Alexander... On Fri, 2004-10-15 at 23:31, Alexander Dalloz wrote: > Am Fr, den 15.10.2004 schrieb John Francis Lee um 12:56: > > > I decided to try with the distributed kernel source... and I got the > > same response. This was with the source and config distributed by redhat > > and presently running the machine! > > Then you repeat the same fault over and over. Did you ever successfully > compile a kernel on Linux? I mean in general. So to know whether you > know what you are doing. I've compiled many kernels since the 2.0 releases. I've never had any trouble like this before. I've just deleted the source tree for 2.6.8.1, untared the binary and begun again... "I repeat the same fault over and over." In any case thanks for your help. > > > So then I decided to delete a lot of modules, which I did bottom up as > > you suggested. Extremely tedious. I had assumed that unchecking ISDN, > > for instance, would have kept all the modules below ISDN from being > > compiled. But I unchecked everything below ISDN by hand as you > > suggested. And so on. > > If you deselect a global section like ISDN in total, it is safe to do so > without unchecking each ISDN module separately. > > > I got the same result again. I have a list of 1193 unique unknown > > symbols from make modules_install! > > Do you have a sane FC2 system or did you exchange any part of the > development tools? I believe to remember that you fiddled around with > the sources of the modutils. > > > I don't understand it. And /lib/modules/2.6.5-1.358custom/ is three > > times the size of the original! > > Because the modules are not stripped and contain all the debug > information. > > > Any insight into what might be wrong here? > > Not really. I did compile the 2.6 kernel more that 1 time on FC2 and had > not any of your difficulties. > > > John Francis Lee <jfl@xxxxxxxxxxxx> > > Alexander -- John Francis Lee <jfl@xxxxxxxxxxxx>
