The Official HWF Kernel Repo

Hey, the working Kernel is my 'old' one - 2.6.22-3-686
I have already un-installed the last 'attempt'...

Just need to get rid of the entry in Lilo. :(
 
All kernels have been updated to a newer build of the same version. I've added a few more optional optimizations to most kernels in order to help you make the most of your newer hardware. Also, lowered the minimum requirements of the generic kernel so it should even run on a Pentium MMX. Please report any issues here!

Best regards,
-AT
 
Sorry AT....

I am getting the same message at boot when installing the new Generic kernel....
I'll just stick to the kernel I have....

Thanks for taking the time to try and sort this out...

:cool:
 
sabashuali said:
I am getting the same message at boot when installing the new Generic kernel....
Click to expand...
Hey Saba, the Generic kernel even boots on a P1, so I have the feeling you may be booting to the P4 Kernel by mistake. LILO is somewhat poorly implemented in Debian, and doesn't clearly label which kernel you're booting to.

This is kinda out of scope for helping you load the kernel, but unless you're booting to a RAID or LVM, I'd recommend switching your bootloader to GRUB. It's much more straight-forward. I have to use LILO on my desktop since I'm booting to a RAID-1, and it can be a pain sometimes, in exactly the ways you've experienced.
 
Thanks AT.

I have decided to start again so there will be a re-install tonight!!!! I have messed up the bootloader good an proper.... :doh:

The debian installer chooses lilo by default (as I am installing over XFS) so I will just have to backup at that stage and choose grub instead...

Can I switch now? While my installation is still going? mmmmm :dry:
Without having to re-install? :confused:
 
sabashuali said:
The debian installer chooses lilo by default (as I am installing over XFS) so I will just have to backup at that stage and choose grub instead...
Click to expand...
Yeah, GRUB is incompatible with XFS as a boot partition. When you plan to use XFS for your root filesystem, it's wise to put the bootloader on a separate partition. For instance, create a 100mb /boot partition and format it as ext2 or ext3. Only kernels go in /boot, so it doesn't have to be very big.

sabashuali said:
Can I switch now? While my installation is still going? mmmmm :dry: Without having to re-install? :confused:
Click to expand...
debian switch from lilo to grub - Google Search

...you can switch, but GRUB is still incompatible with XFS as the boot partition. If you made a 100mb slice, formatted as ext2/3 and put all your boot stuff on it, you could probably make the swap. It won't be super-straightforward, but it's possible.
 
Hi AT

Thanks again - I did not realize there was such a solution as described nor did I realize that GRUB was incompatible with XFS.... (or the other way around)

AAAAnyway, the problem with the Debian installer is that when it come to installing grub, it simply freezes. Now I understand why. :doh:

But! I ended up installing with Lilo and then, with the help of one or two tutorials, swapped to Grub. At the moment Grub is booting the system, reading the kernel from the xfs partition. The test will be when I add a new kernel.... will grub write the new entry properly? It remains to be seen.....

Unfortunately, the link to the repo is down :x:

Will keep you informed. :cool:
 
sabashuali said:
But! I ended up installing with Lilo and then, with the help of one or two tutorials, swapped to Grub. At the moment Grub is booting the system, reading the kernel from the xfs partition. The test will be when I add a new kernel.... will grub write the new entry properly? It remains to be seen.....

Unfortunately, the link to the repo is down :x:

Will keep you informed. :cool:
Click to expand...
Glad to hear you've made progress. Man, you've come so far in learning Linux Saba, you've done me proud. :D *tear*

As for the repo, Sniper IM'ed me about it a short while ago. Apparently some recent changes to the website overwrote our repo config in Apache. I re-added it, so you should be able to get to it without problems. Sorry for the inconvenience.

All the best,
-AT
 
sabashuali said:
The test will be when I add a new kernel.... will grub write the new entry properly? It remains to be seen.....
Click to expand...

As I suspected trying to install a new Kernel envoked lilo once again and installed a fresh (wrong... :dry:) bootloader...... After playing around for a while I thought to myself.... hang on - what's the point? Am I going to have to go through this everytime I update a Kernel? No! So I started again.....

Thsi time I going down the route as described in AT's reply (#27).
I now have three partitions - /boot (ext3); / (xfs) and /home (xfs)
This should make it easy to install grub and maintain a healthy connection between the bootloader and the Kernel.

Will it work? Am I doing it right?
We will find out in the next episode of 'Saba's Linux bash'....

See you then...... :chk:

Thanks, AT for the kind words (It is all down to the teacher... :P ) and for all the help!

Update:
Well everything worked well.... apart from the Kernel. Same message as before and does not boot. This time, thanks to Grub, I know I am attempting to boot the generic kernel and nothing else. Let me make sure I am doing this correctly - I download both headers and image of the generic kernel. Then I install the headers first and then the image. Is this correct? Am I missing a stage? Is there anything I am not doing?
 
New builds today based on vanilla 2.6.24.2 from kernel.org. The naming convention has been modified somewhat to include 'hwf', so it's easier to tell our kernels from the stock ones. This build also includes a fix for a rather serious user privilege escalation vulnerability, so it's recommended that you upgrade as soon as possible.
 
in soviet Russia vmsplices you!

[OT]
i am assuming thats the exploit your talking about, i don't have time to check.

Its only a local exploit, so unless your running ssh (and other people are allowed to login) i doubt you will have had any trouble with it (yet).
[/OT]
 
Impotence said:
in soviet Russia vmsplices you!
Click to expand...
Yeah, that's the one. [ot]It is only a local exploit, meaning that you need a shell on the host to be able to exploit it. Even so, that means that another type of exploit could lead to local privilege escalation. For example, say you're running a game server from your system, and there's a vulnerability in the game software. If it's exploited leading to a local shell, they can go one step further and use the kernel flaw to get root. So, it's wise to upgrade if you're running an earlier kernel.[/ot]
 
The Saba (saga) continues

Hey AT

Thanks for the updated Kernel and all the efforts.
I am afraid I did not get much further with the new kernel either....
Surely I am doing something wrong:
  1. Install the headers
  2. Install the image
  3. Update the environment file
  4. Reboot

This time on boot I do not get an error message, the boot screen simply says:
Decompressing the Kernel..... please wait......
And that's that.

I am sure I am missing something..... :doh:

By the way, just to confirm - I am selecting the HWF kernel in Grub when I boot. :cool:
 
i don't know if i'm right, but, when i threw the kernel in my system i actually installed the image first and then the headers - it entirely possible that i'm completely wrong - but at least i'm trying
 
It seems that the latest generic kernel build for Lenny was indeed broken sabashuali; it wasn't something you didn't do correctly. I'm currently in the process of recompiling that one, and the Gutsy one as well to be safe. I'm also adding a new build to the group just for you saba: the Pentium-M, which does have a few advantages over the P3 such as SSE2 support. I'll add that one also, as soon as both are finished compiling.
 
OK, here's what's going on right now. The 2.6.24 tree has changed so much API stuff that it's really caused a lot of problems for those of us who unfortunately rely on binary blobs like Nvidia or ATI drivers, etc. Even with those being worked around, 2.6.24 just isn't as been stable in my Ubuntu build environment as the 2.6.23 family was. As a result, I'm going to be rolling back to the 2.6.23.16+ kernel family, which does provide a fix for the recent local account escalation bug. I'll also be adding the Pentium M kernel, as promised. :) News will be posted here as soon as I have the newer kernels tested and uploaded to the repo!

Best regards,
-AT
 
[ot]How long can it be until Nvidia release there source code? ATI/AMD have done it... I don't see what they gain by keeping it closed source, it just means they have to fix it themselves![/ot]
 
[ot]
Impotence said:
[ot]How long can it be until Nvidia release there source code? ATI/AMD have done it... I don't see what they gain by keeping it closed source, it just means they have to fix it themselves![/ot]
Click to expand...
ATI has not done this yet, they've only raised a fuss about it.[/ot]
 
Anti-Trend said:
...I'm going to be rolling back to the 2.6.23.16+ kernel family, which does provide a fix for the recent local account escalation bug. I'll also be adding the Pentium M kernel, as promised. :) News will be posted here as soon as I have the newer kernels tested and uploaded to the repo...
Click to expand...
The Lenny kernels are built and uploaded, including the shiny new Pentium M kernel. The new Gutsy kernels are still building right now, but as soon as they're uploaded I'll post again.
 
You must log in or register to reply here.
Back
Top