This series finally makes memory<->device DMA on ep93xx robust/usable. First patch addresses long standing bug where the driver went out of sync with DMA controller and started transfer from buffer1 even though DMA HW expected the transfer to start from buffer0. This lead to frequent hearable audio distortions (on start/stop) and not so frequent system freezes (seems that DMA has overwritten unexpected RAM locations). Both problems are gone now. Second patch addresses high IRQ latency caused by the driver which actually completely prevents usage of EDB93xx audio and/or DMA drivers together with ep93xx watchdog driver. The later resets 250ms HW timer with 200ms period and is not able to do it with just slight load + ~40ms latency caused by DMA driver. Changelog v2: Deleted misleading impossible warning (patch 1/2) Move draining to device_synchronize callback (patch 2/2) Alexander Sverdlin (2): dmaengine: ep93xx: Always start from BASE0 dmaengine: ep93xx: Don't drain the transfers in terminate_all() drivers/dma/ep93xx_dma.c | 39 +++++++++++++++++++++++++++++++++++---- 1 file changed, 35 insertions(+), 4 deletions(-) -- 2.12.2 -- To unsubscribe from this list: send the line "unsubscribe dmaengine" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
