user/sven/linux.git/include/linux/dmaengine.h, branch v4.19.162

dmaengine: Store module owner in dma_device struct

2020-02-24T07:34:44Z

[ Upstream commit dae7a589c18a4d979d5f14b09374e871b995ceb1 ] dma_chan_to_owner() dereferences the driver from the struct device to obtain the owner and call module_[get|put](). However, if the backing device is unbound before the dma_device is unregistered, the driver will be cleared and this will cause a NULL pointer dereference. Instead, store a pointer to the owner module in the dma_device struct so the module reference can be properly put when the channel is put, even if the backing device was destroyed first. This change helps to support a safer unbind of DMA engines. If the dma_device is unregistered in the driver's remove function, there's no guarantee that there are no existing clients and a users action may trigger the WARN_ONCE in dma_async_device_unregister() which is unlikely to leave the system in a consistent state. Instead, a better approach is to allow the backing driver to go away and fail any subsequent requests to it. Signed-off-by: Logan Gunthorpe Link: https://lore.kernel.org/r/20191216190120.21374-2-logang@deltatee.com Signed-off-by: Vinod Koul Signed-off-by: Sasha Levin

dmaengine: Fix access to uninitialized dma_slave_caps

2020-01-09T09:19:00Z

commit 53a256a9b925b47c7e67fc1f16ca41561a7b877c upstream. dmaengine_desc_set_reuse() allocates a struct dma_slave_caps on the stack, populates it using dma_get_slave_caps() and then accesses one of its members. However dma_get_slave_caps() may fail and this isn't accounted for, leading to a legitimate warning of gcc-4.9 (but not newer versions): In file included from drivers/spi/spi-bcm2835.c:19:0: drivers/spi/spi-bcm2835.c: In function 'dmaengine_desc_set_reuse': >> include/linux/dmaengine.h:1370:10: warning: 'caps.descriptor_reuse' is used uninitialized in this function [-Wuninitialized] if (caps.descriptor_reuse) { Fix it, thereby also silencing the gcc-4.9 warning. The issue has been present for 4 years but surfaces only now that the first caller of dmaengine_desc_set_reuse() has been added in spi-bcm2835.c. Another user of reusable DMA descriptors has existed for a while in pxa_camera.c, but it sets the DMA_CTRL_REUSE flag directly instead of calling dmaengine_desc_set_reuse(). Nevertheless, tag this commit for stable in case there are out-of-tree users. Fixes: 272420214d26 ("dmaengine: Add DMA_CTRL_REUSE") Reported-by: kbuild test robot Signed-off-by: Lukas Wunner Cc: stable@vger.kernel.org # v4.3+ Link: https://lore.kernel.org/r/ca92998ccc054b4f2bfd60ef3adbab2913171eac.1575546234.git.lukas@wunner.de Signed-off-by: Vinod Koul Signed-off-by: Greg Kroah-Hartman

dmaengine: add a new helper dmaenginem_async_device_register

2018-07-30T05:20:22Z

This patch adds the dmaenginem_async_device_register for DMA code. Use the Devres to call the release for the DMA engine driver. Signed-off-by: Huang Shijie Signed-off-by: Vinod Koul

dmaengine: add support for reporting pause and resume separately

2018-07-09T17:29:04Z

'cmd_pause' DMA channel capability means that respective DMA engine supports both pausing and resuming given DMA channel. However, in some cases it is important to know if DMA channel can be paused without the need to resume it. This is a typical requirement for proper residue reading on transfer timeout in UART drivers. There are also some DMA engines with limited hardware, which doesn't really support resuming. Reporting pause and resume capabilities separately allows UART drivers to properly check for the really required capabilities and operate in DMA mode also in systems with limited DMA hardware. On the other hand drivers, which rely on full channel suspend/resume support, should now check for both 'pause' and 'resume' features. Existing clients of dma_get_slave_caps() have been checked and the only driver which rely on proper channel resuming is soc-generic-dmaengine-pcm driver, which has been updated to check the newly added capability. Existing 'cmd_pause' now only indicates that DMA engine support pausing given DMA channel. Signed-off-by: Marek Szyprowski Acked-by: Mark Brown Signed-off-by: Vinod Koul

dmaengine: avoid map_cnt overflow with CONFIG_DMA_ENGINE_RAID

2018-02-12T03:48:56Z

When CONFIG_DMA_ENGINE_RAID is enabled, unmap pool size can reach to 256. But in struct dmaengine_unmap_data, map_cnt is only u8, wrapping to 0, if the unmap pool is maximally used. This triggers BUG() when struct dmaengine_unmap_data is freed. Use u16 to fix the problem. Signed-off-by: Zi Yan Signed-off-by: Vinod Koul

dmaengine: Mark struct dma_slave_caps kernel-doc correctly, clarify

2017-09-21T17:04:28Z

struct dma_slave_caps documentation omitted the correct kernel-doc opening comment mark. Document byte granularity and interpretation of the src/dst_addr_widths bit flag fields used by struct dma_slave_caps and struct dma_device. Add punctuation to their "directions" member documentations, and cleanup wording of the description. Signed-off-by: Stefan Brüns Signed-off-by: Vinod Koul

dmaengine: List all allowed values for src/dst_addr_width in kernel doc

2017-09-21T17:04:28Z

Commit 93c6ee94c140 ("dma: Support for 3 bytes word size") and commit 534a729866f9 ("dmaengine: Add 16 bytes, 32 bytes and 64 bytes bus widths") added additional values for the allowed word size, but omitted these from the struct dma_slave_config documentation. Signed-off-by: Stefan Brüns Signed-off-by: Vinod Koul

Merge branch 'topic/dmatest' into for-linus

2017-09-06T16:25:10Z

dmaengine: add DMA_PREP_CMD for non-Data descriptors.

2017-08-28T11:10:18Z

Some of the DMA controllers are capable of issuing the commands to peripheral by the DMA. These commands can be list of register reads/writes and its different from normal data reads/writes. This patch adds new flag DMA_PREP_CMD in DMA API which tells the driver that the data passed to DMA API is command data and DMA controller driver will form descriptor in the required format. This flag can be used by any DMA controller driver which requires the descriptor in different format for non-Data descriptors. Signed-off-by: Abhishek Sahu Signed-off-by: Vinod Koul

dmaengine: remove DMA_SG as it is dead code in kernel

2017-08-22T03:52:11Z

There are no in kernel consumers for DMA_SG op. Removing operation, dead code, and test code in dmatest. Signed-off-by: Dave Jiang Reviewed-by: Linus Walleij Cc: Gary Hook Cc: Ludovic Desroches Cc: Kedareswara rao Appana Cc: Li Yang Cc: Michal Simek Signed-off-by: Vinod Koul