On Fri, 2019-10-18 at 11:13 +0100, Russell King - ARM Linux admin wrote: > On Thu, Oct 17, 2019 at 01:16:00PM +1100, Benjamin Herrenschmidt wrote: > > On Tue, 2019-10-15 at 15:44 +0200, Ulf Hansson wrote: > > > > Hmm, so it looks like PowerPC doesn't mark devices that are dma > > > > coherent with a property that describes them as such. > > > > > > > > I think this opens a wider question - what should > > > > of_dma_is_coherent() > > > > return for PowerPC? It seems right now that it returns false for > > > > devices that are DMA coherent, which seems to me to be a recipe for > > > > future mistakes. > > > > > > Perhaps implement the arch_setup_dma_ops() for PPC, that set > > > "dev->dma_coherent = true" could work? > > > > Only for coherent ops :) > > For those of us who have never touched the PowerPC code before, any > suggestion where you would like to see that? No file in arch/powerpc/mm > stands out as a place for DMA stuff (and dma-noncoherent.c is certainly > not correct.) Anywhere is fine but I think it's easy. Coherent DMA is for historical reasons a function of the processor generation, and as such a CONFIG option. So setting it according to CONFIG_NOT_COHERENT_CACHE will probably work just fine. Well at least I think so ... unless I'm missing some broken HW somewhere I am not aware of. Cheers, Ben.
