Re: Should raw I/O be added to the kernel?

Chris Wedgwood (cw@ix.net.nz)
Tue, 22 Dec 1998 16:02:44 +1300


On Mon, Dec 21, 1998 at 03:06:26PM -0800, Larry McVoy wrote:

> Part of the problem is that OS implementations have sometimes
> forgotten to sync associated metadata as part of fsync(). Given
> that people may have been bitten by this in the past, they don't
> always trust fsync() to do the right thing.

There is code out there that assumes fsync() flushes all the
meta-data too (I think this is a FFS/BSDism more than anything).

To the best of my knowledge, nothing requires meta-data to be flush
upon fsync, and it isn't -- applications wanting this behavior will
have to do other things to ensure this.

-cw

-
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/