| Message ID | 20230112083006.163393-3-wangkefeng.wang@huawei.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | mm: remove cgroup_throttle_swaprate() completely | expand |
On Thu, Jan 12, 2023 at 04:30:01PM +0800, Kefeng Wang wrote: > Convert do_anonymous_page() to use a folio and replace related functions > to folio functions. I think this patch has a prerequisite of sorting out alloc_zeroed_user_highpage_movable(). That way we can get rid of the 'page' variable inside this function altogether. > inc_mm_counter(vma->vm_mm, MM_ANONPAGES); > page_add_new_anon_rmap(page, vma, vmf->address); folio_add_new_anon-rmap(). > - lru_cache_add_inactive_or_unevictable(page, vma); > + folio_add_lru_vma(folio, vma);
On 2023/1/13 23:33, Matthew Wilcox wrote: > On Thu, Jan 12, 2023 at 04:30:01PM +0800, Kefeng Wang wrote: >> Convert do_anonymous_page() to use a folio and replace related functions >> to folio functions. > > I think this patch has a prerequisite of sorting out > alloc_zeroed_user_highpage_movable(). That way we can get rid of > the 'page' variable inside this function altogether. How about provide a wrapper like folio_alloc_zeroed_user_highmem_movable(), but the is a little bit long. static inline struct folio * folio_alloc_zeroed_user_highmem_movable(struct vm_area_struct *vma, unsigned long vaddr) { struct folio *folio = NULL; struct page *page; page = alloc_zeroed_user_highpage_movable(vma, vaddr); if (page) folio = page_folio(page); return folio; } > >> inc_mm_counter(vma->vm_mm, MM_ANONPAGES); >> page_add_new_anon_rmap(page, vma, vmf->address); > > folio_add_new_anon-rmap(). ok, will update. > >> - lru_cache_add_inactive_or_unevictable(page, vma); >> + folio_add_lru_vma(folio, vma);
diff --git a/mm/memory.c b/mm/memory.c index 56b571c83a0e..1cfdb0fd8d79 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -4002,6 +4002,7 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) { struct vm_area_struct *vma = vmf->vma; struct page *page; + struct folio *folio; vm_fault_t ret = 0; pte_t entry; @@ -4055,16 +4056,17 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) if (!page) goto oom; - if (mem_cgroup_charge(page_folio(page), vma->vm_mm, GFP_KERNEL)) + folio = page_folio(page); + if (mem_cgroup_charge(folio, vma->vm_mm, GFP_KERNEL)) goto oom_free_page; - cgroup_throttle_swaprate(page, GFP_KERNEL); + folio_throttle_swaprate(folio, GFP_KERNEL); /* - * The memory barrier inside __SetPageUptodate makes sure that + * The memory barrier inside __folio_mark_uptodate makes sure that * preceding stores to the page contents become visible before * the set_pte_at() write. */ - __SetPageUptodate(page); + __folio_mark_uptodate(folio); entry = mk_pte(page, vma->vm_page_prot); entry = pte_sw_mkyoung(entry); @@ -4085,13 +4087,13 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) /* Deliver the page fault to userland, check inside PT lock */ if (userfaultfd_missing(vma)) { pte_unmap_unlock(vmf->pte, vmf->ptl); - put_page(page); + folio_put(folio); return handle_userfault(vmf, VM_UFFD_MISSING); } inc_mm_counter(vma->vm_mm, MM_ANONPAGES); page_add_new_anon_rmap(page, vma, vmf->address); - lru_cache_add_inactive_or_unevictable(page, vma); + folio_add_lru_vma(folio, vma); setpte: set_pte_at(vma->vm_mm, vmf->address, vmf->pte, entry); @@ -4101,10 +4103,10 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) pte_unmap_unlock(vmf->pte, vmf->ptl); return ret; release: - put_page(page); + folio_put(folio); goto unlock; oom_free_page: - put_page(page); + folio_put(folio); oom: return VM_FAULT_OOM; }
Convert do_anonymous_page() to use a folio and replace related functions to folio functions. Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com> --- mm/memory.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-)