based (loosely) on my 2.4 kernel config file, i've built both a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both of them.
they both freeze right around the "Freeing unused kernel memory" message, sometimes just two or three lines before that, but always pretty much in the same place.
i'm pretty sure this isn't the ubiquitous missing-VT-config problem since that normally stops printing output long before then, and i've verified that i have the vt options configured in.
i'm just digging into that, but if anyone has hints, that would be nice. (obviously, i built those kernels with gcc 3.3).
slowly, but surely.
rday
On Wed, Jul 30, 2003 at 02:57:45PM -0400, Robert P. J. Day wrote:
based (loosely) on my 2.4 kernel config file, i've built both a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both of them.
Try http://people.redhat.com/arjanv/2.5/
Use precompiled or grab .config from kernel-source package and try to modify this one (at lest to do not forget something).
On Wed, 30 Jul 2003, Milan Kerslager wrote:
On Wed, Jul 30, 2003 at 02:57:45PM -0400, Robert P. J. Day wrote:
based (loosely) on my 2.4 kernel config file, i've built both a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both of them.
Try http://people.redhat.com/arjanv/2.5/
Use precompiled or grab .config from kernel-source package and try to modify this one (at lest to do not forget something).
that's my next attempt. i was just wondering whether anyone had a wild guess as to what would cause the kernel boot to just seize after "Freeing unused kernel memory."
back to the whiteboard.
rday
** Reply to message from "Robert P. J. Day" rpjday@mindspring.com on Wed, 30 Jul 2003 20:54:15 -0400 (EDT)
On Wed, 30 Jul 2003, Milan Kerslager wrote:
On Wed, Jul 30, 2003 at 02:57:45PM -0400, Robert P. J. Day wrote:
based (loosely) on my 2.4 kernel config file, i've built both a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both of them.
Try http://people.redhat.com/arjanv/2.5/
Use precompiled or grab .config from kernel-source package and try to modify this one (at lest to do not forget something).
that's my next attempt. i was just wondering whether anyone had a wild guess as to what would cause the kernel boot to just seize after "Freeing unused kernel memory."
Did it to me when I had a couple of vfat Win98 partitions defined in fstab. As soon as I commented them out, the boot could proceed.
jb
On Wed, 30 Jul 2003, Milan Kerslager wrote:
On Wed, Jul 30, 2003 at 02:57:45PM -0400, Robert P. J. Day wrote:
based (loosely) on my 2.4 kernel config file, i've built both a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both of them.
Try http://people.redhat.com/arjanv/2.5/
Use precompiled or grab .config from kernel-source package and try to modify this one (at lest to do not forget something).
depressingly, that worked. which means i have no idea what was the problem with building from the original source. but it's late, and i'm not going to fret at this point.
so far, things are looking good. current config on the inspiron includes:
- severn - 2.6.0 test2 kernel - latest (4496) nvidia driver (patched for 2.6.0) - linksys wireless pcmcia
what to do next? ACPI? LVM2? .... lord, i hope i remember to write down how i did all this.
rday
On Thu, Jul 31, 2003 at 12:48:26AM -0400, Robert P. J. Day wrote:
On Wed, 30 Jul 2003, Milan Kerslager wrote:
On Wed, Jul 30, 2003 at 02:57:45PM -0400, Robert P. J. Day wrote:
based (loosely) on my 2.4 kernel config file, i've built both a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both of them.
Try http://people.redhat.com/arjanv/2.5/
Use precompiled or grab .config from kernel-source package and try to modify this one (at lest to do not forget something).
depressingly, that worked. which means i have no idea what was the problem with building from the original source. but it's late, and i'm not going to fret at this point.
This message is near the point where the kernel is trying to run a program from initrd image or from your root partition (mounted under /). So there is where the possible investigation could start.
You may try to make a diff between the two .config files to see what is different.
--- Milan Kerslager milan.kerslager@pslib.cz wrote: > On Thu, Jul 31, 2003 at 12:48:26AM -0400, Robert P. J. Day wrote:
On Wed, 30 Jul 2003, Milan Kerslager wrote:
On Wed, Jul 30, 2003 at 02:57:45PM -0400, Robert P. J. Day
wrote:
based (loosely) on my 2.4 kernel config file, i've built
both
a 2.6.0-test1 and 2.6.0-test2 kernel, and tried to boot both
of
them.
Try http://people.redhat.com/arjanv/2.5/
Use precompiled or grab .config from kernel-source package and
try to
modify this one (at lest to do not forget something).
depressingly, that worked. which means i have no idea what was the problem with building from the original source. but it's
late,
Just to check which one of these worked.
I know just before I upgraded to severn I had major aggro getting stock 2.5/2.6 to boot, but when I built from arjanv's source everything worked. Bizarrely after I installed this kernel, stock kernel source compiled and booted.
and i'm not going to fret at this point.
This message is near the point where the kernel is trying to run a program from initrd image or from your root partition (mounted under /). So there is where the possible investigation could start.
You may try to make a diff between the two .config files to see what is different.
-- Milan Kerslager E-mail: milan.kerslager@pslib.cz WWW: http://www.pslib.cz/~kerslage/
-- Rhl-beta-list mailing list Rhl-beta-list@redhat.com http://www.redhat.com/mailman/listinfo/rhl-beta-list
On Thu, 31 Jul 2003, Mike Martin wrote:
Just to check which one of these worked.
I know just before I upgraded to severn I had major aggro getting stock 2.5/2.6 to boot, but when I built from arjanv's source everything worked. Bizarrely after I installed this kernel, stock kernel source compiled and booted.
i'm going down that road right now. if i tried building from stock source, neither a 2.6.0-test1 nor a 2.6.0-test2 kernel would boot -- they would both freeze at the "Freeing unused kernel memory" line.
downloaded arjan's 2.6.0-test2 kernel source RPM, and built it with the i386 config file that came with it (which builds eight bazillion modules, but better safe than sorry). that booted, but i did have to go back and build in some things that had been modules (USB, mouse, maybe a couple other things).
right now, i'm in the process of still building with arjan's source, but discarding as many modules as i can that i *know* i don't need.
if that still boots, i'll use that working config file on a stock kernel source tree, and see what happens.
at the very least, i have a working 2.6.0 build that boots, and i'm grateful for small mercies.
rday
Hello Robert,
if i tried building from stock source, neither a 2.6.0-test1 nor a 2.6.0-test2 kernel would boot -- they would both freeze at the "Freeing unused kernel memory" line.
Did you try the suggestion by Jack Bowling to disable vfat mounts in /etc/fstab? Or is that irrelevant to your setup?
Bye, Leonard.
-- How clean is a war when you shoot around nukelar waste? Stop the use of depleted uranium ammo! End all weapons of mass destruction.
On Thu, 31 Jul 2003, Leonard den Ottolander wrote:
Hello Robert,
if i tried building from stock source, neither a 2.6.0-test1 nor a 2.6.0-test2 kernel would boot -- they would both freeze at the "Freeing unused kernel memory" line.
Did you try the suggestion by Jack Bowling to disable vfat mounts in /etc/fstab? Or is that irrelevant to your setup?
ironically, i had a vfat mount in /etc/fstab, but commenting that out didn't make any difference.
i'm curious *how* it would have made a difference, since from what i can see, any mounting of vfat filesystems doesn't occur until later on in the boot process.
very strange.
at this point, i'm trying to clarify how to move over from /etc/modules.conf to /etc/modprobe.conf cleanly, since that appears to be the next obvious step with a 2.6.0 kernel.
rday
p.s. got samba-3.0.0 beta up and verified a simple smbmount. so far, so good.
Hello Robert,
i'm curious *how* it would have made a difference, since from what i can see, any mounting of vfat filesystems doesn't occur until later on in the boot process.
Hm. Right. Dunno. It did make a difference for Jack though...
Bye, Leonard.
-- How clean is a war when you shoot around nukelar waste? Stop the use of depleted uranium ammo! End all weapons of mass destruction.
** Reply to message from "Robert P. J. Day" rpjday@mindspring.com on Thu, 31 Jul 2003 04:46:39 -0400 (EDT)
On Thu, 31 Jul 2003, Mike Martin wrote:
Just to check which one of these worked.
I know just before I upgraded to severn I had major aggro getting stock 2.5/2.6 to boot, but when I built from arjanv's source everything worked. Bizarrely after I installed this kernel, stock kernel source compiled and booted.
i'm going down that road right now. if i tried building from stock source, neither a 2.6.0-test1 nor a 2.6.0-test2 kernel would boot -- they would both freeze at the "Freeing unused kernel memory" line.
downloaded arjan's 2.6.0-test2 kernel source RPM, and built it with the i386 config file that came with it (which builds eight bazillion modules, but better safe than sorry). that booted, but i did have to go back and build in some things that had been modules (USB, mouse, maybe a couple other things).
right now, i'm in the process of still building with arjan's source, but discarding as many modules as i can that i *know* i don't need.
if that still boots, i'll use that working config file on a stock kernel source tree, and see what happens.
at the very least, i have a working 2.6.0 build that boots, and i'm grateful for small mercies.
RH uses different names for the usb drivers from vanilla 2.5/6. One thing to watch for.
jb
** Reply to message from "Robert P. J. Day" rpjday@mindspring.com on Thu, 31 Jul 2003 08:05:25 -0400 (EDT)
On Thu, 31 Jul 2003, Leonard den Ottolander wrote:
Hello Robert,
if i tried building from stock source, neither a 2.6.0-test1 nor a 2.6.0-test2 kernel would boot -- they would both freeze at the "Freeing unused kernel memory" line.
Did you try the suggestion by Jack Bowling to disable vfat mounts in /etc/fstab? Or is that irrelevant to your setup?
ironically, i had a vfat mount in /etc/fstab, but commenting that out didn't make any difference.
i'm curious *how* it would have made a difference, since from what i can see, any mounting of vfat filesystems doesn't occur until later on in the boot process.
very strange.
at this point, i'm trying to clarify how to move over from /etc/modules.conf to /etc/modprobe.conf cleanly, since that appears to be the next obvious step with a 2.6.0 kernel.
I believe Bill Nottingham has already been over this on the list: if you want to play with the test kernels then you are going to have to massage the modprobe.conf yourself since there are currently no tools to do that for you. As you discovered, generate-modprobe only runs once when you install the latest modutils rpm to do the conversion. I suppose you could --force loading modutils after every tweak but that seems a bit heavyhanded.
jb
On Thu, 31 Jul 2003, Jack Bowling wrote:
RH uses different names for the usb drivers from vanilla 2.5/6. One thing to watch for.
well, i started with arjan's 2.6.0-test2, built and booted that, removed a load of modules, that still booted. finally took that .config file, and applied it to a clean 2.6.0-test2 source tree from ftp.kernel.org. that built and booted just fine.
go figure. there's clearly some magic here i don't understand.
rday
-
Jack Bowling -
Leonard den Ottolander -
Mike Martin -
Milan Kerslager -
Robert P. J. Day