CacheFiles: Fix the marking of cached pages
Under some circumstances CacheFiles defers the marking of pages with PG_fscache
so that it can take advantage of pagevecs to reduce the number of calls to
fscache_mark_pages_cached() and the netfs's hook to keep track of this.
There are, however, two problems with this:
(1) It can lead to the PG_fscache mark being applied _after_ the page is set
PG_uptodate and unlocked (by the call to fscache_end_io()).
(2) CacheFiles's ref on the page is dropped immediately following
fscache_end_io() - and so may not still be held when the mark is applied.
This can lead to the page being passed back to the allocator before the
mark is applied.
Fix this by, where appropriate, marking the page before calling
fscache_end_io() and releasing the page. This means that we can't take
advantage of pagevecs and have to make a separate call for each page to the
marking routines.
The symptoms of this are Bad Page state errors cropping up under memory
pressure, for example:
BUG: Bad page state in process tar pfn:002da
page:ffffea0000009fb0 count:0 mapcount:0 mapping: (null) index:0x1447
page flags: 0x1000(private_2)
Pid: 4574, comm: tar Tainted: G W 3.1.0-rc4-fsdevel+ #1064
Call Trace:
[<ffffffff8109583c>] ? dump_page+0xb9/0xbe
[<ffffffff81095916>] bad_page+0xd5/0xea
[<ffffffff81095d82>] get_page_from_freelist+0x35b/0x46a
[<ffffffff810961f3>] __alloc_pages_nodemask+0x362/0x662
[<ffffffff810989da>] __do_page_cache_readahead+0x13a/0x267
[<ffffffff81098942>] ? __do_page_cache_readahead+0xa2/0x267
[<ffffffff81098d7b>] ra_submit+0x1c/0x20
[<ffffffff8109900a>] ondemand_readahead+0x28b/0x29a
[<ffffffff81098ee2>] ? ondemand_readahead+0x163/0x29a
[<ffffffff810990ce>] page_cache_sync_readahead+0x38/0x3a
[<ffffffff81091d8a>] generic_file_aio_read+0x2ab/0x67e
[<ffffffffa008cfbe>] nfs_file_read+0xa4/0xc9 [nfs]
[<ffffffff810c22c4>] do_sync_read+0xba/0xfa
[<ffffffff81177a47>] ? security_file_permission+0x7b/0x84
[<ffffffff810c25dd>] ? rw_verify_area+0xab/0xc8
[<ffffffff810c29a4>] vfs_read+0xaa/0x13a
[<ffffffff810c2a79>] sys_read+0x45/0x6c
[<ffffffff813ac37b>] system_call_fastpath+0x16/0x1b
As can be seen, PG_private_2 (== PG_fscache) is set in the page flags.
Instrumenting fscache_mark_pages_cached() to verify whether page->mapping was
set appropriately showed that sometimes it wasn't. This led to the discovery
that sometimes the page has apparently been reclaimed by the time the marker
got to see it.
Reported-by: M. Stevens <m@tippett.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
diff --git a/fs/cachefiles/rdwr.c b/fs/cachefiles/rdwr.c
index c994691..3367abd 100644
--- a/fs/cachefiles/rdwr.c
+++ b/fs/cachefiles/rdwr.c
@@ -176,9 +176,8 @@
recheck:
if (PageUptodate(monitor->back_page)) {
copy_highpage(monitor->netfs_page, monitor->back_page);
-
- pagevec_add(&pagevec, monitor->netfs_page);
- fscache_mark_pages_cached(monitor->op, &pagevec);
+ fscache_mark_page_cached(monitor->op,
+ monitor->netfs_page);
error = 0;
} else if (!PageError(monitor->back_page)) {
/* the page has probably been truncated */
@@ -335,8 +334,7 @@
backing_page_already_uptodate:
_debug("- uptodate");
- pagevec_add(pagevec, netpage);
- fscache_mark_pages_cached(op, pagevec);
+ fscache_mark_page_cached(op, netpage);
copy_highpage(netpage, backpage);
fscache_end_io(op, netpage, 0);
@@ -448,8 +446,7 @@
&pagevec);
} else if (cachefiles_has_space(cache, 0, 1) == 0) {
/* there's space in the cache we can use */
- pagevec_add(&pagevec, page);
- fscache_mark_pages_cached(op, &pagevec);
+ fscache_mark_page_cached(op, page);
ret = -ENODATA;
} else {
ret = -ENOBUFS;
@@ -465,8 +462,7 @@
*/
static int cachefiles_read_backing_file(struct cachefiles_object *object,
struct fscache_retrieval *op,
- struct list_head *list,
- struct pagevec *mark_pvec)
+ struct list_head *list)
{
struct cachefiles_one_read *monitor = NULL;
struct address_space *bmapping = object->backer->d_inode->i_mapping;
@@ -626,13 +622,13 @@
page_cache_release(backpage);
backpage = NULL;
- if (!pagevec_add(mark_pvec, netpage))
- fscache_mark_pages_cached(op, mark_pvec);
+ fscache_mark_page_cached(op, netpage);
page_cache_get(netpage);
if (!pagevec_add(&lru_pvec, netpage))
__pagevec_lru_add_file(&lru_pvec);
+ /* the netpage is unlocked and marked up to date here */
fscache_end_io(op, netpage, 0);
page_cache_release(netpage);
netpage = NULL;
@@ -775,15 +771,11 @@
/* submit the apparently valid pages to the backing fs to be read from
* disk */
if (nrbackpages > 0) {
- ret2 = cachefiles_read_backing_file(object, op, &backpages,
- &pagevec);
+ ret2 = cachefiles_read_backing_file(object, op, &backpages);
if (ret2 == -ENOMEM || ret2 == -EINTR)
ret = ret2;
}
- if (pagevec_count(&pagevec) > 0)
- fscache_mark_pages_cached(op, &pagevec);
-
_leave(" = %d [nr=%u%s]",
ret, *nr_pages, list_empty(pages) ? " empty" : "");
return ret;
@@ -806,7 +798,6 @@
{
struct cachefiles_object *object;
struct cachefiles_cache *cache;
- struct pagevec pagevec;
int ret;
object = container_of(op->op.object,
@@ -817,13 +808,10 @@
_enter("%p,{%lx},", object, page->index);
ret = cachefiles_has_space(cache, 0, 1);
- if (ret == 0) {
- pagevec_init(&pagevec, 0);
- pagevec_add(&pagevec, page);
- fscache_mark_pages_cached(op, &pagevec);
- } else {
+ if (ret == 0)
+ fscache_mark_page_cached(op, page);
+ else
ret = -ENOBUFS;
- }
_leave(" = %d", ret);
return ret;
diff --git a/fs/fscache/page.c b/fs/fscache/page.c
index 3f7a59b..d7c663c 100644
--- a/fs/fscache/page.c
+++ b/fs/fscache/page.c
@@ -915,6 +915,40 @@
EXPORT_SYMBOL(__fscache_uncache_page);
/**
+ * fscache_mark_page_cached - Mark a page as being cached
+ * @op: The retrieval op pages are being marked for
+ * @page: The page to be marked
+ *
+ * Mark a netfs page as being cached. After this is called, the netfs
+ * must call fscache_uncache_page() to remove the mark.
+ */
+void fscache_mark_page_cached(struct fscache_retrieval *op, struct page *page)
+{
+ struct fscache_cookie *cookie = op->op.object->cookie;
+
+#ifdef CONFIG_FSCACHE_STATS
+ atomic_inc(&fscache_n_marks);
+#endif
+
+ _debug("- mark %p{%lx}", page, page->index);
+ if (TestSetPageFsCache(page)) {
+ static bool once_only;
+ if (!once_only) {
+ once_only = true;
+ printk(KERN_WARNING "FS-Cache:"
+ " Cookie type %s marked page %lx"
+ " multiple times\n",
+ cookie->def->name, page->index);
+ }
+ }
+
+ if (cookie->def->mark_page_cached)
+ cookie->def->mark_page_cached(cookie->netfs_data,
+ op->mapping, page);
+}
+EXPORT_SYMBOL(fscache_mark_page_cached);
+
+/**
* fscache_mark_pages_cached - Mark pages as being cached
* @op: The retrieval op pages are being marked for
* @pagevec: The pages to be marked
@@ -925,32 +959,11 @@
void fscache_mark_pages_cached(struct fscache_retrieval *op,
struct pagevec *pagevec)
{
- struct fscache_cookie *cookie = op->op.object->cookie;
unsigned long loop;
-#ifdef CONFIG_FSCACHE_STATS
- atomic_add(pagevec->nr, &fscache_n_marks);
-#endif
+ for (loop = 0; loop < pagevec->nr; loop++)
+ fscache_mark_page_cached(op, pagevec->pages[loop]);
- for (loop = 0; loop < pagevec->nr; loop++) {
- struct page *page = pagevec->pages[loop];
-
- _debug("- mark %p{%lx}", page, page->index);
- if (TestSetPageFsCache(page)) {
- static bool once_only;
- if (!once_only) {
- once_only = true;
- printk(KERN_WARNING "FS-Cache:"
- " Cookie type %s marked page %lx"
- " multiple times\n",
- cookie->def->name, page->index);
- }
- }
- }
-
- if (cookie->def->mark_pages_cached)
- cookie->def->mark_pages_cached(cookie->netfs_data,
- op->mapping, pagevec);
pagevec_reinit(pagevec);
}
EXPORT_SYMBOL(fscache_mark_pages_cached);
diff --git a/include/linux/fscache-cache.h b/include/linux/fscache-cache.h
index ce31408..9879183 100644
--- a/include/linux/fscache-cache.h
+++ b/include/linux/fscache-cache.h
@@ -504,6 +504,9 @@
extern void fscache_io_error(struct fscache_cache *cache);
+extern void fscache_mark_page_cached(struct fscache_retrieval *op,
+ struct page *page);
+
extern void fscache_mark_pages_cached(struct fscache_retrieval *op,
struct pagevec *pagevec);
diff --git a/include/linux/fscache.h b/include/linux/fscache.h
index 9ec20de..f4b6353 100644
--- a/include/linux/fscache.h
+++ b/include/linux/fscache.h
@@ -135,14 +135,14 @@
*/
void (*put_context)(void *cookie_netfs_data, void *context);
- /* indicate pages that now have cache metadata retained
- * - this function should mark the specified pages as now being cached
- * - the pages will have been marked with PG_fscache before this is
+ /* indicate page that now have cache metadata retained
+ * - this function should mark the specified page as now being cached
+ * - the page will have been marked with PG_fscache before this is
* called, so this is optional
*/
- void (*mark_pages_cached)(void *cookie_netfs_data,
- struct address_space *mapping,
- struct pagevec *cached_pvec);
+ void (*mark_page_cached)(void *cookie_netfs_data,
+ struct address_space *mapping,
+ struct page *page);
/* indicate the cookie is no longer cached
* - this function is called when the backing store currently caching