You are not logged in.
Help!
Has anyone figured-out how to compile 2.4.22 with gcc 3.3.1?
I've successfully applied Con Kolivas -ck2 patches+fixes from http://kernel.kolivas.org but I have multiple 'unresolved symbols' after re-boot on new kernel.
If not, what would you recommend to do to get it compiled properly?
Thanx.
-- Linux! Isn't it time?
Offline
i assume you get these unresolved symbol errors after running depmod -a?
AKA uknowme
I am not your friend
Offline
Exactly.
The 'makepkg' appears to complete without error. I then upgrade to new kernel and run 'lilo'. Re-boot into new kernel and I receive multiple 'unresolved symbol' errors during rc.conf parse. After root login I run 'depmod-a' and the 'unresolved symbols' run page-after-page(mostly drivers it appears).
I'm in the process of downgrading to gcc-3.2.1-1 from arch-0.4 -- whew! Thank god you keep the old repos!
Will let you know how it goes.
Has anyone figured-out how to compile 2.4.22 with gcc 3.3.1?
I compiled today both 2.4.22 & 2.4.23-pre4 with gcc-3.3.1 on a fresh installation without any problems.
I tried both with makepkg & 'the custom way' (since I needed full header files) with no problem either.
Maybe the ck patchset is broken ? I'll try it tomorrow...
Offline
OK -- it worked better with gcc-3.2.1-1. I used the exact same kernel and config. The only unresolved symbols are:
.../drivers/atm/he.o
.../drivers/net/sk98lin/sk98lin.o
I read around on Google that while the kernel will compile under gcc3, not all the drivers will(!?).
I suppose I don't need those two drivers for my SONY PCG-F490 laptop, anyway, so I may just take them out and try again.
Curious, though, does anyone else have ideas/suggestions about this? Should I upgrade to gcc-3.3.1, again, and re-compile to get a second/third result for cross-comparison?
Hmmm......ponderous.
Thanx for your thoughts.
-- Linux! Isn't it time?
Offline
OK ...... really crazy!
I just re-compiled the kernal, once again with the original gcc-3.3.1. I removed the 'he.o" and "sk98lin.o" and everything is fine. 'depmod -a" returns _no_ errors and I'd swear it was only a fluke that I had problems the first time --- except, now I wonder if the system is simply using the pre-compiled modules from the gcc-3.2.1 compilation, previously(!?).
Anybody know if the kernel Make process will simply re-use pre-existing modules from a previous compile if found?
I'm happy---but perplexed. Thanx.
-- Linux! Isn't it time?
Offline
well no i believe it recompiles all of them.
my guess is that if you recompiled this time without patching it was the patch itself that was foobarred
AKA uknowme
I am not your friend
Offline
OK, that makes sense about all the modules, except I did recompile the _same_ patched kernel as before(!). Used 'make clean mrproper distclean' on the sources. I tried to keep the number of changed variables to only things specifically errored. The Con Kolivas patched kernel is what I've been using each time. btw: the patches+fixes from Con's website apply to a vanilla kernel without error and cleanly.
Hmmm, maybe Mars was in my Venus yesterday or something(?).
-- Linux! Isn't it time?
Offline