On Mon, 19 May 2025 10:07:24 -0600 Alex Williamson <alex.williamson@xxxxxxxxxx> wrote: > On Mon, 19 May 2025 15:04:19 +0800 > lizhe.67@xxxxxxxxxxxxx wrote: > > > From: Li Zhe <lizhe.67@xxxxxxxxxxxxx> > > > > When vfio_pin_pages_remote() is called with a range of addresses that > > includes hugetlbfs folios, the function currently performs individual > > statistics counting operations for each page. This can lead to significant > > performance overheads, especially when dealing with large ranges of pages. > > > > This patch optimize this process by batching the statistics counting > > operations. > > > > The performance test results for completing the 8G VFIO IOMMU DMA mapping, > > obtained through trace-cmd, are as follows. In this case, the 8G virtual > > address space has been mapped to physical memory using hugetlbfs with > > pagesize=2M. > > > > Before this patch: > > funcgraph_entry: # 33813.703 us | vfio_pin_map_dma(); > > > > After this patch: > > funcgraph_entry: # 15635.055 us | vfio_pin_map_dma(); > > > > Signed-off-by: Li Zhe <lizhe.67@xxxxxxxxxxxxx> > > --- > > Changelogs: > > > > v1->v2: > > - Fix some issues in comments and formatting. > > - Consolidate vfio_find_vpfn_range() and vfio_find_vpfn(). > > - Move the processing logic for hugetlbfs folio into the while(true) loop > > and use a variable with a default value of 1 to indicate the number of > > consecutive pages. > > > > v1 patch: https://lore.kernel.org/all/20250513035730.96387-1-lizhe.67@xxxxxxxxxxxxx/ > > > > drivers/vfio/vfio_iommu_type1.c | 70 +++++++++++++++++++++++++++------ > > 1 file changed, 58 insertions(+), 12 deletions(-) > > > > diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c > > index 0ac56072af9f..2218ca415366 100644 > > --- a/drivers/vfio/vfio_iommu_type1.c > > +++ b/drivers/vfio/vfio_iommu_type1.c > > @@ -317,17 +317,20 @@ static void vfio_dma_bitmap_free_all(struct vfio_iommu *iommu) > > } > > > > /* > > - * Helper Functions for host iova-pfn list > > + * Find the first vfio_pfn that overlapping the range > > + * [iova, iova + PAGE_SIZE * npage) in rb tree > > */ > > -static struct vfio_pfn *vfio_find_vpfn(struct vfio_dma *dma, dma_addr_t iova) > > +static struct vfio_pfn *vfio_find_vpfn_range(struct vfio_dma *dma, > > + dma_addr_t iova, unsigned long npage) > > { > > struct vfio_pfn *vpfn; > > struct rb_node *node = dma->pfn_list.rb_node; > > + dma_addr_t end_iova = iova + PAGE_SIZE * npage; > > > > while (node) { > > vpfn = rb_entry(node, struct vfio_pfn, node); > > > > - if (iova < vpfn->iova) > > + if (end_iova <= vpfn->iova) > > node = node->rb_left; > > else if (iova > vpfn->iova) > > node = node->rb_right; > > @@ -337,6 +340,14 @@ static struct vfio_pfn *vfio_find_vpfn(struct vfio_dma *dma, dma_addr_t iova) > > return NULL; > > } > > > > +/* > > + * Helper Functions for host iova-pfn list > > + */ > > This comment should still precede the renamed function above, it's in > reference to this section of code related to searching, inserting, and > removing entries from the pfn list. > > > +static inline struct vfio_pfn *vfio_find_vpfn(struct vfio_dma *dma, dma_addr_t iova) > > +{ > > + return vfio_find_vpfn_range(dma, iova, 1); > > +} > > + > > static void vfio_link_pfn(struct vfio_dma *dma, > > struct vfio_pfn *new) > > { > > @@ -681,32 +692,67 @@ static long vfio_pin_pages_remote(struct vfio_dma *dma, unsigned long vaddr, > > * and rsvd here, and therefore continues to use the batch. > > */ > > while (true) { > > + int page_step = 1; > > + long lock_acct_step = 1; > > + struct folio *folio = page_folio(batch->pages[batch->offset]); > > + bool found_vpfn; > > + > > if (pfn != *pfn_base + pinned || > > rsvd != is_invalid_reserved_pfn(pfn)) > > goto out; > > > > + /* Handle hugetlbfs page */ > > + if (folio_test_hugetlb(folio)) { > > Why do we care to specifically test for hugetlb vs > folio_large_nr_pages(), at which point we can just use folio_nr_pages() > directly here. > > > + unsigned long start_pfn = PHYS_PFN(vaddr); > > Using this macro on a vaddr looks wrong. > > > + > > + /* > > + * Note: The current page_step does not achieve the optimal > > + * performance in scenarios where folio_nr_pages() exceeds > > + * batch->capacity. It is anticipated that future enhancements > > + * will address this limitation. > > + */ > > + page_step = min(batch->size, > > + ALIGN(start_pfn + 1, folio_nr_pages(folio)) - start_pfn); > > Why do we assume start_pfn is the beginning of the folio? > > > + found_vpfn = !!vfio_find_vpfn_range(dma, iova, page_step); > > + if (rsvd || !found_vpfn) { > > + lock_acct_step = page_step; > > + } else { > > + dma_addr_t tmp_iova = iova; > > + int i; > > + > > + lock_acct_step = 0; > > + for (i = 0; i < page_step; ++i, tmp_iova += PAGE_SIZE) > > + if (!vfio_find_vpfn(dma, tmp_iova)) > > + lock_acct_step++; > > + if (lock_acct_step) > > + found_vpfn = false; > > Why are we making this so complicated versus falling back to iterating > at page per page? > > > + } > > + } else { > > + found_vpfn = vfio_find_vpfn(dma, iova); > > + } > > + > > /* > > * Reserved pages aren't counted against the user, > > * externally pinned pages are already counted against > > * the user. > > */ > > - if (!rsvd && !vfio_find_vpfn(dma, iova)) { > > + if (!rsvd && !found_vpfn) { > > if (!dma->lock_cap && > > - mm->locked_vm + lock_acct + 1 > limit) { > > + mm->locked_vm + lock_acct + lock_acct_step > limit) { > > pr_warn("%s: RLIMIT_MEMLOCK (%ld) exceeded\n", > > __func__, limit << PAGE_SHIFT); > > ret = -ENOMEM; > > goto unpin_out; > > } > > - lock_acct++; > > + lock_acct += lock_acct_step; > > } > > > > - pinned++; > > - npage--; > > - vaddr += PAGE_SIZE; > > - iova += PAGE_SIZE; > > - batch->offset++; > > - batch->size--; > > + pinned += page_step; > > + npage -= page_step; > > + vaddr += PAGE_SIZE * page_step; > > + iova += PAGE_SIZE * page_step; > > + batch->offset += page_step; > > + batch->size -= page_step; > > > > if (!batch->size) > > break; > > Why is something like below (untested) not sufficient? > > NB. (vaddr - folio_address()) still needs some scrutiny to determine if > it's valid. > > @@ -681,32 +692,40 @@ static long vfio_pin_pages_remote(struct vfio_dma *dma, unsigned long vaddr, > * and rsvd here, and therefore continues to use the batch. > */ > while (true) { > + struct folio *folio = page_folio(batch->pages[batch->offset]); > + long nr_pages; > + > if (pfn != *pfn_base + pinned || > rsvd != is_invalid_reserved_pfn(pfn)) > goto out; > > + nr_pages = min(batch->size, folio_nr_pages(folio) - > + (vaddr - folio_address(folio)) >> PAGE_SHIFT); folio_nr_pages(folio) - folio_page_idx(folio, batch->pages[batch->offset]) might be a better option here. Thanks, Alex > + if (nr_pages > 1 && vfio_find_vpfn_range(dma, iova, nr_pages)) > + nr_pages = 1; > + > /* > * Reserved pages aren't counted against the user, > * externally pinned pages are already counted against > * the user. > */ > - if (!rsvd && !vfio_find_vpfn(dma, iova)) { > + if (!rsvd && (nr_pages > 1 || !vfio_find_vpfn(dma, iova))) { > if (!dma->lock_cap && > - mm->locked_vm + lock_acct + 1 > limit) { > + mm->locked_vm + lock_acct + nr_pages > limit) { > pr_warn("%s: RLIMIT_MEMLOCK (%ld) exceeded\n", > __func__, limit << PAGE_SHIFT); > ret = -ENOMEM; > goto unpin_out; > } > - lock_acct++; > + lock_acct += nr_pages; > } > > - pinned++; > - npage--; > - vaddr += PAGE_SIZE; > - iova += PAGE_SIZE; > - batch->offset++; > - batch->size--; > + pinned += nr_pages; > + npage -= nr_pages; > + vaddr += PAGE_SIZE * nr_pages; > + iova += PAGE_SIZE * nr_pages; > + batch->offset += nr_pages; > + batch->size -= nr_pages; > > if (!batch->size) > break;
