On Tue 10-05-11 22:52:10, OGAWA Hirofumi wrote: > Christoph Hellwig <hch@xxxxxxxxxxxxx> writes: > > > On Tue, May 10, 2011 at 10:59:15AM +0900, OGAWA Hirofumi wrote: > >> I'd like to know those patches are on what state. Waiting in writeback > >> page makes slower, like you mentioned it (I guess it would more > >> noticeable if device was slower that like FAT uses). And I think > >> currently it doesn't help anything others for blk-integrity stuff > >> (without other technic, it doesn't help FS consistency)? > > > > It only makes things slower if we rewrite a region in a file that is > > currently undergoing writeback. I'd be interested to know about real > > life applications doing that, and if they really are badly affect we > > should help them to work around that in userspace, e.g. by adding a > > fadvice will rewrite call that might be used to never write back that > > regions without an explicit fsync call. > > Isn't it reallocated blocks too, and metadata too? Reallocated blocks - not really. For a block to be freed it cannot be under writeback and when it's freed no writeback is started. For metadata - yes. But ext3, ext4, xfs, btrfs have to avoid modifying metadata under writeback anyway (because of journalling / COW constraints) and thus they don't care. For ext2 or vfat it's a different story. But as I wrote to Darrick, I'm not sure about vfat but for ext2 and similar legacy filesystems, I'd rather let them live with their unstable pages under IO ;) because I see a limited use for that. Honza -- Jan Kara <jack@xxxxxxx> SUSE Labs, CR -- 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
