The patch titled
mm: add_to_swap_cache() does not return -EEXIST
has been added to the -mm tree. Its filename is
mm-add_to_swap_cache-does-not-return-eexist.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
out what to do about this
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: mm: add_to_swap_cache() does not return -EEXIST
From: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
After commit 355cfa73 ("mm: modify swap_map and add SWAP_HAS_CACHE flag"),
only the context which have set SWAP_HAS_CACHE flag by swapcache_prepare()
or get_swap_page() would call add_to_swap_cache(). So add_to_swap_cache()
doesn't return -EEXIST any more.
Even though it doesn't return -EEXIST, it's not good behavior conceptually
to call swapcache_prepare() in the -EEXIST case, because it means clearing
SWAP_HAS_CACHE flag while the entry is on swap cache.
This patch removes redundant codes and comments from callers of it, and
adds VM_BUG_ON() in error path of add_to_swap_cache() and some comments.
Signed-off-by: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Cc: Balbir Singh <balbir@xxxxxxxxxxxxxxxxxx>
Cc: Hugh Dickins <hugh.dickins@xxxxxxxxxxxxx>
Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/shmem.c | 4 ++
mm/swap_state.c | 73 ++++++++++++++++++++++------------------------
2 files changed, 40 insertions(+), 37 deletions(-)
diff -puN mm/shmem.c~mm-add_to_swap_cache-does-not-return-eexist mm/shmem.c
--- a/mm/shmem.c~mm-add_to_swap_cache-does-not-return-eexist
+++ a/mm/shmem.c
@@ -1097,6 +1097,10 @@ static int shmem_writepage(struct page *
shmem_swp_unmap(entry);
unlock:
spin_unlock(&info->lock);
+ /*
+ * add_to_swap_cache() doesn't return -EEXIST, so we can safely
+ * clear SWAP_HAS_CACHE flag.
+ */
swapcache_free(swap, NULL);
redirty:
set_page_dirty(page);
diff -puN mm/swap_state.c~mm-add_to_swap_cache-does-not-return-eexist mm/swap_state.c
--- a/mm/swap_state.c~mm-add_to_swap_cache-does-not-return-eexist
+++ a/mm/swap_state.c
@@ -96,6 +96,12 @@ int add_to_swap_cache(struct page *page,
radix_tree_preload_end();
if (unlikely(error)) {
+ /*
+ * Only the context which have set SWAP_HAS_CACHE flag
+ * would call add_to_swap_cache().
+ * So add_to_swap_cache() doesn't returns -EEXIST.
+ */
+ VM_BUG_ON(error == -EEXIST);
set_page_private(page, 0UL);
ClearPageSwapCache(page);
page_cache_release(page);
@@ -137,38 +143,34 @@ int add_to_swap(struct page *page)
VM_BUG_ON(!PageLocked(page));
VM_BUG_ON(!PageUptodate(page));
- for (;;) {
- entry = get_swap_page();
- if (!entry.val)
- return 0;
+ entry = get_swap_page();
+ if (!entry.val)
+ return 0;
+ /*
+ * Radix-tree node allocations from PF_MEMALLOC contexts could
+ * completely exhaust the page allocator. __GFP_NOMEMALLOC
+ * stops emergency reserves from being allocated.
+ *
+ * TODO: this could cause a theoretical memory reclaim
+ * deadlock in the swap out path.
+ */
+ /*
+ * Add it to the swap cache and mark it dirty
+ */
+ err = add_to_swap_cache(page, entry,
+ __GFP_HIGH|__GFP_NOMEMALLOC|__GFP_NOWARN);
+
+ if (!err) { /* Success */
+ SetPageDirty(page);
+ return 1;
+ } else { /* -ENOMEM radix-tree allocation failure */
/*
- * Radix-tree node allocations from PF_MEMALLOC contexts could
- * completely exhaust the page allocator. __GFP_NOMEMALLOC
- * stops emergency reserves from being allocated.
- *
- * TODO: this could cause a theoretical memory reclaim
- * deadlock in the swap out path.
- */
- /*
- * Add it to the swap cache and mark it dirty
+ * add_to_swap_cache() doesn't return -EEXIST, so we can safely
+ * clear SWAP_HAS_CACHE flag.
*/
- err = add_to_swap_cache(page, entry,
- __GFP_HIGH|__GFP_NOMEMALLOC|__GFP_NOWARN);
-
- switch (err) {
- case 0: /* Success */
- SetPageDirty(page);
- return 1;
- case -EEXIST:
- /* Raced with "speculative" read_swap_cache_async */
- swapcache_free(entry, NULL);
- continue;
- default:
- /* -ENOMEM radix-tree allocation failure */
- swapcache_free(entry, NULL);
- return 0;
- }
+ swapcache_free(entry, NULL);
+ return 0;
}
}
@@ -298,14 +300,7 @@ struct page *read_swap_cache_async(swp_e
if (err) /* swp entry is obsolete ? */
break;
- /*
- * Associate the page with swap entry in the swap cache.
- * May fail (-EEXIST) if there is already a page associated
- * with this entry in the swap cache: added by a racing
- * read_swap_cache_async, or add_to_swap or shmem_writepage
- * re-using the just freed swap entry for an existing page.
- * May fail (-ENOMEM) if radix-tree node allocation failed.
- */
+ /* May fail (-ENOMEM) if radix-tree node allocation failed. */
__set_page_locked(new_page);
SetPageSwapBacked(new_page);
err = add_to_swap_cache(new_page, entry, gfp_mask & GFP_KERNEL);
@@ -319,6 +314,10 @@ struct page *read_swap_cache_async(swp_e
}
ClearPageSwapBacked(new_page);
__clear_page_locked(new_page);
+ /*
+ * add_to_swap_cache() doesn't return -EEXIST, so we can safely
+ * clear SWAP_HAS_CACHE flag.
+ */
swapcache_free(entry, NULL);
} while (err != -ENOMEM);
_
Patches currently in -mm which might be from nishimura@xxxxxxxxxxxxxxxxx are
mm-add_to_swap_cache-does-not-return-eexist.patch
memcg-remove-the-overhead-associated-with-the-root-cgroup.patch
memcg-remove-the-overhead-associated-with-the-root-cgroup-fix.patch
memcg-remove-the-overhead-associated-with-the-root-cgroup-fix-2.patch
memcg-add-comments-explaining-memory-barriers.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
