tytso@MIT.EDU said:
> You could potentially do something like that, but it would mean that
> whenever you tried to write to the filesystem (including inode atime
> updates), all modifications would have to stall until the
> EXT2_VALID_FS bit was cleared. It's doable, but it's not clear it's
> worth it.
At least as a mount option, I'd definitely say it's worth it. I've lost count
of the times I've watched my box fsck huge partitions that it hadn't even
touched before the crash - not even for an atime update.
When we mount a filesystem, do we currently wait until the superblock write is
flushed to the media, or do we just queue the superblock write and carry on as
if it's safe?
(If we don't flush the superblock write, what's to stop the block layer from
reordering the writes, so we start writing the actual filesystem before the
superblock write has hit the media?)
-- dwmw2- 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/
This archive was generated by hypermail 2b29 : Sun Jan 23 2000 - 21:00:21 EST