Laptops :: How to install CPU Freq Driver



Looks like a problem with the .config in 2.4.26 - see http://lkml.org/lkml/2004/4/19/202

Although I'm guessing you may not need to recompile that module - it's good to have a fully successful build anyways.

Hmmm. This is a bit OT but is it anyway possible to load different kernel (from usbstick or hd) when using live cd?
^thehatsrule^ - thanks for the link (I didn't think Google would turn up anything so obscure).

This confirms I'm not going crazy and there is an issue that stops things when compiling /drivers/char/drm in 2.4.26 - the link offers two solutions:

1. Build for more than CONFIG_M386=y - menuconfig won't allow me to do this and if I manually set more than one processor then menuconfig will not open the .config file...

2. Apply the patch listed in the link - I tried this but there must be an error introduced by the author cutting & pasting a long line(s) of text into a narrow web page and the patch gives an error message about unexpected end of line (but does not give a line number). I guessed a couple of corrections but it's all gobbledegook to me...

I got the compile to run fully by disabling DRM in .config - just unchecking the DRM modules does not help - and cpufreq looked to compile properly. 'Trouble is I need DRM in the kernel image in order to use my monitor with XFree...

1.  What did you try to do for "build for more than ..."?

2.  To manually do it, edit that .h file: the '+' in the first column means to add that line, '-' means to delete that line.

Zucca: probably, depends on what you changed, etc.

1.  What did you try to do for "build for more than ..."? - CONFIG_M386=y & CONFIG_MPENTIUMIII=y

2.  To manually do it, edit that .h file: the '+' in the first column means to add that line, '-' means to delete that line. - Aaah, so that's how it works. Thanks for the tip, I did this in the drmP.h and it almost worked. For some reason the compile didn't like the reference to system.h added in the diff, after I removed this, the compile worked fine.

As a test, since I don't know if cpufreq works yet, I compiled bluetooth in and it works (after apt-get install bluez-utils) - therefore I assume the compile "bug" is fixed.

# insmod bluez
Using /lib/modules/2.4.26/kernel/net/bluetooth/bluez.o
# insmod hci_usb
Using /lib/modules/2.4.26/kernel/net/bluetooth/rfcomm/rfcomm.o
# insmod l2cap
Using /lib/modules/2.4.26/kernel/net/bluetooth/l2cap.o
# insmod rfcomm
Using /lib/modules/2.4.26/kernel/net/bluetooth/rfcomm/rfcomm.o
# hciconfig hci0 up
# hcitool scan
Scanning ...
       00:0A:D9:E8:4A:65       .P900 [my mobile phone]

Next Page...
original here.