tmpfs: simplify prealloc_page
Hugh Dickins [Tue, 26 Jul 2011 00:12:35 +0000 (17:12 -0700)]
The prealloc_page handling in shmem_getpage_gfp() is unnecessarily
complicated: first simplify that before going on to filepage/swappage.

That's right, don't report ENOMEM when the preallocation fails: we may or
may not need the page.  But simply report ENOMEM once we find we do need
it, instead of dropping lock, repeating allocation, unwinding on failure
etc.  And leave the out label on the fast path, don't goto.

Fix something that looks like a bug but turns out not to be: set
PageSwapBacked on prealloc_page before its mem_cgroup_cache_charge(), as
the removed case was doing.  That's important before adding to LRU
(determines which LRU the page goes on), and does affect which path it
takes through memcontrol.c, but in the end MEM_CGROUP_CHANGE_TYPE_ SHMEM
is handled no differently from CACHE.

Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Shaohua Li <shaohua.li@intel.com>
Cc: "Zhang, Yanmin" <yanmin.zhang@intel.com>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/shmem.c

index ff6713a..8f8534f 100644 (file)
@@ -1269,9 +1269,9 @@ repeat:
                        goto failed;
                radix_tree_preload_end();
                if (sgp != SGP_READ && !prealloc_page) {
-                       /* We don't care if this fails */
                        prealloc_page = shmem_alloc_page(gfp, info, idx);
                        if (prealloc_page) {
+                               SetPageSwapBacked(prealloc_page);
                                if (mem_cgroup_cache_charge(prealloc_page,
                                                current->mm, GFP_KERNEL)) {
                                        page_cache_release(prealloc_page);
@@ -1403,7 +1403,8 @@ repeat:
                        goto repeat;
                }
                spin_unlock(&info->lock);
-       } else {
+
+       } else if (prealloc_page) {
                shmem_swp_unmap(entry);
                sbinfo = SHMEM_SB(inode->i_sb);
                if (sbinfo->max_blocks) {
@@ -1419,41 +1420,8 @@ repeat:
                if (!filepage) {
                        int ret;
 
-                       if (!prealloc_page) {
-                               spin_unlock(&info->lock);
-                               filepage = shmem_alloc_page(gfp, info, idx);
-                               if (!filepage) {
-                                       spin_lock(&info->lock);
-                                       shmem_unacct_blocks(info->flags, 1);
-                                       shmem_free_blocks(inode, 1);
-                                       spin_unlock(&info->lock);
-                                       error = -ENOMEM;
-                                       goto failed;
-                               }
-                               SetPageSwapBacked(filepage);
-
-                               /*
-                                * Precharge page while we can wait, compensate
-                                * after
-                                */
-                               error = mem_cgroup_cache_charge(filepage,
-                                       current->mm, GFP_KERNEL);
-                               if (error) {
-                                       page_cache_release(filepage);
-                                       spin_lock(&info->lock);
-                                       shmem_unacct_blocks(info->flags, 1);
-                                       shmem_free_blocks(inode, 1);
-                                       spin_unlock(&info->lock);
-                                       filepage = NULL;
-                                       goto failed;
-                               }
-
-                               spin_lock(&info->lock);
-                       } else {
-                               filepage = prealloc_page;
-                               prealloc_page = NULL;
-                               SetPageSwapBacked(filepage);
-                       }
+                       filepage = prealloc_page;
+                       prealloc_page = NULL;
 
                        entry = shmem_swp_alloc(info, idx, sgp, gfp);
                        if (IS_ERR(entry))
@@ -1492,11 +1460,20 @@ repeat:
                SetPageUptodate(filepage);
                if (sgp == SGP_DIRTY)
                        set_page_dirty(filepage);
+       } else {
+               spin_unlock(&info->lock);
+               error = -ENOMEM;
+               goto out;
        }
 done:
        *pagep = filepage;
        error = 0;
-       goto out;
+out:
+       if (prealloc_page) {
+               mem_cgroup_uncharge_cache_page(prealloc_page);
+               page_cache_release(prealloc_page);
+       }
+       return error;
 
 nospace:
        /*
@@ -1520,12 +1497,7 @@ failed:
                unlock_page(filepage);
                page_cache_release(filepage);
        }
-out:
-       if (prealloc_page) {
-               mem_cgroup_uncharge_cache_page(prealloc_page);
-               page_cache_release(prealloc_page);
-       }
-       return error;
+       goto out;
 }
 
 static int shmem_fault(struct vm_area_struct *vma, struct vm_fault *vmf)