Re: FIBMAP ioctl restrictions: why?

Jamie Lokier (lkd@tantalophile.demon.co.uk)
Wed, 10 Nov 1999 20:33:54 +0100


Alan Cox wrote:
> > While we're here, I would like to request FIBMAP on directories, thanks.
>
> Just fix all of the race conditions with ftruncate, the sanity checking
> and other missing map checks in the fs layers and send me a diff 8)

The result of FIBMAP can always be out of date by the time it reaches
user space. So you can only use it as a hint for optimising read order,
or like Lilo, you have to assume the underlying file is not being
changed.

If you use it as a hint there's no big deal if the information is
occasionally wrong due to a race. Are there any real race conditions
with FIBMAP that cause real problems?

(It's not a great hint for a lot of reasons. One of them is the
filesystem's tendancy to read ahead in a file, which tends to break your
carefully planned read sequence...)

-- Jamie

-
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/