mm/hugetlb: use a folio in hugetlb_fault()
authorZhangPeng <zhangpeng362@huawei.com>
Tue, 6 Jun 2023 06:20:13 +0000 (14:20 +0800)
committerAndrew Morton <akpm@linux-foundation.org>
Mon, 19 Jun 2023 23:19:04 +0000 (16:19 -0700)
We can replace seven implicit calls to compound_head() with one by using
folio.

[akpm@linux-foundation.org: update comment, per Sidhartha]
Link: https://lkml.kernel.org/r/20230606062013.2947002-4-zhangpeng362@huawei.com
Signed-off-by: ZhangPeng <zhangpeng362@huawei.com>
Reviewed-by Sidhartha Kumar <sidhartha.kumar@oracle.com>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Muchun Song <songmuchun@bytedance.com>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Nanyong Sun <sunnanyong@huawei.com>
Cc: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
mm/hugetlb.c

index e58f800..dfa412d 100644 (file)
@@ -6062,7 +6062,7 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma,
        vm_fault_t ret;
        u32 hash;
        pgoff_t idx;
-       struct page *page = NULL;
+       struct folio *folio = NULL;
        struct folio *pagecache_folio = NULL;
        struct hstate *h = hstate_vma(vma);
        struct address_space *mapping;
@@ -6179,16 +6179,16 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma,
        /*
         * hugetlb_wp() requires page locks of pte_page(entry) and
         * pagecache_folio, so here we need take the former one
-        * when page != pagecache_folio or !pagecache_folio.
+        * when folio != pagecache_folio or !pagecache_folio.
         */
-       page = pte_page(entry);
-       if (page_folio(page) != pagecache_folio)
-               if (!trylock_page(page)) {
+       folio = page_folio(pte_page(entry));
+       if (folio != pagecache_folio)
+               if (!folio_trylock(folio)) {
                        need_wait_lock = 1;
                        goto out_ptl;
                }
 
-       get_page(page);
+       folio_get(folio);
 
        if (flags & (FAULT_FLAG_WRITE|FAULT_FLAG_UNSHARE)) {
                if (!huge_pte_write(entry)) {
@@ -6204,9 +6204,9 @@ vm_fault_t hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma,
                                                flags & FAULT_FLAG_WRITE))
                update_mmu_cache(vma, haddr, ptep);
 out_put_page:
-       if (page_folio(page) != pagecache_folio)
-               unlock_page(page);
-       put_page(page);
+       if (folio != pagecache_folio)
+               folio_unlock(folio);
+       folio_put(folio);
 out_ptl:
        spin_unlock(ptl);
 
@@ -6225,7 +6225,7 @@ out_mutex:
         * here without taking refcount.
         */
        if (need_wait_lock)
-               wait_on_page_locked(page);
+               folio_wait_locked(folio);
        return ret;
 }