> On Tue, 26 Aug 2003, Kevin Coffman wrote: > > > > > gcc -O2 -fomit-frame-pointer -fno-strength-reduce -fno-strict-aliasing > > > > -fno-common -pipe -march=pentium -D__KERNEL__ -DCPU=586 -DKERNEL > > > > -D_KERNEL -DMODULE -DAFS_SMP -I. -I../ -I/home/kwc/build/openafs-1.2 > . > > > > 10/src/config -c ../afs/osi_module.c > > > > ../afs/osi_module.c: In function `afs_init': > > > > ../afs/osi_module.c:325: `sys_exit' undeclared (first use in this > > > > function) > > > > ../afs/osi_module.c:325: (Each undeclared identifier is reported only > > > > once > > > > ../afs/osi_module.c:325: for each function it appears in.) > > > > ../afs/osi_module.c:327: warning: assignment from incompatible pointer > > > > type > > > > make[4]: *** [osi_module.o] Error 1 > > > > make[4]: Leaving directory `/home/kwc/build/openafs-1.2.10/src/libafs/M > O > > > > DLOAD-2.4.20-MP' > > > > make[3]: *** [linux_compdirs] Error 2 > > > > make[3]: Leaving directory `/home/kwc/build/openafs-1.2.10/src/libafs' > > > > make[2]: *** [libafs] Error 2 > > > > make[2]: Leaving directory `/home/kwc/build/openafs-1.2.10' > > > > make[1]: *** [build] Error 2 > > > > make[1]: Leaving directory `/home/kwc/build/openafs-1.2.10' > > > > make: *** [all] Error 2 > > > > -bash-2.05b$ ls > > > > > > > > The kernel header files are located in /usr/include/linux and I have > > > > > > It would be nicer if you could point at e.g. /usr/src/linux-whatever, > > > where the kernel was compiled. > > > > Yeah, it would be nice ... > > Well, the "is sys_call_table exported" stuff might actually work right.
I can probably build this on the machine where the kernel was actually built. Is there an equivalent of a 'make dest' for a non-transarc-paths build? (I can't do a 'make install' on that machine.) > > > including the modversions crap which would be in > > > include/linux/modules/*.ver? > > > > OK, that reminds me. /usr/include/linux/modules is empty on this > > system. I applied this patch previously to the cvs head version and to > > 1.2.10. > > Well, if nothing else is including it, it's needed. Are you using a > versioned kernel? Without this, you should get a module with refs to > sys_call_table and not sys_call_table_R12345678, which means the kernel > module just won't load, unless you use insmod -f. We don't want people to > use insmod -f. > > Is something else including it now? It doesn't exist, so no-one can be including it. All I can tell you is that the kernel module built from cvs head loaded without insmod -f. This was on a different, but 'identical' machine. > > -bash-2.05b$ diff -u src/config/param.i386_linux24.h.orig src/config/param. > i386_linux24.h > > --- src/config/param.i386_linux24.h.orig 2003-07-16 16:41:17.000000000 -0 > 400 > > +++ src/config/param.i386_linux24.h 2003-07-16 16:32:03.000000000 -0 > 400 > > @@ -47,10 +47,12 @@ > > #define AFS_GLOBAL_SUNLOCK > > #endif > > > > +/* KWC: This doesn't seem to be needed ??? > > #if defined(MODULE) && defined(CONFIG_MODVERSIONS) > > #define MODVERSIONS > > #include <linux/modversions.h> > > #endif > > + KWC */ > > > > #endif /* __KERNEL__ && !DUMP_KERNEL*/ > > > > _______________________________________________ > OpenAFS-info mailing list > [EMAIL PROTECTED] > https://lists.openafs.org/mailman/listinfo/openafs-info > _______________________________________________ OpenAFS-info mailing list [EMAIL PROTECTED] https://lists.openafs.org/mailman/listinfo/openafs-info
