Actually, both errors are *not* the same
Dan's is due to something missing in the kernel sources; I've read the thread in the kernel developpers' mailing list, I can't recall exactly what is missing, but I know that a patch is available, either for 2.4.20 or the soon-to-come 2.4.21. An acceptable workaround for now is to downgrade your compiler (3.2.x seems to be OK with it; I recompiled my kernel using 2.95 and it worked fine). As for Peter's, this kind of segfaulting has been around since gcc-3.x AFAIK. I've came across quite a few times, everytime it involved compilation of large projects (the kernel, QT3, etc). All you have to do is simply to re-start make (make, make bzImage, make modules, whatever) and all works fine. For both of you, I'm not saying that the solutions I offer is *THE* solutions, but at least these worked for me. Someone might come with a better one, you decide :) Jeff On Mon, 2003-06-09 at 23:16, Peter S Galbraith wrote: > Dan Jones <[EMAIL PROTECTED]> wrote: > > > On Mon, 09 Jun 2003 18:26:46 GMT, Matthias Leopold wrote: > > >hi > > > > > >i tried to recompile my 2.4.20 kernel with make-kpkg (as i have > > >successfully done before) incorporating the xfs and ck patches. this > > >time the process stopped with the following error message: > > > > > >net/network.o(.text+0xe117): In function `rtnetlink_rcv': : > > >undefined reference to `rtnetlink_rcv_skb > > > > > >in google i found one reference to this problem (but no solution) > > >which linked the problem to the newly introduced gcc-3.3. how do i > > >avoid this? > > I had the exact same error. > > > Two ways. One is to use gcc-3.2.3. You can install the gcc-3.2.3 > > without uninstalling the gcc-3.3 package. Just change the symbolic > > link in /usr/bin to point to the older compiler after install. > > > > Alternately, you can edit /usr/src/linux/net/core/rtnetlink.c and > > change the function declaration from extern __inline__ int > > rtlink_rcv_skb(struct sk_buf *skb); to static inline int > > rtlink_rcv_skb(struct sk_buf *skb); (The leading and trailing > > underscores around inline are no longer needed.) Changing the > > declaration from extern inline to static inline causes gcc-3.3 to > > actually inline the function and fix the problem. > > I'm now compiling with gcc-3.2 (by editing the top-level Makefile > value for CC) but I'm getting erors all over the place. Then I config > out the problematic file only to have it fail later. :-( > > Latest is: > > gcc-3.2 -D__KERNEL__ -I/usr/src/kernel-source-2.4.20/include -Wall > -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common > -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=athlon -DMODULE > -DMODVERSIONS -include /usr/src/kernel-source-2.4.20/include/linux/modversions.h > -nostdinc -iwithprefix include -DKBUILD_BASENAME=cciss -c -o cciss.o cciss.c > cciss.c: In function `cciss_ioctl': > cciss.c:422: internal error: Segmentation fault > Please submit a full bug report, > with preprocessed source if appropriate. > See <URL:http://gcc.gnu.org/bugs.html> for instructions. > make[3]: *** [cciss.o] Error 1 > make[3]: Leaving directory `/usr/src/kernel-source-2.4.20/drivers/block' > make[2]: *** [_modsubdir_block] Error 2 > make[2]: Leaving directory `/usr/src/kernel-source-2.4.20/drivers' > make[1]: *** [_mod_drivers] Error 2 > make[1]: Leaving directory `/usr/src/kernel-source-2.4.20' > make: *** [stamp-build] Error 2 > > Peter > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
