On Tue, Mar 26, 2013 at 01:54:19PM +0100, Lukáš Czerner wrote: > > + /* Wait all existing dio workers, newcomers will block on i_mutex */ > > + ext4_inode_block_unlocked_dio(inode); > > This was not present in the indirect punch hole code, does it means > that there was a bug ? If so maybe it's worth mentioning in the > description? Yes, I'm pretty sure it was a bug. The ext4_inode_block_unlocked_dio() call was added in the extents code path by commit 02d262dffcf4c. The problem is that i_mutex will not block DIO readers in dioread_nolock mode. One of the problems with not having done the code refactorization earlier was that a bug fixed in one code path doesn't necessarily get fixed in another. I'll add a comment to this effect in the commit description. > > + if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) > > + credits = ext4_writepage_trans_blocks(inode); > > + else > > + credits = ext4_blocks_for_truncate(inode); > > + handle = ext4_journal_start(inode, EXT4_HT_TRUNCATE, > > + ext4_blocks_for_truncate(inode)); > > Hmm, shouldn't we be using 'credits' instead of > ext4_blocks_for_truncate(inode) here ? Yes, good catch! Thanks, - Ted -- 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
