2.4.19-pre7: rootfs mounted twice

2.4.19-pre7: rootfs mounted twice

[Michael Dreher]

Hello,

on 2.4.18 I get the following:

dreher@karpfen:~ > uname -a
Linux karpfen 2.4.18 #2 Sun Apr 21 23:27:03 CEST 2002 i586 unknown

dreher@karpfen:~ > df
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/root              7060308   6276080    425580  94% /
/dev/hda4              3794936   3043812    558344  84% /home

dreher@karpfen:~ > cat /proc/\bmounts
/dev/root / ext3 rw,noatime,nodiratime 0 0
proc /proc proc rw 0 0
/dev/hda4 /home ext3 rw,noatime,nodiratime 0 0
usbdevfs /proc/bus/usb usbdevfs rw 0 0
devpts /dev/pts devpts rw 0 0


On the other hand, 2.4.19-pre7 gives:

dreher@karpfen:~ > uname -a
Linux karpfen 2.4.19-pre7 #6 Sun Apr 21 20:27:47 CEST 2002 i586 unknown

dreher@karpfen:~ > df
Filesystem           1k-blocks      Used Available Use% Mounted on
rootfs                 7060308   6276188    425472  94% /
/dev/root              7060308   6276188    425472  94% /
/dev/hda4              3794936   3042316    559840  84% /home

dreher@karpfen:~ > cat /proc/mounts
rootfs / rootfs rw 0 0
/dev/root / ext3 rw,noatime,nodiratime 0 0
proc /proc proc rw 0 0
/dev/hda4 /home ext3 rw,noatime,nodiratime 0 0
usbdevfs /proc/bus/usb usbdevfs rw 0 0
devpts /dev/pts devpts rw 0 0


I don't know if it is related, but in my bootscripts I link
/etc/mtab to /proc/mounts.


Regards,
Michael

Re: 2.4.19-pre7: rootfs mounted twice

[Alexander Viro]

On Fri, 26 Apr 2002, Michael Dreher wrote:

> dreher@karpfen:~ > df
> Filesystem           1k-blocks      Used Available Use% Mounted on
> rootfs                 7060308   6276188    425472  94% /

df(1) is wrong.  There is (see /proc/mounts below) rootfs mounted as
root (ramfs, actually) and ext3 mounted over it.  df sees two entries
in /etc/mtab (on your box - /proc/mounts) with mountpoint "/" and
does statfs("/", &buf); for both.  Surprise, surprise, results of
two calls of statf2(2) are identical - what with arguments being
the same both times - and refer to the filesystem where your "/"
lives.  I.e. to ext3.

> /dev/root              7060308   6276188    425472  94% /
> /dev/hda4              3794936   3042316    559840  84% /home
> 
> dreher@karpfen:~ > cat /proc/mounts
> rootfs / rootfs rw 0 0
> /dev/root / ext3 rw,noatime,nodiratime 0 0
> proc /proc proc rw 0 0
> /dev/hda4 /home ext3 rw,noatime,nodiratime 0 0
> usbdevfs /proc/bus/usb usbdevfs rw 0 0
> devpts /dev/pts devpts rw 0 0

Re: 2.4.19-pre7: root filesystem issues

[Daniel Gryniewicz]

On Fri, 26 Apr 2002 02:32:23 -0400 (EDT)
Alexander Viro <viro@math.psu.edu> wrote:

> 
> 
> On Fri, 26 Apr 2002, Michael Dreher wrote:
> 
> > dreher@karpfen:~ > df
> > Filesystem           1k-blocks      Used Available Use% Mounted on
> > rootfs                 7060308   6276188    425472  94% /
> 
> df(1) is wrong.  There is (see /proc/mounts below) rootfs mounted as
> root (ramfs, actually) and ext3 mounted over it.  df sees two entries
> in /etc/mtab (on your box - /proc/mounts) with mountpoint "/" and
> does statfs("/", &buf); for both.  Surprise, surprise, results of
> two calls of statf2(2) are identical - what with arguments being
> the same both times - and refer to the filesystem where your "/"
> lives.  I.e. to ext3.
> 
> > /dev/root              7060308   6276188    425472  94% /
> > /dev/hda4              3794936   3042316    559840  84% /home
> > 
> > dreher@karpfen:~ > cat /proc/mounts
> > rootfs / rootfs rw 0 0
> > /dev/root / ext3 rw,noatime,nodiratime 0 0
> > proc /proc proc rw 0 0
> > /dev/hda4 /home ext3 rw,noatime,nodiratime 0 0
> > usbdevfs /proc/bus/usb usbdevfs rw 0 0
> > devpts /dev/pts devpts rw 0 0

On a related note (the listing of /dev/root reminded me), as of
2.4.19-pre7-ac2, I can no longer boot with a "root=/dev/discs/disc0/part1"
command line, I have to use "root=/dev/ide/host2/bus0/target0/lun0/part1".  My
previous kernel (2.4.19-pre4-ac2-radix) and all ones before that worked fine
with the first version of the command line.  Is this an intended change or is
it a bug?  I can give more information if necessary, and it's happening on two
different AMD based systems with different chipsets.

Daniel
--- 
Recursion n.:
        See Recursion.
                        -- Random Shack Data Processing Dictionary

Re: 2.4.19-pre7: rootfs mounted twice

[Pavel Machek]

Hi!

> > dreher@karpfen:~ > df
> > Filesystem           1k-blocks      Used Available Use% Mounted on
> > rootfs                 7060308   6276188    425472  94% /
> 
> df(1) is wrong.  There is (see /proc/mounts below) rootfs mounted as
> root (ramfs, actually) and ext3 mounted over it.  df sees two entries
> in /etc/mtab (on your box - /proc/mounts) with mountpoint "/" and
> does statfs("/", &buf); for both.  Surprise, surprise, results of
> two calls of statf2(2) are identical - what with arguments being
> the same both times - and refer to the filesystem where your "/"
> lives.  I.e. to ext3.

df might be wrong, but lets say that this /proc/mounts become
interesting. This could not have happened in the past. That means you
changed kernel interface in stable series, and that's wrong.

> > /dev/root              7060308   6276188    425472  94% /
> > /dev/hda4              3794936   3042316    559840  84% /home
> > 
> > dreher@karpfen:~ > cat /proc/mounts
> > rootfs / rootfs rw 0 0
> > /dev/root / ext3 rw,noatime,nodiratime 0 0
> > proc /proc proc rw 0 0
> > /dev/hda4 /home ext3 rw,noatime,nodiratime 0 0
> > usbdevfs /proc/bus/usb usbdevfs rw 0 0
> > devpts /dev/pts devpts rw 0 0

									Pavel
-- 
(about SSSCA) "I don't say this lightly.  However, I really think that the U.S.
no longer is classifiable as a democracy, but rather as a plutocracy." --hpa

Re: 2.4.19-pre7: rootfs mounted twice

[Alexander Viro]

On Fri, 26 Apr 2002, Pavel Machek wrote:

> > does statfs("/", &buf); for both.  Surprise, surprise, results of
> > two calls of statf2(2) are identical - what with arguments being
> > the same both times - and refer to the filesystem where your "/"
> > lives.  I.e. to ext3.
> 
> df might be wrong, but lets say that this /proc/mounts become
> interesting. This could not have happened in the past. That means you

This _could_ happen in past - as the matter of fact, I can reproduce it
on any 2.4 kernel.  Mount something over the root of already mounted
filesystem and watch the show.

Now, we could disable showing rootfs in /proc/mounts and it might be a
good idea for 2.4,  I'm not all that sure that it's a right thing, though.

Re: 2.4.19-pre7: rootfs mounted twice

[Jeff Chua]

On Fri, 26 Apr 2002, Alexander Viro wrote:

>
>
> On Fri, 26 Apr 2002, Pavel Machek wrote:
>
> > > does statfs("/", &buf); for both.  Surprise, surprise, results of
> > > two calls of statf2(2) are identical - what with arguments being
> > > the same both times - and refer to the filesystem where your "/"
> > > lives.  I.e. to ext3.
> >
> > df might be wrong, but lets say that this /proc/mounts become
> > interesting. This could not have happened in the past. That means you
>
> This _could_ happen in past - as the matter of fact, I can reproduce it
> on any 2.4 kernel.  Mount something over the root of already mounted
> filesystem and watch the show.
>
> Now, we could disable showing rootfs in /proc/mounts and it might be a
> good idea for 2.4,  I'm not all that sure that it's a right thing, though.


This happens all the time if you use initrd ramdisk and switch to hard
disk during boot up.

2.4.19-pre6 is ok, but 2.4.19-pre7 is not.

Jeff.

Re: 2.4.19-pre7: rootfs mounted twice

[Alexander Viro]

On Sat, 27 Apr 2002, Jeff Chua wrote:

> This happens all the time if you use initrd ramdisk and switch to hard
> disk during boot up.

This happens all the time with rootfs, period.  For exactly the same
reasons.

Again, we could hide that fact (we _do_ have a filesystem mounted
underneath the final root), but I'm not sure that it's worth hiding.
The real problem is in the interaction between df(1) and /proc/mounts
and it occurs every time when mountpoint happens to be a root of
already mounted filesystem.  Which could happen at any time since
the beginning of 2.4 (actually, way back in 2.3).

Notice that the same bug happens in _all_ version for chroot environment.
Namely, df(1) will try to call statfs() for mountpoints outside chroot
jail and that will give very interesting output, since results were for
corresponsing directories _inside_ the jail.

Arguably, correct way to handle that is to leave everything outside the
process' root out of /proc/mounts.  Which would solve both the chroot
problems and rootfs one (all processes are chrooted into the "final"
root).  It _still_ doesn't solve the old problem with overmounts, but
that one could be only handled in userland (by letting df(1) notice
that several getmntent(3) have the same mountpoint and ignore all
but the last one).

Unless somebody has a good reason why /proc/mounts in chroot jail should
show objects outside it, I'm going to do it that way.

Re: 2.4.19-pre7: rootfs mounted twice

[Richard B. Johnson]

On Sat, 27 Apr 2002, Jeff Chua wrote:

> On Fri, 26 Apr 2002, Alexander Viro wrote:
> 
> >
> >
> > On Fri, 26 Apr 2002, Pavel Machek wrote:
> >
> > > > does statfs("/", &buf); for both.  Surprise, surprise, results of
> > > > two calls of statf2(2) are identical - what with arguments being
> > > > the same both times - and refer to the filesystem where your "/"
> > > > lives.  I.e. to ext3.
> > >
> > > df might be wrong, but lets say that this /proc/mounts become
> > > interesting. This could not have happened in the past. That means you
> >
> > This _could_ happen in past - as the matter of fact, I can reproduce it
> > on any 2.4 kernel.  Mount something over the root of already mounted
> > filesystem and watch the show.
> >
> > Now, we could disable showing rootfs in /proc/mounts and it might be a
> > good idea for 2.4,  I'm not all that sure that it's a right thing, though.
> 
> 
> This happens all the time if you use initrd ramdisk and switch to hard
> disk during boot up.
> 
> 2.4.19-pre6 is ok, but 2.4.19-pre7 is not.
> 
> Jeff.
> 

In 2.4.18, it looks like this when booting initrd and switching
to a hard disk. This looks okay.

/dev/root.old /initrd ext2 rw 0 0
/dev/root / ext2 rw 0 0
/dev/sdc1 /alt ext2 rw 0 0
/dev/sdc3 /home/users ext2 rw 0 0
none /proc proc rw 0 0
/dev/sda1 /dos/drive_C msdos rw 0 0
/dev/sda5 /dos/drive_D msdos rw 0 0


Cheers,
Dick Johnson

Penguin : Linux version 2.4.18 on an i686 machine (797.90 BogoMips).

                 Windows-2000/Professional isn't.