As I mentioned in the bug-report, when the file size is 
not n * BUFFERSIZE, this file cannot be read correctly 
and the system will crash or hang up.

After investigation, I fount that data_len should be 
checked before lock_extent, in order to prevent the 
situation that when it comes to the file end, 
diocb->lockstart changes to lockend+1, then lock_extent
 will cause system's crash or hang.

So move the data_len's check ahead.

Signed-off-by: Liu Bo <[email protected]>
Signed-off-by: Miao Xie <[email protected]>
---
 fs/btrfs/dio.c |   10 +++++-----
 1 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/dio.c b/fs/btrfs/dio.c
index c930ff5..7178a25 100644
--- a/fs/btrfs/dio.c
+++ b/fs/btrfs/dio.c
@@ -517,11 +517,6 @@ static void btrfs_dio_read(struct btrfs_diocb *diocb)
        lockend = ALIGN(end, blocksize) - 1;
 
 getlock:
-       /* writeout everything we read for checksum or compressed extents */
-       filemap_write_and_wait_range(diocb->inode->i_mapping,
-                               diocb->lockstart, lockend);
-       lock_extent(io_tree, diocb->lockstart, lockend, GFP_NOFS);
-
        data_len = min_t(u64, end, i_size_read(diocb->inode));
        if (data_len <= diocb->start) {
                /* whatever we finished (or 0) is returned past EOF */
@@ -529,6 +524,11 @@ getlock:
        }
        data_len -= diocb->start;
 
+       /* writeout everything we read for checksum or compressed extents */
+       filemap_write_and_wait_range(diocb->inode->i_mapping,
+                               diocb->lockstart, lockend);
+       lock_extent(io_tree, diocb->lockstart, lockend, GFP_NOFS);
+
        safe_to_read = 0;
        while (data_len && !diocb->error) { /* error in reaper stops submit */
                struct extent_map *em;
-- 
1.6.2.5


--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to