto out; diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c index ed33c6cce08362..7dcfffed260e6b 100644 --- a/drivers/iommu/s390-iommu.c +++ b/drivers/iommu/s390-iommu.c @@ -52,7 +52,7 @@ static struct iommu_domain *s390_domain_alloc(unsigned domain_type) if (!s390_domain) return NULL; - s390_domain->dma_table = dma_alloc_cpu_table(); + s390_domain->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!s390_domain->dma_table) { kfree(s390_domain); return NULL; @@ -260,7 +260,8 @@ static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain, static int s390_iommu_validate_trans(struct s390_domain *s390_domain, phys_addr_t pa, dma_addr_t dma_addr, - unsigned long nr_pages, int flags) + unsigned long nr_pages, int flags, + gfp_t gfp) { phys_addr_t page_addr = pa & PAGE_MASK; unsigned long *entry; @@ -268,7 +269,8 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, int rc; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + gfp); if (unlikely(!entry)) { rc = -ENOMEM; goto undo_cpu_trans; @@ -284,7 +286,7 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, while (i-- > 0) { dma_addr -= PAGE_SIZE; entry = dma_walk_cpu_trans(s390_domain->dma_table, - dma_addr); + dma_addr, gfp); if (!entry) break; dma_update_cpu_trans(entry, 0, ZPCI_PTE_INVALID); @@ -301,7 +303,8 @@ static int s390_iommu_invalidate_trans(struct s390_domain *s390_domain, int rc = 0; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + GFP_ATOMIC); if (unlikely(!entry)) { rc = -EINVAL; break; @@ -339,7 +342,7 @@ static int s390_iommu_map_pages(struct iommu_domain *domain, flags |= ZPCI_TABLE_PROTECTED; rc = s390_iommu_validate_trans(s390_domain, paddr, iova, - pgcount, flags); + pgcount, flags, gfp); if (!rc) *mapped = size; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by iommufd through s390_iommu_map_pages() and it should not be forced to atomic. Thread the gfp parameter through the call chain starting from s390_iommu_map_pages(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/s390/include/asm/pci_dma.h | 5 +++-- arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- drivers/iommu/s390-iommu.c | 15 +++++++++------ 3 files changed, 29 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/pci_dma.h b/arch/s390/include/asm/pci_dma.h index 91e63426bdc53f..7119c04c51c5c8 100644 --- a/arch/s390/include/asm/pci_dma.h +++ b/arch/s390/include/asm/pci_dma.h @@ -186,9 +186,10 @@ static inline unsigned long *get_st_pto(unsigned long entry) /* Prototypes */ void dma_free_seg_table(unsigned long); -unsigned long *dma_alloc_cpu_table(void); +unsigned long *dma_alloc_cpu_table(gfp_t gfp); void dma_cleanup_tables(unsigned long *); -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr); +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp); void dma_update_cpu_trans(unsigned long *entry, phys_addr_t page_addr, int flags); extern const struct dma_map_ops s390_pci_dma_ops; diff --git a/arch/s390/pci/pci_dma.c b/arch/s390/pci/pci_dma.c index ea478d11fbd132..2d9b01d7ca4c5c 100644 --- a/arch/s390/pci/pci_dma.c +++ b/arch/s390/pci/pci_dma.c @@ -27,11 +27,11 @@ static int zpci_refresh_global(struct zpci_dev *zdev) zdev->iommu_pages * PAGE_SIZE); } -unsigned long *dma_alloc_cpu_table(void) +unsigned long *dma_alloc_cpu_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_region_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_region_table_cache, gfp); if (!table) return NULL; @@ -45,11 +45,11 @@ static void dma_free_cpu_table(void *table) kmem_cache_free(dma_region_table_cache, table); } -static unsigned long *dma_alloc_page_table(void) +static unsigned long *dma_alloc_page_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_page_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_page_table_cache, gfp); if (!table) return NULL; @@ -63,7 +63,7 @@ static void dma_free_page_table(void *table) kmem_cache_free(dma_page_table_cache, table); } -static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) +static unsigned long *dma_get_seg_table_origin(unsigned long *rtep, gfp_t gfp) { unsigned long old_rte, rte; unsigned long *sto; @@ -72,7 +72,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) if (reg_entry_isvalid(rte)) { sto = get_rt_sto(rte); } else { - sto = dma_alloc_cpu_table(); + sto = dma_alloc_cpu_table(gfp); if (!sto) return NULL; @@ -90,7 +90,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) return sto; } -static unsigned long *dma_get_page_table_origin(unsigned long *step) +static unsigned long *dma_get_page_table_origin(unsigned long *step, gfp_t gfp) { unsigned long old_ste, ste; unsigned long *pto; @@ -99,7 +99,7 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) if (reg_entry_isvalid(ste)) { pto = get_st_pto(ste); } else { - pto = dma_alloc_page_table(); + pto = dma_alloc_page_table(gfp); if (!pto) return NULL; set_st_pto(&ste, virt_to_phys(pto)); @@ -116,18 +116,19 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) return pto; } -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr) +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp) { unsigned long *sto, *pto; unsigned int rtx, sx, px; rtx = calc_rtx(dma_addr); - sto = dma_get_seg_table_origin(&rto[rtx]); + sto = dma_get_seg_table_origin(&rto[rtx], gfp); if (!sto) return NULL; sx = calc_sx(dma_addr); - pto = dma_get_page_table_origin(&sto[sx]); + pto = dma_get_page_table_origin(&sto[sx], gfp); if (!pto) return NULL; @@ -170,7 +171,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, return -EINVAL; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) { rc = -ENOMEM; goto undo_cpu_trans; @@ -186,7 +188,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, while (i-- > 0) { page_addr -= PAGE_SIZE; dma_addr -= PAGE_SIZE; - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) break; dma_update_cpu_trans(entry, page_addr, flags); @@ -576,7 +579,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev) spin_lock_init(&zdev->iommu_bitmap_lock); - zdev->dma_table = dma_alloc_cpu_table(); + zdev->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!zdev->dma_table) { rc = -ENOMEM; goto out; diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c index ed33c6cce08362..7dcfffed260e6b 100644 --- a/drivers/iommu/s390-iommu.c +++ b/drivers/iommu/s390-iommu.c @@ -52,7 +52,7 @@ static struct iommu_domain *s390_domain_alloc(unsigned domain_type) if (!s390_domain) return NULL; - s390_domain->dma_table = dma_alloc_cpu_table(); + s390_domain->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!s390_domain->dma_table) { kfree(s390_domain); return NULL; @@ -260,7 +260,8 @@ static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain, static int s390_iommu_validate_trans(struct s390_domain *s390_domain, phys_addr_t pa, dma_addr_t dma_addr, - unsigned long nr_pages, int flags) + unsigned long nr_pages, int flags, + gfp_t gfp) { phys_addr_t page_addr = pa & PAGE_MASK; unsigned long *entry; @@ -268,7 +269,8 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, int rc; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + gfp); if (unlikely(!entry)) { rc = -ENOMEM; goto undo_cpu_trans; @@ -284,7 +286,7 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, while (i-- > 0) { dma_addr -= PAGE_SIZE; entry = dma_walk_cpu_trans(s390_domain->dma_table, - dma_addr); + dma_addr, gfp); if (!entry) break; dma_update_cpu_trans(entry, 0, ZPCI_PTE_INVALID); @@ -301,7 +303,8 @@ static int s390_iommu_invalidate_trans(struct s390_domain *s390_domain, int rc = 0; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + GFP_ATOMIC); if (unlikely(!entry)) { rc = -EINVAL; break; @@ -339,7 +342,7 @@ static int s390_iommu_map_pages(struct iommu_domain *domain, flags |= ZPCI_TABLE_PROTECTED; rc = s390_iommu_validate_trans(s390_domain, paddr, iova, - pgcount, flags); + pgcount, flags, gfp); if (!rc) *mapped = size; -- 2.39.0 ^ permalink raw reply related [flat|nested] 144+ messages in thread
* [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-06 16:42 ` Jason Gunthorpe 0 siblings, 0 replies; 144+ messages in thread From: Jason Gunthorpe @ 2023-01-06 16:42 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, Niklas Schnelle, virtualization dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by iommufd through s390_iommu_map_pages() and it should not be forced to atomic. Thread the gfp parameter through the call chain starting from s390_iommu_map_pages(). Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> --- arch/s390/include/asm/pci_dma.h | 5 +++-- arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- drivers/iommu/s390-iommu.c | 15 +++++++++------ 3 files changed, 29 insertions(+), 22 deletions(-) diff --git a/arch/s390/include/asm/pci_dma.h b/arch/s390/include/asm/pci_dma.h index 91e63426bdc53f..7119c04c51c5c8 100644 --- a/arch/s390/include/asm/pci_dma.h +++ b/arch/s390/include/asm/pci_dma.h @@ -186,9 +186,10 @@ static inline unsigned long *get_st_pto(unsigned long entry) /* Prototypes */ void dma_free_seg_table(unsigned long); -unsigned long *dma_alloc_cpu_table(void); +unsigned long *dma_alloc_cpu_table(gfp_t gfp); void dma_cleanup_tables(unsigned long *); -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr); +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp); void dma_update_cpu_trans(unsigned long *entry, phys_addr_t page_addr, int flags); extern const struct dma_map_ops s390_pci_dma_ops; diff --git a/arch/s390/pci/pci_dma.c b/arch/s390/pci/pci_dma.c index ea478d11fbd132..2d9b01d7ca4c5c 100644 --- a/arch/s390/pci/pci_dma.c +++ b/arch/s390/pci/pci_dma.c @@ -27,11 +27,11 @@ static int zpci_refresh_global(struct zpci_dev *zdev) zdev->iommu_pages * PAGE_SIZE); } -unsigned long *dma_alloc_cpu_table(void) +unsigned long *dma_alloc_cpu_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_region_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_region_table_cache, gfp); if (!table) return NULL; @@ -45,11 +45,11 @@ static void dma_free_cpu_table(void *table) kmem_cache_free(dma_region_table_cache, table); } -static unsigned long *dma_alloc_page_table(void) +static unsigned long *dma_alloc_page_table(gfp_t gfp) { unsigned long *table, *entry; - table = kmem_cache_alloc(dma_page_table_cache, GFP_ATOMIC); + table = kmem_cache_alloc(dma_page_table_cache, gfp); if (!table) return NULL; @@ -63,7 +63,7 @@ static void dma_free_page_table(void *table) kmem_cache_free(dma_page_table_cache, table); } -static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) +static unsigned long *dma_get_seg_table_origin(unsigned long *rtep, gfp_t gfp) { unsigned long old_rte, rte; unsigned long *sto; @@ -72,7 +72,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) if (reg_entry_isvalid(rte)) { sto = get_rt_sto(rte); } else { - sto = dma_alloc_cpu_table(); + sto = dma_alloc_cpu_table(gfp); if (!sto) return NULL; @@ -90,7 +90,7 @@ static unsigned long *dma_get_seg_table_origin(unsigned long *rtep) return sto; } -static unsigned long *dma_get_page_table_origin(unsigned long *step) +static unsigned long *dma_get_page_table_origin(unsigned long *step, gfp_t gfp) { unsigned long old_ste, ste; unsigned long *pto; @@ -99,7 +99,7 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) if (reg_entry_isvalid(ste)) { pto = get_st_pto(ste); } else { - pto = dma_alloc_page_table(); + pto = dma_alloc_page_table(gfp); if (!pto) return NULL; set_st_pto(&ste, virt_to_phys(pto)); @@ -116,18 +116,19 @@ static unsigned long *dma_get_page_table_origin(unsigned long *step) return pto; } -unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr) +unsigned long *dma_walk_cpu_trans(unsigned long *rto, dma_addr_t dma_addr, + gfp_t gfp) { unsigned long *sto, *pto; unsigned int rtx, sx, px; rtx = calc_rtx(dma_addr); - sto = dma_get_seg_table_origin(&rto[rtx]); + sto = dma_get_seg_table_origin(&rto[rtx], gfp); if (!sto) return NULL; sx = calc_sx(dma_addr); - pto = dma_get_page_table_origin(&sto[sx]); + pto = dma_get_page_table_origin(&sto[sx], gfp); if (!pto) return NULL; @@ -170,7 +171,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, return -EINVAL; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) { rc = -ENOMEM; goto undo_cpu_trans; @@ -186,7 +188,8 @@ static int __dma_update_trans(struct zpci_dev *zdev, phys_addr_t pa, while (i-- > 0) { page_addr -= PAGE_SIZE; dma_addr -= PAGE_SIZE; - entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr); + entry = dma_walk_cpu_trans(zdev->dma_table, dma_addr, + GFP_ATOMIC); if (!entry) break; dma_update_cpu_trans(entry, page_addr, flags); @@ -576,7 +579,7 @@ int zpci_dma_init_device(struct zpci_dev *zdev) spin_lock_init(&zdev->iommu_bitmap_lock); - zdev->dma_table = dma_alloc_cpu_table(); + zdev->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!zdev->dma_table) { rc = -ENOMEM; goto out; diff --git a/drivers/iommu/s390-iommu.c b/drivers/iommu/s390-iommu.c index ed33c6cce08362..7dcfffed260e6b 100644 --- a/drivers/iommu/s390-iommu.c +++ b/drivers/iommu/s390-iommu.c @@ -52,7 +52,7 @@ static struct iommu_domain *s390_domain_alloc(unsigned domain_type) if (!s390_domain) return NULL; - s390_domain->dma_table = dma_alloc_cpu_table(); + s390_domain->dma_table = dma_alloc_cpu_table(GFP_KERNEL); if (!s390_domain->dma_table) { kfree(s390_domain); return NULL; @@ -260,7 +260,8 @@ static void s390_iommu_iotlb_sync_map(struct iommu_domain *domain, static int s390_iommu_validate_trans(struct s390_domain *s390_domain, phys_addr_t pa, dma_addr_t dma_addr, - unsigned long nr_pages, int flags) + unsigned long nr_pages, int flags, + gfp_t gfp) { phys_addr_t page_addr = pa & PAGE_MASK; unsigned long *entry; @@ -268,7 +269,8 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, int rc; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + gfp); if (unlikely(!entry)) { rc = -ENOMEM; goto undo_cpu_trans; @@ -284,7 +286,7 @@ static int s390_iommu_validate_trans(struct s390_domain *s390_domain, while (i-- > 0) { dma_addr -= PAGE_SIZE; entry = dma_walk_cpu_trans(s390_domain->dma_table, - dma_addr); + dma_addr, gfp); if (!entry) break; dma_update_cpu_trans(entry, 0, ZPCI_PTE_INVALID); @@ -301,7 +303,8 @@ static int s390_iommu_invalidate_trans(struct s390_domain *s390_domain, int rc = 0; for (i = 0; i < nr_pages; i++) { - entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr); + entry = dma_walk_cpu_trans(s390_domain->dma_table, dma_addr, + GFP_ATOMIC); if (unlikely(!entry)) { rc = -EINVAL; break; @@ -339,7 +342,7 @@ static int s390_iommu_map_pages(struct iommu_domain *domain, flags |= ZPCI_TABLE_PROTECTED; rc = s390_iommu_validate_trans(s390_domain, paddr, iova, - pgcount, flags); + pgcount, flags, gfp); if (!rc) *mapped = size; -- 2.39.0 -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply related [flat|nested] 144+ messages in thread
* Re: [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s 2023-01-06 16:42 ` Jason Gunthorpe ` (3 preceding siblings ...) (?) @ 2023-01-17 8:49 ` Niklas Schnelle -1 siblings, 0 replies; 144+ messages in thread From: Niklas Schnelle @ 2023-01-17 8:49 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, virtualization On Fri, 2023-01-06 at 12:42 -0400, Jason Gunthorpe wrote: > dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by > iommufd through s390_iommu_map_pages() and it should not be forced to > atomic. Thread the gfp parameter through the call chain starting from > s390_iommu_map_pages(). > > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/s390/include/asm/pci_dma.h | 5 +++-- > arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- > drivers/iommu/s390-iommu.c | 15 +++++++++------ > 3 files changed, 29 insertions(+), 22 deletions(-) > ---8<--- > Looks good to me and I have no objections. Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com> _______________________________________________ ath10k mailing list ath10k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath10k ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-17 8:49 ` Niklas Schnelle 0 siblings, 0 replies; 144+ messages in thread From: Niklas Schnelle @ 2023-01-17 8:49 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media On Fri, 2023-01-06 at 12:42 -0400, Jason Gunthorpe wrote: > dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by > iommufd through s390_iommu_map_pages() and it should not be forced to > atomic. Thread the gfp parameter through the call chain starting from > s390_iommu_map_pages(). > > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/s390/include/asm/pci_dma.h | 5 +++-- > arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- > drivers/iommu/s390-iommu.c | 15 +++++++++------ > 3 files changed, 29 insertions(+), 22 deletions(-) > ---8<--- > Looks good to me and I have no objections. Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com> ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-17 8:49 ` Niklas Schnelle 0 siblings, 0 replies; 144+ messages in thread From: Niklas Schnelle @ 2023-01-17 8:49 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, virtualization On Fri, 2023-01-06 at 12:42 -0400, Jason Gunthorpe wrote: > dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by > iommufd through s390_iommu_map_pages() and it should not be forced to > atomic. Thread the gfp parameter through the call chain starting from > s390_iommu_map_pages(). > > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/s390/include/asm/pci_dma.h | 5 +++-- > arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- > drivers/iommu/s390-iommu.c | 15 +++++++++------ > 3 files changed, 29 insertions(+), 22 deletions(-) > ---8<--- > Looks good to me and I have no objections. Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com> _______________________________________________ linux-arm-kernel mailing list linux-arm-kernel@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-arm-kernel ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [Nouveau] [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-17 8:49 ` Niklas Schnelle 0 siblings, 0 replies; 144+ messages in thread From: Niklas Schnelle @ 2023-01-17 8:49 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: linux-s390, kvm, nouveau, linux-rdma, linux-arm-msm, linux-remoteproc, iommu, dri-devel, linux-stm32, Alex Williamson, netdev, ath10k, linux-wireless, linux-tegra, Christian Borntraeger, virtualization, ath11k, linux-arm-kernel, linux-media On Fri, 2023-01-06 at 12:42 -0400, Jason Gunthorpe wrote: > dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by > iommufd through s390_iommu_map_pages() and it should not be forced to > atomic. Thread the gfp parameter through the call chain starting from > s390_iommu_map_pages(). > > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/s390/include/asm/pci_dma.h | 5 +++-- > arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- > drivers/iommu/s390-iommu.c | 15 +++++++++------ > 3 files changed, 29 insertions(+), 22 deletions(-) > ---8<--- > Looks good to me and I have no objections. Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com> ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-17 8:49 ` Niklas Schnelle 0 siblings, 0 replies; 144+ messages in thread From: Niklas Schnelle @ 2023-01-17 8:49 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, virtualization On Fri, 2023-01-06 at 12:42 -0400, Jason Gunthorpe wrote: > dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by > iommufd through s390_iommu_map_pages() and it should not be forced to > atomic. Thread the gfp parameter through the call chain starting from > s390_iommu_map_pages(). > > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/s390/include/asm/pci_dma.h | 5 +++-- > arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- > drivers/iommu/s390-iommu.c | 15 +++++++++------ > 3 files changed, 29 insertions(+), 22 deletions(-) > ---8<--- > Looks good to me and I have no objections. Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com> ^ permalink raw reply [flat|nested] 144+ messages in thread
* Re: [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s @ 2023-01-17 8:49 ` Niklas Schnelle 0 siblings, 0 replies; 144+ messages in thread From: Niklas Schnelle @ 2023-01-17 8:49 UTC (permalink / raw) To: Jason Gunthorpe, Lu Baolu, Joerg Roedel, Kevin Tian, Matthew Rosato, Robin Murphy Cc: Alex Williamson, ath10k, ath11k, Christian Borntraeger, dri-devel, iommu, kvm, linux-arm-kernel, linux-arm-msm, linux-media, linux-rdma, linux-remoteproc, linux-s390, linux-stm32, linux-tegra, linux-wireless, netdev, nouveau, virtualization On Fri, 2023-01-06 at 12:42 -0400, Jason Gunthorpe wrote: > dma_alloc_cpu_table() and dma_alloc_page_table() are eventually called by > iommufd through s390_iommu_map_pages() and it should not be forced to > atomic. Thread the gfp parameter through the call chain starting from > s390_iommu_map_pages(). > > Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> > --- > arch/s390/include/asm/pci_dma.h | 5 +++-- > arch/s390/pci/pci_dma.c | 31 +++++++++++++++++-------------- > drivers/iommu/s390-iommu.c | 15 +++++++++------ > 3 files changed, 29 insertions(+), 22 deletions(-) > ---8<--- > Looks good to me and I have no objections. Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com> -- ath11k mailing list ath11k@lists.infradead.org http://lists.infradead.org/mailman/listinfo/ath11k ^ permalink raw reply [flat|nested] 144+ messages in thread
end of thread, other threads:[~2023-05-04 12:32 UTC | newest] Thread overview: 144+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2023-01-06 16:42 [PATCH 0/8] Let iommufd charge IOPTE allocations to the memory cgroup Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 1/8] iommu: Add a gfp parameter to iommu_map() Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 17:15 ` Robin Murphy 2023-01-06 17:15 ` Robin Murphy 2023-01-06 17:15 ` Robin Murphy 2023-01-06 17:15 ` Robin Murphy 2023-01-06 17:15 ` [Nouveau] " Robin Murphy 2023-01-06 17:15 ` Robin Murphy 2023-01-06 17:15 ` Robin Murphy 2023-01-06 17:24 ` Jason Gunthorpe 2023-01-06 17:24 ` Jason Gunthorpe 2023-01-06 17:24 ` Jason Gunthorpe 2023-01-06 17:24 ` [Nouveau] " Jason Gunthorpe 2023-01-06 17:24 ` Jason Gunthorpe 2023-01-06 17:24 ` Jason Gunthorpe 2023-01-20 9:24 ` Joerg Roedel 2023-01-20 9:24 ` Joerg Roedel 2023-01-20 9:24 ` Joerg Roedel 2023-01-20 9:24 ` Joerg Roedel 2023-01-20 9:24 ` [Nouveau] " Joerg Roedel 2023-01-20 9:24 ` Joerg Roedel 2023-01-20 9:24 ` Joerg Roedel 2023-01-20 17:53 ` Jason Gunthorpe 2023-01-20 17:53 ` Jason Gunthorpe 2023-01-20 17:53 ` Jason Gunthorpe 2023-01-20 17:53 ` [Nouveau] " Jason Gunthorpe 2023-01-20 17:53 ` Jason Gunthorpe 2023-01-20 17:53 ` Jason Gunthorpe 2023-01-23 9:59 ` Joerg Roedel 2023-01-23 9:59 ` Joerg Roedel 2023-01-23 9:59 ` Joerg Roedel 2023-01-23 9:59 ` Joerg Roedel 2023-01-23 9:59 ` [Nouveau] " Joerg Roedel 2023-01-23 9:59 ` Joerg Roedel 2023-01-23 9:59 ` Joerg Roedel 2023-01-16 18:12 ` Jason Gunthorpe 2023-01-16 18:12 ` Jason Gunthorpe 2023-01-16 18:12 ` Jason Gunthorpe 2023-01-16 18:12 ` [Nouveau] " Jason Gunthorpe 2023-01-16 18:12 ` Jason Gunthorpe 2023-01-16 18:12 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 2/8] iommu: Remove iommu_map_atomic() Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 3/8] iommu: Add a gfp parameter to iommu_map_sg() Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 4/8] iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous() Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 5/8] iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map() Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 6/8] iommu/intel: Add a gfp parameter to alloc_pgtable_page() Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-17 3:35 ` Tian, Kevin 2023-01-17 3:35 ` Tian, Kevin 2023-01-17 3:35 ` Tian, Kevin 2023-01-17 3:35 ` Tian, Kevin 2023-01-17 3:35 ` [Nouveau] " Tian, Kevin 2023-01-17 3:35 ` Tian, Kevin 2023-01-17 3:35 ` Tian, Kevin 2023-01-17 13:30 ` Jason Gunthorpe 2023-01-17 13:30 ` Jason Gunthorpe 2023-01-17 13:30 ` Jason Gunthorpe 2023-01-17 13:30 ` [Nouveau] " Jason Gunthorpe 2023-01-17 13:30 ` Jason Gunthorpe 2023-01-17 13:30 ` Jason Gunthorpe 2023-01-18 1:18 ` Tian, Kevin 2023-01-18 1:18 ` Tian, Kevin 2023-01-18 1:18 ` Tian, Kevin 2023-01-18 1:18 ` Tian, Kevin 2023-01-18 1:18 ` [Nouveau] " Tian, Kevin 2023-01-18 1:18 ` Tian, Kevin 2023-01-18 1:18 ` Tian, Kevin 2023-01-18 15:15 ` Jason Gunthorpe 2023-01-18 15:15 ` Jason Gunthorpe 2023-01-18 15:15 ` Jason Gunthorpe 2023-01-18 15:15 ` [Nouveau] " Jason Gunthorpe 2023-01-18 15:15 ` Jason Gunthorpe 2023-01-18 15:15 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 7/8] iommu/intel: Support the gfp argument to the map_pages op Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-17 3:38 ` Tian, Kevin 2023-01-17 3:38 ` Tian, Kevin 2023-01-17 3:38 ` Tian, Kevin 2023-01-17 3:38 ` Tian, Kevin 2023-01-17 3:38 ` [Nouveau] " Tian, Kevin 2023-01-17 3:38 ` Tian, Kevin 2023-01-17 3:38 ` Tian, Kevin 2023-01-17 8:35 ` Baolu Lu 2023-01-17 8:35 ` Baolu Lu 2023-01-17 8:35 ` Baolu Lu 2023-01-17 8:35 ` [Nouveau] " Baolu Lu 2023-01-17 8:35 ` Baolu Lu 2023-01-17 8:35 ` Baolu Lu 2023-01-17 13:28 ` Jason Gunthorpe 2023-01-17 13:28 ` Jason Gunthorpe 2023-01-17 13:28 ` Jason Gunthorpe 2023-01-17 13:28 ` [Nouveau] " Jason Gunthorpe 2023-01-17 13:28 ` Jason Gunthorpe 2023-01-17 13:28 ` Jason Gunthorpe 2023-01-06 16:42 ` [PATCH 8/8] iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` [Nouveau] " Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-06 16:42 ` Jason Gunthorpe 2023-01-17 8:49 ` Niklas Schnelle 2023-01-17 8:49 ` Niklas Schnelle 2023-01-17 8:49 ` Niklas Schnelle 2023-01-17 8:49 ` [Nouveau] " Niklas Schnelle 2023-01-17 8:49 ` Niklas Schnelle 2023-01-17 8:49 ` Niklas Schnelle
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.