From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.6) Gecko/20011120 Description of problem: Attempts to build from kernel-source-2.4.9-13.i386.rpm using the supplied config file: configs/kernel-2.4.9-athlon.config fail with multiple undefined references e.g.:create_bounce, kmap_high, kunmap_high, mmx_clear_page, mmx_copy_page, _mmx_memcpy, noathlon, pkmap_page_table, rwsem_down_read_failed, rwsem_down_write_failed, rwsem_wake are ALL undefined as far as the build process is concerned Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. cd /usr/src/ 2. rpm -e kernel-source 3. rm -rf /usr/src/linux-2.4.9-13 4. rpm -ivh kernel-source-2.4.9-13.i386.rpm 5. cd /usr/src/linux-2.4 6. cp configs/kernel-2.4.9-athlon.config .config 7. make dep 8. make bzImage Actual Results: Kernel fails to build with errors such as: arch/i386/kernel/kernel.o(.text+0x111): undefined reference to `_mmx_memcpy' See the attached file for a full listing of the output from "make bzImage" Expected Results: Expected a nice bzImage of the kernel for an athlon processor. Additional info:
Created attachment 40967 [details] output from "make bzImage"
Looks like you missed 'make oldconfig' to actually activate the changes to .config (and technically you also need to run 'make mrproper' very first) Note that if you don't change the .config, we also ship binaries compiled with it....
Thanks for the follow up. There is nothing more pleasant than making an error in public. 'make oldconfig' resolved the build problem. The status can be changed to "not a bug". Re: the binaries compiled with configs/kernel-2.4.9-athlon.config I assume this config was used to create: kernel-2.4.9-13.athlon.rpm My main asus/athlon system is so unstable using this package that is unusable for any real work. (It regularly dies with: Unable to handle kernel paging request at virtual address ...) That's what prompted me start working with compiling my own.
The following might be worth a shot: http://people.redhat.com/arjanv/testkernels it has at least one workaround for a bug in some athlon bioses....