btrfs: fix page leaks after failure to lock page for delalloc

When locking pages for delalloc, we check if it's dirty and mapping still
matches. If it does not match, we need to return -EAGAIN and release all
pages. Only the current page was put though, iterate over all the
remaining pages too.

CC: [email protected] # 4.14+
Reviewed-by: Filipe Manana <[email protected]>
Reviewed-by: Nikolay Borisov <[email protected]>
Signed-off-by: Robbie Ko <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 95313bb7fe40d2cdb8f79e76fcfdf696cdc9c83e..72c836b54efc0d687b64fd8e6165a331002d6756 100644 (file)
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -1999,7 +1999,8 @@ static int __process_pages_contig(struct address_space *mapping,
                                if (!PageDirty(pages[i]) ||
                                    pages[i]->mapping != mapping) {
                                        unlock_page(pages[i]);
-                                       put_page(pages[i]);
+                                       for (; i < ret; i++)
+                                               put_page(pages[i]);
                                        err = -EAGAIN;
                                        goto out;
                                }