Skip to Content.
Sympa Menu

sm-users - Re: [SM-Users] Please help with 2.6.14.2

sm-users AT lists.ibiblio.org

Subject: Sourcemage Users List

List archive

Chronological Thread  
  • From: David Brown <dmlb2000 AT gmail.com>
  • To: Tanel Tammet <tammet AT staff.ttu.ee>
  • Cc: sm-users AT lists.ibiblio.org
  • Subject: Re: [SM-Users] Please help with 2.6.14.2
  • Date: Tue, 29 Nov 2005 15:50:30 -0800

> I have been using SMGL for almost two years: so far I have managed to
> upgrade
> stuff, but now I' ve run into a serious problem with the kernel. I am
> running
> 2.6.11 and decided to upgrade to to 2.6.14.2. I know that the dev
> filesystem
> has been removed from 2.6.13 and above - and perhaps this is my problem -
> but I cannot find a cure.

Yeah you can't use devfs with 2.6.13 and above make sure you dispel
devfsd and cast udev

> The symptoms: the 2.6.14 kernel comes up until it has managed to find
> the keyboard,
> touchpad and mouse (succesfully mounting fs readonly) but refuses to go
> any further.
> It does not freeze - ctrl alt del works, each time I remove the mouse it
> says that
> the mouse has been removed, the cursor blinks happily - but it won' t go
> any further.
> Just before finding the keyboard it warns that it did not manage to
> create or use
> the initial console or terminal.

Probably because it's not there, the init.d spell will create a static
/dev which will create the inital console and lots of other device
nodes in /dev

> Needless to say, I have experimented a lot with various config options:
> using my
> previous config, using default, using different combinations etc: does
> not make
> a difference.

can't think of anything that would need to be changed...

> I am running ordinary intel on a Fujitu Siemens laptop and my SMGL is
> otherwise up to date, but the /etc stuff and lilo have no been upgraded
> for a while.
>
> In addition to the obvious suspicion that I have mislooked something in the
> kernel config, I suspect that the issue might have something to do with
> my lilo
> conf or fstab: I have attached both (they are pretty old).

Yes they are ;)


> ------------- lilo.conf ----------------------
>
> boot=/dev/discs/disc0/disc
> root=/dev/discs/disc0/part7
> prompt
> # lba32
> delay=128
> timeout=128
> install=boot-bmp.b
> bitmap=/boot/boot.bmp
> bmp-table=227p,233p,1,7
> bmp-colors=13,0,,11,0,13
> bmp-timer=76,30,11,0
> map=/boot/map
> image = /boot/vmlinubz-2.6.11.6
> label = linux-2.6.11.6
> read-only
> append=" devfs=mount"
> image = /boot/vmlinuz-2.6.14.2
> label = linux-2.6.14.2
> read-only
> image = /boot/vmlinubz-2.6.9
> label = linux-2.6.9
> read-only
> append=" devfs=mount"
> image = /boot/vmlinubz-2.4.26
> label = linux-2.4.26
> read-only
> append=" devfs=mount"
> image = /boot/vmlinubz-2.4.21
> label = linux-2.4.21
> read-only
> append=" devfs=mount"
> other=/dev/discs/disc0/part1
> label=Windows

try changing all the /dev/discs/... to /dev/hda{1-7} respectively
udev doesn't have any of the old devfs style directory structures to
it, if you have used devfsd the symlinks that daemon create are now
the actual devices.

> ---------------------- fstab -----------------
> proc /proc proc defaults 0 0
> devfs /dev devfs defaults 0 0
> #tmpfs /var/lock tmpfs defaults 0 0
> #tmpfs /var/run tmpfs defaults 0 0
> #tmpfs /tmp tmpfs defaults,size=256m,nr_inodes=64m 0 0
> devpts /dev/pts devpts defaults 0 0
>
> /dev/discs/disc0/part7 / ext3 defaults,noatime 0 1
> /dev/discs/disc0/part6 none swap defaults 0 2
> /dev/discs/disc0/part5 /mnt/windows vfat defaults 0 0
> /dev/cdroms/cdrom0 /mnt/cdrom iso9960 noauto,users,ro,exec 0 0
> ------------------------------------------
>
>
> There is a side issue which - perhaps - has something to do with
> the problem. Namely, I cannot run mount -f in mountall.sh in a normal
> way: mount just crashes, with the error caught by glibc. Instead, I have
> to use the following ugly hack:
>
> ---------------- normal mountall.sh snippet -----------------
>
> echo "Mounting root file system read/write..."
> {
> mount -n -o remount,rw / &&
> echo > /etc/mtab &&
> mount -f -o remount,rw /
> } || exit 1
>
> ----------- mountall.sh snippet which works for me ----------
>
> echo "Mounting root file system read/write, as modified by Tanel..."
> {
> mount -n -o remount,rw /dev/discs/disc0/part7 / &&
> cp /etc/mtab.initial /etc/mtab # &&
> #echo > /etc/mtab &&
> #mount -f -o remount,rw /dev/discs/disc0/part7 /
> } || exit 1

I've seen this before as well but I've never figured out why...
I run a jfs root filesystem and it also throws a glibc double free
error on the mount -f -o remount,rw /
iirc you can uncomment out the mount -f -o command and it will work,
something about having an empty /etc/mtab causes the double free
error.

> Any ideas are most welcome.

Plenty! They all lead to certain doom. ;)

- David Brown




Archive powered by MHonArc 2.6.24.

Top of Page