The patch titled
mm: add_to_swap_cache() must not sleep
has been added to the -mm tree. Its filename is
mm-add_to_swap_cache-must-not-sleep.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() must not sleep
From: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
After commit 355cfa73 ("mm: modify swap_map and add SWAP_HAS_CACHE flag"),
read_swap_cache_async() will busy-wait while a entry doesn't exist in swap
cache but it has SWAP_HAS_CACHE flag.
Such entries can exist on add/delete path of swap cache. On add path,
add_to_swap_cache() is called soon after SWAP_HAS_CACHE flag is set, and
on delete path, swapcache_free() will be called (SWAP_HAS_CACHE flag is
cleared) soon after __delete_from_swap_cache() is called. So, the
busy-wait works well in most cases.
But this mechanism can cause soft lockup if add_to_swap_cache() sleeps and
read_swap_cache_async() tries to swap-in the same entry on the same cpu.
This patch calls radix_tree_preload() before swapcache_prepare() and
divides add_to_swap_cache() into two part: radix_tree_preload() part and
radix_tree_insert() part(define it as __add_to_swap_cache()).
Signed-off-by: Daisuke Nishimura <nishimura@xxxxxxxxxxxxxxxxx>
Cc: 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/swap_state.c | 70 ++++++++++++++++++++++++++++++----------------
1 file changed, 46 insertions(+), 24 deletions(-)
diff -puN mm/swap_state.c~mm-add_to_swap_cache-must-not-sleep mm/swap_state.c
--- a/mm/swap_state.c~mm-add_to_swap_cache-must-not-sleep
+++ a/mm/swap_state.c
@@ -67,10 +67,10 @@ void show_swap_cache_info(void)
}
/*
- * add_to_swap_cache resembles add_to_page_cache_locked on swapper_space,
+ * __add_to_swap_cache resembles add_to_page_cache_locked on swapper_space,
* but sets SwapCache flag and private instead of mapping and index.
*/
-int add_to_swap_cache(struct page *page, swp_entry_t entry, gfp_t gfp_mask)
+static int __add_to_swap_cache(struct page *page, swp_entry_t entry)
{
int error;
@@ -78,28 +78,37 @@ int add_to_swap_cache(struct page *page,
VM_BUG_ON(PageSwapCache(page));
VM_BUG_ON(!PageSwapBacked(page));
+ page_cache_get(page);
+ SetPageSwapCache(page);
+ set_page_private(page, entry.val);
+
+ spin_lock_irq(&swapper_space.tree_lock);
+ error = radix_tree_insert(&swapper_space.page_tree, entry.val, page);
+ if (likely(!error)) {
+ total_swapcache_pages++;
+ __inc_zone_page_state(page, NR_FILE_PAGES);
+ INC_CACHE_INFO(add_total);
+ }
+ spin_unlock_irq(&swapper_space.tree_lock);
+
+ if (unlikely(error)) {
+ set_page_private(page, 0UL);
+ ClearPageSwapCache(page);
+ page_cache_release(page);
+ }
+
+ return error;
+}
+
+
+int add_to_swap_cache(struct page *page, swp_entry_t entry, gfp_t gfp_mask)
+{
+ int error;
+
error = radix_tree_preload(gfp_mask);
if (!error) {
- page_cache_get(page);
- SetPageSwapCache(page);
- set_page_private(page, entry.val);
-
- spin_lock_irq(&swapper_space.tree_lock);
- error = radix_tree_insert(&swapper_space.page_tree,
- entry.val, page);
- if (likely(!error)) {
- total_swapcache_pages++;
- __inc_zone_page_state(page, NR_FILE_PAGES);
- INC_CACHE_INFO(add_total);
- }
- spin_unlock_irq(&swapper_space.tree_lock);
+ error = __add_to_swap_cache(page, entry);
radix_tree_preload_end();
-
- if (unlikely(error)) {
- set_page_private(page, 0UL);
- ClearPageSwapCache(page);
- page_cache_release(page);
- }
}
return error;
}
@@ -290,13 +299,24 @@ struct page *read_swap_cache_async(swp_e
}
/*
+ * call radix_tree_preload() while we can wait.
+ */
+ err = radix_tree_preload(gfp_mask & GFP_KERNEL);
+ if (err)
+ break;
+
+ /*
* Swap entry may have been freed since our caller observed it.
*/
err = swapcache_prepare(entry);
- if (err == -EEXIST) /* seems racy */
+ if (err == -EEXIST) { /* seems racy */
+ radix_tree_preload_end();
continue;
- if (err) /* swp entry is obsolete ? */
+ }
+ if (err) { /* swp entry is obsolete ? */
+ radix_tree_preload_end();
break;
+ }
/*
* Associate the page with swap entry in the swap cache.
@@ -308,8 +328,9 @@ struct page *read_swap_cache_async(swp_e
*/
__set_page_locked(new_page);
SetPageSwapBacked(new_page);
- err = add_to_swap_cache(new_page, entry, gfp_mask & GFP_KERNEL);
+ err = __add_to_swap_cache(new_page, entry);
if (likely(!err)) {
+ radix_tree_preload_end();
/*
* Initiate read into locked page and return.
*/
@@ -317,6 +338,7 @@ struct page *read_swap_cache_async(swp_e
swap_readpage(new_page);
return new_page;
}
+ radix_tree_preload_end();
ClearPageSwapBacked(new_page);
__clear_page_locked(new_page);
swapcache_free(entry, NULL);
_
Patches currently in -mm which might be from nishimura@xxxxxxxxxxxxxxxxx are
mm-add_to_swap_cache-must-not-sleep.patch
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
