We have hit the same BUG_ON as described in https://bugzilla.kernel.org/show_bug.cgi?id=31222 on some of our servers that have disk failures or corrupted inodes. After looking at the code, I think the problem is that we are not freeing inode's preallocation list when check_eofblocks_fl fails in ext4_ext_map_blocks(), which leaves the inode's preallocation list in an inconsistent state. Below is a proposed patch to fix the bug. I have tested it by manually inserting a random failure in check_eofblocks_fl() and run a test that creates and uses an inode's preallocated blocks. Without the fix, the kernel crashes after a few runs. With the fix, no crash is observed. ext4: free allocated and pre-allocated blocks when check_eofblocks_fl fails Upon corrupted inode or disk failures, we may fail after we already allocate some blocks from the inode or take some blocks from the inode's preallocation list, but before we successfully insert the corresponding extent to the extent tree. In this case, we should free any allocated blocks and discard the inode's preallocated blocks because the entries in the inode's preallocation list may be in an inconsistent state. Signed-off-by: Jiaying Zhang <jiayingz@xxxxxxxxxx> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 5199bac..8cf6ec9 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -3596,10 +3596,8 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode, } err = check_eofblocks_fl(handle, inode, map->m_lblk, path, ar.len); - if (err) - goto out2; - - err = ext4_ext_insert_extent(handle, inode, path, &newex, flags); + if (!err) + err = ext4_ext_insert_extent(handle, inode, path, &newex, flags); if (err) { /* free data blocks we just allocated */ /* not a good idea to call discard here directly, -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
