[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: df not showing ext3 partition

One thing I think i've found, with mount, is that the kernel is probably
reading the superblock information from the partition whilst making it's
decisions. I've noticed behaviour in mount that does a similar check...
I'm in the process of patching mount to do this check if the fstype
reported is 'auto'... i haven't actually patched the writeout to
/etc/mtab yet, but so far during boot, it properly detects that the
partition type was ext3... All I will have to do after this, is modify the
-f /etc/mtab writeout to use the value garnered from the code I added, so
/etc/mtab will always reflect actual fs type, not 'auto'.

Anyone have a better idea on this?

Curtis Hogg [buckminst inconnu isu edu]

Above all things, reverence yourself.
Email 1 - buckminst inconnu isu edu
Email 2 - buckminst hotmail com
WWW - [in transit]

On Wed, 29 Aug 2001, Theodore Tso wrote:

> On Tue, Aug 28, 2001 at 05:36:45PM -0600, Curtis Hogg wrote:
> > mount 2.11h also puts 'auto' in /etc/mtab...
> I see what's going on.  The problem is that mount is putting "auto" in
> /etc/mtab when the -f flag is used.  It doesn't do this for a standard
> mount; but in the case of the root filesystem, after the root is
> remounted read-write, /etc/mtab is zeroed out, and "mount -f" is used
> to put the entry for the root filesystem into /etc/mtab.  Mount -f
> doesn't actually do the mount, so it never tries to instantiate what
> "auto" means, so "auto" is ending up in /etc/mtab.
> Actually, this makes a certain amount sense; since mount didn't
> actually do the mount (it let the kernel do that), short of having it
> query /proc/mounts (and /proc isn't mounted yet at that stage of the
> boot process), there's no way for it to really know how the kernel
> decided to mount the filesystem when it was booting.
> 						- Ted

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]