> On Mon, Aug 24, 2009 at 05:43:36PM -0600, Andreas Dilger wrote: > > Without transaction checksums waiting on all of the blocks together > > is NOT safe. If the commit record is on disk, but the rest of the > > transaction's blocks are not then during replay it may cause garbage > > to be written from the journal into the filesystem metadata. > > Yes, I *said* that we can only wait on all of the blocks together with > the commit record when doing journal checksums. Sorry if I didn't > make that clear enough. I suppose we talk about the case when write caches are turned off and we use barrier=0 (because case barrier=1 does not really care about when we wait for the blocks from the correctness POV - at least according to Documentation/block/barrier.txt). In that case, you have to wait for *data* blocks before writing the commit block because you have no checksumming of those... Honza -- Jan Kara <jack@xxxxxxx> SuSE CR Labs -- 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
