Messages in this thread | | | Date | Fri, 14 Jul 2000 14:39:10 +0100 | From | "Stephen C. Tweedie" <> | Subject | Re: ext3-0.0.2e released |
| |
Hi,
On Thu, Jul 06, 2000 at 09:30:10AM -0700, Andre Hedrick wrote: > > > Yeah, I know, but then again, why do they all want to fsck? We have so > > many nice journaling file systems now. > > Maybe because if the damn drivers are not cool with ext2, they will not > work with journaling, you think that is valid enough reason?
With one exception (the munging of buffer_heads directly by the device driver layer, which is something that only the raid1/5 drivers do for now), journaling filesystems have no special device driver requirements that are not also shared by applications such as mailers (which use fsync for ordering guarantees) or databases (which use O_SYNC, typically).
If the drivers aren't cool with ext2, then the kernel's whole synchronous IO API breaks too, and all user space apps which rely on write ordering will break. This is generally considered a Bad Idea!
--Stephen
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |