On 05/18/2018 06:23 AM, hch at lst.de wrote: > Fri, May 18, 2018 at 01:03:46PM +0000, Alexey Brodkin wrote: >> Note mmc_get_dma_dir() is just "data->flags & MMC_DATA_WRITE ? DMA_TO_DEVICE : DMA_FROM_DEVICE". >> I.e. if we're preparing for sending data dma_noncoherent_map_sg() will have DMA_TO_DEVICE which >> is quite OK for passing to dma_noncoherent_sync_sg_for_device() but in case of reading we'll have >> DMA_FROM_DEVICE which we'll pass to dma_noncoherent_sync_sg_for_device() in dma_noncoherent_map_sg(). >> >> I'd say this is not entirely correct because IMHO arch_sync_dma_for_cpu() is supposed to only be used >> in case of DMA_FROM_DEVICE and arch_sync_dma_for_device() only in case of DMA_TO_DEVICE. > arc overrides the dir paramter of the dma_sync_single_for_device/ > dma_sync_single_for_cpu calls. My patches dropped that, and I have > restored that, and audit for the other architectures is pending. Right, for now lets retain that and do a sweeping audit of @direction - to me it seems extraneous (as it did 10 years ago), but I'm not an expert in this are so perhaps it is needed for some device / arches and it would be good to understand that finally. > That being said the existing arc code still looks rather odd as it > didn't do the same thing for the scatterlist versions of the calls. > I've thrown in a few patches into my new tree to make the sg versions > make the normal calls, and to clean up the area a bit. Not calling names or anything here, but it doesn't exist for sg variants, because I didn't write that code :-) It was introduced by your commi: 2016-01-20 052c96dbe33b arc: convert to dma_map_ops
