| Message ID | 20231128204938.1453583-10-pasha.tatashin@soleen.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | IOMMU memory observability | expand |
On Tue, Nov 28, 2023 at 08:49:31PM +0000, Pasha Tatashin wrote: > Convert iommu/iommufd/* files to use the new page allocation functions > provided in iommu-pages.h. > > Signed-off-by: Pasha Tatashin <pasha.tatashin@soleen.com> > --- > drivers/iommu/iommufd/iova_bitmap.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) This is a short term allocation, it should not be counted, that is why it is already not using GFP_KERNEL_ACCOUNT. Jason
On Tue, Nov 28, 2023 at 6:52 PM Jason Gunthorpe <jgg@ziepe.ca> wrote: > > On Tue, Nov 28, 2023 at 08:49:31PM +0000, Pasha Tatashin wrote: > > Convert iommu/iommufd/* files to use the new page allocation functions > > provided in iommu-pages.h. > > > > Signed-off-by: Pasha Tatashin <pasha.tatashin@soleen.com> > > --- > > drivers/iommu/iommufd/iova_bitmap.c | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > > This is a short term allocation, it should not be counted, that is why > it is already not using GFP_KERNEL_ACCOUNT. I made this change for completeness. I changed all calls to get_free_page/alloc_page etc under driver/iommu to use the iommu_alloc_* variants, this also helps future developers in this area to use the right allocation functions. The accounting is implemented using cheap per-cpu counters, so should not affect the performance, I think it is OK to keep them here.
On Wed, Nov 29, 2023 at 04:59:43PM -0500, Pasha Tatashin wrote: > On Tue, Nov 28, 2023 at 6:52 PM Jason Gunthorpe <jgg@ziepe.ca> wrote: > > > > On Tue, Nov 28, 2023 at 08:49:31PM +0000, Pasha Tatashin wrote: > > > Convert iommu/iommufd/* files to use the new page allocation functions > > > provided in iommu-pages.h. > > > > > > Signed-off-by: Pasha Tatashin <pasha.tatashin@soleen.com> > > > --- > > > drivers/iommu/iommufd/iova_bitmap.c | 4 ++-- > > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > This is a short term allocation, it should not be counted, that is why > > it is already not using GFP_KERNEL_ACCOUNT. > > I made this change for completeness. I changed all calls to > get_free_page/alloc_page etc under driver/iommu to use the > iommu_alloc_* variants, this also helps future developers in this area > to use the right allocation functions. > The accounting is implemented using cheap per-cpu counters, so should > not affect the performance, I think it is OK to keep them here. Except it is a mis use of an API that should only be used for page table memory :( Jason
diff --git a/drivers/iommu/iommufd/iova_bitmap.c b/drivers/iommu/iommufd/iova_bitmap.c index 0a92c9eeaf7f..6b8575b93f17 100644 --- a/drivers/iommu/iommufd/iova_bitmap.c +++ b/drivers/iommu/iommufd/iova_bitmap.c @@ -253,7 +253,7 @@ struct iova_bitmap *iova_bitmap_alloc(unsigned long iova, size_t length, bitmap->iova = iova; bitmap->length = length; mapped->iova = iova; - mapped->pages = (struct page **)__get_free_page(GFP_KERNEL); + mapped->pages = iommu_alloc_page(GFP_KERNEL); if (!mapped->pages) { rc = -ENOMEM; goto err; @@ -284,7 +284,7 @@ void iova_bitmap_free(struct iova_bitmap *bitmap) iova_bitmap_put(bitmap); if (mapped->pages) { - free_page((unsigned long)mapped->pages); + iommu_free_page(mapped->pages); mapped->pages = NULL; }
Convert iommu/iommufd/* files to use the new page allocation functions provided in iommu-pages.h. Signed-off-by: Pasha Tatashin <pasha.tatashin@soleen.com> --- drivers/iommu/iommufd/iova_bitmap.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)