Gilles Espinasse
2008-Dec-29 15:04 UTC
[klibc] klibc broken on ppc with linux-2.6.27 and workaround
Compilation was ok with klibc-1.5.10 on linux-2.6.25. Distrib does not matter as it is inside a chroot a la LSF With the upgrade to linux-2.6.27 it broke on 1.5.10 (I try too the last 1.5.14) with make[1]: Entering directory `/usr/src/klibc-1.5.14' GEN klcc/klibc.config GEN klcc/klcc HOSTCC scripts/basic/fixdep KLIBCCC usr/klibc/__static_init.o KLIBCCC usr/klibc/vsnprintf.o In file included from usr/klibc/../include/sys/types.h:15, from usr/klibc/../include/unistd.h:11, from usr/klibc/../include/stdio.h:11, from usr/klibc/vsnprintf.c:13: linux/include/linux/posix_types.h:47:29: error: asm/posix_types.h: No such file or directory In file included from usr/klibc/../include/unistd.h:11, from usr/klibc/../include/stdio.h:11, from usr/klibc/vsnprintf.c:13: usr/klibc/../include/sys/types.h:16:23: error: asm/types.h: No such file or directory In file included from usr/klibc/../include/unistd.h:11, from usr/klibc/../include/stdio.h:11, from usr/klibc/vsnprintf.c:13: usr/klibc/../include/sys/types.h:23: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'ino_t' ... As a workaround I have symlinked the directories removed in 2.6.27 and it now compile cd /usr/src/linux-2.6.27/arch && ln -sf powerpc ppc cd /usr/src/linux-2.6.27/include && ln -sf arch-powerpc arch-ppc I have seen that Gentoo use a different system and a klibcasmarch.patch http://sources.gentoo.org/viewcvs.py/gentoo-x86/dev-libs/klibc/klibc-1.5.12.ebuild?rev=1.1&view=markup A bit complicated for me. Gilles
H. Peter Anvin
2008-Dec-29 17:53 UTC
[klibc] klibc broken on ppc with linux-2.6.27 and workaround
Gilles Espinasse wrote:> Compilation was ok with klibc-1.5.10 on linux-2.6.25. > Distrib does not matter as it is inside a chroot a la LSF > > As a workaround I have symlinked the directories removed in 2.6.27 and it > now compile > cd /usr/src/linux-2.6.27/arch && ln -sf powerpc ppc > cd /usr/src/linux-2.6.27/include && ln -sf arch-powerpc arch-ppc > > I have seen that Gentoo use a different system and a klibcasmarch.patch > http://sources.gentoo.org/viewcvs.py/gentoo-x86/dev-libs/klibc/klibc-1.5.12.ebuild?rev=1.1&view=markup > A bit complicated for me.I guess I really should spend some TLC on klibc... I haven't given it too much time, and with all the header changing around lately it gets tricky. The Right Thing[TM] is probably to use the exported header stuff for klibc, assuming it contains sufficient information. -hpa -- H. Peter Anvin, Intel Open Source Technology Center I work for Intel. I don't speak on their behalf.
Seemingly Similar Threads
- [BUG] Choice between compile errors
- [klibc:riscv64-enable-relax] Kbuild, klcc: Support multiple objects in KLIBCCRTSHARED
- Problems with klibc 1.1.8
- Bug#653790: klibc/s390x: searches for include in the wrong path
- [PATCH klibc 1/3] Revert " Kbuild: Tell gas we don't want executable stacks"