From fe89277c9ceb0d6af0aa665bcf24a41d8b1b79cd Mon Sep 17 00:00:00 2001 From: Guanghui Feng Date: Mon, 16 Mar 2026 15:16:39 +0800 Subject: iommu/vt-d: Fix intel iommu iotlb sync hardlockup and retry During the qi_check_fault process after an IOMMU ITE event, requests at odd-numbered positions in the queue are set to QI_ABORT, only satisfying single-request submissions. However, qi_submit_sync now supports multiple simultaneous submissions, and can't guarantee that the wait_desc will be at an odd-numbered position. Therefore, if an item times out, IOMMU can't re-initiate the request, resulting in an infinite polling wait. This modifies the process by setting the status of all requests already fetched by IOMMU and recorded as QI_IN_USE status (including wait_desc requests) to QI_ABORT, thus enabling multiple requests to be resubmitted. Fixes: 8a1d82462540 ("iommu/vt-d: Multiple descriptors per qi_submit_sync()") Cc: stable@vger.kernel.org Signed-off-by: Guanghui Feng Tested-by: Shuai Xue Reviewed-by: Shuai Xue Reviewed-by: Samiullah Khawaja Link: https://lore.kernel.org/r/20260306101516.3885775-1-guanghuifeng@linux.alibaba.com Signed-off-by: Lu Baolu Fixes: 8a1d82462540 ("iommu/vt-d: Multiple descriptors per qi_submit_sync()") Signed-off-by: Joerg Roedel --- drivers/iommu/intel/dmar.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c index d68c06025cac..69222dbd2af0 100644 --- a/drivers/iommu/intel/dmar.c +++ b/drivers/iommu/intel/dmar.c @@ -1314,7 +1314,6 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index) if (fault & DMA_FSTS_ITE) { head = readl(iommu->reg + DMAR_IQH_REG); head = ((head >> shift) - 1 + QI_LENGTH) % QI_LENGTH; - head |= 1; tail = readl(iommu->reg + DMAR_IQT_REG); tail = ((tail >> shift) - 1 + QI_LENGTH) % QI_LENGTH; @@ -1331,7 +1330,7 @@ static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index) do { if (qi->desc_status[head] == QI_IN_USE) qi->desc_status[head] = QI_ABORT; - head = (head - 2 + QI_LENGTH) % QI_LENGTH; + head = (head - 1 + QI_LENGTH) % QI_LENGTH; } while (head != tail); /* -- cgit v1.2.3 From 39c20c4e83b9f78988541d829aa34668904e54a0 Mon Sep 17 00:00:00 2001 From: Lu Baolu Date: Mon, 16 Mar 2026 15:16:40 +0800 Subject: iommu/vt-d: Only handle IOPF for SVA when PRI is supported In intel_svm_set_dev_pasid(), the driver unconditionally manages the IOPF handling during a domain transition. However, commit a86fb7717320 ("iommu/vt-d: Allow SVA with device-specific IOPF") introduced support for SVA on devices that handle page faults internally without utilizing the PCI PRI. On such devices, the IOMMU-side IOPF infrastructure is not required. Calling iopf_for_domain_replace() on these devices is incorrect and can lead to unexpected failures during PASID attachment or unwinding. Add a check for info->pri_supported to ensure that the IOPF queue logic is only invoked for devices that actually rely on the IOMMU's PRI-based fault handling. Fixes: 17fce9d2336d ("iommu/vt-d: Put iopf enablement in domain attach path") Cc: stable@vger.kernel.org Suggested-by: Kevin Tian Reviewed-by: Kevin Tian Signed-off-by: Lu Baolu Link: https://lore.kernel.org/r/20260310075520.295104-1-baolu.lu@linux.intel.com Signed-off-by: Joerg Roedel --- drivers/iommu/intel/svm.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/iommu/intel/svm.c b/drivers/iommu/intel/svm.c index fea10acd4f02..57cd1db7207a 100644 --- a/drivers/iommu/intel/svm.c +++ b/drivers/iommu/intel/svm.c @@ -164,9 +164,12 @@ static int intel_svm_set_dev_pasid(struct iommu_domain *domain, if (IS_ERR(dev_pasid)) return PTR_ERR(dev_pasid); - ret = iopf_for_domain_replace(domain, old, dev); - if (ret) - goto out_remove_dev_pasid; + /* SVA with non-IOMMU/PRI IOPF handling is allowed. */ + if (info->pri_supported) { + ret = iopf_for_domain_replace(domain, old, dev); + if (ret) + goto out_remove_dev_pasid; + } /* Setup the pasid table: */ sflags = cpu_feature_enabled(X86_FEATURE_LA57) ? PASID_FLAG_FL5LP : 0; @@ -181,7 +184,8 @@ static int intel_svm_set_dev_pasid(struct iommu_domain *domain, return 0; out_unwind_iopf: - iopf_for_domain_replace(old, domain, dev); + if (info->pri_supported) + iopf_for_domain_replace(old, domain, dev); out_remove_dev_pasid: domain_remove_dev_pasid(domain, dev, pasid); return ret; -- cgit v1.2.3 From 0a4d00e2e99a39a5698e4b63c394415dcbb39d90 Mon Sep 17 00:00:00 2001 From: Antheas Kapenekakis Date: Fri, 27 Feb 2026 09:06:37 +0100 Subject: iommu: Fix mapping check for 0x0 to avoid re-mapping it Commit 789a5913b29c ("iommu/amd: Use the generic iommu page table") introduces the shared iommu page table for AMD IOMMU. Some bioses contain an identity mapping for address 0x0, which is not parsed properly (e.g., certain Strix Halo devices). This causes the DMA components of the device to fail to initialize (e.g., the NVMe SSD controller), leading to a failed post. Specifically, on the GPD Win 5, the NVME and SSD GPU fail to mount, making collecting errors difficult. While debugging, it was found that a -EADDRINUSE error was emitted and its source was traced to iommu_iova_to_phys(). After adding some debug prints, it was found that phys_addr becomes 0, which causes the code to try to re-map the 0 address and fail, causing a cascade leading to a failed post. This is because the GPD Win 5 contains a 0x0-0x1 identity mapping for DMA devices, causing it to be repeated for each device. The cause of this failure is the following check in iommu_create_device_direct_mappings(), where address aliasing is handled via the following check: ``` phys_addr = iommu_iova_to_phys(domain, addr); if (!phys_addr) { map_size += pg_size; continue; } ```` Obviously, the iommu_iova_to_phys() signature is faulty and aliases unmapped and 0 together, causing the allocation code to try to re-allocate the 0 address per device. However, it has too many instantiations to fix. Therefore, use a ternary so that when addr is 0, the check is done for address 1 instead. Suggested-by: Robin Murphy Fixes: 789a5913b29c ("iommu/amd: Use the generic iommu page table") Signed-off-by: Antheas Kapenekakis Reviewed-by: Vasant Hegde Reviewed-by: Jason Gunthorpe Signed-off-by: Joerg Roedel --- drivers/iommu/iommu.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 35db51780954..50718ab810a4 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -1213,7 +1213,11 @@ static int iommu_create_device_direct_mappings(struct iommu_domain *domain, if (addr == end) goto map_end; - phys_addr = iommu_iova_to_phys(domain, addr); + /* + * Return address by iommu_iova_to_phys for 0 is + * ambiguous. Offset to address 1 if addr is 0. + */ + phys_addr = iommu_iova_to_phys(domain, addr ? addr : 1); if (!phys_addr) { map_size += pg_size; continue; -- cgit v1.2.3 From 45c6a2dc7ec8339052666b06065c521a10cc29bb Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 1 Mar 2026 16:52:14 -0800 Subject: iommu/io-pgtable: fix all kernel-doc warnings in io-pgtable.h Avoid kernel-doc warnings in io-pgtable.h: - use the correct struct member names or kernel-doc format - add a missing struct member description - add a missing function return comment section Warning: include/linux/io-pgtable.h:187 struct member 'coherent_walk' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:187 struct member 'arm_lpae_s1_cfg' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:187 struct member 'arm_lpae_s2_cfg' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:187 struct member 'arm_v7s_cfg' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:187 struct member 'arm_mali_lpae_cfg' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:187 struct member 'apple_dart_cfg' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:187 struct member 'amd' not described in 'io_pgtable_cfg' Warning: include/linux/io-pgtable.h:223 struct member 'read_and_clear_dirty' not described in 'io_pgtable_ops' Warning: include/linux/io-pgtable.h:237 No description found for return value of 'alloc_io_pgtable_ops' Signed-off-by: Randy Dunlap Reviewed-by: Jason Gunthorpe Signed-off-by: Joerg Roedel --- include/linux/io-pgtable.h | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/include/linux/io-pgtable.h b/include/linux/io-pgtable.h index 7a1516011ccf..e19872e37e06 100644 --- a/include/linux/io-pgtable.h +++ b/include/linux/io-pgtable.h @@ -53,7 +53,7 @@ struct iommu_flush_ops { * tables. * @ias: Input address (iova) size, in bits. * @oas: Output address (paddr) size, in bits. - * @coherent_walk A flag to indicate whether or not page table walks made + * @coherent_walk: A flag to indicate whether or not page table walks made * by the IOMMU are coherent with the CPU caches. * @tlb: TLB management callbacks for this set of tables. * @iommu_dev: The device representing the DMA configuration for the @@ -136,6 +136,7 @@ struct io_pgtable_cfg { void (*free)(void *cookie, void *pages, size_t size); /* Low-level data specific to the table format */ + /* private: */ union { struct { u64 ttbr; @@ -203,6 +204,9 @@ struct arm_lpae_io_pgtable_walk_data { * @unmap_pages: Unmap a range of virtually contiguous pages of the same size. * @iova_to_phys: Translate iova to physical address. * @pgtable_walk: (optional) Perform a page table walk for a given iova. + * @read_and_clear_dirty: Record dirty info per IOVA. If an IOVA is dirty, + * clear its dirty state from the PTE unless the + * IOMMU_DIRTY_NO_CLEAR flag is passed in. * * These functions map directly onto the iommu_ops member functions with * the same names. @@ -231,7 +235,9 @@ struct io_pgtable_ops { * the configuration actually provided by the allocator (e.g. the * pgsize_bitmap may be restricted). * @cookie: An opaque token provided by the IOMMU driver and passed back to - * the callback routines in cfg->tlb. + * the callback routines. + * + * Returns: Pointer to the &struct io_pgtable_ops for this set of page tables. */ struct io_pgtable_ops *alloc_io_pgtable_ops(enum io_pgtable_fmt fmt, struct io_pgtable_cfg *cfg, -- cgit v1.2.3 From 06e14c36e20b48171df13d51b89fe67c594ed07a Mon Sep 17 00:00:00 2001 From: Lizhi Hou Date: Wed, 4 Mar 2026 22:18:42 -0800 Subject: iommu/sva: Fix crash in iommu_sva_unbind_device() domain->mm->iommu_mm can be freed by iommu_domain_free(): iommu_domain_free() mmdrop() __mmdrop() mm_pasid_drop() After iommu_domain_free() returns, accessing domain->mm->iommu_mm may dereference a freed mm structure, leading to a crash. Fix this by moving the code that accesses domain->mm->iommu_mm to before the call to iommu_domain_free(). Fixes: e37d5a2d60a3 ("iommu/sva: invalidate stale IOTLB entries for kernel address space") Signed-off-by: Lizhi Hou Reviewed-by: Jason Gunthorpe Reviewed-by: Yi Liu Reviewed-by: Vasant Hegde Reviewed-by: Lu Baolu Signed-off-by: Joerg Roedel --- drivers/iommu/iommu-sva.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/iommu/iommu-sva.c b/drivers/iommu/iommu-sva.c index 07d64908a05f..bc7c7232a43e 100644 --- a/drivers/iommu/iommu-sva.c +++ b/drivers/iommu/iommu-sva.c @@ -182,13 +182,13 @@ void iommu_sva_unbind_device(struct iommu_sva *handle) iommu_detach_device_pasid(domain, dev, iommu_mm->pasid); if (--domain->users == 0) { list_del(&domain->next); - iommu_domain_free(domain); - } + if (list_empty(&iommu_mm->sva_domains)) { + list_del(&iommu_mm->mm_list_elm); + if (list_empty(&iommu_sva_mms)) + iommu_sva_present = false; + } - if (list_empty(&iommu_mm->sva_domains)) { - list_del(&iommu_mm->mm_list_elm); - if (list_empty(&iommu_sva_mms)) - iommu_sva_present = false; + iommu_domain_free(domain); } mutex_unlock(&iommu_sva_lock); -- cgit v1.2.3 From ba17de98545d07285d15ce4fe2afe98283338fb0 Mon Sep 17 00:00:00 2001 From: Joe Damato Date: Mon, 9 Mar 2026 16:52:33 -0700 Subject: iommu/amd: Block identity domain when SNP enabled Previously, commit 8388f7df936b ("iommu/amd: Do not support IOMMU_DOMAIN_IDENTITY after SNP is enabled") prevented users from changing the IOMMU domain to identity if SNP was enabled. This resulted in an error when writing to sysfs: # echo "identity" > /sys/kernel/iommu_groups/50/type -bash: echo: write error: Cannot allocate memory However, commit 4402f2627d30 ("iommu/amd: Implement global identity domain") changed the flow of the code, skipping the SNP guard and allowing users to change the IOMMU domain to identity after a machine has booted. Once the user does that, they will probably try to bind and the device/driver will start to do DMA which will trigger errors: iommu ivhd3: AMD-Vi: Event logged [ILLEGAL_DEV_TABLE_ENTRY device=0000:43:00.0 pasid=0x00000 address=0x3737b01000 flags=0x0020] iommu ivhd3: AMD-Vi: Control Reg : 0xc22000142148d AMD-Vi: DTE[0]: 6000000000000003 AMD-Vi: DTE[1]: 0000000000000001 AMD-Vi: DTE[2]: 2000003088b3e013 AMD-Vi: DTE[3]: 0000000000000000 bnxt_en 0000:43:00.0 (unnamed net_device) (uninitialized): Error (timeout: 500015) msg {0x0 0x0} len:0 iommu ivhd3: AMD-Vi: Event logged [ILLEGAL_DEV_TABLE_ENTRY device=0000:43:00.0 pasid=0x00000 address=0x3737b01000 flags=0x0020] iommu ivhd3: AMD-Vi: Control Reg : 0xc22000142148d AMD-Vi: DTE[0]: 6000000000000003 AMD-Vi: DTE[1]: 0000000000000001 AMD-Vi: DTE[2]: 2000003088b3e013 AMD-Vi: DTE[3]: 0000000000000000 bnxt_en 0000:43:00.0: probe with driver bnxt_en failed with error -16 To prevent this from happening, create an attach wrapper for identity_domain_ops which returns EINVAL if amd_iommu_snp_en is true. With this commit applied: # echo "identity" > /sys/kernel/iommu_groups/62/type -bash: echo: write error: Invalid argument Fixes: 4402f2627d30 ("iommu/amd: Implement global identity domain") Signed-off-by: Joe Damato Reviewed-by: Vasant Hegde Reviewed-by: Jason Gunthorpe Signed-off-by: Joerg Roedel --- drivers/iommu/amd/iommu.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index 81c4d7733872..760d5f4623b5 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -2909,8 +2909,21 @@ static struct iommu_domain blocked_domain = { static struct protection_domain identity_domain; +static int amd_iommu_identity_attach(struct iommu_domain *dom, struct device *dev, + struct iommu_domain *old) +{ + /* + * Don't allow attaching a device to the identity domain if SNP is + * enabled. + */ + if (amd_iommu_snp_en) + return -EINVAL; + + return amd_iommu_attach_device(dom, dev, old); +} + static const struct iommu_domain_ops identity_domain_ops = { - .attach_dev = amd_iommu_attach_device, + .attach_dev = amd_iommu_identity_attach, }; void amd_iommu_init_identity_domain(void) -- cgit v1.2.3