| Message ID | 20190422175942.18788-3-hch@lst.de (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [01/26] arm64/iommu: handle non-remapped addresses in ->mmap and ->get_sgtable | expand |
On 22/04/2019 18:59, Christoph Hellwig wrote: > The nr_pages checks should be done for all mmap requests, not just those > using remap_pfn_range. I think it probably makes sense now to just squash this with #22 one way or the other, but if you really really still want to keep it as a separate patch with a misleading commit message then I'm willing to keep my complaints to myself :) Robin. > Signed-off-by: Christoph Hellwig <hch@lst.de> > --- > arch/arm64/mm/dma-mapping.c | 21 ++++++++------------- > 1 file changed, 8 insertions(+), 13 deletions(-) > > diff --git a/arch/arm64/mm/dma-mapping.c b/arch/arm64/mm/dma-mapping.c > index 674860e3e478..604c638b2787 100644 > --- a/arch/arm64/mm/dma-mapping.c > +++ b/arch/arm64/mm/dma-mapping.c > @@ -73,19 +73,9 @@ static int __swiotlb_get_sgtable_page(struct sg_table *sgt, > static int __swiotlb_mmap_pfn(struct vm_area_struct *vma, > unsigned long pfn, size_t size) > { > - int ret = -ENXIO; > - unsigned long nr_vma_pages = vma_pages(vma); > - unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; > - unsigned long off = vma->vm_pgoff; > - > - 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; > + return remap_pfn_range(vma, vma->vm_start, pfn + vma->vm_pgoff, > + vma->vm_end - vma->vm_start, > + vma->vm_page_prot); > } > #endif /* CONFIG_IOMMU_DMA */ > > @@ -241,6 +231,8 @@ static int __iommu_mmap_attrs(struct device *dev, struct vm_area_struct *vma, > void *cpu_addr, dma_addr_t dma_addr, size_t size, > unsigned long attrs) > { > + unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; > + unsigned long off = vma->vm_pgoff; > struct vm_struct *area; > int ret; > > @@ -249,6 +241,9 @@ static int __iommu_mmap_attrs(struct device *dev, struct vm_area_struct *vma, > if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret)) > return ret; > > + if (off >= nr_pages || vma_pages(vma) > nr_pages - off) > + return -ENXIO; > + > if (!is_vmalloc_addr(cpu_addr)) { > unsigned long pfn = page_to_pfn(virt_to_page(cpu_addr)); > return __swiotlb_mmap_pfn(vma, pfn, size); >
On Mon, Apr 29, 2019 at 01:35:46PM +0100, Robin Murphy wrote: > On 22/04/2019 18:59, Christoph Hellwig wrote: >> The nr_pages checks should be done for all mmap requests, not just those >> using remap_pfn_range. > > I think it probably makes sense now to just squash this with #22 one way or > the other, but if you really really still want to keep it as a separate > patch with a misleading commit message then I'm willing to keep my > complaints to myself :) Well, I split this out in response to your earlier comments, so if you prefer it squasheѕ back in I can do that..
On 29/04/2019 20:01, Christoph Hellwig wrote: > On Mon, Apr 29, 2019 at 01:35:46PM +0100, Robin Murphy wrote: >> On 22/04/2019 18:59, Christoph Hellwig wrote: >>> The nr_pages checks should be done for all mmap requests, not just those >>> using remap_pfn_range. >> >> I think it probably makes sense now to just squash this with #22 one way or >> the other, but if you really really still want to keep it as a separate >> patch with a misleading commit message then I'm willing to keep my >> complaints to myself :) > > Well, I split this out in response to your earlier comments, so if you > prefer it squasheѕ back in I can do that.. AFAICS I only ever suggested splitting the original "fix and refactor" commit into the fix (patch #1) and the refactor - I think we've just ended up adding more "refactor" on top in the evolution of the series :) Robin.
diff --git a/arch/arm64/mm/dma-mapping.c b/arch/arm64/mm/dma-mapping.c index 674860e3e478..604c638b2787 100644 --- a/arch/arm64/mm/dma-mapping.c +++ b/arch/arm64/mm/dma-mapping.c @@ -73,19 +73,9 @@ static int __swiotlb_get_sgtable_page(struct sg_table *sgt, static int __swiotlb_mmap_pfn(struct vm_area_struct *vma, unsigned long pfn, size_t size) { - int ret = -ENXIO; - unsigned long nr_vma_pages = vma_pages(vma); - unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; - unsigned long off = vma->vm_pgoff; - - 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; + return remap_pfn_range(vma, vma->vm_start, pfn + vma->vm_pgoff, + vma->vm_end - vma->vm_start, + vma->vm_page_prot); } #endif /* CONFIG_IOMMU_DMA */ @@ -241,6 +231,8 @@ static int __iommu_mmap_attrs(struct device *dev, struct vm_area_struct *vma, void *cpu_addr, dma_addr_t dma_addr, size_t size, unsigned long attrs) { + unsigned long nr_pages = PAGE_ALIGN(size) >> PAGE_SHIFT; + unsigned long off = vma->vm_pgoff; struct vm_struct *area; int ret; @@ -249,6 +241,9 @@ static int __iommu_mmap_attrs(struct device *dev, struct vm_area_struct *vma, if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret)) return ret; + if (off >= nr_pages || vma_pages(vma) > nr_pages - off) + return -ENXIO; + if (!is_vmalloc_addr(cpu_addr)) { unsigned long pfn = page_to_pfn(virt_to_page(cpu_addr)); return __swiotlb_mmap_pfn(vma, pfn, size);
The nr_pages checks should be done for all mmap requests, not just those using remap_pfn_range. Signed-off-by: Christoph Hellwig <hch@lst.de> --- arch/arm64/mm/dma-mapping.c | 21 ++++++++------------- 1 file changed, 8 insertions(+), 13 deletions(-)