Linaro-mm-sig April 2026

linaro-mm-sig@lists.linaro.org

91 participants
216 discussions

[PATCH] drm/amdgpu: fix recursive ww_mutex acquire in amdgpu_devcoredump_format

by Mikhail Gavrilov

When dumping IB contents from a hung job, amdgpu_devcoredump_format() acquires the VM root PD's reservation lock via amdgpu_vm_lock_by_pasid() and then, for each IB referenced by the job, calls amdgpu_bo_reserve() on the BO that backs the IB. Both reservations are taken on reservation_ww_class_mutex objects but neither uses a ww_acquire_ctx, which trips lockdep: WARNING: possible recursive locking detected -------------------------------------------- kworker/u128:0 is trying to acquire lock: ffff88838b16e1f0 (reservation_ww_class_mutex){+.+.}-{4:4}, at: amdgpu_devcoredump_format+0x1594/0x23f0 [amdgpu] but task is already holding lock: ffff8882f82681f0 (reservation_ww_class_mutex){+.+.}-{4:4}, at: amdgpu_devcoredump_format+0x1594/0x23f0 [amdgpu] Possible unsafe locking scenario: CPU0 ---- lock(reservation_ww_class_mutex); lock(reservation_ww_class_mutex); *** DEADLOCK *** May be due to missing lock nesting notation Workqueue: events_unbound amdgpu_devcoredump_deferred_work [amdgpu] Call Trace: __ww_mutex_lock.constprop.0 ww_mutex_lock amdgpu_bo_reserve amdgpu_devcoredump_format+0x1594 [amdgpu] amdgpu_devcoredump_deferred_work+0xea [amdgpu] process_one_work worker_thread kthread The two reservations are on different BOs in the captured trace, so the splat is a lockdep-correctness warning, not an observed deadlock. It becomes a real self-deadlock whenever the IB BO shares its dma_resv with the root PD (the always-valid case, see amdgpu_vm_is_bo_always_valid()): amdgpu_bo_reserve(abo) re-acquires the same ww_mutex without a ticket and blocks forever. With amdgpu.gpu_recovery=0 the timeout handler refires every ~2 s and each invocation produces this splat, drowning the kernel ring buffer. Fix it by collecting the per-IB BO references under the root PD's reservation, then releasing the root before reserving each IB BO individually. The walk over the VM mapping tree must remain under the root lock (mappings can be torn down without it), but the actual content copies do not need to nest inside it. Each per-IB reservation is now an independent top-level acquire, eliminating the nested ww_mutex. The collect/release logic is factored out into two small helpers (amdgpu_devcoredump_collect_ib_refs / amdgpu_devcoredump_release_ib_refs) to keep the main function's indentation reasonable. This also fixes a BO refcount leak in the original code: when amdgpu_bo_reserve() failed, control jumped to free_ib_content without running amdgpu_bo_unref(). In the new structure the per-IB BO refs are released unconditionally in the cleanup helper. Reproducer (~150 LoC libdrm_amdgpu): submit a single GFX IB containing PACKET3_INDIRECT_BUFFER chained at GPU VA 0 and wait for the fence. The TDR fires within ~10 s and the deferred coredump worker produces the splat above on every invocation. Fixes: 7b15fc2d1f1a ("drm/amdgpu: dump job ibs in the devcoredump") Cc: stable(a)vger.kernel.org # 7.1 Signed-off-by: Mikhail Gavrilov <mikhail.v.gavrilov(a)gmail.com> --- .../gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c | 147 +++++++++++++----- 1 file changed, 110 insertions(+), 37 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c index d386bc775d03..f6bb968de756 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dev_coredump.c @@ -207,6 +207,72 @@ static void amdgpu_devcoredump_fw_info(struct amdgpu_device *adev, } } +struct amdgpu_devcoredump_ib_ref { + struct amdgpu_bo *bo; + u64 offset; +}; + +/* + * Walk the VM's mapping tree under the root PD's reservation to obtain the BO + * that backs each IB and pin it with a refcount. The root PD reservation is + * dropped before this function returns; the caller can then reserve each IB + * BO individually without nesting ww_mutex acquires on + * reservation_ww_class_mutex. + * + * Returns an array of num_ibs entries (each ib_refs[i].bo may be NULL if its + * mapping was not found), or NULL on allocation failure / VM lookup failure. + * The caller must release the BO refs and free the array. + */ +static struct amdgpu_devcoredump_ib_ref * +amdgpu_devcoredump_collect_ib_refs(struct amdgpu_device *adev, + struct amdgpu_coredump_info *coredump) +{ + struct amdgpu_devcoredump_ib_ref *ib_refs; + struct amdgpu_bo_va_mapping *mapping; + struct amdgpu_bo *root; + struct amdgpu_vm *vm; + u64 va_start; + + ib_refs = kcalloc(coredump->num_ibs, sizeof(*ib_refs), GFP_KERNEL); + if (!ib_refs) + return NULL; + + vm = amdgpu_vm_lock_by_pasid(adev, &root, coredump->pasid); + if (!vm) { + kfree(ib_refs); + return NULL; + } + + for (int i = 0; i < coredump->num_ibs; i++) { + va_start = coredump->ibs[i].gpu_addr & AMDGPU_GMC_HOLE_MASK; + mapping = amdgpu_vm_bo_lookup_mapping(vm, va_start / AMDGPU_GPU_PAGE_SIZE); + if (!mapping) + continue; + + ib_refs[i].bo = amdgpu_bo_ref(mapping->bo_va->base.bo); + ib_refs[i].offset = va_start - + mapping->start * AMDGPU_GPU_PAGE_SIZE; + } + + amdgpu_bo_unreserve(root); + amdgpu_bo_unref(&root); + + return ib_refs; +} + +static void +amdgpu_devcoredump_release_ib_refs(struct amdgpu_devcoredump_ib_ref *ib_refs, + int num_ibs) +{ + if (!ib_refs) + return; + + for (int i = 0; i < num_ibs; i++) + if (ib_refs[i].bo) + amdgpu_bo_unref(&ib_refs[i].bo); + kfree(ib_refs); +} + static ssize_t amdgpu_devcoredump_format(char *buffer, size_t count, struct amdgpu_coredump_info *coredump) { @@ -214,13 +280,11 @@ amdgpu_devcoredump_format(char *buffer, size_t count, struct amdgpu_coredump_inf struct drm_printer p; struct drm_print_iterator iter; struct amdgpu_vm_fault_info *fault_info; - struct amdgpu_bo_va_mapping *mapping; struct amdgpu_ip_block *ip_block; struct amdgpu_res_cursor cursor; - struct amdgpu_bo *abo, *root; - uint64_t va_start, offset; + struct amdgpu_bo *abo; + uint64_t offset; struct amdgpu_ring *ring; - struct amdgpu_vm *vm; u32 *ib_content; uint8_t *kptr; int ver, i, j, r; @@ -343,43 +407,52 @@ amdgpu_devcoredump_format(char *buffer, size_t count, struct amdgpu_coredump_inf drm_printf(&p, "VRAM is lost due to GPU reset!\n"); if (coredump->num_ibs) { - /* Don't try to lookup the VM or map the BOs when calculating the - * size required to store the devcoredump. + struct amdgpu_devcoredump_ib_ref *ib_refs = NULL; + + /* + * Snapshot per-IB BO references under the root PD's reservation, + * then release the root before reserving each IB BO individually + * to copy its contents. + * + * Reserving an IB BO while the root PD is still reserved would + * be a nested ww_mutex acquire on reservation_ww_class_mutex + * without a ww_acquire_ctx, which trips lockdep's recursive- + * locking check and self-deadlocks for IB BOs that share their + * dma_resv with the root PD (always-valid BOs). + * + * Skip lookup/reservation entirely on the sizing pass: it does + * not write IB content, and the size estimate doesn't depend on + * whether the BOs are reachable. */ - if (sizing_pass) - vm = NULL; - else - vm = amdgpu_vm_lock_by_pasid(adev, &root, coredump->pasid); + if (!sizing_pass) + ib_refs = amdgpu_devcoredump_collect_ib_refs(adev, coredump); - for (int i = 0; i < coredump->num_ibs && (sizing_pass || vm); i++) { + for (int i = 0; i < coredump->num_ibs; i++) { ib_content = kvmalloc_array(coredump->ibs[i].ib_size_dw, 4, GFP_KERNEL); if (!ib_content) continue; - /* vm=NULL can only happen when 'sizing_pass' is true. Skip to the - * drm_printf() calls (ib_content doesn't need to be initialized - * as its content won't be written anywhere). - */ - if (!vm) + if (sizing_pass) goto output_ib_content; - va_start = coredump->ibs[i].gpu_addr & AMDGPU_GMC_HOLE_MASK; - mapping = amdgpu_vm_bo_lookup_mapping(vm, va_start / AMDGPU_GPU_PAGE_SIZE); - if (!mapping) - goto free_ib_content; + if (!ib_refs || !ib_refs[i].bo) + goto output_ib_content; + + abo = ib_refs[i].bo; + offset = ib_refs[i].offset; - offset = va_start - (mapping->start * AMDGPU_GPU_PAGE_SIZE); - abo = amdgpu_bo_ref(mapping->bo_va->base.bo); r = amdgpu_bo_reserve(abo, false); if (r) - goto free_ib_content; + goto output_ib_content; if (abo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS) { off = 0; - if (abo->tbo.resource->mem_type != TTM_PL_VRAM) - goto unreserve_abo; + if (abo->tbo.resource->mem_type != TTM_PL_VRAM) { + amdgpu_bo_unreserve(abo); + goto output_ib_content; + } amdgpu_res_first(abo->tbo.resource, offset, coredump->ibs[i].ib_size_dw * 4, @@ -395,8 +468,10 @@ amdgpu_devcoredump_format(char *buffer, size_t count, struct amdgpu_coredump_inf r = ttm_bo_kmap(&abo->tbo, 0, PFN_UP(abo->tbo.base.size), &abo->kmap); - if (r) - goto unreserve_abo; + if (r) { + amdgpu_bo_unreserve(abo); + goto output_ib_content; + } kptr = amdgpu_bo_kptr(abo); kptr += offset; @@ -406,21 +481,19 @@ amdgpu_devcoredump_format(char *buffer, size_t count, struct amdgpu_coredump_inf amdgpu_bo_kunmap(abo); } + amdgpu_bo_unreserve(abo); + output_ib_content: drm_printf(&p, "\nIB #%d 0x%llx %d dw\n", i, coredump->ibs[i].gpu_addr, coredump->ibs[i].ib_size_dw); - for (int j = 0; j < coredump->ibs[i].ib_size_dw; j++) - drm_printf(&p, "0x%08x\n", ib_content[j]); -unreserve_abo: - if (vm) - amdgpu_bo_unreserve(abo); -free_ib_content: + if (!sizing_pass && ib_refs && ib_refs[i].bo) { + for (int j = 0; j < coredump->ibs[i].ib_size_dw; j++) + drm_printf(&p, "0x%08x\n", ib_content[j]); + } kvfree(ib_content); } - if (vm) { - amdgpu_bo_unreserve(root); - amdgpu_bo_unref(&root); - } + + amdgpu_devcoredump_release_ib_refs(ib_refs, coredump->num_ibs); } return count - iter.remain; -- 2.54.0

1 month, 1 week

[PATCH v3 00/10] Add dmabuf read/write via io_uring

by Pavel Begunkov

The patch set allows to register a dmabuf to an io_uring instance for a specified file and use it with io_uring read / write requests. The infrastructure is not tied to io_uring and there could be more users in the future. A similar idea was attempted some years ago by Keith [1], from where I borrowed a good number of changes, and later was brough up by Tushar and Vishal from Intel. It's an opt-in feature for files, and they need to implement a new file operation to use it. Only NVMe block devices are supported in this series. The user API is built on top of io_uring's "registered buffers", where a dmabuf is registered in a special way, but after it can be used as any other "registered buffer" with IORING_OP_{READ,WRITE}_FIXED requests. It's created via a new file operation and the resulted map is then passed through the I/O stack in a new iterator type. There is some additional infrastructure to bind it all, which also counts requests using a dmabuf map and managing lifetimes, which is used to implement map invalidation. It was tested for GPU <-> NVMe transfers. Also, as it maintains a long-term dma mapping, it helps with the IOMMU cost. The numbers below are for udmabuf reads previously run by Anuj for different IOMMU modes: - STRICT: before = 570 KIOPS, after = 5.01 MIOPS - LAZY: before = 1.93 MIOPS, after = 5.01 MIOPS - PASSTHROUGH: before = 5.01 MIOPS, after = 5.01 MIOPS There are some liburing tests that can serve as an example: git: https://github.com/isilence/liburing.git rw-dmabuf-tests-v3 url: https://github.com/isilence/liburing/tree/rw-dmabuf-tests-v3 [1] https://lore.kernel.org/io-uring/20220805162444.3985535-1-kbusch@fb.com/ v3: - Rework io_uring registration - Move token/map infrastructure code out of blk-mq - Simplify callbacks: remove a separate blk-mq table, which was mostly just forwarding calls (to nvme). - Don't skip dma sync depending on request direction - Fix a couple of hangs - Rename s/dma/dmabuf/ - Other small changes v2: - Don't pass raw dma addresses, wrap it into a driver specific object - Split into two objects: token and map - Implement move_notify Pavel Begunkov (10): file: add callback for creating long-term dmabuf maps iov_iter: add iterator type for dmabuf maps block: move bvec init into __bio_clone block: introduce dma map backed bio type lib: add dmabuf token infrastructure block: forward create_dmabuf_token to drivers nvme-pci: implement dma_token backed requests io_uring/rsrc: introduce buf registration structure io_uring/rsrc: extend buffer update io_uring/rsrc: add dmabuf backed registered buffers block/bio.c | 28 +++- block/blk-merge.c | 14 ++ block/blk.h | 3 +- block/fops.c | 16 ++ drivers/nvme/host/pci.c | 282 ++++++++++++++++++++++++++++++++ include/linux/bio.h | 19 ++- include/linux/blk-mq.h | 9 + include/linux/blk_types.h | 8 +- include/linux/fs.h | 2 + include/linux/io_dmabuf_token.h | 92 +++++++++++ include/linux/io_uring_types.h | 5 + include/linux/uio.h | 11 ++ include/uapi/linux/io_uring.h | 31 +++- io_uring/io_uring.c | 3 +- io_uring/rsrc.c | 266 +++++++++++++++++++++++++----- io_uring/rsrc.h | 30 +++- io_uring/rw.c | 4 +- lib/Kconfig | 4 + lib/Makefile | 2 + lib/io_dmabuf_token.c | 272 ++++++++++++++++++++++++++++++ lib/iov_iter.c | 29 +++- 21 files changed, 1071 insertions(+), 59 deletions(-) create mode 100644 include/linux/io_dmabuf_token.h create mode 100644 lib/io_dmabuf_token.c -- 2.53.0

1 month, 2 weeks

[PATCH v5 0/2] dma-buf: heaps: system: add an option to allocate explicitly shared/decrypted memory

by Jiri Pirko

From: Jiri Pirko <jiri(a)nvidia.com> Confidential computing (CoCo) VMs/guests, such as AMD SEV and Intel TDX, run with private/encrypted memory which creates a challenge for devices that do not support DMA to it (no TDISP support). For kernel-only DMA operations, swiotlb bounce buffering provides a transparent solution by copying data through shared memory. However, the only way to get this memory into userspace is via the DMA API's dma_alloc_pages()/dma_mmap_pages() type interfaces which limits the use of the memory to a single DMA device, and is incompatible with pin_user_pages(). These limitations are particularly problematic for the RDMA subsystem which makes heavy use of pin_user_pages() and expects flexible memory usage between many different DMA devices. This patch series enables userspace to explicitly request shared (decrypted) memory allocations from new dma-buf system_cc_shared heap. Userspace can mmap this memory and pass the dma-buf fd to other existing importers such as RDMA or DRM devices to access the memory. The DMA API is improved to allow the dma heap exporter to DMA map the shared memory to each importing device. Based on dma-mapping-for-next e7442a68cd1ee797b585f045d348781e9c0dde0d Jiri Pirko (2): dma-mapping: introduce DMA_ATTR_CC_SHARED for shared memory dma-buf: heaps: system: add system_cc_shared heap for explicitly shared memory drivers/dma-buf/heaps/system_heap.c | 103 ++++++++++++++++++++++++++-- include/linux/dma-mapping.h | 10 +++ include/trace/events/dma.h | 3 +- kernel/dma/direct.h | 14 +++- kernel/dma/mapping.c | 13 +++- 5 files changed, 132 insertions(+), 11 deletions(-) -- 2.51.1

1 month, 2 weeks

[PATCH v5 0/4] dma-buf: heaps: Turn heaps into modules

by Maxime Ripard

Hi, The recent introduction of heaps in the optee driver [1] made possible the creation of heaps as modules. It's generally a good idea if possible, including for the already existing system and CMA heaps. The system one is pretty trivial, the CMA is now easy too with the reworks we got in 7.1-r1. Let me know what you think, Maxime 1: https://lore.kernel.org/dri-devel/20250911135007.1275833-4-jens.wiklander@l… Signed-off-by: Maxime Ripard <mripard(a)kernel.org> --- Changes in v5: - Rebase on 7.1-rc1 - Add a patch to enable the heaps in arm64 defconfig - Link to v4: https://lore.kernel.org/r/20260331-dma-buf-heaps-as-modules-v4-0-e18fda5044… Changes in v4: - Fix compilation failure - Rework to take into account OF_RESERVED_MEM - Fix regression making the default CMA area disappear if not created through the DT - Added some documentation and comments - Link to v3: https://lore.kernel.org/r/20260303-dma-buf-heaps-as-modules-v3-0-24344812c7… Changes in v3: - Squashed cma_get_name and cma_alloc/release patches - Fixed typo in Export dev_get_cma_area commit title - Fixed compilation failure with DMA_CMA but not OF_RESERVED_MEM - Link to v2: https://lore.kernel.org/r/20260227-dma-buf-heaps-as-modules-v2-0-454aee7e06… Changes in v2: - Collect tags - Don't export dma_contiguous_default_area anymore, but export dev_get_cma_area instead - Mentioned that heap modules can't be removed - Link to v1: https://lore.kernel.org/r/20260225-dma-buf-heaps-as-modules-v1-0-2109225a09… --- Maxime Ripard (4): dma-buf: heaps: Export mem_accounting parameter dma-buf: heaps: cma: Turn the heap into a module dma-buf: heaps: system: Turn the heap into a module arm64: defconfig: Enable dma-buf heaps arch/arm64/configs/defconfig | 3 +++ drivers/dma-buf/dma-heap.c | 1 + drivers/dma-buf/heaps/Kconfig | 4 ++-- drivers/dma-buf/heaps/cma_heap.c | 3 +++ drivers/dma-buf/heaps/system_heap.c | 5 +++++ 5 files changed, 14 insertions(+), 2 deletions(-) --- base-commit: 5e9b7d093f3f77cb0af4409559e3d139babfb443 change-id: 20260225-dma-buf-heaps-as-modules-1034b3ec9f2a Best regards, -- Maxime Ripard <mripard(a)kernel.org>

1 month, 2 weeks

[PATCH 0/9] vfio/pci: Add mmap() for DMABUFs

by Matt Evans

Hi all, This series is based on previous RFCs/discussions: Tech topic: https://lore.kernel.org/linux-iommu/20250918214425.2677057-1-amastro@fb.com/ RFCv1: https://lore.kernel.org/all/20260226202211.929005-1-mattev@meta.com/ RFCv2: https://lore.kernel.org/kvm/20260312184613.3710705-1-mattev@meta.com/ The background/rationale is covered in more detail in the RFC cover letters. The TL;DR is: The goal is to enable userspace driver designs that use VFIO to export DMABUFs representing subsets of PCI device BARs, and "vend" those buffers from a primary process to other subordinate processes by fd. These processes then mmap() the buffers and their access to the device is isolated to the exported ranges. This is an improvement on sharing the VFIO device fd to subordinate processes, which would allow unfettered access . This is achieved by enabling mmap() of vfio-pci DMABUFs. Second, a new ioctl()-based revocation mechanism is added to allow the primary process to forcibly revoke access to previously-shared BAR spans, even if the subordinate processes haven't cleanly exited. (The related topic of safe delegation of iommufd control to the subordinate processes is not addressed here, and is follow-up work.) As well as isolation and revocation, another advantage to accessing a BAR through a VMA backed by a DMABUF is that it's straightforward to create the buffer with access attributes, such as write-combining. Notes on patches ================ Feedback from the RFCs requested that, instead of creating DMABUF-specific vm_ops and .fault paths, to go the whole way and migrate the existing VFIO PCI BAR mmap() to be backed by a DMABUF too, resulting in a common vm_ops and fault handler for mmap()s of both the VFIO device and explicitly-exported DMABUFs. This has been done for vfio-pci, but not sub-drivers (nvgrace-gpu's special-case mappings are unchanged). vfio/pci: Fix vfio_pci_dma_buf_cleanup() double-put A bug fix to a related are, whose context is a depdency for later patches. vfio/pci: Add a helper to look up PFNs for DMABUFs vfio/pci: Add a helper to create a DMABUF for a BAR-map VMA The first is for a DMABUF VMA fault handler to determine arbitrary-sized PFNs from ranges in DMABUF. Secondly, refactor DMABUF export for use by the existing export feature and a new helper that creates a DMABUF corresponding to a VFIO BAR mmap() request. vfio/pci: Convert BAR mmap() to use a DMABUF The vfio-pci core mmap() creates a DMABUF with the helper, and the vm_ops fault handler uses the other helper to resolve the fault. Because this depends on DMABUF structs/code, CONFIG_VFIO_PCI_CORE needs to depend on CONFIG_DMA_SHARED_BUFFER. The CONFIG_VFIO_PCI_DMABUF still conditionally enables the export support code. NOTE: The user mmap()s a device fd, but the resulting VMA's vm_file becomes that of the DMABUF which takes ownership of the device and puts it on release. This maintains the existing behaviour of a VMA keeping the VFIO device open. BAR zapping then happens via the existing vfio_pci_dma_buf_move() path, which now needs to unmap PTEs in the DMABUF's address_space. vfio/pci: Provide a user-facing name for BAR mappings There was a request for decent debug naming in /proc/<pid>/maps etc. comparable to the existing VFIO names: since the VMAs are DMABUFs, they have a "dmabuf:" prefix and can't be 100% identical to before. This is a user-visible change, but this patch at least now gives us extra info on the BDF & BAR being mapped. vfio/pci: Clean up BAR zap and revocation In general (see NOTE!) the vfio_pci_zap_bars() is now obsolete, since it unmaps PTEs in the VFIO device address_space which is now unused. This consolidates all calls (e.g. around reset) with the neighbouring vfio_pci_dma_buf_move()s into new functions, to revoke-zap/unrevoke. NOTE: the nvgrace-gpu driver continues to use its own private vm_ops, fault handler, etc. for its special memregions, and these DO still add PTEs to the VFIO device address_space. So, a temporary flag, vdev->bar_needs_zap, maintains the old behaviour for this use. At least this patch's consolidation makes it easy to remove the remaining zap when this need goes away. A FIXME is added: if nvgrace-gpu is converted to DMABUFs, remove the flag and final zap. vfio/pci: Support mmap() of a VFIO DMABUF Adds mmap() for a DMABUF fd exported from vfio-pci. It was a goal to keep the VFIO device fd lifetime behaviour unchanged with respect to the DMABUFs. An application can close all device fds, and this will revoke/clean up all DMABUFs; no mappings or other access can be performed now. When enabling mmap() of the DMABUFs, this means access through the VMA is also revoked. This complicates the fault handler because whilst the DMABUF exists, it has no guarantee that the corresponding VFIO device is still alive. Adds synchronisation ensuring the vdev is available before vdev->memory_lock is touched. (I decided against the alternative of preventing cleanup by holding the VFIO device open if any DMABUFs exist, because it's both a change of behaviour and less clean overall.) I've added a chonky comment in place, happy to clarify more if you have ideas. vfio/pci: Permanently revoke a DMABUF on request By weight, this is mostly a rename of revoked to an enum, status. There are now 3 states for a buffer, usable and revoked temporary/permanent. A new VFIO device ioctl is added, VFIO_DEVICE_PCI_DMABUF_REVOKE, which passes a DMABUF (exported from that device) and permanently revokes it. Thus a userspace driver can guarantee any downstream consumers of a shared fd are prevented from accessing a BAR range, and that range can be reused. The code doing revocation in vfio_pci_dma_buf_move() is moved, unchanged, to a common function for use by _move() and the new ioctl path. Q: I can't think of a good reason to temporarily revoke/unrevoke buffers from userspace, so didn't add a 'flags' field to the ioctl struct. Easy to add if people think it's worthwhile for future use. vfio/pci: Add mmap() attributes to DMABUF feature Reserves bits [31:28] in vfio_device_feature_dma_buf to allow a (CPU) mapping attribute to be specified for an exported set of ranges. The default is the current UC, and a new flag can specify CPU access as WC. Q: I've taken 4 bits; the intention is for this field to be a scalar not a bitmap (i.e. mutually-exclusive access properties). Perhaps 4 is a bit too many? Testing ======= (The [RFC ONLY] userspace test program, for QEMU edu-plus, has been dropped, but can be found in the GitHub branch below.) This code has been tested in mapping DMABUFs of single/multiple ranges, aliasing mmap()s, aliasing ranges across DMABUFs, vm_pgoff > 0, revocation, shutdown/cleanup scenarios, and hugepage mappings seem to work correctly. I've lightly tested WC mappings also (by observing resulting PTEs as having the correct attributes...). No regressions observed on the VFIO selftests, or on our internal vfio-pci applications. End === This is based on -next (next-20260414 but will merge earlier), as it depends on Leon's series "vfio: Wait for dma-buf invalidation to complete": https://lore.kernel.org/linux-iommu/20260205-nocturnal-poetic-chamois-f566a… These commits are on GitHub, along with "[RFC ONLY] selftests: vfio: Add standalone vfio_dmabuf_mmap_test": https://github.com/metamev/linux/compare/next-20260414...metamev:linux:dev/… Thanks for reading, Matt ================================================================================ Change log: v1: - Cleanup of the common DMABUF-aware VMA vm_ops fault handler and export code. - Fixed a lot of races, particularly faults racing with DMABUF cleanup (if the VFIO device fds close, for example). - Added nicer human-readable names for VFIO mmap() VMAs RFCv2: Respin based on the feedback/suggestions: https://lore.kernel.org/kvm/20260312184613.3710705-1-mattev@meta.com/ - Transform the existing VFIO BAR mmap path to also use DMABUFs behind the scenes, and then simply share that code for explicitly-mapped DMABUFs. Jason wanted to go that direction to enable iommufd VFIO type 1 emulation to pick up a DMABUF for an IO mapping. - Revoke buffers using a VFIO device fd ioctl RFCv1: https://lore.kernel.org/all/20260226202211.929005-1-mattev@meta.com/ Matt Evans (9): vfio/pci: Fix vfio_pci_dma_buf_cleanup() double-put vfio/pci: Add a helper to look up PFNs for DMABUFs vfio/pci: Add a helper to create a DMABUF for a BAR-map VMA vfio/pci: Convert BAR mmap() to use a DMABUF vfio/pci: Provide a user-facing name for BAR mappings vfio/pci: Clean up BAR zap and revocation vfio/pci: Support mmap() of a VFIO DMABUF vfio/pci: Permanently revoke a DMABUF on request vfio/pci: Add mmap() attributes to DMABUF feature drivers/vfio/pci/Kconfig | 3 +- drivers/vfio/pci/Makefile | 3 +- drivers/vfio/pci/nvgrace-gpu/main.c | 5 + drivers/vfio/pci/vfio_pci_config.c | 30 +- drivers/vfio/pci/vfio_pci_core.c | 224 ++++++++++--- drivers/vfio/pci/vfio_pci_dmabuf.c | 500 +++++++++++++++++++++++----- drivers/vfio/pci/vfio_pci_priv.h | 49 ++- include/linux/vfio_pci_core.h | 1 + include/uapi/linux/vfio.h | 42 ++- 9 files changed, 690 insertions(+), 167 deletions(-) -- 2.47.3

1 month, 2 weeks

[PATCH v2] dma-fence: Clarify external lock use case in dma_fence_init() docs

by Maíra Canal

The kerneldoc comment on dma_fence_init() and dma_fence_init64() describe the legacy reason to pass an external lock as a need to prevent multiple fences "from signaling out of order". However, this wording is a bit misleading: a shared spinlock does not (and cannot) prevent the signaler from signaling out of order. Signaling order is the driver's responsibility regardless of whether the lock is shared or per-fence. What a shared lock actually provides is serialization of signaling and observation across fences in a given context, so that observers never see a later fence signaled while an earlier one is not. Reword both comments to describe this more accurately. Signed-off-by: Maíra Canal <mcanal(a)igalia.com> --- v1 -> v2: https://lore.kernel.org/dri-devel/20260411185756.1887119-4-mcanal@igalia.co… - Be more explicit about not allowing new users to use an external lock. - De-duplicate the explanation in dma_fence_init64() by pointing to the dma_fence_init() documentation. --- drivers/dma-buf/dma-fence.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c index 1c1eaecaf1b0..63b349ba9a34 100644 --- a/drivers/dma-buf/dma-fence.c +++ b/drivers/dma-buf/dma-fence.c @@ -1102,9 +1102,11 @@ __dma_fence_init(struct dma_fence *fence, const struct dma_fence_ops *ops, * context and seqno are used for easy comparison between fences, allowing * to check which fence is later by simply using dma_fence_later(). * - * It is strongly discouraged to provide an external lock because this couples - * lock and fence life time. This is only allowed for legacy use cases when - * multiple fences need to be prevented from signaling out of order. + * External locks are a relic from legacy use cases that needed a shared lock + * to serialize signaling and observation of fences within a context, so that + * observers never see a later fence signaled while an earlier one isn't. New + * users MUST NOT use external locks, as they force the issuer to outlive all + * fences that reference the lock. */ void dma_fence_init(struct dma_fence *fence, const struct dma_fence_ops *ops, @@ -1129,9 +1131,8 @@ EXPORT_SYMBOL(dma_fence_init); * Context and seqno are used for easy comparison between fences, allowing * to check which fence is later by simply using dma_fence_later(). * - * It is strongly discouraged to provide an external lock because this couples - * lock and fence life time. This is only allowed for legacy use cases when - * multiple fences need to be prevented from signaling out of order. + * New users MUST NOT use external locks. Check the documentation in + * dma_fence_init() to understand the motives behind the legacy use cases. */ void dma_fence_init64(struct dma_fence *fence, const struct dma_fence_ops *ops, -- 2.53.0

1 month, 3 weeks

[PATCH 00/13] Add DMA support for LINFlexD UART driver

by Larisa Grigore

This patchset enhances the LINFlexD UART driver and its device tree bindings to support DMA transfers, configurable clock inputs, dynamic baudrate changes, and termios features. It also includes a series of fixes and improvements to ensure reliable operation across various modes and configurations. The changes added can be summarized as follows: 1. Fixes with respect to FIFO handling, locking, interrupt related registers and INITM mode transition. 2. Removal of the earlycon workaround, as proper FIFO handling and INITM transitions now ensure stable behavior. 3. Support for configurable stop bits and dynamic baudrate changes based on clock inputs and termios settings. 4. Optional DMA support for RX and TX paths, preventing character loss during high-throughput operations like copy-paste. Cyclic DMA is used for RX to avoid gaps between transactions. Larisa Grigore (8): serial: linflexuart: Clean SLEEP bit in LINCR1 after suspend serial: linflexuart: Check FIFO full before writing serial: linflexuart: Correctly clear UARTSR in buffer mode serial: linflexuart: Update RXEN/TXEN outside INITM mode serial: linflexuart: Ensure FIFO is empty when entering INITM serial: linflexuart: Revert earlycon workaround serial: linflexuart: Add support for configurable stop bits serial: linflexuart: Add DMA support Radu Pirea (5): serial: linflexuart: Fix locking in set_termios dt-bindings: serial: fsl-linflexuart: add clock input properties dt-bindings: serial: fsl-linflexuart: add dma properties serial: linflexuart: Add support for changing baudrate serial: linflexuart: Avoid stopping DMA during receive operations .../bindings/serial/fsl,s32-linflexuart.yaml | 31 + drivers/tty/serial/fsl_linflexuart.c | 972 +++++++++++++++--- 2 files changed, 846 insertions(+), 157 deletions(-) -- 2.47.0

1 month, 3 weeks

[PATCH v2 00/10] dma-bug: heaps: Add Tegra VPR support

by Thierry Reding

From: Thierry Reding <treding(a)nvidia.com> Hi, This series adds support for the video protection region (VPR) used on Tegra SoC devices. It's a special region of memory that is protected from accesses by the CPU and used to store DRM protected content (both decrypted stream data as well as decoded video frames). Patches 1 and 2 add DT binding documentation for the VPR and add the VPR to the list of memory-region items for display and host1x. Patch 3 adds bitmap_allocate(), which is like bitmap_allocate_region() but works on sizes that are not a power of two. Patch 4 introduces new APIs needed by the Tegra VPR implementation that allow CMA areas to be dynamically created at runtime rather than using the fixed, system-wide list. This is used in this driver specifically because it can use an arbitrary number of these areas (though they are currently limited to 4). Patch 5 adds some infrastructure for DMA heap implementations to provide information through debugfs. The Tegra VPR implementation is added in patch 6. See its commit message for more details about the specifics of this implementation. Finally, patches 7-10 add the VPR placeholder node on Tegra234 and hook it up to the host1x and GPU nodes so that they can make use of this region. Changes in v2: - Tegra VPR implementation is now more optimized to reduce the number of (very slow) resize operations, and allows cross-chunk allocations - dynamic CMA areas are now trackd separately from static ones, but the global number of CMA pages accounts for all areas Thierry Thierry Reding (10): dt-bindings: reserved-memory: Document Tegra VPR dt-bindings: display: tegra: Document memory regions bitmap: Add bitmap_allocate() function mm/cma: Allow dynamically creating CMA areas dma-buf: heaps: Add debugfs support dma-buf: heaps: Add support for Tegra VPR arm64: tegra: Add VPR placeholder node on Tegra234 arm64: tegra: Add GPU node on Tegra234 arm64: tegra: Hook up VPR to host1x arm64: tegra: Hook up VPR to the GPU .../display/tegra/nvidia,tegra186-dc.yaml | 10 + .../display/tegra/nvidia,tegra20-dc.yaml | 10 +- .../display/tegra/nvidia,tegra20-host1x.yaml | 7 + .../nvidia,tegra-video-protection-region.yaml | 55 + arch/arm/mm/dma-mapping.c | 2 +- arch/arm64/boot/dts/nvidia/tegra234.dtsi | 60 + arch/s390/mm/init.c | 2 +- drivers/dma-buf/dma-heap.c | 56 + drivers/dma-buf/heaps/Kconfig | 7 + drivers/dma-buf/heaps/Makefile | 1 + drivers/dma-buf/heaps/cma_heap.c | 2 +- drivers/dma-buf/heaps/tegra-vpr.c | 1265 +++++++++++++++++ include/linux/bitmap.h | 25 +- include/linux/cma.h | 7 +- include/linux/dma-heap.h | 2 + include/trace/events/tegra_vpr.h | 57 + mm/cma.c | 187 ++- mm/cma.h | 5 +- 18 files changed, 1713 insertions(+), 47 deletions(-) create mode 100644 Documentation/devicetree/bindings/reserved-memory/nvidia,tegra-video-protection-region.yaml create mode 100644 drivers/dma-buf/heaps/tegra-vpr.c create mode 100644 include/trace/events/tegra_vpr.h -- 2.52.0

2 months

[PATCH] dma-buf: system_heap: Optimize sg_table-to-pages conversion in vmap

by Barry Song (Xiaomi)

From: Xueyuan Chen <Xueyuan.chen21(a)gmail.com> Replace the heavy for_each_sgtable_page() iterator in system_heap_do_vmap() with a more efficient nested loop approach. Instead of iterating page by page, we now iterate through the scatterlist entries via for_each_sgtable_sg(). Because pages within a single sg entry are physically contiguous, we can populate the page array with a in an inner loop using simple pointer math. This save a lot of time. The WARN_ON check is also pulled out of the loop to save branch instructions. Performance results mapping a 2GB buffer on Radxa O6: - Before: ~1440000 ns - After: ~232000 ns (~84% reduction in iteration time, or ~6.2x faster) Cc: Sumit Semwal <sumit.semwal(a)linaro.org> Cc: Benjamin Gaignard <benjamin.gaignard(a)collabora.com> Cc: Brian Starkey <Brian.Starkey(a)arm.com> Cc: John Stultz <jstultz(a)google.com> Cc: T.J. Mercier <tjmercier(a)google.com> Cc: Christian König <christian.koenig(a)amd.com> Signed-off-by: Xueyuan Chen <Xueyuan.chen21(a)gmail.com> Signed-off-by: Barry Song (Xiaomi) <baohua(a)kernel.org> --- drivers/dma-buf/heaps/system_heap.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/drivers/dma-buf/heaps/system_heap.c b/drivers/dma-buf/heaps/system_heap.c index b3650d8fd651..769f01f0cc96 100644 --- a/drivers/dma-buf/heaps/system_heap.c +++ b/drivers/dma-buf/heaps/system_heap.c @@ -224,16 +224,21 @@ static void *system_heap_do_vmap(struct system_heap_buffer *buffer) int npages = PAGE_ALIGN(buffer->len) / PAGE_SIZE; struct page **pages = vmalloc(sizeof(struct page *) * npages); struct page **tmp = pages; - struct sg_page_iter piter; void *vaddr; + u32 i, j, count; + struct page *base_page; + struct scatterlist *sg; if (!pages) return ERR_PTR(-ENOMEM); - for_each_sgtable_page(table, &piter, 0) { - WARN_ON(tmp - pages >= npages); - *tmp++ = sg_page_iter_page(&piter); + for_each_sgtable_sg(table, sg, i) { + base_page = sg_page(sg); + count = sg->length >> PAGE_SHIFT; + for (j = 0; j < count; j++) + *tmp++ = base_page + j; } + WARN_ON(tmp - pages != npages); vaddr = vmap(pages, npages, VM_MAP, PAGE_KERNEL); vfree(pages); -- 2.39.3 (Apple Git-146)

2 months

[PATCH 1/2] dma-buf/dma_fence_array: remove unused functionality v3

by Christian König

Amdgpu was the only user of the signal on any feature and we dropped that use case recently, so we can remove that functionality. v2: update num_pending only after the fence is signaled v3: separate out simplifying dma_fence_array implementation Signed-off-by: Christian König <christian.koenig(a)amd.com> --- drivers/dma-buf/dma-fence-array.c | 13 ++++--------- drivers/dma-buf/dma-fence-unwrap.c | 3 +-- drivers/dma-buf/dma-resv.c | 3 +-- drivers/dma-buf/st-dma-fence-unwrap.c | 2 +- drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c | 3 +-- drivers/gpu/drm/xe/xe_sync.c | 2 +- drivers/gpu/drm/xe/xe_vm.c | 4 ++-- include/linux/dma-fence-array.h | 6 ++---- 8 files changed, 13 insertions(+), 23 deletions(-) diff --git a/drivers/dma-buf/dma-fence-array.c b/drivers/dma-buf/dma-fence-array.c index 089f69469524..5e10e8df372f 100644 --- a/drivers/dma-buf/dma-fence-array.c +++ b/drivers/dma-buf/dma-fence-array.c @@ -190,15 +190,13 @@ EXPORT_SYMBOL(dma_fence_array_alloc); * @fences: [in] array containing the fences * @context: [in] fence context to use * @seqno: [in] sequence number to use - * @signal_on_any: [in] signal on any fence in the array * * Implementation of @dma_fence_array_create without allocation. Useful to init * a preallocated dma fence array in the path of reclaim or dma fence signaling. */ void dma_fence_array_init(struct dma_fence_array *array, int num_fences, struct dma_fence **fences, - u64 context, unsigned seqno, - bool signal_on_any) + u64 context, unsigned seqno) { static struct lock_class_key dma_fence_array_lock_key; @@ -222,7 +220,7 @@ void dma_fence_array_init(struct dma_fence_array *array, */ lockdep_set_class(&array->base.inline_lock, &dma_fence_array_lock_key); - atomic_set(&array->num_pending, signal_on_any ? 1 : num_fences); + atomic_set(&array->num_pending, num_fences); array->fences = fences; array->base.error = PENDING_ERROR; @@ -249,7 +247,6 @@ EXPORT_SYMBOL(dma_fence_array_init); * @fences: [in] array containing the fences * @context: [in] fence context to use * @seqno: [in] sequence number to use - * @signal_on_any: [in] signal on any fence in the array * * Allocate a dma_fence_array object and initialize the base fence with * dma_fence_init(). @@ -264,8 +261,7 @@ EXPORT_SYMBOL(dma_fence_array_init); */ struct dma_fence_array *dma_fence_array_create(int num_fences, struct dma_fence **fences, - u64 context, unsigned seqno, - bool signal_on_any) + u64 context, unsigned seqno) { struct dma_fence_array *array; @@ -273,8 +269,7 @@ struct dma_fence_array *dma_fence_array_create(int num_fences, if (!array) return NULL; - dma_fence_array_init(array, num_fences, fences, - context, seqno, signal_on_any); + dma_fence_array_init(array, num_fences, fences, context, seqno); return array; } diff --git a/drivers/dma-buf/dma-fence-unwrap.c b/drivers/dma-buf/dma-fence-unwrap.c index 07fe9bf45aea..53bb40e70b27 100644 --- a/drivers/dma-buf/dma-fence-unwrap.c +++ b/drivers/dma-buf/dma-fence-unwrap.c @@ -180,8 +180,7 @@ struct dma_fence *__dma_fence_unwrap_merge(unsigned int num_fences, if (count > 1) { result = dma_fence_array_create(count, array, - dma_fence_context_alloc(1), - 1, false); + dma_fence_context_alloc(1), 1); if (!result) { for (i = 0; i < count; i++) dma_fence_put(array[i]); diff --git a/drivers/dma-buf/dma-resv.c b/drivers/dma-buf/dma-resv.c index ce9e6c04897f..39a92d9f2413 100644 --- a/drivers/dma-buf/dma-resv.c +++ b/drivers/dma-buf/dma-resv.c @@ -648,8 +648,7 @@ int dma_resv_get_singleton(struct dma_resv *obj, enum dma_resv_usage usage, } array = dma_fence_array_create(count, fences, - dma_fence_context_alloc(1), - 1, false); + dma_fence_context_alloc(1), 1); if (!array) { while (count--) dma_fence_put(fences[count]); diff --git a/drivers/dma-buf/st-dma-fence-unwrap.c b/drivers/dma-buf/st-dma-fence-unwrap.c index 51c87869b7b8..4e7ee25372ba 100644 --- a/drivers/dma-buf/st-dma-fence-unwrap.c +++ b/drivers/dma-buf/st-dma-fence-unwrap.c @@ -64,7 +64,7 @@ static struct dma_fence *mock_array(unsigned int num_fences, ...) array = dma_fence_array_create(num_fences, fences, dma_fence_context_alloc(1), - 1, false); + 1); if (!array) goto error_free; return &array->base; diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c index 942f4eed817f..4a1a9031f9db 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c @@ -3205,8 +3205,7 @@ eb_composite_fence_create(struct i915_execbuffer *eb, int out_fence_fd) fence_array = dma_fence_array_create(eb->num_batches, fences, eb->context->parallel.fence_context, - eb->context->parallel.seqno++, - false); + eb->context->parallel.seqno++); if (!fence_array) { kfree(fences); return ERR_PTR(-ENOMEM); diff --git a/drivers/gpu/drm/xe/xe_sync.c b/drivers/gpu/drm/xe/xe_sync.c index 24d6d9af20d6..37866768d64c 100644 --- a/drivers/gpu/drm/xe/xe_sync.c +++ b/drivers/gpu/drm/xe/xe_sync.c @@ -376,7 +376,7 @@ xe_sync_in_fence_get(struct xe_sync_entry *sync, int num_sync, xe_assert(vm->xe, current_fence == num_fence); cf = dma_fence_array_create(num_fence, fences, dma_fence_context_alloc(1), - 1, false); + 1); if (!cf) goto err_out; diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 56e2db50bb36..8f472911469d 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -3370,7 +3370,7 @@ static struct dma_fence *ops_execute(struct xe_vm *vm, goto err_trace; } - cf = dma_fence_array_alloc(n_fence); + cf = dma_fence_array_alloc(); if (!cf) { fence = ERR_PTR(-ENOMEM); goto err_out; @@ -3414,7 +3414,7 @@ static struct dma_fence *ops_execute(struct xe_vm *vm, xe_assert(vm->xe, current_fence == n_fence); dma_fence_array_init(cf, n_fence, fences, dma_fence_context_alloc(1), - 1, false); + 1); fence = &cf->base; for_each_tile(tile, vm->xe, id) { diff --git a/include/linux/dma-fence-array.h b/include/linux/dma-fence-array.h index 370b3d2bba37..1b1d87579c38 100644 --- a/include/linux/dma-fence-array.h +++ b/include/linux/dma-fence-array.h @@ -81,13 +81,11 @@ to_dma_fence_array(struct dma_fence *fence) struct dma_fence_array *dma_fence_array_alloc(int num_fences); void dma_fence_array_init(struct dma_fence_array *array, int num_fences, struct dma_fence **fences, - u64 context, unsigned seqno, - bool signal_on_any); + u64 context, unsigned seqno); struct dma_fence_array *dma_fence_array_create(int num_fences, struct dma_fence **fences, - u64 context, unsigned seqno, - bool signal_on_any); + u64 context, unsigned seqno); bool dma_fence_match_context(struct dma_fence *fence, u64 context); -- 2.43.0

2 months

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

Linaro-mm-sig April 2026