* lilo and system maps?
@ 2003-12-04 17:53 Robert L. Harris
2003-12-04 18:09 ` Andre Tomt
2003-12-04 18:25 ` Maciej Zenczykowski
0 siblings, 2 replies; 6+ messages in thread
From: Robert L. Harris @ 2003-12-04 17:53 UTC (permalink / raw)
To: Linux-Kernel
[-- Attachment #1: Type: text/plain, Size: 1441 bytes --]
I'm messing around on one of my dev machines which has 4 possible
kernels installed. 2.4, 2.4-stable, 2.6, 2.6-stable (stable is the last
known good kernel). I currently have my System.map files laid out as:
/boot/System.map-2.6.0-test11-bk2
/boot/System.map-2.6.0-test10-bk4
etc.
Would it be possible, necessary, sane to have a lilo.conf similar to
this:
image=/boot/vmlinuz-2.4
label=current
map=System.map-2.4
read-only
image=/boot/vmlinuz-2.4.stable
label=stable
map=System.map-2.4.stable
read-only
optional
image=/boot/vmlinuz-2.6
label=2.6
map=System.map-2.6
read-only
optional
image=/boot/vmlinuz-2.6.stable
label=2.6-stable
map=System.map-2.6.stable
read-only
optional
This way when I install a new kernel I can copy the System.map to
/boot/System.map-2.6 instead of keeping up with all the version numbers?
lilo doesn't seem to like the map= arguements. Does the kernel need the
System.map in a single place, can it figure out where it's at for a
multiple config?
Robert
:wq!
---------------------------------------------------------------------------
Robert L. Harris | GPG Key ID: E344DA3B
@ x-hkp://pgp.mit.edu
DISCLAIMER:
These are MY OPINIONS ALONE. I speak for no-one else.
Life is not a destination, it's a journey.
Microsoft produces 15 car pileups on the highway.
Don't stop traffic to stand and gawk at the tragedy.
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: lilo and system maps?
2003-12-04 17:53 lilo and system maps? Robert L. Harris
@ 2003-12-04 18:09 ` Andre Tomt
2003-12-04 18:15 ` Robert L. Harris
2003-12-04 18:25 ` Maciej Zenczykowski
1 sibling, 1 reply; 6+ messages in thread
From: Andre Tomt @ 2003-12-04 18:09 UTC (permalink / raw)
To: Robert L. Harris; +Cc: Linux-Kernel
On Thu, 2003-12-04 at 18:53, Robert L. Harris wrote:
> I'm messing around on one of my dev machines which has 4 possible
> kernels installed. 2.4, 2.4-stable, 2.6, 2.6-stable (stable is the last
> known good kernel). I currently have my System.map files laid out as:
>
> /boot/System.map-2.6.0-test11-bk2
> /boot/System.map-2.6.0-test10-bk4
> etc.
^^^
> This way when I install a new kernel I can copy the System.map to
> /boot/System.map-2.6 instead of keeping up with all the version numbers?
> lilo doesn't seem to like the map= arguements. Does the kernel need the
> System.map in a single place, can it figure out where it's at for a
> multiple config?
Just stick with the System.map-$(uname -r) variant and it will just work
automaticly. map= in lilo is not for System.map's.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: lilo and system maps?
2003-12-04 18:09 ` Andre Tomt
@ 2003-12-04 18:15 ` Robert L. Harris
2003-12-04 18:41 ` Maciej Zenczykowski
0 siblings, 1 reply; 6+ messages in thread
From: Robert L. Harris @ 2003-12-04 18:15 UTC (permalink / raw)
To: Andre Tomt; +Cc: Linux-Kernel
[-- Attachment #1: Type: text/plain, Size: 2089 bytes --]
Thus spake Andre Tomt (lkml@tomt.net):
> On Thu, 2003-12-04 at 18:53, Robert L. Harris wrote:
> > I'm messing around on one of my dev machines which has 4 possible
> > kernels installed. 2.4, 2.4-stable, 2.6, 2.6-stable (stable is the last
> > known good kernel). I currently have my System.map files laid out as:
> >
> > /boot/System.map-2.6.0-test11-bk2
> > /boot/System.map-2.6.0-test10-bk4
> > etc.
> ^^^
>
> > This way when I install a new kernel I can copy the System.map to
> > /boot/System.map-2.6 instead of keeping up with all the version numbers?
> > lilo doesn't seem to like the map= arguements. Does the kernel need the
> > System.map in a single place, can it figure out where it's at for a
> > multiple config?
>
> Just stick with the System.map-$(uname -r) variant and it will just work
> automaticly. map= in lilo is not for System.map's.
>
Upon reading the man page on lilo.conf again I realized that was the
wrong map, but was hoping there was some notation that'd allow for maps
other than "/boot/System.map" for the multiple kernels.
Ok, I've tried rdev and file, is there a "cleaner" way of getting a
kernel version other than:
{0}:/usr/share/doc/lire>strings /boot/vmlinuz-2.6 | grep -i 2.[46] | head
2.6.0-test11-bk2 (root@wally) #3 SMP Thu Dec 4 12:41:42 EST 2003
M2#6gbQ+
{2 6B
.2'4]
AG2{6
b2)42
2y6&LD
2U6;
214<|7
2"6:
Yes it's functional, I can live with it, just wondering if theres a
"better" way for when I'm updating servers at work and have no idea what
kernels are in /boot on the machine (some have been up longer than the
year+ I've been here).
:wq!
---------------------------------------------------------------------------
Robert L. Harris | GPG Key ID: E344DA3B
@ x-hkp://pgp.mit.edu
DISCLAIMER:
These are MY OPINIONS ALONE. I speak for no-one else.
Life is not a destination, it's a journey.
Microsoft produces 15 car pileups on the highway.
Don't stop traffic to stand and gawk at the tragedy.
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: lilo and system maps?
2003-12-04 17:53 lilo and system maps? Robert L. Harris
2003-12-04 18:09 ` Andre Tomt
@ 2003-12-04 18:25 ` Maciej Zenczykowski
1 sibling, 0 replies; 6+ messages in thread
From: Maciej Zenczykowski @ 2003-12-04 18:25 UTC (permalink / raw)
To: Robert L. Harris; +Cc: Linux-Kernel
If I'm not mistaken the map file specified in map= is -not- the System.map
file of the kernel, it's a lilo specific map file and has nothing to do
with System.map. Indeed it'll probably overwrite the System.map file.
you don't want to specify map= anything other than /boot/map unless your
/boot directory is somewhere else. As for system.map I have my boot
scripts deleting and symbolic linking /boot/System.map to
/boot/System.map-`uname -r` (these files actually exist) right after boot.
(not sure maybe this isalready done by the standard distribution of RH, I
do remember fooling around with it though...)
Cheers,
MaZe.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: lilo and system maps?
2003-12-04 18:15 ` Robert L. Harris
@ 2003-12-04 18:41 ` Maciej Zenczykowski
2003-12-04 18:45 ` Robert L. Harris
0 siblings, 1 reply; 6+ messages in thread
From: Maciej Zenczykowski @ 2003-12-04 18:41 UTC (permalink / raw)
To: Robert L. Harris; +Cc: Andre Tomt, Linux-Kernel
>
> {0}:/usr/share/doc/lire>strings /boot/vmlinuz-2.6 | grep -i 2.[46] | head
> 2.6.0-test11-bk2 (root@wally) #3 SMP Thu Dec 4 12:41:42 EST 2003
> M2#6gbQ+
> {2 6B
Of course the correct solution is to have the kernel version in the file
name... and have linux-current or whatever as a symlink. Besides truth
be told the kernel version is far too little to identify a kernel anyway,
there's also compilation options - they can change a lot - and all the
patches which were/are applied to it. I keep System.map-`uname -r`,
vmlinuz-`uname -r`, .config-`uname -r`, descr-`uname -r` in my /boot dirs
- the first three come from the kernel and the last is a text file
containing notes about what patches were applied (I keep an up to date
descr file in each kernel source dir).
Cheers,
MaZe.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: lilo and system maps?
2003-12-04 18:41 ` Maciej Zenczykowski
@ 2003-12-04 18:45 ` Robert L. Harris
0 siblings, 0 replies; 6+ messages in thread
From: Robert L. Harris @ 2003-12-04 18:45 UTC (permalink / raw)
To: Maciej Zenczykowski; +Cc: Andre Tomt, Linux-Kernel
[-- Attachment #1: Type: text/plain, Size: 1754 bytes --]
Thus spake Maciej Zenczykowski (maze@cela.pl):
> >
> > {0}:/usr/share/doc/lire>strings /boot/vmlinuz-2.6 | grep -i 2.[46] | head
> > 2.6.0-test11-bk2 (root@wally) #3 SMP Thu Dec 4 12:41:42 EST 2003
> > M2#6gbQ+
> > {2 6B
>
> Of course the correct solution is to have the kernel version in the file
> name... and have linux-current or whatever as a symlink. Besides truth
> be told the kernel version is far too little to identify a kernel anyway,
> there's also compilation options - they can change a lot - and all the
> patches which were/are applied to it. I keep System.map-`uname -r`,
> vmlinuz-`uname -r`, .config-`uname -r`, descr-`uname -r` in my /boot dirs
> - the first three come from the kernel and the last is a text file
> containing notes about what patches were applied (I keep an up to date
> descr file in each kernel source dir).
>
> Cheers,
> MaZe.
>
I'm maintaing machines that pre-exist me and many don't have
vmlinuz-`uname -r` naming and can't quite reboot the machine on
different kernels just to find out what it is.
Did the symlink think but it got ugly after a while when I had 6 kernel,
6 System.map files and a bunch of symlinks. Once I fully drop the 2.4
on this set of machines I can move back but it was just ugly for a
while.
:wq!
---------------------------------------------------------------------------
Robert L. Harris | GPG Key ID: E344DA3B
@ x-hkp://pgp.mit.edu
DISCLAIMER:
These are MY OPINIONS ALONE. I speak for no-one else.
Life is not a destination, it's a journey.
Microsoft produces 15 car pileups on the highway.
Don't stop traffic to stand and gawk at the tragedy.
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2003-12-04 18:46 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-12-04 17:53 lilo and system maps? Robert L. Harris
2003-12-04 18:09 ` Andre Tomt
2003-12-04 18:15 ` Robert L. Harris
2003-12-04 18:41 ` Maciej Zenczykowski
2003-12-04 18:45 ` Robert L. Harris
2003-12-04 18:25 ` Maciej Zenczykowski
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).