On Tue, Feb 05, 2019 at 03:02:23PM +0000, Robin Murphy wrote: > On 14/01/2019 09:41, Christoph Hellwig wrote: >> The current iommu_dma_mmap code does not properly handle memory from the >> page allocator that hasn't been remapped, which can happen in the rare >> case of allocations for a coherent device that aren't allowed to block. >> >> Fix this by replacing iommu_dma_mmap with a slightly tweaked copy of >> dma_common_mmap with special handling for the remapped array of >> pages allocated from __iommu_dma_alloc. > > If there's an actual bugfix here, can we make that before all of the other > code movement? If it's at all related to other reports of weird mmap > behaviour it might warrant backporting, and either way I'm finding it > needlessly tough to follow what's going on in this patch :(
The bug fix is to handle non-vmalloc pages. I'll see if I can do a smaller and more bandaid-y fix first.
