On Mon, Jul 29, 2019 at 11:57:19AM +0200, Geert Uytterhoeven wrote: > Hi Christoph, > > On Thu, Jul 25, 2019 at 8:35 AM Christoph Hellwig <hch@xxxxxx> wrote: > > Most dma_map_ops instances are IOMMUs that work perfectly fine in 32-bits > > of IOVA space, and the generic direct mapping code already provides its > > own routines that is intelligent based on the amount of memory actually > > present. Wire up the dma-direct routine for the ARM direct mapping code > > as well, and otherwise default to the constant 32-bit mask. This way > > we only need to override it for the occasional odd IOMMU that requires > > 64-bit IOVA support, or IOMMU drivers that are more efficient if they > > can fall back to the direct mapping. > > As I know you like diving into cans of worms ;-) > > Does 64-bit IOVA support actually work in general? Or only on 64-bit > platforms, due to dma_addr_t to unsigned long truncation on 32-bit? Most IOMMUs use 32-bit IOVAs, and thus we default to the 32-bit mask because it is common and failsafe vs the normal linux assumptions. However the ia64 SGI SN2 platform, and the powerpc IBM ebus implementations seem to require a 64-bit mask already, so we keep that behavior as is.
