| Message ID | 1394732716-11507-1-git-send-email-lauraa@codeaurora.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Mar 13, 2014 at 05:45:15PM +0000, Laura Abbott wrote: > +/* vma->vm_page_prot must be set appropriately before calling this function */ > +static int __dma_common_mmap(struct device *dev, struct vm_area_struct *vma, > + void *cpu_addr, dma_addr_t dma_addr, size_t size) > +{ > + int ret = -ENXIO; > + unsigned long nr_vma_pages = (vma->vm_end - vma->vm_start) >> > + PAGE_SHIFT; > + unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; > + unsigned long pfn = dma_to_phys(dev, dma_addr) >> PAGE_SHIFT; > + unsigned long off = vma->vm_pgoff; > + > + if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret)) > + return ret; > + > + if (off < nr_pages && nr_vma_pages <= (nr_pages - off)) { > + ret = remap_pfn_range(vma, vma->vm_start, > + pfn + off, > + vma->vm_end - vma->vm_start, > + vma->vm_page_prot); > + } > + > + return ret; > +} > + > +static int arm64_swiotlb_mmap_noncoherent(struct device *dev, > + struct vm_area_struct *vma, > + void *cpu_addr, dma_addr_t dma_addr, size_t size, > + struct dma_attrs *attrs) > +{ > + /* Just use whatever page_prot attributes were specified */ > + return __dma_common_mmap(dev, vma, cpu_addr, dma_addr, size); > +} > + > +static int arm64_swiotlb_mmap_coherent(struct device *dev, > + struct vm_area_struct *vma, > + void *cpu_addr, dma_addr_t dma_addr, size_t size, > + struct dma_attrs *attrs) > +{ > + vma->vm_page_prot = pgprot_dmacoherent(vma->vm_page_prot); > + return __dma_common_mmap(dev, vma, cpu_addr, dma_addr, size); > +} So I think we got the naming the other way around. The noncoherent_dma_ops assume that the DMA is non-coherent and therefore must change the pgprot.
On 3/13/2014 10:49 AM, Catalin Marinas wrote: > On Thu, Mar 13, 2014 at 05:45:15PM +0000, Laura Abbott wrote: >> +/* vma->vm_page_prot must be set appropriately before calling this function */ >> +static int __dma_common_mmap(struct device *dev, struct vm_area_struct *vma, >> + void *cpu_addr, dma_addr_t dma_addr, size_t size) >> +{ >> + int ret = -ENXIO; >> + unsigned long nr_vma_pages = (vma->vm_end - vma->vm_start) >> >> + PAGE_SHIFT; >> + unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; >> + unsigned long pfn = dma_to_phys(dev, dma_addr) >> PAGE_SHIFT; >> + unsigned long off = vma->vm_pgoff; >> + >> + if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret)) >> + return ret; >> + >> + if (off < nr_pages && nr_vma_pages <= (nr_pages - off)) { >> + ret = remap_pfn_range(vma, vma->vm_start, >> + pfn + off, >> + vma->vm_end - vma->vm_start, >> + vma->vm_page_prot); >> + } >> + >> + return ret; >> +} >> + >> +static int arm64_swiotlb_mmap_noncoherent(struct device *dev, >> + struct vm_area_struct *vma, >> + void *cpu_addr, dma_addr_t dma_addr, size_t size, >> + struct dma_attrs *attrs) >> +{ >> + /* Just use whatever page_prot attributes were specified */ >> + return __dma_common_mmap(dev, vma, cpu_addr, dma_addr, size); >> +} >> + >> +static int arm64_swiotlb_mmap_coherent(struct device *dev, >> + struct vm_area_struct *vma, >> + void *cpu_addr, dma_addr_t dma_addr, size_t size, >> + struct dma_attrs *attrs) >> +{ >> + vma->vm_page_prot = pgprot_dmacoherent(vma->vm_page_prot); >> + return __dma_common_mmap(dev, vma, cpu_addr, dma_addr, size); >> +} > > So I think we got the naming the other way around. The > noncoherent_dma_ops assume that the DMA is non-coherent and therefore > must change the pgprot. I spent way too long convincing myself this was right and I was still wrong. I'll fix it up for v2 (I have one more patch I need to put in for CMA integration as well) Thanks, Laura
diff --git a/arch/arm64/mm/dma-mapping.c b/arch/arm64/mm/dma-mapping.c index e923a5b..9a639bf 100644 --- a/arch/arm64/mm/dma-mapping.c +++ b/arch/arm64/mm/dma-mapping.c @@ -194,9 +194,52 @@ static void arm64_swiotlb_sync_sg_for_device(struct device *dev, sg->length, dir); } +/* vma->vm_page_prot must be set appropriately before calling this function */ +static int __dma_common_mmap(struct device *dev, struct vm_area_struct *vma, + void *cpu_addr, dma_addr_t dma_addr, size_t size) +{ + int ret = -ENXIO; + unsigned long nr_vma_pages = (vma->vm_end - vma->vm_start) >> + PAGE_SHIFT; + unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; + unsigned long pfn = dma_to_phys(dev, dma_addr) >> PAGE_SHIFT; + unsigned long off = vma->vm_pgoff; + + if (dma_mmap_from_coherent(dev, vma, cpu_addr, size, &ret)) + return ret; + + if (off < nr_pages && nr_vma_pages <= (nr_pages - off)) { + ret = remap_pfn_range(vma, vma->vm_start, + pfn + off, + vma->vm_end - vma->vm_start, + vma->vm_page_prot); + } + + return ret; +} + +static int arm64_swiotlb_mmap_noncoherent(struct device *dev, + struct vm_area_struct *vma, + void *cpu_addr, dma_addr_t dma_addr, size_t size, + struct dma_attrs *attrs) +{ + /* Just use whatever page_prot attributes were specified */ + return __dma_common_mmap(dev, vma, cpu_addr, dma_addr, size); +} + +static int arm64_swiotlb_mmap_coherent(struct device *dev, + struct vm_area_struct *vma, + void *cpu_addr, dma_addr_t dma_addr, size_t size, + struct dma_attrs *attrs) +{ + vma->vm_page_prot = pgprot_dmacoherent(vma->vm_page_prot); + return __dma_common_mmap(dev, vma, cpu_addr, dma_addr, size); +} + struct dma_map_ops noncoherent_swiotlb_dma_ops = { .alloc = arm64_swiotlb_alloc_noncoherent, .free = arm64_swiotlb_free_noncoherent, + .mmap = arm64_swiotlb_mmap_noncoherent, .map_page = arm64_swiotlb_map_page, .unmap_page = arm64_swiotlb_unmap_page, .map_sg = arm64_swiotlb_map_sg_attrs, @@ -213,6 +256,7 @@ EXPORT_SYMBOL(noncoherent_swiotlb_dma_ops); struct dma_map_ops coherent_swiotlb_dma_ops = { .alloc = arm64_swiotlb_alloc_coherent, .free = arm64_swiotlb_free_coherent, + .mmap = arm64_swiotlb_mmap_coherent, .map_page = swiotlb_map_page, .unmap_page = swiotlb_unmap_page, .map_sg = swiotlb_map_sg_attrs,
The current dma_ops do not specify an mmap function so maping falls back to the default implementation. There are at least two issues with using the default implementation: 1) The pgprot is always pgprot_noncached (strongly ordered) memory even with coherent operations 2) dma_common_mmap calls virt_to_page on the remapped non-coherent address which leads to invalid memory being mapped. Fix both these issue by implementing a custom mmap function which correctly accounts for remapped addresses and sets vm_pg_prot appropriately. Signed-off-by: Laura Abbott <lauraa@codeaurora.org> --- arch/arm64/mm/dma-mapping.c | 44 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 44 insertions(+)