JFS: fsync wrong behavior when I/O failure occurs

This is half of a patch that Qu Fuping submitted in April.  The first part
was applied to fs/mpage.c in 2.6.12-rc4.

jfs_fsync should return error, but it doesn't wait for the metadata page to
be uptodate, e.g.:
jfs_fsync->jfs_commit_inode->txCommit->diWrite->read_metapage->
__get_metapage->read_cache_page reads a page from disk. Because read is
async, when read_cache_page: err = filler(data, page), filler will not
return error, it just submits I/O request and returns. So, page is not
uptodate.  Checking only if(IS_ERROR(mp->page)) is not enough, we should
add "|| !PageUptodate(mp->page)"

Signed-off-by: Dave Kleikamp <shaggy@austin.ibm.com>

diff --git a/fs/jfs/jfs_metapage.c b/fs/jfs/jfs_metapage.c
index 8cb803b54bc12080b6f5dba1f57b953707c77901..02add5d8ff8988aa9f34bf320593beadf6c24d7f 100644 (file)
--- a/fs/jfs/jfs_metapage.c
+++ b/fs/jfs/jfs_metapage.c
@@ -640,7 +640,7 @@ struct metapage *__get_metapage(struct inode *inode, unsigned long lblock,
        } else {
                page = read_cache_page(mapping, page_index,
                            (filler_t *)mapping->a_ops->readpage, NULL);
-               if (IS_ERR(page)) {
+               if (IS_ERR(page) || !PageUptodate(mp->page)) {
                        jfs_err("read_cache_page failed!");
                        return NULL;
                }