[NTLUG:Discuss] (weird) framebuffer disable breaks kernel?

[Darin W. Smith]

I've encountered something weird, that I have no explanation for.

After seeing Chris' post the other day, I decided to try building a 
framebuffer-less kernel to see if that helped with my memory problems with 
the Nvidia driver.

So I took my current kernel config and turned off framebuffer support, then 
made it and tried it.  It hung immediately after finding my hard drives.  I 
figured it must have been some issue related to not making a clean compile. 
 So I made clean and made mrproper (just to be safe), copied my .config 
back in and did a make menuconfig to check the settings, rebuilt 
EVERYTHING, including modules, installed them, tried it out--same thing.

This was on a 2.4.18 kernel.

So then I tried a 2.4.20 kernel.  It got further, but hung after 
"Installing keymap: us".  What gives?  Why would removing framebuffer 
support make my kernel not boot all the way?  Back in the 2.2 kernel days, 
even after framebuffer support was introduced, I never used it.  I was 
under the impression that it was still supposed to be safer to go without 
it (though it would be nice if it worked correctly--though I don't yet know 
if it is my problem, as there are plenty of gee-whiz things that could be 
done with it).

Has anybody seen removal of framebuffer support break their kernel?

On a slightly unrelated note, does anyone have any opinions--positive or 
negative--about the preemptive kernel patch from Robert Love?  One guy at 
work said he has noticed some strange blocking of tasks after using it, and 
was going to turn it off.


-- 
D!
Darin W. Smith
AIM: JediGrover