kernel/linux.git/drivers/dma-buf/heaps, branch v6.18.21

dma-buf: heaps: Give default CMA heap a fixed name

2025-07-09T10:21:40+00:00

The CMA heap's name in devtmpfs can vary depending on how the heap is defined. Its name defaults to "reserved", but if a CMA area is defined in the devicetree, the heap takes on the devicetree node's name, such as "default-pool" or "linux,cma". To simplify naming, unconditionally name it "default_cma_region", but keep a legacy node in place backed by the same underlying allocator for backwards compatibility. Reviewed-by: Maxime Ripard Signed-off-by: Jared Kangas Signed-off-by: Sumit Semwal Link: https://lore.kernel.org/r/20250610131231.1724627-4-jkangas@redhat.com

dma-buf: heaps: Parameterize heap name in __add_cma_heap()

2025-07-09T10:21:39+00:00

Prepare for the introduction of a fixed-name CMA heap by replacing the unused void pointer parameter in __add_cma_heap() with the heap name. Reviewed-by: Maxime Ripard Acked-by: John Stultz Signed-off-by: Jared Kangas Signed-off-by: Sumit Semwal Link: https://lore.kernel.org/r/20250610131231.1724627-3-jkangas@redhat.com

dma-buf: system_heap: No separate allocation for attachment sg_tables

2025-07-09T10:21:39+00:00

struct dma_heap_attachment is a separate allocation from the struct sg_table it contains, but there is no reason for this. Let's use the slab allocator just once instead of twice for dma_heap_attachment. Signed-off-by: T.J. Mercier Reviewed-by: Christian König Signed-off-by: Sumit Semwal Link: https://lore.kernel.org/r/20250417180943.1559755-1-tjmercier@google.com

dma-buf: heaps: system: Remove global variable

2025-04-08T09:48:39+00:00

The system heap is storing its struct dma_heap pointer in a global variable but isn't using it anywhere. Let's move the global variable into system_heap_create() to make it local. Signed-off-by: Maxime Ripard Reviewed-by: Christian König Reviewed-by: Mattijs Korpershoek Link: https://patchwork.freedesktop.org/patch/msgid/20250407-dma-buf-ecc-heap-v3-1-97cdd36a5f29@kernel.org Signed-off-by: Christian König

dma-buf/heaps: replace kmap_atomic with kmap_local_page

2024-10-14T14:46:03+00:00

Use of kmap_atomic/kunmap_atomic is deprecated, use kmap_local_page/kunmap_local instead. This is reported by checkpatch. Also fix repeated word issue. WARNING: Deprecated use of 'kmap_atomic', prefer 'kmap_local_page' instead + void *vaddr = kmap_atomic(page); WARNING: Deprecated use of 'kunmap_atomic', prefer 'kunmap_local' instead + kunmap_atomic(vaddr); WARNING: Possible repeated word: 'by' + * has been killed by by SIGKILL total: 0 errors, 3 warnings, 405 lines checked Signed-off-by: Pintu Kumar Reviewed-by: T.J. Mercier Signed-off-by: Sumit Semwal Link: https://patchwork.freedesktop.org/patch/msgid/20241001175057.27172-1-quic_pintu@quicinc.com

Merge drm/drm-next into drm-misc-next

2024-09-30T08:50:54+00:00

Get drm-misc-next to up v6.12-rc1. Signed-off-by: Thomas Zimmermann

dma-buf: heaps: Add __init to CMA and system heap module_init functions

2024-09-09T10:21:54+00:00

Shrink the kernel .text a bit after successful initialization of the heaps. Signed-off-by: T.J. Mercier Acked-by: John Stultz Signed-off-by: Sumit Semwal Link: https://patchwork.freedesktop.org/patch/msgid/20240906000314.2368749-1-tjmercier@google.com

dma-buf: heaps: Fix off-by-one in CMA heap fault handler

2024-09-09T10:07:20+00:00

Until VM_DONTEXPAND was added in commit 1c1914d6e8c6 ("dma-buf: heaps: Don't track CMA dma-buf pages under RssFile") it was possible to obtain a mapping larger than the buffer size via mremap and bypass the overflow check in dma_buf_mmap_internal. When using such a mapping to attempt to fault past the end of the buffer, the CMA heap fault handler also checks the fault offset against the buffer size, but gets the boundary wrong by 1. Fix the boundary check so that we don't read off the end of the pages array and insert an arbitrary page in the mapping. Reported-by: Xingyu Jin Fixes: a5d2d29e24be ("dma-buf: heaps: Move heap-helper logic into the cma_heap implementation") Cc: stable@vger.kernel.org # Applicable >= 5.10. Needs adjustments only for 5.10. Signed-off-by: T.J. Mercier Acked-by: John Stultz Signed-off-by: Sumit Semwal Link: https://patchwork.freedesktop.org/patch/msgid/20240830192627.2546033-1-tjmercier@google.com

dma-buf/heaps: Correct the types of fd_flags and heap_flags

2024-06-19T14:35:34+00:00

dma_heap_allocation_data defines the UAPI as follows: struct dma_heap_allocation_data { __u64 len; __u32 fd; __u32 fd_flags; __u64 heap_flags; }; But dma heaps are casting both fd_flags and heap_flags into unsigned long. This patch makes dma heaps - cma heap and system heap have consistent types with UAPI. Signed-off-by: Barry Song Acked-by: John Stultz Reviewed-by: Carlos Llamas Signed-off-by: Sumit Semwal Link: https://patchwork.freedesktop.org/patch/msgid/20240606020213.49854-1-21cnbao@gmail.com

dma-buf: heaps: Don't track CMA dma-buf pages under RssFile

2024-01-31T14:24:58+00:00

DMA buffers allocated from the CMA dma-buf heap get counted under RssFile for processes that map them and trigger page faults. In addition to the incorrect accounting reported to userspace, reclaim behavior was influenced by the MM_FILEPAGES counter until linux 6.8, but this memory is not reclaimable. [1] Change the CMA dma-buf heap to set VM_PFNMAP on the VMA so MM does not poke at the memory managed by this dma-buf heap, and use vmf_insert_pfn to correct the RSS accounting. The system dma-buf heap does not suffer from this issue since remap_pfn_range is used during the mmap of the buffer, which also sets VM_PFNMAP on the VMA. [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/mm/vmscan.c?id=fb46e22a9e3863e08aef8815df9f17d0f4b9aede Fixes: b61614ec318a ("dma-buf: heaps: Add CMA heap to dmabuf heaps") Signed-off-by: T.J. Mercier Acked-by: Christian König Signed-off-by: Sumit Semwal Link: https://patchwork.freedesktop.org/patch/msgid/20240117181141.286383-1-tjmercier@google.com