wireguard-linux/drivers/crypto/marvell/cesa, branch stable

crypto: cesa - clear cesa_dev on _remove

2026-08-10T08:28:56Z

This is already done on probe failure. cesa_dev should be set to NULL as the driver checks that it is not and errors saying that only one device is authorized. Signed-off-by: Rosen Penev Signed-off-by: Herbert Xu

crypto: cesa - manage SRAM teardown with devm

2026-07-30T07:36:41Z

mv_cesa_put_sram() is called explicitly from both the probe error path and mv_cesa_remove(). The non-pool ioremap is already devm-managed, but dma_map_resource() and gen_pool_dma_alloc() have no devm helpers, so the mapping is released by hand. This is error-prone: the error path iterates over every engine and can dma_unmap_resource() an uninitialized/zero address for engines that were never set up. Convert the teardown into a devm_add_action_or_reset() callback registered only after a mapping is successfully established. The callback fires automatically on probe failure (devres rollback) and on device detach, after mv_cesa_remove() has already stopped the engine and freed the IRQ, so the unmap still happens in a safe order. This deletes the explicit mv_cesa_put_sram() calls and the uninitialized-engine bug at once. Add a struct mv_cesa_dev back-pointer to struct mv_cesa_engine so the callback can reach cesa->dev and cesa->sram_size from the engine alone. Assisted-by: opencode:hy3-free Signed-off-by: Rosen Penev Signed-off-by: Herbert Xu

crypto: cesa - use max to simplify mv_cesa_probe

2026-05-22T12:25:28Z

Use max() to simplify mv_cesa_probe() and improve its readability. Signed-off-by: Thorsten Blum Signed-off-by: Herbert Xu

crypto: cesa - allocate engines with main struct

2026-05-07T08:10:02Z

Use a flexible array member to combine and simplify allocation. Move struct mv_cesa_dev down as flexible array members require full definitions. Signed-off-by: Rosen Penev Signed-off-by: Herbert Xu

crypto: marvell/cesa - use memcpy_and_pad in mv_cesa_ahash_export

2026-03-27T09:52:25Z

Replace memset() followed by memcpy() with memcpy_and_pad() to simplify the code and to write to 'cache' only once. Signed-off-by: Thorsten Blum Signed-off-by: Herbert Xu

crypto: cesa - Simplify return statement in mv_cesa_dequeue_req_locked

2026-02-06T10:52:22Z

Return the result of calling crypto_dequeue_request() directly and remove the local return variable. Signed-off-by: Thorsten Blum Signed-off-by: Herbert Xu

crypto: cesa - Simplify with of_device_get_match_data()

2025-11-14T10:15:49Z

Driver's probe function matches against driver's of_device_id table, where each entry has non-NULL match data, so of_match_node() can be simplified with of_device_get_match_data(). Signed-off-by: Krzysztof Kozlowski Signed-off-by: Herbert Xu

crypto: marvell/cesa - Fix engine load inaccuracy

2025-06-13T09:26:16Z

If an error occurs during queueing the engine load will never be decremented. Fix this by moving the engine load adjustment into the cleanup function. Fixes: bf8f91e71192 ("crypto: marvell - Add load balancing between engines") Signed-off-by: Herbert Xu

crypto: marvell/cesa - Remove unnecessary state setting on final

2025-06-13T09:26:16Z

There is no point in setting the hash state after finalisation since the hash state must never be used again. Remove that code. Signed-off-by: Herbert Xu

crypto: marvell/cesa - Do not chain submitted requests

2025-05-21T05:27:46Z

This driver tries to chain requests together before submitting them to hardware in order to reduce completion interrupts. However, it even extends chains that have already been submitted to hardware. This is dangerous because there is no way of knowing whether the hardware has already read the DMA memory in question or not. Fix this by splitting the chain list into two. One for submitted requests and one for requests that have not yet been submitted. Only extend the latter. Reported-by: Klaus Kudielka Fixes: 85030c5168f1 ("crypto: marvell - Add support for chaining crypto requests in TDMA mode") Cc: Signed-off-by: Herbert Xu