kernel/linux.git/drivers/iommu/amd, branch v6.13.6

iommu/amd: Expicitly enable CNTRL.EPHEn bit in resume path

2025-02-21T13:10:51+00:00

[ Upstream commit ef75966abf950c0539534effa4960caa29fb7167 ] With recent kernel, AMDGPU failed to resume after suspend on certain laptop. Sample log: ----------- Nov 14 11:52:19 Thinkbook kernel: iommu ivhd0: AMD-Vi: Event logged [ILLEGAL_DEV_TABLE_ENTRY device=0000:06:00.0 pasid=0x00000 address=0x135300000 flags=0x0080] Nov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[0]: 7d90000000000003 Nov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[1]: 0000100103fc0009 Nov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[2]: 2000000117840013 Nov 14 11:52:19 Thinkbook kernel: AMD-Vi: DTE[3]: 0000000000000000 This is because in resume path, CNTRL[EPHEn] is not set. Fix this by setting CNTRL[EPHEn] to 1 in resume path if EFR[EPHSUP] is set. Note May be better approach is to save the control register in suspend path and restore it in resume path instead of trying to set indivisual bits. We will have separate patch for that. Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219499 Fixes: c4cb23111103 ("iommu/amd: Add support for enable/disable IOPF") Tested-by: Hamish McIntyre-Bhatty Signed-off-by: Vasant Hegde Link: https://lore.kernel.org/r/20250127094411.5931-1-vasant.hegde@amd.com Signed-off-by: Joerg Roedel Signed-off-by: Sasha Levin

iommu/amd: Fully decode all combinations of alloc_paging_flags

2025-02-08T09:01:40+00:00

[ Upstream commit 082f1bcae8d1b5f76e92e369091176b8d61120ec ] Currently AMD does not support IOMMU_HWPT_ALLOC_PASID | IOMMU_HWPT_ALLOC_DIRTY_TRACKING It should be rejected. Instead it creates a V1 domain without dirty tracking support. Use a switch to fully decode the flags. Fixes: ce2cd175469f ("iommu/amd: Enhance amd_iommu_domain_alloc_user()") Reviewed-by: Vasant Hegde Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/7-v2-9776c53c2966+1c7-amd_paging_flags_jgg@nvidia.com Signed-off-by: Joerg Roedel Signed-off-by: Sasha Levin

iommu/amd: Change amd_iommu_pgtable to use enum protection_domain_mode

2025-02-08T09:01:40+00:00

[ Upstream commit 13b4ec749163710e3d188d2fed7405308b1b1e73 ] Currently it uses enum io_pgtable_fmt which is from the io pagetable code and most of the enum values are invalid. protection_domain_mode is internal the driver and has the only two valid values. Fix some signatures and variables to use the right type as well. Reviewed-by: Vasant Hegde Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/5-v2-9776c53c2966+1c7-amd_paging_flags_jgg@nvidia.com Signed-off-by: Joerg Roedel Stable-dep-of: 082f1bcae8d1 ("iommu/amd: Fully decode all combinations of alloc_paging_flags") Signed-off-by: Sasha Levin

iommu/amd: Remove type argument from do_iommu_domain_alloc() and related

2025-02-08T09:01:40+00:00

[ Upstream commit 55b237dd7f7ec2ee9c7986e0fc28c5867bf63282 ] do_iommu_domain_alloc() is only called from amd_iommu_domain_alloc_paging_flags() so type is always IOMMU_DOMAIN_UNMANAGED. Remove type and all the dead conditionals checking it. IOMMU_DOMAIN_IDENTITY checks are similarly obsolete as the conversion to the global static identity domain removed those call paths. The caller of protection_domain_alloc() should set the type, fix the miss in the SVA code. Reviewed-by: Vasant Hegde Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/4-v2-9776c53c2966+1c7-amd_paging_flags_jgg@nvidia.com Signed-off-by: Joerg Roedel Stable-dep-of: 082f1bcae8d1 ("iommu/amd: Fully decode all combinations of alloc_paging_flags") Signed-off-by: Sasha Levin

iommu/amd: Remove dev == NULL checks

2025-02-08T09:01:40+00:00

[ Upstream commit 02bcd1a8b991c6fc29271fa02250bea1b61fb742 ] This is no longer possible, amd_iommu_domain_alloc_paging_flags() is never called with dev = NULL from the core code. Similarly get_amd_iommu_from_dev() can never be NULL either. Reviewed-by: Vasant Hegde Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/3-v2-9776c53c2966+1c7-amd_paging_flags_jgg@nvidia.com Signed-off-by: Joerg Roedel Stable-dep-of: 082f1bcae8d1 ("iommu/amd: Fully decode all combinations of alloc_paging_flags") Signed-off-by: Sasha Levin

iommu/amd: Remove domain_alloc()

2025-02-08T09:01:40+00:00

[ Upstream commit f9b80f941e0e68c3347c5d22a17a0f636a064e2c ] IOMMU drivers should not be sensitive to the domain type, a paging domain should be created based only on the flags passed in, the same for all callers. AMD was using the domain_alloc() path to force VFIO into a v1 domain type, because v1 gives higher performance. However now that IOMMU_HWPT_ALLOC_PASID is present, and a NULL device is not possible, domain_alloc_paging_flags() will do the right thing for VFIO. When invoked from VFIO flags will be 0 and the amd_iommu_pgtable type of domain will be selected. This is v1 by default unless the kernel command line has overridden it to v2. If the admin is forcing v2 assume they know what they are doing so force it everywhere, including for VFIO. Reviewed-by: Vasant Hegde Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/2-v2-9776c53c2966+1c7-amd_paging_flags_jgg@nvidia.com Signed-off-by: Joerg Roedel Stable-dep-of: 082f1bcae8d1 ("iommu/amd: Fully decode all combinations of alloc_paging_flags") Signed-off-by: Sasha Levin

iommu/amd: Remove unused amd_iommu_domain_update()

2025-02-08T09:01:40+00:00

[ Upstream commit 1a684b099fac9a37e6fe2f0e594adbb1eff5181a ] All the callers have been removed by the below commit, remove the implementation and prototypes. Fixes: 322d889ae7d3 ("iommu/amd: Remove amd_iommu_domain_update() from page table freeing") Reviewed-by: Vasant Hegde Signed-off-by: Alejandro Jimenez Signed-off-by: Jason Gunthorpe Link: https://lore.kernel.org/r/1-v2-9776c53c2966+1c7-amd_paging_flags_jgg@nvidia.com Signed-off-by: Joerg Roedel Signed-off-by: Sasha Levin

iommu/amd: Add lockdep asserts for domain->dev_list

2024-12-10T09:12:06+00:00

Add an assertion to all the iteration points that don't obviously have the lock held already. These all take the locker higher in their call chains. Signed-off-by: Jason Gunthorpe Reviewed-by: Vasant Hegde Link: https://lore.kernel.org/r/2-v1-3b9edcf8067d+3975-amd_dev_list_locking_jgg@nvidia.com Signed-off-by: Joerg Roedel

iommu/amd: Put list_add/del(dev_data) back under the domain->lock

2024-12-10T09:12:05+00:00

The list domain->dev_list is protected by the domain->lock spinlock. Any iteration, addition or removal must be under the lock. Move the list_del() up into the critical section. pdom_is_sva_capable(), and destroy_gcr3_table() do not interact with the list element. Wrap the list_add() in a lock, it would make more sense if this was under the same critical section as adjusting the refcounts earlier, but that requires more complications. Fixes: d6b47dec3684 ("iommu/amd: Reduce domain lock scope in attach device path") Signed-off-by: Jason Gunthorpe Reviewed-by: Vasant Hegde Link: https://lore.kernel.org/r/1-v1-3b9edcf8067d+3975-amd_dev_list_locking_jgg@nvidia.com Signed-off-by: Joerg Roedel

iommu: Rename ops->domain_alloc_user() to domain_alloc_paging_flags()

2024-11-22T18:43:45+00:00

Now that the main domain allocating path is calling this function it doesn't make sense to leave it named _user. Change the name to alloc_paging_flags() to mirror the new iommu_paging_domain_alloc_flags() function. A driver should implement only one of ops->domain_alloc_paging() or ops->domain_alloc_paging_flags(). The former is a simpler interface with less boiler plate that the majority of drivers use. The latter is for drivers with a greater feature set (PASID, multiple page table support, advanced iommufd support, nesting, etc). Additional patches will be needed to achieve this. Link: https://patch.msgid.link/r/2-v1-c252ebdeb57b+329-iommu_paging_flags_jgg@nvidia.com Reviewed-by: Lu Baolu Signed-off-by: Jason Gunthorpe