I'm building OpenAFS 1.4.2 (or 1.5.10, doesn't matter) on SLED 10. As documented in /usr/src/linux/README.SUSE, I want to use /lib/modules/<version>/build instead of /usr/src/linux as my kernel linux kernel headers directory, as one would do on RHEL or Fedora Core.
Unlike RHEL or FC, however, SUSE's /lib/modules/<version>/build contains only a skeleton with the autogenerated files from the kernel build. I've included the output of "find" there for reference. In particular, scripts/Makefile.build is not present, so AC_TRY_KBUILD erroneously uses AC_TRY_KBUILD24, causing all of the configure tests to return negatively. (The build then bombs out because it doesn't know what to do with rlim.) As far as I know at this time, that's the *only* thing wrong with what the build system is doing; if I create a dummy /lib/modules/<version>/build/scripts/Makefile.build, the AC_TRY_KBUILD26 macro works beautifully and the build succeeds. So, I think AC_TRY_KBUILD needs to be looking for a different file in order to detect a 2.6 tree in order to work out of the box on SUSE. I'm not immediately sure what would work. ---- error-messages:/lib/modules/2.6.16.21-0.25-smp/build # find . ./arch ./arch/i386 ./arch/i386/kernel ./include ./include/asm-i386 ./include/asm-i386/asm-offsets.h ./include/asm ./include/linux ./include/linux/autoconf.h ./include/linux/version.h ./Module.symvers ./.config ./.kernelrelease ./Makefile ./include2 ./include2/asm ./scripts ./scripts/basic ./scripts/basic/split-include ./scripts/basic/fixdep ./scripts/basic/docproc ./scripts/genksyms ./scripts/genksyms/keywords.c ./scripts/genksyms/parse.h ./scripts/genksyms/parse.c ./scripts/genksyms/lex.c ./scripts/genksyms/genksyms ./scripts/kconfig ./scripts/kconfig/zconf.hash.c ./scripts/kconfig/zconf.tab.c ./scripts/kconfig/lex.zconf.c ./scripts/kconfig/conf ./scripts/mod ./scripts/mod/mk_elfconfig ./scripts/mod/modpost ./scripts/mod/elfconfig.h ./scripts/kallsyms ./scripts/conmakehash ./scripts/bin2c _______________________________________________ OpenAFS-devel mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-devel
