Re: [PATCH RFC 2/5] ARM: introduce arm_dma_direct

2019-10-15 Thread Christoph Hellwig
I think we just need to byte the bullet and move over arm to dma-direct entirely. This needs a careful audit of what differs, but the biggest item is that we need to ensure dmabounce keeps working (or is replaced with swiotlb in a suitable way, but that might be a lot more work). _

[PATCH RFC 2/5] ARM: introduce arm_dma_direct

2019-10-14 Thread Nicolas Saenz Julienne
ARM devices might use the arch's custom dma-mapping implementation or dma-direct/swiotlb depending on how the kernel is built. This is not good enough as we need to be able to control the device's DMA ops based on the specific machine configuration. Centralise control over DMA ops with arm_dma_dir