[redhat8 on an ASUS L8400C laptop] I'm playing, I'm learning, I'm having fun...but I need some help...

grepcomputers

Senior member
Apr 9, 2002
209
0
0
So, I continue on my 2.4.20 kernel compiling quest. I'm a hell of a lot further on than I used to be, much thanks to some very nice people.

But I am still having some problems...firstly, some data:
http://smithec.resnet.bucknell.edu/linux_kernel/ - you can get my lsmod output and config file for the 2.4.20 kernel I've compiled...

I copied my distro's config file - /boot/config-2.4.18-14 - into the kernel source directory - /usr/src/linux-2.4.20/.config - and used it as a base set of options, it made things a lot easier. But I still have a few problems.

On reboots - ie, I'm in linux and I type reboot at the console, the system hangs right after reaching this stage:
ACPI Namespace successfully loaded at root c0321ddc
evxfevnt-0074 [04] Acpi_enable : Transition to ACPI mode successful
evgpe-0263: *** Info: GPE Block0 defined as GPE0 to GPE15

Now, this isn't an error message, it appears to be part of the normal boot sequence. it's just on reboots, once this stage is reached, the boot hangs. Starting up from a system halt (cold startup) doesn't reproduce this problem. Merely rebooting from within linux.

My suspend key no longer works. Neither does the suspend command. Not only that, it doesn't work in *either* kernel. I used to be able to suspend when booted into the 2.4.18 kernel Redhat8 installed with, but now I cannot suspend at all. Typing suspend in the cli (inside of xwindows) now just sits there, as if waiting for input or something. And control-c does not exit out of it, nor can I close the console window. I've fooled around in the BIOS, but nothing has made it work. I *know* it worked before.
I feel kind of stupid, because I don't remember quite when it stopped working, or I'd have a better idea of how to fix it.
Aiieee! I was just booted in the 2.4.18 kernel and the suspend key worked. I rebooted into 2.4.20 to see if it worked, and it didn't, rebooted back into 2.4.18...and it didn't work again!

And one last interesting thing I noticed is that, even with the default settings from my distro, I still get errors when running depmod -ae 2.4.20. They are along the lines of this:
***Unresolved symbols in /lib/modules/2.4.20/kernel/net/wanrouter/wanrouter.o, but are not specific to wanrouter, or even to net modules. There were two or so compiles where I didn't get this, but I cannot seem to repeat it. Even if I completely delete /lib/modules/2.4.20 and allow the install to recreate and repopulate it, I still get these errors.

thanks!
...grep
 

cleverhandle

Diamond Member
Dec 17, 2001
3,566
3
81
Are you compiling 2.4.20 from vanilla sources or RH sources? RH adds lots of patches to their kernel and at least some of them are ACPI related, which might explain why 2.4.18 works somewhat better. As for why 18 would work sometimes and not others, I have no idea, but given the state of Linux ACPI and the variation in laptops' implementations, it doesn't really surprise me. Perhaps if you stick with 18 for more than one reboot, it will keep working?

The unresolved symbols could be caused by a couple things. You already mentioned cleaning out the install directory. Also, especially with RH-patched sources, you'll want to do a "make mrproper" before "make dep," etc. If you can't figure it out, just delete or unconfigure the modules you don't need.
 

grepcomputers

Senior member
Apr 9, 2002
209
0
0
I got the source from kernel.org. I got the acpi patch from acpi.sourceforge.net.

Actually, I noticed that when I ran make xconfig within the 2.4.18 kernel source directory, there was an option to enable intel speedstep technology...think that'll do what it looks like?

That, and the fact that I like the ACPI battery monitor better are two of the main reasons for the new kernel. See, using the APM battery monitor wmapmload (or any dockapp battery monitor) causes slight stuttering in video playback (mplayer and xine). I get the same thing with the ACPI based dockapp battery monitors, but it is barely noticeable. It is really noticeable with the APM based ones...

cheers...
...grep
 

grepcomputers

Senior member
Apr 9, 2002
209
0
0
If I do a cold boot into 2.4.18, the suspend button works. If I reboot into 2.4.20 it stops working, and rebooting back into 2.4.18 does not fix it. If I cold boot into 2.4.18 and then reboot into 2.4.18 it works. Even in a cold boot into 2.4.20, it does not work. There is something about ACPI and 2.4.20 that is globally affecting the BIOS/suspend button.

I'm beginning to think I should just stick with 2.4.18 until there is a stable kernel released with built in, working ACPI support.

And I have flashed my laptop BIOS to the latest revision...

cheers...
...grep
 

cleverhandle

Diamond Member
Dec 17, 2001
3,566
3
81
Originally posted by: McGrep
If I cold boot into 2.4.18 and then reboot into 2.4.18 it works.
Interesting - I suspected that might happen, but I have no idea why. Any ACPI experts out there? Does an ACPI system maintain some memory space for its settings that are not overwritten by a reset?
I'm beginning to think I should just stick with 2.4.18 until there is a stable kernel released with built in, working ACPI support.
Yup. Was there a specific reason you were upgrading to 20 in the first place? Unless there's a feature you need, I wouldn't recommend changing a working kernel on a laptop. Just too many issues with non-standard hardware to justify any small increase in performance.