kernel/linux.git/drivers/base/dma-mapping.c, branch v4.17.1

dma-mapping: postpone cpu addr translation on mmap

2018-04-23T12:44:24+00:00

Postpone calling virt_to_page() translation on memory locations not guaranteed to be backed by a struct page. Try first to map memory from the device coherent memory pool, then perform translation if that fails. On some architectures, specifically SH when configured with the SPARSEMEM memory model, assuming a struct page is always assigned to a memory address lead to unexpected hangs during the virtual to page address translation. This patch fixes that specific issue but applies in the general case too. Suggested-by: Laurent Pinchart Signed-off-by: Jacopo Mondi Reviewed-by: Robin Murphy Signed-off-by: Christoph Hellwig

driver core: Remove redundant license text

2017-12-07T17:36:44+00:00

Now that the SPDX tag is in all driver core files, that identifies the license in a specific and legally-defined manner. So the extra GPL text wording can be removed as it is no longer needed at all. This is done on a quest to remove the 700+ different ways that files in the kernel describe the GPL license text. And there's unneeded stuff like the address (sometimes incorrect) for the FSF which is never needed. No copyright headers or other non-license-description text was removed. Cc: Johannes Berg Cc: "Luis R. Rodriguez" Signed-off-by: Greg Kroah-Hartman

driver core: add SPDX identifiers to all driver core files

2017-12-07T17:36:43+00:00

It's good to have SPDX identifiers in all files to make it easier to audit the kernel tree for correct licenses. Update the driver core files files with the correct SPDX license identifier based on the license text in the file itself. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This work is based on a script and data from Thomas Gleixner, Philippe Ombredanne, and Kate Stewart. Cc: Johannes Berg Cc: "Luis R. Rodriguez" Cc: William Breathitt Gray Cc: Thomas Gleixner Cc: Kate Stewart Cc: Philippe Ombredanne Signed-off-by: Greg Kroah-Hartman

dma-coherent: remove the DMA_MEMORY_MAP and DMA_MEMORY_IO flags

2017-09-01T09:59:17+00:00

DMA_MEMORY_IO was never used in the tree, so remove it. That means there is no need for the DMA_MEMORY_MAP flag either now, so remove it as well and change dma_declare_coherent_memory to return a normal errno value. Signed-off-by: Christoph Hellwig Reviewed-by: Marek Szyprowski

dma-coherent: introduce interface for default DMA pool

2017-07-20T14:09:10+00:00

Christoph noticed [1] that default DMA pool in current form overload the DMA coherent infrastructure. In reply, Robin suggested [2] to split the per-device vs. global pool interfaces, so allocation/release from default DMA pool is driven by dma ops implementation. This patch implements Robin's idea and provide interface to allocate/release/mmap the default (aka global) DMA pool. To make it clear that existing *_from_coherent routines work on per-device pool rename them to *_from_dev_coherent. [1] https://lkml.org/lkml/2017/7/7/370 [2] https://lkml.org/lkml/2017/7/7/431 Cc: Vineet Gupta Cc: Russell King Cc: Catalin Marinas Cc: Will Deacon Cc: Ralf Baechle Suggested-by: Robin Murphy Tested-by: Andras Szemzo Reviewed-by: Robin Murphy Signed-off-by: Vladimir Murzin Signed-off-by: Christoph Hellwig

Merge tag 'dma-mapping-4.13' of git://git.infradead.org/users/hch/dma-mapping

2017-07-07T02:20:54+00:00

Pull dma-mapping infrastructure from Christoph Hellwig: "This is the first pull request for the new dma-mapping subsystem In this new subsystem we'll try to properly maintain all the generic code related to dma-mapping, and will further consolidate arch code into common helpers. This pull request contains: - removal of the DMA_ERROR_CODE macro, replacing it with calls to ->mapping_error so that the dma_map_ops instances are more self contained and can be shared across architectures (me) - removal of the ->set_dma_mask method, which duplicates the ->dma_capable one in terms of functionality, but requires more duplicate code. - various updates for the coherent dma pool and related arm code (Vladimir) - various smaller cleanups (me)" * tag 'dma-mapping-4.13' of git://git.infradead.org/users/hch/dma-mapping: (56 commits) ARM: dma-mapping: Remove traces of NOMMU code ARM: NOMMU: Set ARM_DMA_MEM_BUFFERABLE for M-class cpus ARM: NOMMU: Introduce dma operations for noMMU drivers: dma-mapping: allow dma_common_mmap() for NOMMU drivers: dma-coherent: Introduce default DMA pool drivers: dma-coherent: Account dma_pfn_offset when used with device tree dma: Take into account dma_pfn_offset dma-mapping: replace dmam_alloc_noncoherent with dmam_alloc_attrs dma-mapping: remove dmam_free_noncoherent crypto: qat - avoid an uninitialized variable warning au1100fb: remove a bogus dma_free_nonconsistent call MAINTAINERS: add entry for dma mapping helpers powerpc: merge __dma_set_mask into dma_set_mask dma-mapping: remove the set_dma_mask method powerpc/cell: use the dma_supported method for ops switching powerpc/cell: clean up fixed mapping dma_ops initialization tile: remove dma_supported and mapping_error methods xen-swiotlb: remove xen_swiotlb_set_dma_mask arm: implement ->dma_supported instead of ->set_dma_mask mips/loongson64: implement ->dma_supported instead of ->set_dma_mask ...

drivers: dma-mapping: allow dma_common_mmap() for NOMMU

2017-06-30T17:03:07+00:00

Currently, internals of dma_common_mmap() is compiled out if build is done for either NOMMU or target which explicitly says it does not have/want coherent DMA mmap. It turned out that dma_common_mmap() can be handy in NOMMU setup (at least for ARM). This patch converts exitent NOMMU targets to use ARCH_NO_COHERENT_DMA_MMAP, thus when CONFIG_MMU is gone from dma_common_mmap() their behaviour stays unchanged. ARM is not converted to ARCH_NO_COHERENT_DMA_MMAP because it 1) already has mmap callback which can handle (at some extent) NOMMU 2) already defines dummy pgprot_noncached() for NOMMU build. c6x and frv stay untouched since they already have ARCH_NO_COHERENT_DMA_MMAP. Cc: Steven Miao Cc: Geert Uytterhoeven Cc: Michal Simek Cc: Yoshinori Sato Cc: Rich Felker Cc: Chris Zankel Cc: Max Filippov Suggested-by: Christoph Hellwig Signed-off-by: Vladimir Murzin Tested-by: Benjamin Gaignard

dma-mapping: replace dmam_alloc_noncoherent with dmam_alloc_attrs

2017-06-28T13:55:00+00:00

dmam_alloc_noncoherent is a trivial wrapper around dmam_alloc_attrs, that hardcodes one particular flag. Make the devres code more flexible by allowing the callers to pass arbitrary flags. Signed-off-by: Christoph Hellwig Acked-by: Tejun Heo

dma-mapping: remove dmam_free_noncoherent

2017-06-28T13:54:59+00:00

This function was never used since it was added. Signed-off-by: Christoph Hellwig Acked-by: Tejun Heo

drivers: dma-mapping: Do not leave an invalid area->pages pointer in dma_common_contiguous_remap()

2017-06-03T09:51:12+00:00

The dma_common_pages_remap() function allocates a vm_struct object and initialises the pages pointer to value passed as argument. However, when this function is called dma_common_contiguous_remap(), the pages array is only temporarily allocated, being freed shortly after dma_common_contiguous_remap() returns. Architecture code checking the validity of an area->pages pointer would incorrectly dereference already freed pointers. This has been exposed by the arm64 commit 44176bb38fa4 ("arm64: Add support for DMA_ATTR_FORCE_CONTIGUOUS to IOMMU"). Fixes: 513510ddba96 ("common: dma-mapping: introduce common remapping functions") Cc: Greg Kroah-Hartman Reported-by: Andrzej Hajda Acked-by: Laura Abbott Reviewed-by: Robin Murphy Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman