Re: statuscheck patch..

[Andrew Clausen]

On Tue, Aug 17, 2004 at 01:26:48PM +0200, B.Hakvoort wrote:
> So you're proposing to do it the other way around? Instead of getting a
> path and look for it in /proc/mounts, you take paths from /proc/mounts,
> get their major/minor and compare this to the major/minor from the
> selected partition.

Right.

> This sounds nice!

Well, it's rather inefficient, because you end up having to do lots
of stat() syscalls.  Paranoia is my primary motivation ;)

> very efficient, but wouldn't we be getting problems
> with some strange entrys like your /dev2/root2 or rootfs ? 
> rootdevice again :)

Yes, we still need to do whatever we can to find out what /dev2/root2
actually is.  (Whether it be following symlinks or real-root-dev)
But, after following all the leads, we should end up with a
major/minor at the end, right?

> Of course we could always check the rootdevice indepedently from the
> other partitions.

Right.

Cheers,
Andrew