[$] Device-to-device memory-transfer offload with P2PDMA

One of the most common tasks carried out by device drivers is settingup DMA operations for data transfers between main memory and the device. Often,data read into memory from one device will be immediately written, unchanged,to another device. Common examples include carrying the image between thecamera and screen on a mobile phone, or downloading files to be saved on adisk. Those transfers have an impact on the CPU even if it does not use thedata directly, due to higher memory use and effects likecache trashing. There are cases where it is possible to avoid usage of thesystem memory completely, though. A patch set (posted by Logan Gunthorpe withcontributions by Christoph Hellwig and Steve Wise)has been in the works for some time that addresses this case for PCIdevices using peer-to-peer (P2P) transfers, with a focus on offering anoffload option for the NVMe fabrics target subsystem.