[gentoo-releng] x86-livecd - bugs and suggestions

[gentoo-releng] x86-livecd - bugs and suggestions

[Benjamin Judas]

[-- Attachment #1: Type: text/plain, Size: 1029 bytes --]

Hi folks,

Just a quick information-update about the bugs and suggestions for the
final-version of the x86-livecd for 2004.0

Bugs:
- Calculation of module dependencies fails at bootup
- selection of keymap not possible
- Inability to halt/reboot (system just stops when all processes are    
  shut-down)
- Hardware-detection not included

Suggestions:
- by spyderous: Make one of the two kernels an SMP-Kernel (for x86, this
would be the 
  2.6.1-gentoo-dev-kernel, kernel-name "experimental")
- by Strider: add a simple firewall on the cd that is automatically
started

If you have any suggestions or bugs, please tell me via _mail_, since I
am not always logging IRC (i.e. when I am at work) and my brain has a
limited capacity ;)

Kind regards,
  Benjamin

-- 
Benjamin Judas             http://dev.gentoo.org/~beejay
Gentoo-developer
Germany

GPG-Public-Key : http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xC31DEDD8
Key-Fingerprint : 4E65 AAFE 785B 61D8 E4D9  1671 E017 87B7 C31D EDD8

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 968 bytes --]

On Fri, 2004-01-23 at 17:04, Benjamin Judas wrote:

> Bugs:
> - Calculation of module dependencies fails at bootup

Anybody have a log?

> - selection of keymap not possible

We could add an alternative keymap script, that do that, and only
install with USE=livecd ... anybody got a template, etc to start
off?

> - Inability to halt/reboot (system just stops when all processes are    
>   shut-down)

Logs if possible?

> - Hardware-detection not included
> 

Sure will be nifty, but I guess catalyst should just install appropriate
package and add to boot runlevel?  Do not see any special handling, or
am I missing something?


As for 2.6 - think we can start to use module-init-tools as default for
ALL profiles, or is there ARCH issues I am not aware of?  Should just
be needed to adjust the default virtual.


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 890 bytes --]

On Fri, 2004-01-23 at 19:51, Benjamin Judas wrote:
> I can provide you "logs" - but only as a .png-file, since I test this
> livecd in a vmware session (errors also appear when booting off livecd
> "in reality"
> 
> > On Fri, 2004-01-23 at 17:04, Benjamin Judas wrote:
> > 
> > > Bugs:
> > > - Calculation of module dependencies fails at bootup
> > 
> > Anybody have a log?
> 
> see attachment modules.png.tar.bz2
> 

Is /lib/modules/<ver>/ writable?  It needs to be so ...

> 
> > > - Inability to halt/reboot (system just stops when all processes are    
> > >   shut-down)
> > 
> > Logs if possible?
> 
> see attachment halt.png.tar.bz2

Could you add a 'cat /proc/mounts' to the end of halt.sh, also include
one before calling 'halt'.


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

Commenting inline below.

>> Bugs:
>> - Calculation of module dependencies fails at bootup
>
> Anybody have a log?

reason for this is that /lib/modules/${KV} is off a read-only
filesystem.  It should not need to depmod that directory off
a livecd, as we won't have removed modules, etc.  An alternative
approach would be to mount /lib/modules/${KV} on a tmpfs,
make sure the sub directories are symlinks to the read-only
fs, and copy the files like  modules.pcimap to the tmpfs
so they are writable...

The first approach of just skipping the depmod if booting off
a livecd is the better approach.  Also, this failure really
doesn't mean anything or do anything harmful, which is one
reason I didn't fix it in my baselayout.

>
>> - selection of keymap not possible
>
> We could add an alternative keymap script, that do that, and only
> install with USE=livecd ... anybody got a template, etc to start
> off?

this I don't have a clue about, so i'm not commenting on.

>
>> - Inability to halt/reboot (system just stops when all processes are
>>   shut-down)
>
> Logs if possible?

The reason for this is quite simple.  The livecd was built with
the cdfs/loop type as 'normal' which creates a loopback filesystem.
The loopback filesystem gets unmounted '_before_' everything else
is remounted readonly.  But since the loopback is readonly as well,
it should never actually be being umounted before the reboot.  The
actual / filesystem is a tmpfs with symlinks back to the mounted
loopback in this scenerio, so when you umount the loopback, the /
filesystem has symlinks to nothing, hence the hang.

This is not an issue if you use cdfs/loop type of zisofs or noloop
as there is no loop device to be umounted.

>
>> - Hardware-detection not included
>>
>
> Sure will be nifty, but I guess catalyst should just install appropriate
> package and add to boot runlevel?  Do not see any special handling, or
> am I missing something?

I had not gotten a chance to put together the  livecd-tools package.
That script, as well as net-setup is in there.  Basically, it will just
be a package that needs to be a dep in the livecd-stage1 packages,
and a quick one-liner in the default-runscript.sh will get it all
up and running, really a simple thing.

>
>
> As for 2.6 - think we can start to use module-init-tools as default for
> ALL profiles, or is there ARCH issues I am not aware of?  Should just
> be needed to adjust the default virtual.

Right, as long as the current module-init-tools ebuild installs the
insmod.old for 2.4 kernels.  I believe at one time it did not.


Even though I've stepped away from livecd stuff, I need to fix this
stuff to build my amd64 2004.0 release anyhow.  I'd take care of this
since I'm more familiar with the problems if anyone wants me to. I can
do that this weekend.


-Brad House
brad_mssw@gentoo.org

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 3593 bytes --]

On Fri, 2004-01-23 at 20:43, Brad House wrote:
> Commenting inline below.
> 
> >> Bugs:
> >> - Calculation of module dependencies fails at bootup
> >
> > Anybody have a log?
> 
> reason for this is that /lib/modules/${KV} is off a read-only
> filesystem.


Right, as I suggested in my mail to beejay.

>   It should not need to depmod that directory off
> a livecd, as we won't have removed modules, etc.  An alternative
> approach would be to mount /lib/modules/${KV} on a tmpfs,
> make sure the sub directories are symlinks to the read-only
> fs, and copy the files like  modules.pcimap to the tmpfs
> so they are writable...
> 
> The first approach of just skipping the depmod if booting off
> a livecd is the better approach.  Also, this failure really
> doesn't mean anything or do anything harmful, which is one
> reason I didn't fix it in my baselayout.
> 

Easiest will be with a check to CDBOOT.

> >
> >> - selection of keymap not possible
> >
> > We could add an alternative keymap script, that do that, and only
> > install with USE=livecd ... anybody got a template, etc to start
> > off?
> 
> this I don't have a clue about, so i'm not commenting on.
> 

Same - I use the defaults 8)

> >
> >> - Inability to halt/reboot (system just stops when all processes are
> >>   shut-down)
> >
> > Logs if possible?
> 
> The reason for this is quite simple.  The livecd was built with
> the cdfs/loop type as 'normal' which creates a loopback filesystem.
> The loopback filesystem gets unmounted '_before_' everything else
> is remounted readonly.  But since the loopback is readonly as well,
> it should never actually be being umounted before the reboot.  The
> actual / filesystem is a tmpfs with symlinks back to the mounted
> loopback in this scenerio, so when you umount the loopback, the /
> filesystem has symlinks to nothing, hence the hang.
> 
> This is not an issue if you use cdfs/loop type of zisofs or noloop
> as there is no loop device to be umounted.
> 

Ok, I changed things a bit, so might have expected I guess.  I do
however have an issue with the naming.  Some time ago it was decided
that it should be /mnt/livecd.  Now it changed again.  Can we decide
on something now?  If so, fix is easy.

> >
> >> - Hardware-detection not included
> >>
> >
> > Sure will be nifty, but I guess catalyst should just install appropriate
> > package and add to boot runlevel?  Do not see any special handling, or
> > am I missing something?
> 
> I had not gotten a chance to put together the  livecd-tools package.
> That script, as well as net-setup is in there.  Basically, it will just
> be a package that needs to be a dep in the livecd-stage1 packages,
> and a quick one-liner in the default-runscript.sh will get it all
> up and running, really a simple thing.
> 
> >
> >
> > As for 2.6 - think we can start to use module-init-tools as default for
> > ALL profiles, or is there ARCH issues I am not aware of?  Should just
> > be needed to adjust the default virtual.
> 
> Right, as long as the current module-init-tools ebuild installs the
> insmod.old for 2.4 kernels.  I believe at one time it did not.
> 
> 
> Even though I've stepped away from livecd stuff, I need to fix this
> stuff to build my amd64 2004.0 release anyhow.  I'd take care of this
> since I'm more familiar with the problems if anyone wants me to. I can
> do that this weekend.
> 

Sound good.


-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

>> This is not an issue if you use cdfs/loop type of zisofs or noloop
>> as there is no loop device to be umounted.
>>
>
> Ok, I changed things a bit, so might have expected I guess.  I do
> however have an issue with the naming.  Some time ago it was decided
> that it should be /mnt/livecd.  Now it changed again.  Can we decide
> on something now?  If so, fix is easy.
>

Well, it's mounting to /mnt/loop right now.  I had no idea there was
a current naming convention (I did not have access to the prior 2.4
kernel livecd scripts that did any sort of loop mounting).  This would be
an easy change to make to whatever you want it to be, as it's just a
couple of lines in the genkernel linuxrc that need to be changed, just
let me know, as I needed to roll another genkernel that had some
fixes people posted in bugs.gentoo.org (along with some suggested
doc changes), before I totally abandon that project (as previously
discussed).

-Brad
brad_mssw@gentoo.org


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 1410 bytes --]

On Fri, 2004-01-23 at 23:38, Brad House wrote:
> >> This is not an issue if you use cdfs/loop type of zisofs or noloop
> >> as there is no loop device to be umounted.
> >>
> >
> > Ok, I changed things a bit, so might have expected I guess.  I do
> > however have an issue with the naming.  Some time ago it was decided
> > that it should be /mnt/livecd.  Now it changed again.  Can we decide
> > on something now?  If so, fix is easy.
> >
> 
> Well, it's mounting to /mnt/loop right now.  I had no idea there was
> a current naming convention (I did not have access to the prior 2.4
> kernel livecd scripts that did any sort of loop mounting).  This would be
> an easy change to make to whatever you want it to be, as it's just a
> couple of lines in the genkernel linuxrc that need to be changed, just
> let me know, as I needed to roll another genkernel that had some
> fixes people posted in bugs.gentoo.org (along with some suggested
> doc changes),

I do not have an issue, but '/mnt/livecd' or similar will be more
generic between different approaches.  I will leave it up to you
guys for final decision though.

> before I totally abandon that project (as previously
> discussed).
> 

No need to be rash.  If you still feel this way in a week, sure.


Cheers,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Daniel Robbins]

On Fri, 2004-01-23 at 15:26, Martin Schlemmer wrote:
> I do not have an issue, but '/mnt/livecd' or similar will be more
> generic between different approaches.  I will leave it up to you
> guys for final decision though.

I'd actually recommend not using anything in /mnt, as people have a
habit of mounting stuff to /mnt. When you do this, the loopback fs is
covered up and the LiveCD dies:

mount /dev/hda3 /mnt
(explosion!)

So it would be better to create a special mountpoint in /var somewhere.
I'm talking about the ideal solution, not necessarily what we need for
2004.0.

Regards,

Daniel




--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1.1: Type: text/plain, Size: 1085 bytes --]

On Fri, 2004-01-23 at 22:22, Benjamin Judas wrote:
> Am Fr, den 23.01.2004 schrieb Martin Schlemmer um 19:01:
> 
> > Is /lib/modules/<ver>/ writable?  It needs to be so ...
> 
> Gotcha! - This directory isn't writeable, because:
> 
> --8<-------------------------------------------------------
> livecd / # ls -al
> ...
> lrwxrwxrwx    1 root     root   21 Jan 23 11:29 lib ->
> mnt/cdrom/zisofs/lib
> ...
> -->8--------------------------------------------------------
> 
> 
> and:
> 
> --8<--------------------------------------------------------
> livecd / # mount
> ...
> /newroot/dev/cdroms/cdrom0 on /mnt/cdrom type is9660 (ro)
> ...
> -->8--------------------------------------------------------
> Additional Info : There is no /newroot in livecd's /
> 
> 
> > Could you add a 'cat /proc/mounts' to the end of halt.sh, also include
> > one before calling 'halt'.
> 
> see attached .png-file

Could you try attached patch?


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #1.2: rc-scripts-livecd.patch --]
[-- Type: text/x-patch, Size: 1887 bytes --]

Index: init.d/halt.sh
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/halt.sh,v
retrieving revision 1.46
diff -u -r1.46 halt.sh
--- init.d/halt.sh	26 Nov 2003 19:16:18 -0000	1.46
+++ init.d/halt.sh	24 Jan 2004 16:37:52 -0000
@@ -48,7 +48,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+            sort -r | sed -e 's:/mnt/livecd::g'`"
 [ -n "${remaining}" ] && {
 	sig=
 	retry=3
@@ -66,7 +67,8 @@
 			eend $? "Failed to unmount filesystems"
 		fi
 		
-		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+		            sort -r | sed -e 's:/mnt/livecd::g'`"
 		[ -z "${remaining}" ] && break
 		
 		/bin/fuser -k -m ${sig} ${remaining} &>/dev/null
Index: init.d/modules
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/modules,v
retrieving revision 1.28
diff -u -r1.28 modules
--- init.d/modules	11 Aug 2003 01:27:16 -0000	1.28
+++ init.d/modules	24 Jan 2004 16:37:52 -0000
@@ -82,9 +82,12 @@
 		return 1
 	fi
 	
-	ebegin "Calculating module dependencies"
-	/sbin/modules-update &>/dev/null
-	eend $? "Failed to calculate dependencies"
+	if [ -z "${CDBOOT}" ]
+	then
+		ebegin "Calculating module dependencies"
+		/sbin/modules-update &>/dev/null
+		eend $? "Failed to calculate dependencies"
+	fi
 
 	if [ -f /etc/modules.autoload -a ! -L /etc/modules.autoload ]
 	then

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1.1: Type: text/plain, Size: 1239 bytes --]

On Sat, 2004-01-24 at 18:38, Martin Schlemmer wrote:
> On Fri, 2004-01-23 at 22:22, Benjamin Judas wrote:
> > Am Fr, den 23.01.2004 schrieb Martin Schlemmer um 19:01:
> > 
> > > Is /lib/modules/<ver>/ writable?  It needs to be so ...
> > 
> > Gotcha! - This directory isn't writeable, because:
> > 
> > --8<-------------------------------------------------------
> > livecd / # ls -al
> > ...
> > lrwxrwxrwx    1 root     root   21 Jan 23 11:29 lib ->
> > mnt/cdrom/zisofs/lib
> > ...
> > -->8--------------------------------------------------------
> > 
> > 
> > and:
> > 
> > --8<--------------------------------------------------------
> > livecd / # mount
> > ...
> > /newroot/dev/cdroms/cdrom0 on /mnt/cdrom type is9660 (ro)
> > ...
> > -->8--------------------------------------------------------
> > Additional Info : There is no /newroot in livecd's /
> > 
> > 
> > > Could you add a 'cat /proc/mounts' to the end of halt.sh, also include
> > > one before calling 'halt'.
> > 
> > see attached .png-file
> 
> Could you try attached patch?
> 

Err, should be this one.

> 
> Thanks,
-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #1.2: rc-scripts-livecd.patch --]
[-- Type: text/x-patch, Size: 1881 bytes --]

Index: init.d/halt.sh
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/halt.sh,v
retrieving revision 1.46
diff -u -r1.46 halt.sh
--- init.d/halt.sh	26 Nov 2003 19:16:18 -0000	1.46
+++ init.d/halt.sh	24 Jan 2004 16:40:05 -0000
@@ -48,7 +48,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+            sort -r | sed -e 's:/newroot::g'`"
 [ -n "${remaining}" ] && {
 	sig=
 	retry=3
@@ -66,7 +67,8 @@
 			eend $? "Failed to unmount filesystems"
 		fi
 		
-		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+		            sort -r | sed -e 's:/newroot::g'`"
 		[ -z "${remaining}" ] && break
 		
 		/bin/fuser -k -m ${sig} ${remaining} &>/dev/null
Index: init.d/modules
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/modules,v
retrieving revision 1.28
diff -u -r1.28 modules
--- init.d/modules	11 Aug 2003 01:27:16 -0000	1.28
+++ init.d/modules	24 Jan 2004 16:40:05 -0000
@@ -82,9 +82,12 @@
 		return 1
 	fi
 	
-	ebegin "Calculating module dependencies"
-	/sbin/modules-update &>/dev/null
-	eend $? "Failed to calculate dependencies"
+	if [ -z "${CDBOOT}" ]
+	then
+		ebegin "Calculating module dependencies"
+		/sbin/modules-update &>/dev/null
+		eend $? "Failed to calculate dependencies"
+	fi
 
 	if [ -f /etc/modules.autoload -a ! -L /etc/modules.autoload ]
 	then

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 848 bytes --]

On Sat, 2004-01-24 at 01:55, Daniel Robbins wrote:
> On Fri, 2004-01-23 at 15:26, Martin Schlemmer wrote:
> > I do not have an issue, but '/mnt/livecd' or similar will be more
> > generic between different approaches.  I will leave it up to you
> > guys for final decision though.
> 
> I'd actually recommend not using anything in /mnt, as people have a
> habit of mounting stuff to /mnt. When you do this, the loopback fs is
> covered up and the LiveCD dies:
> 
> mount /dev/hda3 /mnt
> (explosion!)
> 
> So it would be better to create a special mountpoint in /var somewhere.
> I'm talking about the ideal solution, not necessarily what we need for
> 2004.0.
> 

What about '/var/lib/livecd' or something similar?


-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Daniel Robbins]

On Sat, 2004-01-24 at 09:40, Martin Schlemmer wrote:
> What about '/var/lib/livecd' or something similar?

Some inconspicuous location like that would work best long-term.

Regards,

Daniel


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Benjamin Judas]

[-- Attachment #1: Type: text/plain, Size: 1618 bytes --]

Am Sa, den 24.01.2004 schrieb Martin Schlemmer um 17:40:
> On Sat, 2004-01-24 at 18:38, Martin Schlemmer wrote:
> > On Fri, 2004-01-23 at 22:22, Benjamin Judas wrote:
> > > Am Fr, den 23.01.2004 schrieb Martin Schlemmer um 19:01:
> > > 
> > > > Is /lib/modules/<ver>/ writable?  It needs to be so ...
> > > 
> > > Gotcha! - This directory isn't writeable, because:
> > > 
> > > --8<-------------------------------------------------------
> > > livecd / # ls -al
> > > ...
> > > lrwxrwxrwx    1 root     root   21 Jan 23 11:29 lib ->
> > > mnt/cdrom/zisofs/lib
> > > ...
> > > -->8--------------------------------------------------------
> > > 
> > > 
> > > and:
> > > 
> > > --8<--------------------------------------------------------
> > > livecd / # mount
> > > ...
> > > /newroot/dev/cdroms/cdrom0 on /mnt/cdrom type is9660 (ro)
> > > ...
> > > -->8--------------------------------------------------------
> > > Additional Info : There is no /newroot in livecd's /
> > > 
> > > 
> > > > Could you add a 'cat /proc/mounts' to the end of halt.sh, also include
> > > > one before calling 'halt'.
> > > 
> > > see attached .png-file
> > 
> > Could you try attached patch?
> > 
> 
> Err, should be this one.
> 
> > 
> > Thanks,

Patch for modules works fabulous, no error-message anymore. But complete
shutdown still fails.

beejay
-- 
Benjamin Judas             http://dev.gentoo.org/~beejay
Gentoo-developer
Germany

GPG-Public-Key : http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xC31DEDD8
Key-Fingerprint : 4E65 AAFE 785B 61D8 E4D9  1671 E017 87B7 C31D EDD8

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1.1: Type: text/plain, Size: 1859 bytes --]

On Sun, 2004-01-25 at 00:51, Benjamin Judas wrote:
> Am Sa, den 24.01.2004 schrieb Martin Schlemmer um 17:40:
> > On Sat, 2004-01-24 at 18:38, Martin Schlemmer wrote:
> > > On Fri, 2004-01-23 at 22:22, Benjamin Judas wrote:
> > > > Am Fr, den 23.01.2004 schrieb Martin Schlemmer um 19:01:
> > > > 
> > > > > Is /lib/modules/<ver>/ writable?  It needs to be so ...
> > > > 
> > > > Gotcha! - This directory isn't writeable, because:
> > > > 
> > > > --8<-------------------------------------------------------
> > > > livecd / # ls -al
> > > > ...
> > > > lrwxrwxrwx    1 root     root   21 Jan 23 11:29 lib ->
> > > > mnt/cdrom/zisofs/lib
> > > > ...
> > > > -->8--------------------------------------------------------
> > > > 
> > > > 
> > > > and:
> > > > 
> > > > --8<--------------------------------------------------------
> > > > livecd / # mount
> > > > ...
> > > > /newroot/dev/cdroms/cdrom0 on /mnt/cdrom type is9660 (ro)
> > > > ...
> > > > -->8--------------------------------------------------------
> > > > Additional Info : There is no /newroot in livecd's /
> > > > 
> > > > 
> > > > > Could you add a 'cat /proc/mounts' to the end of halt.sh, also include
> > > > > one before calling 'halt'.
> > > > 
> > > > see attached .png-file
> > > 
> > > Could you try attached patch?
> > > 
> > 
> > Err, should be this one.
> > 
> > > 
> > > Thanks,
> 
> Patch for modules works fabulous, no error-message anymore. But complete
> shutdown still fails.
> 

Ok, I missed a spot - try this one.  One thing I am not too sure about,
is the sed's added on the loop section ... you might add a 
'cat /proc/mounts' just after to verify that it did not unmount it if
there are still issues ...


-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #1.2: rc-scripts-livecd.patch --]
[-- Type: text/x-patch, Size: 2120 bytes --]

Index: init.d/halt.sh
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/halt.sh,v
retrieving revision 1.46
diff -u -r1.46 halt.sh
--- init.d/halt.sh	26 Nov 2003 19:16:18 -0000	1.46
+++ init.d/halt.sh	25 Jan 2004 06:24:00 -0000
@@ -48,7 +48,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+            sort -r | sed -e 's:/newroot::g'`"
 [ -n "${remaining}" ] && {
 	sig=
 	retry=3
@@ -66,7 +67,8 @@
 			eend $? "Failed to unmount filesystems"
 		fi
 		
-		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+		            sort -r | sed -e 's:/newroot::g'`"
 		[ -z "${remaining}" ] && break
 		
 		/bin/fuser -k -m ${sig} ${remaining} &>/dev/null
@@ -95,7 +97,7 @@
 		[ "${x}" = "${y}" ] && do_unmount="no"
 	done
 	
-	if [ "${do_unmount}" = "yes" -a "${x}" != "/mnt/livecd" ]
+	if [ "${do_unmount}" = "yes" -a "${x}" != "/newroot" ]
 	then
 		umount ${x} &>/dev/null || {
 		
Index: init.d/modules
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/modules,v
retrieving revision 1.28
diff -u -r1.28 modules
--- init.d/modules	11 Aug 2003 01:27:16 -0000	1.28
+++ init.d/modules	25 Jan 2004 06:24:00 -0000
@@ -82,9 +82,12 @@
 		return 1
 	fi
 	
-	ebegin "Calculating module dependencies"
-	/sbin/modules-update &>/dev/null
-	eend $? "Failed to calculate dependencies"
+	if [ -z "${CDBOOT}" ]
+	then
+		ebegin "Calculating module dependencies"
+		/sbin/modules-update &>/dev/null
+		eend $? "Failed to calculate dependencies"
+	fi
 
 	if [ -f /etc/modules.autoload -a ! -L /etc/modules.autoload ]
 	then

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

[-- Attachment #1: Type: text/plain, Size: 1147 bytes --]

Attached is a patch based of yours that is tested and works fine.
This applies against your -r5 rc-scripts

I've also updated my  -r4  ebuild to pull a fully patched version
that includes these patches.

The CD reboots just fine with root mounted on a loop device. Also,
I took care of quite a few other little errors (LVM/RAID).

A few deps you'll need for livecds now:
baselayout-1.8.6.12-r4 (as of this evening, or a patched version
                        that may be released later by Azarah)
livecd-tools-1.0.0_beta2 (autoconfig and net-setup scripts)
genkernel-3.0.1_beta7

Works well, though we seem to be getting an error when it tries to start
hotplug from within the autoconfig script about needing to run
/sbin/depscan.sh.  I made it run /sbin/depscan during the livecd creation
to try to get rid of this error, but it did not seem to help, so I'm not
sure under what circumstances this can occur, but it does seem that
hotplug does start, and works properly, so I'm somewhat at a loss, other
than that, the livecd seems to be 100% in sync with the previous livecds
at this point.

(My testing was on AMD64 if anyone is curious)

-Brad

[-- Attachment #2: rc-scripts-livecd.patch --]
[-- Type: application/octet-stream, Size: 4092 bytes --]

diff -ruN rc-scripts-1.4.3.12p3/init.d/checkfs rc-scripts-1.4.3.12p7/init.d/checkfs
--- rc-scripts-1.4.3.12p3/init.d/checkfs	2004-01-05 18:50:42.000000000 -0500
+++ rc-scripts-1.4.3.12p7/init.d/checkfs	2004-01-25 02:16:11.000000000 -0500
@@ -14,7 +14,7 @@
 	#
 	# EVMS2 summport for /usr, /var ....
 	#
-	if [ -f /sbin/evms_activate ]
+	if [ "${CDBOOT}" = "" -a -f /sbin/evms_activate ]
 	then
 		ebegin "Activating EVMS2"
 		evms_activate
@@ -29,7 +29,7 @@
 	# NOTE: Add needed modules for LVM or RAID, etc
 	#       to /etc/modules.autoload if needed
 	
-	if [ -x /sbin/vgscan ] && \
+	if [ "${CDBOOT}" = "" -a -x /sbin/vgscan ] && \
 	   [ -d /proc/lvm -o "$(grep device-mapper /proc/misc 2>/dev/null)" ]
 	then
 		ebegin "Setting up the Logical Volume Manager"
@@ -48,7 +48,7 @@
 	# properly configured /etc/mdadm.conf for mdadm usage. Devices in
 	# /etc/mdadm.conf are initialized first, so any duplicate devices in
 	# /etc/raidtab will not get initialized.
-	if [ -f /proc/mdstat ]
+	if [ "${CDBOOT}" = "" -a -f /proc/mdstat ]
 	then
 		local mdadm_devices=
 		local raidtools_devices=
diff -ruN rc-scripts-1.4.3.12p3/init.d/halt.sh rc-scripts-1.4.3.12p7/init.d/halt.sh
--- rc-scripts-1.4.3.12p3/init.d/halt.sh	2003-11-26 14:16:18.000000000 -0500
+++ rc-scripts-1.4.3.12p7/init.d/halt.sh	2004-01-25 02:17:27.000000000 -0500
@@ -48,7 +48,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+	   sort -r | sed -e 's:/newroot::g' | grep -v mnt/livecd`"
 [ -n "${remaining}" ] && {
 	sig=
 	retry=3
@@ -66,7 +67,8 @@
 			eend $? "Failed to unmount filesystems"
 		fi
 		
-		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+			   sort -r | sed -e 's:/newroot::g' | grep -v mnt/livecd`"
 		[ -z "${remaining}" ] && break
 		
 		/bin/fuser -k -m ${sig} ${remaining} &>/dev/null
@@ -110,7 +112,7 @@
 eend 0
 
 # Stop LVM
-if [ -x /sbin/vgchange ] && [ -f /etc/lvmtab -o -d /etc/lvm ] && \
+if [ "${CDBOOT}" = "" ] && [ -x /sbin/vgchange ] && [ -f /etc/lvmtab -o -d /etc/lvm ] && \
    [ -d /proc/lvm  -o "`grep device-mapper /proc/misc 2>/dev/null`" ]
 then
 	ebegin "Shutting down the Logical Volume Manager"
diff -ruN rc-scripts-1.4.3.12p3/init.d/modules rc-scripts-1.4.3.12p7/init.d/modules
--- rc-scripts-1.4.3.12p3/init.d/modules	2003-08-10 21:27:16.000000000 -0400
+++ rc-scripts-1.4.3.12p7/init.d/modules	2004-01-25 02:12:55.000000000 -0500
@@ -81,11 +81,12 @@
 		eerror "ERROR:  system is missing /sbin/depmod !"
 		return 1
 	fi
-	
-	ebegin "Calculating module dependencies"
-	/sbin/modules-update &>/dev/null
-	eend $? "Failed to calculate dependencies"
-
+	if [ "${CDBOOT}" = "" ]
+	then
+		ebegin "Calculating module dependencies"
+		/sbin/modules-update &>/dev/null
+		eend $? "Failed to calculate dependencies"
+	fi
 	if [ -f /etc/modules.autoload -a ! -L /etc/modules.autoload ]
 	then
 		# Loop over every line in /etc/modules.autoload.
diff -ruN rc-scripts-1.4.3.12p3/sbin/rc rc-scripts-1.4.3.12p7/sbin/rc
--- rc-scripts-1.4.3.12p3/sbin/rc	2004-01-21 13:03:11.000000000 -0500
+++ rc-scripts-1.4.3.12p7/sbin/rc	2004-01-24 22:57:11.000000000 -0500
@@ -168,7 +168,7 @@
 	# Read off the kernel commandline to see if there's any special settings
 	# especially check to see if we need to set the  CDBOOT environment variable
 	# Note: /proc MUST be mounted
-	livecd_read_commandline
+	[ -f /sbin/livecd-functions.sh ] && livecd_read_commandline
 
 	if [ -d /sys -a "$(get_KV)" -ge "$(KV_to_int '2.6.0')" ]
 	then
@@ -418,7 +418,7 @@
 	fi
 
 	# If booting off CD, we want to update inittab before setting the runlevel
-	if [ "${CDBOOT}" != "" ]
+	if [ -f "/sbin/livecd-functions.sh" -a "${CDBOOT}" != "" ]
 	then
 		ebegin "Updating inittab"
 		livecd_fix_inittab

[-- Attachment #3: Type: text/plain, Size: 40 bytes --]

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 1434 bytes --]

On Sun, 2004-01-25 at 10:20, Brad House wrote:
> Attached is a patch based of yours that is tested and works fine.
> This applies against your -r5 rc-scripts
> 

Are we sure we do not want lvm, etc setup?

> I've also updated my  -r4  ebuild to pull a fully patched version
> that includes these patches.
> 
> The CD reboots just fine with root mounted on a loop device. Also,
> I took care of quite a few other little errors (LVM/RAID).
> 
> A few deps you'll need for livecds now:
> baselayout-1.8.6.12-r4 (as of this evening, or a patched version
>                         that may be released later by Azarah)
> livecd-tools-1.0.0_beta2 (autoconfig and net-setup scripts)
> genkernel-3.0.1_beta7
> 
> Works well, though we seem to be getting an error when it tries to start
> hotplug from within the autoconfig script about needing to run
> /sbin/depscan.sh.  I made it run /sbin/depscan during the livecd creation
> to try to get rid of this error, but it did not seem to help, so I'm not
> sure under what circumstances this can occur, but it does seem that
> hotplug does start, and works properly, so I'm somewhat at a loss, other
> than that, the livecd seems to be 100% in sync with the previous livecds
> at this point.
> 
> (My testing was on AMD64 if anyone is curious)
> 
> -Brad
-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

> On Sun, 2004-01-25 at 10:20, Brad House wrote:
>> Attached is a patch based of yours that is tested and works fine.
>> This applies against your -r5 rc-scripts
>>
>
> Are we sure we do not want lvm, etc setup?
>

well, for someone without LVM it seems as though it just pops up
errors, which will tend to provide a crapload of bug reports.
We're still providing utilities, etc, so someone can manually specify
vgscan and vgchange.  If you have a way to prevent it from popping
up the errors (and I don't remember what the errors were, just that
they were annoying), and still keep the LVM detection stuff, that
would be fine.


-Brad

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 968 bytes --]

On Sun, 2004-01-25 at 17:52, Brad House wrote:
> > On Sun, 2004-01-25 at 10:20, Brad House wrote:
> >> Attached is a patch based of yours that is tested and works fine.
> >> This applies against your -r5 rc-scripts
> >>
> >
> > Are we sure we do not want lvm, etc setup?
> >
> 
> well, for someone without LVM it seems as though it just pops up
> errors, which will tend to provide a crapload of bug reports.
> We're still providing utilities, etc, so someone can manually specify
> vgscan and vgchange.  If you have a way to prevent it from popping
> up the errors (and I don't remember what the errors were, just that
> they were annoying), and still keep the LVM detection stuff, that
> would be fine.

I am not very familiar with raid/lvm, so was not sure.  I assume without
lvmtab and raidtab, it means nothing anyhow?.


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 347 bytes --]

On Sun, 2004-01-25 at 10:20, Brad House wrote:
> Attached is a patch based of yours that is tested and works fine.
> This applies against your -r5 rc-scripts
> 

BTW, think we should do the /var/lib/livecd change now, or later?


-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 789 bytes --]

On Sun, 2004-01-25 at 10:20, Brad House wrote:

> Works well, though we seem to be getting an error when it tries to start
> hotplug from within the autoconfig script about needing to run
> /sbin/depscan.sh.  I made it run /sbin/depscan during the livecd creation
> to try to get rid of this error, but it did not seem to help, so I'm not
> sure under what circumstances this can occur, but it does seem that
> hotplug does start, and works properly, so I'm somewhat at a loss, other
> than that, the livecd seems to be 100% in sync with the previous livecds
> at this point.
> 

What type of fs is / once booted ?  Also, when abouts is the autoconfig
script run?

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

>
> I am not very familiar with raid/lvm, so was not sure.  I assume without
> lvmtab and raidtab, it means nothing anyhow?.
>

pretty much :/ So we should be fine disabling it on livecds.

-Brad

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

>
> BTW, think we should do the /var/lib/livecd change now, or later?

Let's do it later. I'd have to go back and change genkernel now
to make that change, and it works as is, as long as anyone doesn't
overwrite /mnt (/me sighs).

-Brad

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

>> Works well, though we seem to be getting an error when it tries to start
>> hotplug from within the autoconfig script about needing to run
>> /sbin/depscan.sh.  I made it run /sbin/depscan during the livecd
>> creation
>> to try to get rid of this error, but it did not seem to help, so I'm not
>> sure under what circumstances this can occur, but it does seem that
>> hotplug does start, and works properly, so I'm somewhat at a loss, other
>> than that, the livecd seems to be 100% in sync with the previous livecds
>> at this point.
>>
>
> What type of fs is / once booted ?  Also, when abouts is the autoconfig
> script run?

/ ends up on a tmpfs, with symlinks back to the read-only livecd stuff
for the main directories, and /etc is also on a tmpfs so it's writable.
Autoconfig I made have a dep on   'need modules', so it gets run sometime
after the manually specified modules are loaded, but before local which
is  'after *'.  It seemed it started _right_ before local on my system
though. (the autoconfig script I'm using is in this small tarball:
http://dev.gentoo.org/~brad_mssw/livecd-tools-1.0.0_beta2.tar.bz2
incase you wanted to look at how it calls hotplug)

-Brad


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 1544 bytes --]

On Sun, 2004-01-25 at 18:29, Brad House wrote:
> >> Works well, though we seem to be getting an error when it tries to start
> >> hotplug from within the autoconfig script about needing to run
> >> /sbin/depscan.sh.  I made it run /sbin/depscan during the livecd
> >> creation
> >> to try to get rid of this error, but it did not seem to help, so I'm not
> >> sure under what circumstances this can occur, but it does seem that
> >> hotplug does start, and works properly, so I'm somewhat at a loss, other
> >> than that, the livecd seems to be 100% in sync with the previous livecds
> >> at this point.
> >>
> >
> > What type of fs is / once booted ?  Also, when abouts is the autoconfig
> > script run?
> 
> / ends up on a tmpfs, with symlinks back to the read-only livecd stuff
> for the main directories, and /etc is also on a tmpfs so it's writable.
> Autoconfig I made have a dep on   'need modules', so it gets run sometime
> after the manually specified modules are loaded, but before local which
> is  'after *'.  It seemed it started _right_ before local on my system
> though. (the autoconfig script I'm using is in this small tarball:
> http://dev.gentoo.org/~brad_mssw/livecd-tools-1.0.0_beta2.tar.bz2
> incase you wanted to look at how it calls hotplug)
> 


Could you try to change it to:

  env -i /etc/init.d/hotplug start

just in case its an env full or whatever issue?


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

>
> Could you try to change it to:
>
>   env -i /etc/init.d/hotplug start
>
> just in case its an env full or whatever issue?
>

yep, I can try that.  Also, it seems when using zisofs, it's
killing all processes, or something, so it won't reboot. I think
it's because it's trying to umount the cdrom, so I tried to do
the patch pasted onto the end of this e-mail.  It didn't seem
to help.  I need to add in some debugging, but I'm 99% sure
the error exists in this region as I see the  fuser  command
sending a signal 9 to all processes, so it needs to be prevented
I would assume :/

Any help on that would be appreciated as well.  Thanks!

Oh yeah, if you use 'normal' for the cdfs type now, it works
great, and does not umount the loopback (reboots fine). It's just
zisofs I'm having an issue with now.

-Brad


--- rc-scripts-1.4.3.12p3/init.d/halt.sh        2003-11-26
14:16:18.000000000 -0500
+++ rc-scripts-1.4.3.12p8/init.d/halt.sh        2004-01-25
22:42:15.000000000 -0500
@@ -48,7 +48,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}'
/proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}'
/proc/mounts | \
+          sort -r | sed -e 's:/newroot::g' | grep -v mnt/livecd`"
 [ -n "${remaining}" ] && {
        sig=
        retry=3
@@ -66,7 +67,8 @@
                        eend $? "Failed to unmount filesystems"
                fi

-               remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/"
{print $2}' /proc/mounts | sort -r`"
+               remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/"
{print $2}' /proc/mounts | \
+                          sort -r | sed -e 's:/newroot::g' | grep -v
mnt/livecd`"
                [ -z "${remaining}" ] && break

                /bin/fuser -k -m ${sig} ${remaining} &>/dev/null
@@ -94,8 +96,20 @@
        do
                [ "${x}" = "${y}" ] && do_unmount="no"
        done
-
-       if [ "${do_unmount}" = "yes" -a "${x}" != "/mnt/livecd" ]
+
+       if [ "${CDBOOT}" != "" ]
+       then
+               case ${x} in
+                       *mnt/cdrom*)
+                               do_unmount="no"
+                       ;;
+                       *mnt/livecd*)
+                               do_unmount="no"
+                       ;;
+               esac
+       fi
+
+       if [ "${do_unmount}" = "yes" ]
        then
                umount ${x} &>/dev/null || {

@@ -110,7 +124,7 @@
 eend 0

 # Stop LVM
-if [ -x /sbin/vgchange ] && [ -f /etc/lvmtab -o -d /etc/lvm ] && \
+if [ "${CDBOOT}" = "" ] && [ -x /sbin/vgchange ] && [ -f /etc/lvmtab -o
-d /etc/lvm ] && \
    [ -d /proc/lvm  -o "`grep device-mapper /proc/misc 2>/dev/null`" ]
 then
        ebegin "Shutting down the Logical Volume Manager"


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 928 bytes --]

On Mon, 2004-01-26 at 20:02, Brad House wrote:
> >
> > Could you try to change it to:
> >
> >   env -i /etc/init.d/hotplug start
> >
> > just in case its an env full or whatever issue?
> >
> 
> yep, I can try that.  Also, it seems when using zisofs, it's
> killing all processes, or something, so it won't reboot. I think
> it's because it's trying to umount the cdrom, so I tried to do
> the patch pasted onto the end of this e-mail.  It didn't seem
> to help.  I need to add in some debugging, but I'm 99% sure
> the error exists in this region as I see the  fuser  command
> sending a signal 9 to all processes, so it needs to be prevented
> I would assume :/
> 

Do you have a smallish iso I can try with?  My bandwidth for the month
is nearly done, so nothing too large please.


Thanks!

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

Err, do you have an AMD64 box? My livecd is amd64, but same stuff
should apply to x86 ... you'd have to contact beejay on that though.
The livecd is only like 50-55MB though.

-Brad

>
> Do you have a smallish iso I can try with?  My bandwidth for the month
> is nearly done, so nothing too large please.
>
>
> Thanks!
>
> --
>
> Martin Schlemmer
> Gentoo Linux Developer, Desktop/System Team Developer
> Cape Town, South Africa
>
>
>


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1.1: Type: text/plain, Size: 649 bytes --]

Hi

Ok, attached works for me (when edited in).  For the livecd I used
at least (zisofs) it still uses /newroot, so that was not caught
properly with that silly sed.  With hotplug not starting, its just
variables that should not be exported (although the 'env -i ...' works
but is ugly and not expected thing to try and do).

Oh, and I left the lvm shutdown stuff - I guess it will be safer, and
we could &>/dev/null if there are any bugs ... thoughts?

Please test and let me know - its against last baselayout.


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #1.2: rc-scripts-livecd.patch --]
[-- Type: text/x-patch, Size: 5244 bytes --]

Index: init.d/checkfs
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/checkfs,v
retrieving revision 1.32
diff -u -r1.32 checkfs
--- init.d/checkfs	5 Jan 2004 23:50:42 -0000	1.32
+++ init.d/checkfs	27 Jan 2004 17:08:25 -0000
@@ -14,7 +14,7 @@
 	#
 	# EVMS2 summport for /usr, /var ....
 	#
-	if [ -f /sbin/evms_activate ]
+	if [ -z "${CDBOOT}" -a -f /sbin/evms_activate ]
 	then
 		ebegin "Activating EVMS2"
 		evms_activate
@@ -29,7 +29,7 @@
 	# NOTE: Add needed modules for LVM or RAID, etc
 	#       to /etc/modules.autoload if needed
 	
-	if [ -x /sbin/vgscan ] && \
+	if [ -z "${CDBOOT}" -a -x /sbin/vgscan ] && \
 	   [ -d /proc/lvm -o "$(grep device-mapper /proc/misc 2>/dev/null)" ]
 	then
 		ebegin "Setting up the Logical Volume Manager"
@@ -48,7 +48,7 @@
 	# properly configured /etc/mdadm.conf for mdadm usage. Devices in
 	# /etc/mdadm.conf are initialized first, so any duplicate devices in
 	# /etc/raidtab will not get initialized.
-	if [ -f /proc/mdstat ]
+	if [ -z "${CDBOOT}" -a -f /proc/mdstat ]
 	then
 		local mdadm_devices=
 		local raidtools_devices=
Index: init.d/halt.sh
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/halt.sh,v
retrieving revision 1.46
diff -u -r1.46 halt.sh
--- init.d/halt.sh	26 Nov 2003 19:16:18 -0000	1.46
+++ init.d/halt.sh	27 Jan 2004 17:08:26 -0000
@@ -48,7 +48,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+            sort -r | grep -v '/newroot' | grep -v '/mnt/livecd'`"
 [ -n "${remaining}" ] && {
 	sig=
 	retry=3
@@ -66,7 +67,8 @@
 			eend $? "Failed to unmount filesystems"
 		fi
 		
-		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+		            sort -r | grep -v '/newroot' | grep -v '/mnt/livecd'`"
 		[ -z "${remaining}" ] && break
 		
 		/bin/fuser -k -m ${sig} ${remaining} &>/dev/null
@@ -95,7 +97,8 @@
 		[ "${x}" = "${y}" ] && do_unmount="no"
 	done
 	
-	if [ "${do_unmount}" = "yes" -a "${x}" != "/mnt/livecd" ]
+	if [ "${do_unmount}" = "yes" -a \
+	     "${x}" != "/newroot" -a "${x}" != "/mnt/livecd" ]
 	then
 		umount ${x} &>/dev/null || {
 		
Index: init.d/modules
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/init.d/modules,v
retrieving revision 1.28
diff -u -r1.28 modules
--- init.d/modules	11 Aug 2003 01:27:16 -0000	1.28
+++ init.d/modules	27 Jan 2004 17:08:26 -0000
@@ -82,9 +82,12 @@
 		return 1
 	fi
 	
-	ebegin "Calculating module dependencies"
-	/sbin/modules-update &>/dev/null
-	eend $? "Failed to calculate dependencies"
+	if [ -z "${CDBOOT}" ]
+	then
+		ebegin "Calculating module dependencies"
+		/sbin/modules-update &>/dev/null
+		eend $? "Failed to calculate dependencies"
+	fi
 
 	if [ -f /etc/modules.autoload -a ! -L /etc/modules.autoload ]
 	then
Index: sbin/rc
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/sbin/rc,v
retrieving revision 1.71
diff -u -r1.71 rc
--- sbin/rc	21 Jan 2004 18:03:11 -0000	1.71
+++ sbin/rc	27 Jan 2004 17:08:26 -0000
@@ -168,7 +168,7 @@
 	# Read off the kernel commandline to see if there's any special settings
 	# especially check to see if we need to set the  CDBOOT environment variable
 	# Note: /proc MUST be mounted
-	livecd_read_commandline
+	[ -f /sbin/livecd-functions.sh ] && livecd_read_commandline
 
 	if [ -d /sys -a "$(get_KV)" -ge "$(KV_to_int '2.6.0')" ]
 	then
@@ -418,7 +418,7 @@
 	fi
 
 	# If booting off CD, we want to update inittab before setting the runlevel
-	if [ "${CDBOOT}" != "" ]
+	if [ -f "/sbin/livecd-functions.sh" -a -n "${CDBOOT}" ]
 	then
 		ebegin "Updating inittab"
 		livecd_fix_inittab
Index: src/awk/gendepends.awk
===================================================================
RCS file: /home/cvsroot/gentoo-src/rc-scripts/src/awk/gendepends.awk,v
retrieving revision 1.19
diff -u -r1.19 gendepends.awk
--- src/awk/gendepends.awk	29 Dec 2003 01:00:27 -0000	1.19
+++ src/awk/gendepends.awk	27 Jan 2004 17:08:26 -0000
@@ -466,14 +466,14 @@
 		print "" >> (CACHEDTREE)
 	}
 
-	print "export RC_DEPEND_TREE" >> (CACHEDTREE)
-	print "export RC_GOT_DEPTREE_INFO=\"yes\"" >> (CACHEDTREE)
+	# Do not export these, as we want them local
+	print "RC_GOT_DEPTREE_INFO=\"yes\"" >> (CACHEDTREE)
 	print "" >> (CACHEDTREE)
 
 	if (check_provide("logger"))
-		print "export LOGGER_SERVICE=\"" get_provide("logger") "\"" >> (CACHEDTREE)
+		print "LOGGER_SERVICE=\"" get_provide("logger") "\"" >> (CACHEDTREE)
 	else
-		print "export LOGGER_SERVICE=" >> (CACHEDTREE)
+		print "LOGGER_SERVICE=" >> (CACHEDTREE)
 		
 	close(CACHEDTREE)
 }

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

should be able to do that tonight when I get home.

-Brad

> On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
>
> Brad, did you see/could you test this?  Benjamin?
>
>> Hi
>>
>> Ok, attached works for me (when edited in).  For the livecd I used
>> at least (zisofs) it still uses /newroot, so that was not caught
>> properly with that silly sed.  With hotplug not starting, its just
>> variables that should not be exported (although the 'env -i ...' works
>> but is ugly and not expected thing to try and do).
>>
>> Oh, and I left the lvm shutdown stuff - I guess it will be safer, and
>> we could &>/dev/null if there are any bugs ... thoughts?
>>
>> Please test and let me know - its against last baselayout.
>>
>>
>> Thanks,
> --
> Martin Schlemmer
>


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Benjamin Judas]

[-- Attachment #1: Type: text/plain, Size: 802 bytes --]

Am Mi, den 28.01.2004 schrieb Martin Schlemmer um 18:44:
> On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
> 
> Brad, did you see/could you test this?  Benjamin?
> 

Hi Martin,

I tried you patch today and it looks like bootup is running fine now -
no errors anymore in this part. All my hardware got detected and is
working (we might need some kernel-tweaking -- but that shouldn't affect
baselayout in the startup-stage anymore). Rebooting and halting however
still don't work - I am also using zisofs. But it really looks good so
far :)

-- 
Benjamin Judas             http://dev.gentoo.org/~beejay
Gentoo-developer
Germany

GPG-Public-Key : http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xC31DEDD8
Key-Fingerprint : 4E65 AAFE 785B 61D8 E4D9  1671 E017 87B7 C31D EDD8

[-- Attachment #2: Dies ist ein digital signierter Nachrichtenteil --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 822 bytes --]

On Thu, 2004-01-29 at 19:19, Benjamin Judas wrote:
> Am Mi, den 28.01.2004 schrieb Martin Schlemmer um 18:44:
> > On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
> > 
> > Brad, did you see/could you test this?  Benjamin?
> > 
> 
> Hi Martin,
> 
> I tried you patch today and it looks like bootup is running fine now -
> no errors anymore in this part. All my hardware got detected and is
> working (we might need some kernel-tweaking -- but that shouldn't affect
> baselayout in the startup-stage anymore). Rebooting and halting however
> still don't work - I am also using zisofs. But it really looks good so
> far :)

Could you please put the new version of the iso up again?


Thanks,
-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

I have attached a new patch against the p3 rc-scripts.
This incorporate's azarah's fixes, as well as another patch
for halt.sh which resolves the no-reboot issue for zisofs
cds.  As usual, I have updated the baselayout-1.8.6.12-r4.ebuild
with my patches, so anyone building livecd's needs to mark
that stable for their arch for testing.

-Brad

> On Thu, 2004-01-29 at 19:19, Benjamin Judas wrote:
>> Am Mi, den 28.01.2004 schrieb Martin Schlemmer um 18:44:
>> > On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
>> >
>> > Brad, did you see/could you test this?  Benjamin?
>> >
>>
>> Hi Martin,
>>
>> I tried you patch today and it looks like bootup is running fine now -
>> no errors anymore in this part. All my hardware got detected and is
>> working (we might need some kernel-tweaking -- but that shouldn't affect
>> baselayout in the startup-stage anymore). Rebooting and halting however
>> still don't work - I am also using zisofs. But it really looks good so
>> far :)
>
> Could you please put the new version of the iso up again?
>
>
> Thanks,
> --
>
> Martin Schlemmer
> Gentoo Linux Developer, Desktop/System Team Developer
> Cape Town, South Africa
>
>
>


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

figures I forgot to attach it.


> I have attached a new patch against the p3 rc-scripts.
> This incorporate's azarah's fixes, as well as another patch
> for halt.sh which resolves the no-reboot issue for zisofs
> cds.  As usual, I have updated the baselayout-1.8.6.12-r4.ebuild
> with my patches, so anyone building livecd's needs to mark
> that stable for their arch for testing.
>
> -Brad
>
>> On Thu, 2004-01-29 at 19:19, Benjamin Judas wrote:
>>> Am Mi, den 28.01.2004 schrieb Martin Schlemmer um 18:44:
>>> > On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
>>> >
>>> > Brad, did you see/could you test this?  Benjamin?
>>> >
>>>
>>> Hi Martin,
>>>
>>> I tried you patch today and it looks like bootup is running fine now -
>>> no errors anymore in this part. All my hardware got detected and is
>>> working (we might need some kernel-tweaking -- but that shouldn't
>>> affect
>>> baselayout in the startup-stage anymore). Rebooting and halting however
>>> still don't work - I am also using zisofs. But it really looks good so
>>> far :)
>>
>> Could you please put the new version of the iso up again?
>>
>>
>> Thanks,
>> --
>>
>> Martin Schlemmer
>> Gentoo Linux Developer, Desktop/System Team Developer
>> Cape Town, South Africa
>>
>>
>>
>
>
> --
> gentoo-releng@gentoo.org mailing list
>
>
>


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

[-- Attachment #1: Type: text/plain, Size: 1312 bytes --]

figures I forgot to attach it.
-Brad

> I have attached a new patch against the p3 rc-scripts.
> This incorporate's azarah's fixes, as well as another patch
> for halt.sh which resolves the no-reboot issue for zisofs
> cds.  As usual, I have updated the baselayout-1.8.6.12-r4.ebuild
> with my patches, so anyone building livecd's needs to mark
> that stable for their arch for testing.
>
> -Brad
>
>> On Thu, 2004-01-29 at 19:19, Benjamin Judas wrote:
>>> Am Mi, den 28.01.2004 schrieb Martin Schlemmer um 18:44:
>>> > On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
>>> >
>>> > Brad, did you see/could you test this?  Benjamin?
>>> >
>>>
>>> Hi Martin,
>>>
>>> I tried you patch today and it looks like bootup is running fine now -
>>> no errors anymore in this part. All my hardware got detected and is
>>> working (we might need some kernel-tweaking -- but that shouldn't
>>> affect
>>> baselayout in the startup-stage anymore). Rebooting and halting however
>>> still don't work - I am also using zisofs. But it really looks good so
>>> far :)
>>
>> Could you please put the new version of the iso up again?
>>
>>
>> Thanks,
>> --
>>
>> Martin Schlemmer
>> Gentoo Linux Developer, Desktop/System Team Developer
>> Cape Town, South Africa
>>
>>
>>
>
>
> --
> gentoo-releng@gentoo.org mailing list
>
>
>

[-- Attachment #2: rc-scripts-livecd-02012004.diff --]
[-- Type: application/octet-stream, Size: 5852 bytes --]

diff -ruN rc-scripts-1.4.3.12p3/init.d/checkfs rc-scripts-1.4.3.12p9/init.d/checkfs
--- rc-scripts-1.4.3.12p3/init.d/checkfs	2004-01-05 18:50:42.000000000 -0500
+++ rc-scripts-1.4.3.12p9/init.d/checkfs	2004-01-25 02:16:11.000000000 -0500
@@ -14,7 +14,7 @@
 	#
 	# EVMS2 summport for /usr, /var ....
 	#
-	if [ -f /sbin/evms_activate ]
+	if [ "${CDBOOT}" = "" -a -f /sbin/evms_activate ]
 	then
 		ebegin "Activating EVMS2"
 		evms_activate
@@ -29,7 +29,7 @@
 	# NOTE: Add needed modules for LVM or RAID, etc
 	#       to /etc/modules.autoload if needed
 	
-	if [ -x /sbin/vgscan ] && \
+	if [ "${CDBOOT}" = "" -a -x /sbin/vgscan ] && \
 	   [ -d /proc/lvm -o "$(grep device-mapper /proc/misc 2>/dev/null)" ]
 	then
 		ebegin "Setting up the Logical Volume Manager"
@@ -48,7 +48,7 @@
 	# properly configured /etc/mdadm.conf for mdadm usage. Devices in
 	# /etc/mdadm.conf are initialized first, so any duplicate devices in
 	# /etc/raidtab will not get initialized.
-	if [ -f /proc/mdstat ]
+	if [ "${CDBOOT}" = "" -a -f /proc/mdstat ]
 	then
 		local mdadm_devices=
 		local raidtools_devices=
diff -ruN rc-scripts-1.4.3.12p3/init.d/halt.sh rc-scripts-1.4.3.12p9/init.d/halt.sh
--- rc-scripts-1.4.3.12p3/init.d/halt.sh	2003-11-26 14:16:18.000000000 -0500
+++ rc-scripts-1.4.3.12p9/init.d/halt.sh	2004-02-01 17:37:07.000000000 -0500
@@ -2,6 +2,10 @@
 # Distributed under the terms of the GNU General Public License v2
 # $Header: /home/cvsroot/gentoo-src/rc-scripts/init.d/halt.sh,v 1.46 2003/11/26 19:16:18 azarah Exp $
 
+# Check to see if this is a livecd, if it is read the commandline
+# this mainly makes sure $CDBOOT is defined if it's a livecd
+[ -f "/sbin/livecd-functions.sh" ] && source /sbin/livecd-functions.sh && \
+	livecd_read_commandline
 
 # Reset pam_console permissions
 [ -x /sbin/pam_console_apply -a ! -c /dev/.devfsd ] && \
@@ -48,7 +52,8 @@
 # Unmount file systems, killing processes if we have to.
 # Unmount loopback stuff first
 # Use `umount -d` to detach the loopback device
-remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+	   sort -r | sed -e 's:/newroot::g' | grep -v mnt/livecd`"
 [ -n "${remaining}" ] && {
 	sig=
 	retry=3
@@ -66,7 +71,8 @@
 			eend $? "Failed to unmount filesystems"
 		fi
 		
-		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | sort -r`"
+		remaining="`awk '!/^#/ && $1 ~ /^\/dev\/loop/ && $2 != "/" {print $2}' /proc/mounts | \
+			   sort -r | sed -e 's:/newroot::g' | grep -v mnt/livecd`"
 		[ -z "${remaining}" ] && break
 		
 		/bin/fuser -k -m ${sig} ${remaining} &>/dev/null
@@ -94,8 +100,20 @@
 	do
 		[ "${x}" = "${y}" ] && do_unmount="no"
 	done
-	
-	if [ "${do_unmount}" = "yes" -a "${x}" != "/mnt/livecd" ]
+
+	if [ "${CDBOOT}" != "" ]
+	then
+		case ${x} in
+			*mnt/cdrom*)
+				do_unmount="no"
+			;;
+			*mnt/livecd*)
+				do_unmount="no"
+			;;
+		esac
+	fi
+
+	if [ "${do_unmount}" = "yes" ]
 	then
 		umount ${x} &>/dev/null || {
 		
@@ -110,7 +128,7 @@
 eend 0
 
 # Stop LVM
-if [ -x /sbin/vgchange ] && [ -f /etc/lvmtab -o -d /etc/lvm ] && \
+if [ "${CDBOOT}" = "" ] && [ -x /sbin/vgchange ] && [ -f /etc/lvmtab -o -d /etc/lvm ] && \
    [ -d /proc/lvm  -o "`grep device-mapper /proc/misc 2>/dev/null`" ]
 then
 	ebegin "Shutting down the Logical Volume Manager"
diff -ruN rc-scripts-1.4.3.12p3/init.d/modules rc-scripts-1.4.3.12p9/init.d/modules
--- rc-scripts-1.4.3.12p3/init.d/modules	2003-08-10 21:27:16.000000000 -0400
+++ rc-scripts-1.4.3.12p9/init.d/modules	2004-01-25 02:12:55.000000000 -0500
@@ -81,11 +81,12 @@
 		eerror "ERROR:  system is missing /sbin/depmod !"
 		return 1
 	fi
-	
-	ebegin "Calculating module dependencies"
-	/sbin/modules-update &>/dev/null
-	eend $? "Failed to calculate dependencies"
-
+	if [ "${CDBOOT}" = "" ]
+	then
+		ebegin "Calculating module dependencies"
+		/sbin/modules-update &>/dev/null
+		eend $? "Failed to calculate dependencies"
+	fi
 	if [ -f /etc/modules.autoload -a ! -L /etc/modules.autoload ]
 	then
 		# Loop over every line in /etc/modules.autoload.
diff -ruN rc-scripts-1.4.3.12p3/sbin/rc rc-scripts-1.4.3.12p9/sbin/rc
--- rc-scripts-1.4.3.12p3/sbin/rc	2004-01-21 13:03:11.000000000 -0500
+++ rc-scripts-1.4.3.12p9/sbin/rc	2004-01-24 22:57:11.000000000 -0500
@@ -168,7 +168,7 @@
 	# Read off the kernel commandline to see if there's any special settings
 	# especially check to see if we need to set the  CDBOOT environment variable
 	# Note: /proc MUST be mounted
-	livecd_read_commandline
+	[ -f /sbin/livecd-functions.sh ] && livecd_read_commandline
 
 	if [ -d /sys -a "$(get_KV)" -ge "$(KV_to_int '2.6.0')" ]
 	then
@@ -418,7 +418,7 @@
 	fi
 
 	# If booting off CD, we want to update inittab before setting the runlevel
-	if [ "${CDBOOT}" != "" ]
+	if [ -f "/sbin/livecd-functions.sh" -a "${CDBOOT}" != "" ]
 	then
 		ebegin "Updating inittab"
 		livecd_fix_inittab
diff -ruN rc-scripts-1.4.3.12p3/src/awk/gendepends.awk rc-scripts-1.4.3.12p9/src/awk/gendepends.awk
--- rc-scripts-1.4.3.12p3/src/awk/gendepends.awk	2003-12-28 20:00:27.000000000 -0500
+++ rc-scripts-1.4.3.12p9/src/awk/gendepends.awk	2004-02-01 17:34:36.000000000 -0500
@@ -466,14 +466,14 @@
 		print "" >> (CACHEDTREE)
 	}
 
-	print "export RC_DEPEND_TREE" >> (CACHEDTREE)
-	print "export RC_GOT_DEPTREE_INFO=\"yes\"" >> (CACHEDTREE)
+	# Do not export these, as we want them local
+	print "RC_GOT_DEPTREE_INFO=\"yes\"" >> (CACHEDTREE)
 	print "" >> (CACHEDTREE)
 
 	if (check_provide("logger"))
-		print "export LOGGER_SERVICE=\"" get_provide("logger") "\"" >> (CACHEDTREE)
+		print "LOGGER_SERVICE=\"" get_provide("logger") "\"" >> (CACHEDTREE)
 	else
-		print "export LOGGER_SERVICE=" >> (CACHEDTREE)
+		print "LOGGER_SERVICE=" >> (CACHEDTREE)
 		
 	close(CACHEDTREE)
 }

[-- Attachment #3: Type: text/plain, Size: 40 bytes --]

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 2577 bytes --]

On Mon, 2004-02-02 at 06:48, Brad House wrote:
> figures I forgot to attach it.
> -Brad
> 

1) the right bash version of [ "$foo" == "" ] is [ -z "$foo" ], or that
is my opinion at least.  The rest of baselayout uses this.

2) So now we have '/newroot', '/mnt/cdrom' and possibly '/mnt/livecd'
for older livecd scripts, or '/var/lib/livecd' for future.  Basically,
can we really get this sorted out _now_ ?  I am _not_ going to add
20 tests.

3) I still think shutting lvm down is the safer way for uses that uses
lvm at install.  Anybody else want to comment on this? (Note I am not
saying to start it ...).

4) I say this for the last time - If you commit to baselayout, you go
through _me_.  If anybody took long to respond to patches or testing
there of, it was not me.  You do not know if your changes could have
any more implications than you might see/think of.  I do not see you
doing this with portage.  I do not see me removing all the broken
'ppc64' keywords (and breaking in on what you consider your domain)?
Questions?

> > I have attached a new patch against the p3 rc-scripts.
> > This incorporate's azarah's fixes, as well as another patch
> > for halt.sh which resolves the no-reboot issue for zisofs
> > cds.  As usual, I have updated the baselayout-1.8.6.12-r4.ebuild
> > with my patches, so anyone building livecd's needs to mark
> > that stable for their arch for testing.
> >
> > -Brad
> >
> >> On Thu, 2004-01-29 at 19:19, Benjamin Judas wrote:
> >>> Am Mi, den 28.01.2004 schrieb Martin Schlemmer um 18:44:
> >>> > On Tue, 2004-01-27 at 19:13, Martin Schlemmer wrote:
> >>> >
> >>> > Brad, did you see/could you test this?  Benjamin?
> >>> >
> >>>
> >>> Hi Martin,
> >>>
> >>> I tried you patch today and it looks like bootup is running fine now -
> >>> no errors anymore in this part. All my hardware got detected and is
> >>> working (we might need some kernel-tweaking -- but that shouldn't
> >>> affect
> >>> baselayout in the startup-stage anymore). Rebooting and halting however
> >>> still don't work - I am also using zisofs. But it really looks good so
> >>> far :)
> >>
> >> Could you please put the new version of the iso up again?
> >>
> >>
> >> Thanks,
> >> --
> >>
> >> Martin Schlemmer
> >> Gentoo Linux Developer, Desktop/System Team Developer
> >> Cape Town, South Africa
> >>
> >>
> >>
> >
> >
> > --
> > gentoo-releng@gentoo.org mailing list
> >
> >
> >
-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

> 1) the right bash version of [ "$foo" == "" ] is [ -z "$foo" ], or that
> is my opinion at least.  The rest of baselayout uses this.

be my guest to change it.  It was just for clarification for myself
as I always mix up -z vs -n ... did that too many times.

>
> 2) So now we have '/newroot', '/mnt/cdrom' and possibly '/mnt/livecd'
> for older livecd scripts, or '/var/lib/livecd' for future.  Basically,
> can we really get this sorted out _now_ ?  I am _not_ going to add
> 20 tests.

/mnt/cdrom and /mnt/livecd will both be mounted at the SAME time.
One for the cdrom itself, the other for the loopback.  Don't worry about
/var/lib/livecd for now.

>
> 3) I still think shutting lvm down is the safer way for uses that uses
> lvm at install.  Anybody else want to comment on this? (Note I am not
> saying to start it ...).

Well, as long as it's silent on shutdown if it doesn't find something
to unmount/bind, I'm fine with that, otherwise we'll get a flurry of
bug reports.

>
> 4) I say this for the last time - If you commit to baselayout, you go
> through _me_.  If anybody took long to respond to patches or testing
> there of, it was not me.  You do not know if your changes could have
> any more implications than you might see/think of.  I do not see you
> doing this with portage.  I do not see me removing all the broken
> 'ppc64' keywords (and breaking in on what you consider your domain)?
> Questions?

dude, I took over -r4, and many people are using that for building
livecds right now.  I just updated what I already had. I didn't screw
up anything else.  Don't get your panties in a wad. I'm just doing
what needs to be done to get stuff done without the bureaucracy of
everything else. My ebuild is keyworded -*, and it was NOT committed
to the rc-scripts CVS.  Heck I even have RESTRICT="nomirror".  Can't
get much more harmless than that. Calm down dude, not touching a damn
thing you had done.  I tried to resign from doing any livecd work at
all, but that's just not feasible apparently b/c stuff has to get done
cut me some slack, I work a full-time job (more than fulltime actually)
and I just don't have time to waste.

-Brad


--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 1610 bytes --]

On Mon, 2004-02-02 at 21:54, Brad House wrote:

> >
> > 2) So now we have '/newroot', '/mnt/cdrom' and possibly '/mnt/livecd'
> > for older livecd scripts, or '/var/lib/livecd' for future.  Basically,
> > can we really get this sorted out _now_ ?  I am _not_ going to add
> > 20 tests.
> 
> /mnt/cdrom and /mnt/livecd will both be mounted at the SAME time.
> One for the cdrom itself, the other for the loopback.  Don't worry about
> /var/lib/livecd for now.

And /newroot?  Lets clarify:

1) /mnt/livecd will be your zisofs image/whater
2) /mnt/cdrom will be the cdrom itself ?

So basically, the looback stuff just need to worry about /mnt/livecd,
and both for the non-critical unmount stuff?

Also, is /newroot now totally not used anymore?  And why does /mnt/cdrom
need to be mounted - i thought / was in there?  I guess a quick
explanation on all the types used (zisofs, etc) and what they mount for
what will be great ...

> >
> > 3) I still think shutting lvm down is the safer way for uses that uses
> > lvm at install.  Anybody else want to comment on this? (Note I am not
> > saying to start it ...).
> 
> Well, as long as it's silent on shutdown if it doesn't find something
> to unmount/bind, I'm fine with that, otherwise we'll get a flurry of
> bug reports.

Hmm, not sure - that is why I asked for comments from others.  I guess
the truth will still hold - the conf files tested for should not be
there if the user did not setup it ...


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Brad House]

>
> And /newroot?  Lets clarify:

/newroot should never be visible as it's strictly in the
initrd, and pivot_root will pretty much clobber that, it's
something you don't need to handle.

> 1) /mnt/livecd will be your zisofs image/whater

It is the loopback mount location for normal or gcloop filesystems.
zisofs is not a loopback device, so it won't make use of that.

> 2) /mnt/cdrom will be the cdrom itself ?

yes.

>
> So basically, the looback stuff just need to worry about /mnt/livecd,
> and both for the non-critical unmount stuff?

not sure what you mean by this, but yes, you should never umount
/mnt/livecd.

>
> Also, is /newroot now totally not used anymore?  And why does /mnt/cdrom
> need to be mounted - i thought / was in there?  I guess a quick
> explanation on all the types used (zisofs, etc) and what they mount for
> what will be great ...

/mnt/cdrom is the cdrom
/ is a tmpfs filesystem
/mnt/livecd is where a loopback gets mounted
/mnt/cdrom/zisofs  is the zisofs location that symlinks get
                   made to to fill in the root, but it's not a
                   loop device, so /mnt/livecd is not used if the
                   cd is zisofs
/etc is a tmpfs filesystem
/newroot is only used in the initrd, and should not be visible once
         /sbin/init is called on bootup.

I think that about covers it.

>
> Hmm, not sure - that is why I asked for comments from others.  I guess
> the truth will still hold - the conf files tested for should not be
> there if the user did not setup it ...

ok, as long as that's true. ..


> Tough, we all have 12 hour or close working days.  As for calming down,
> I _am_ calm - you have not seen me on the war path yet =)  And I do not
> imagine myself that the -r4 changes was not the only.
>
> Also, 'resigning' from the livecd project because you overstep
> boundaries with no visible regard for the impact or feelings of others,
> is to be frank, silly.  Ask me - any change that usually cannot wait a
> few hours to be verified is usually a wrong one.  I did this more than
> once myself - overstepping boundaries for this 'critical' fix, and the
> result was worse than what the few extra hours of waiting could ever
> have done.
>
> We do not ask your soul - we ask some sort of regard.  Work with us,
> and you will see that most of us will work with you.

Seeing as I had already 'overstepped' my bounds as you put it and
committed a -r4 in the first place (it was never anything you had
committed), what was the harm at all in me updating my own ebuild?
Anyhow, we use CVS for collaboration, etc, this send me a diff and I'll
approve it thing is BS as long as we don't mess anything up, which I
have taken steps to make sure I do not.

Oh also, just as a little wood for your fire, I committed a patch to
-r2,-r3,-r4, and -r5 that fixes a bug where /dev was not being created
_again_. And I did view this as fairly critical, and it was simplistic.
We are a team, I don't believe ONE person should hold up bug fixes, etc
when things are easily fixed without disturbance.  I view myself as a
competant developer, if I were not, perhaps it would be a different
situation.  Again, I bring up the point, why do we use CVS if not to
collaborate ?

-Brad

--
gentoo-releng@gentoo.org mailing list

Re: [gentoo-releng] x86-livecd - bugs and suggestions

[Martin Schlemmer]

[-- Attachment #1: Type: text/plain, Size: 3668 bytes --]

On Mon, 2004-02-02 at 23:17, Brad House wrote:

> > Tough, we all have 12 hour or close working days.  As for calming down,
> > I _am_ calm - you have not seen me on the war path yet =)  And I do not
> > imagine myself that the -r4 changes was not the only.
> >
> > Also, 'resigning' from the livecd project because you overstep
> > boundaries with no visible regard for the impact or feelings of others,
> > is to be frank, silly.  Ask me - any change that usually cannot wait a
> > few hours to be verified is usually a wrong one.  I did this more than
> > once myself - overstepping boundaries for this 'critical' fix, and the
> > result was worse than what the few extra hours of waiting could ever
> > have done.
> >
> > We do not ask your soul - we ask some sort of regard.  Work with us,
> > and you will see that most of us will work with you.
> 
> Seeing as I had already 'overstepped' my bounds as you put it and
> committed a -r4 in the first place (it was never anything you had
> committed), what was the harm at all in me updating my own ebuild?

I did not talk this time about -r4, I talked about the other commit.

> Anyhow, we use CVS for collaboration, etc, this send me a diff and I'll
> approve it thing is BS as long as we don't mess anything up, which I
> have taken steps to make sure I do not.
> 

See below.

> Oh also, just as a little wood for your fire, I committed a patch to
> -r2,-r3,-r4, and -r5 that fixes a bug where /dev was not being created
> _again_. And I did view this as fairly critical, and it was simplistic.

If you want to get into it, then technically you 'fixed' it in the
wrong place.  The real culprit in this situation (and others that might
pop up besides /dev if $ROOT != '/'), is keepdir_mount().

> We are a team, I don't believe ONE person should hold up bug fixes, etc
> when things are easily fixed without disturbance.  I view myself as a
> competant developer, if I were not, perhaps it would be a different
> situation.  Again, I bring up the point, why do we use CVS if not to
> collaborate ?
> 

Ok, so please go and commit to portage and see where that gets you.  As
far as I can see, the only 'team' in your vocabulary, is the one that
gets you your way.  You obviously fail to see why it is that some core
packages have, lets call it 'supervising maintainers' - its to make sure
that all changes are regulated and will not break something due to
limited understanding from less active developers when it comes to that
package.

Go and have a look at the kernel for instance.  All major parts have
their respective maintainers who concur changes, and then send it off
to Linus.  Same thing here.

The biggest issue here - or reason why I get 'my panties in a wad' - is
that you do not even have the decency to let me know about a change.  I
never said I was the all mighty committer of baselayout - check the
ChangeLog .. others _do_ commit there ... _after_ they check with me.
Its not what you do, but _how_ you do it.

What you also seemingly seem to miss, is this is not about me getting my
undies tied up, but you repeatedly disregarding any type of system in
place, and doing as _you_ please.  Yes, _you_.  Great hey, its still all
about you - just get this ... its not in a good way? 

So please, go get an ego adjustment; un-wad _your_ panties; and play
some ball.  Just for the record - you did see that I did not CC releng?
I see I should apologise for underestimating your righteousness,
sorry :/


Thanks,

-- 

Martin Schlemmer
Gentoo Linux Developer, Desktop/System Team Developer
Cape Town, South Africa



[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]