| Message ID | 20210107044401.19828-7-zhukeqian1@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | vfio/iommu_type1: Some optimizations about dirty tracking | expand |
On Thu, 7 Jan 2021 12:44:01 +0800 Keqian Zhu <zhukeqian1@huawei.com> wrote: > We always use the smallest supported page size of vfio_iommu as > pgsize. Drop parameter "pgsize" of update_user_bitmap. > > Signed-off-by: Keqian Zhu <zhukeqian1@huawei.com> > --- > drivers/vfio/vfio_iommu_type1.c | 9 ++++----- > 1 file changed, 4 insertions(+), 5 deletions(-) > > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > index 82649a040148..bceda5e8baaa 100644 > --- a/drivers/vfio/vfio_iommu_type1.c > +++ b/drivers/vfio/vfio_iommu_type1.c > @@ -978,10 +978,9 @@ static void vfio_update_pgsize_bitmap(struct vfio_iommu *iommu) > } > > static int update_user_bitmap(u64 __user *bitmap, struct vfio_iommu *iommu, > - struct vfio_dma *dma, dma_addr_t base_iova, > - size_t pgsize) > + struct vfio_dma *dma, dma_addr_t base_iova) > { > - unsigned long pgshift = __ffs(pgsize); > + unsigned long pgshift = __ffs(iommu->pgsize_bitmap); > unsigned long nbits = dma->size >> pgshift; > unsigned long bit_offset = (dma->iova - base_iova) >> pgshift; > unsigned long copy_offset = bit_offset / BITS_PER_LONG; > @@ -1046,7 +1045,7 @@ static int vfio_iova_dirty_bitmap(u64 __user *bitmap, struct vfio_iommu *iommu, > if (dma->iova > iova + size - 1) > break; > > - ret = update_user_bitmap(bitmap, iommu, dma, iova, pgsize); > + ret = update_user_bitmap(bitmap, iommu, dma, iova); > if (ret) > return ret; > > @@ -1192,7 +1191,7 @@ static int vfio_dma_do_unmap(struct vfio_iommu *iommu, > > if (unmap->flags & VFIO_DMA_UNMAP_FLAG_GET_DIRTY_BITMAP) { > ret = update_user_bitmap(bitmap->data, iommu, dma, > - unmap->iova, pgsize); > + unmap->iova); > if (ret) > break; > } Same as the previous, both call sites already have both pgsize and pgshift, pass both rather than recalculate. Thanks, Alex
On 2021/1/16 7:44, Alex Williamson wrote: > On Thu, 7 Jan 2021 12:44:01 +0800 > Keqian Zhu <zhukeqian1@huawei.com> wrote: > >> We always use the smallest supported page size of vfio_iommu as >> pgsize. Drop parameter "pgsize" of update_user_bitmap. >> >> Signed-off-by: Keqian Zhu <zhukeqian1@huawei.com> >> --- >> drivers/vfio/vfio_iommu_type1.c | 9 ++++----- >> 1 file changed, 4 insertions(+), 5 deletions(-) >> >> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c >> index 82649a040148..bceda5e8baaa 100644 >> --- a/drivers/vfio/vfio_iommu_type1.c >> +++ b/drivers/vfio/vfio_iommu_type1.c >> @@ -978,10 +978,9 @@ static void vfio_update_pgsize_bitmap(struct vfio_iommu *iommu) >> } >> >> static int update_user_bitmap(u64 __user *bitmap, struct vfio_iommu *iommu, >> - struct vfio_dma *dma, dma_addr_t base_iova, >> - size_t pgsize) >> + struct vfio_dma *dma, dma_addr_t base_iova) >> { >> - unsigned long pgshift = __ffs(pgsize); >> + unsigned long pgshift = __ffs(iommu->pgsize_bitmap); >> unsigned long nbits = dma->size >> pgshift; >> unsigned long bit_offset = (dma->iova - base_iova) >> pgshift; >> unsigned long copy_offset = bit_offset / BITS_PER_LONG; >> @@ -1046,7 +1045,7 @@ static int vfio_iova_dirty_bitmap(u64 __user *bitmap, struct vfio_iommu *iommu, >> if (dma->iova > iova + size - 1) >> break; >> >> - ret = update_user_bitmap(bitmap, iommu, dma, iova, pgsize); >> + ret = update_user_bitmap(bitmap, iommu, dma, iova); >> if (ret) >> return ret; >> >> @@ -1192,7 +1191,7 @@ static int vfio_dma_do_unmap(struct vfio_iommu *iommu, >> >> if (unmap->flags & VFIO_DMA_UNMAP_FLAG_GET_DIRTY_BITMAP) { >> ret = update_user_bitmap(bitmap->data, iommu, dma, >> - unmap->iova, pgsize); >> + unmap->iova); >> if (ret) >> break; >> } > > Same as the previous, both call sites already have both pgsize and > pgshift, pass both rather than recalculate. Thanks, > My idea is that the "pgsize" parameter goes here and there, disturbs our sight when read code. To be frankly, the recalculate is negligible. Or we can add new fields in vfio_iommu, which are updated in vfio_update_pgsize_bitmap(). Thanks, Keqian > Alex > > . >
diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c index 82649a040148..bceda5e8baaa 100644 --- a/drivers/vfio/vfio_iommu_type1.c +++ b/drivers/vfio/vfio_iommu_type1.c @@ -978,10 +978,9 @@ static void vfio_update_pgsize_bitmap(struct vfio_iommu *iommu) } static int update_user_bitmap(u64 __user *bitmap, struct vfio_iommu *iommu, - struct vfio_dma *dma, dma_addr_t base_iova, - size_t pgsize) + struct vfio_dma *dma, dma_addr_t base_iova) { - unsigned long pgshift = __ffs(pgsize); + unsigned long pgshift = __ffs(iommu->pgsize_bitmap); unsigned long nbits = dma->size >> pgshift; unsigned long bit_offset = (dma->iova - base_iova) >> pgshift; unsigned long copy_offset = bit_offset / BITS_PER_LONG; @@ -1046,7 +1045,7 @@ static int vfio_iova_dirty_bitmap(u64 __user *bitmap, struct vfio_iommu *iommu, if (dma->iova > iova + size - 1) break; - ret = update_user_bitmap(bitmap, iommu, dma, iova, pgsize); + ret = update_user_bitmap(bitmap, iommu, dma, iova); if (ret) return ret; @@ -1192,7 +1191,7 @@ static int vfio_dma_do_unmap(struct vfio_iommu *iommu, if (unmap->flags & VFIO_DMA_UNMAP_FLAG_GET_DIRTY_BITMAP) { ret = update_user_bitmap(bitmap->data, iommu, dma, - unmap->iova, pgsize); + unmap->iova); if (ret) break; }
We always use the smallest supported page size of vfio_iommu as pgsize. Drop parameter "pgsize" of update_user_bitmap. Signed-off-by: Keqian Zhu <zhukeqian1@huawei.com> --- drivers/vfio/vfio_iommu_type1.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)