[gcc-4.4] Re: arch/x86/events/intel/uncore_snbep.o: warning: objtool: snbep_uncore_msr_enable_event()+0x2e: function has unreachable instruction
On Thu, Jul 21, 2016 at 11:06:06AM +0800, Liang, Kan wrote: Hi Fengguang, I located the unreachable instruction which is ud2. This instruction will raise invalid opcode exception. So I think normally it should not be reached. Also ud2 should be generated by compiler, not our codes. It's great to know that! I have no idea how to fix it. Have you met similar issue before? Do you have any suggestions? The warning shows up when we enabled gcc-4.4 for kernel build tests. It emits this "function has unreachable instruction" warning for ~1 kernel functions all over the places. So it must be a toolchain specific problem. Let's CC LKML and gcc lists for possible solutions. 0e0f : e0f: 55 push %rbp e10: 48 89 e5mov%rsp,%rbp e13: 53 push %rbx e14: 83 be 80 01 00 00 ffcmpl $0x,0x180(%rsi) e1b: 48 89 f3mov%rsi,%rbx e1e: 74 2b je e4b e20: 31 f6 xor%esi,%esi e22: e8 00 00 00 00 callq e27 e27: 48 89 c2mov%rax,%rdx e2a: 8b b3 78 01 00 00 mov0x178(%rbx),%esi e30: 48 c1 ea 20 shr$0x20,%rdx e34: 48 83 3d 00 00 00 00cmpq $0x0,0x0(%rip)# e3c e3b: 00 e3c: 75 02 jnee40 e3e: 0f 0b ud2 e40: 89 f7 mov%esi,%edi e42: 89 c6 mov%eax,%esi e44: ff 14 25 00 00 00 00callq *0x0 Thanks, Kan tree: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master head: 47ef4ad2684d380dd6d596140fb79395115c3950 commit: 3b94a891667c30fb4624221497d77fc65d950345 perf/x86/intel/uncore: Remove SBOX support for Broadwell server date: 7 weeks ago config: x86_64-randconfig-s0-07191857 (attached as .config) compiler: gcc-4.4 (Debian 4.4.7-8) 4.4.7 reproduce: git checkout 3b94a891667c30fb4624221497d77fc65d950345 # save the attached .config to linux build tree make ARCH=x86_64 All warnings (new ones prefixed by >>): >> arch/x86/events/intel/uncore_snbep.o: warning: objtool: snbep_uncore_msr_enable_event()+0x2e: function has unreachable instruction --- 0-DAY kernel test infrastructureOpen Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation
[net-next:master 488/665] verifier.c:undefined reference to `__multi3'
CC gcc list. According to Alexei: This is a known issue with gcc 7 on mips that is "optimizing" normal 64-bit multiply into 128-bit variant. Nothing to fix on the kernel side. Digging further, this function was added to arch/sparc/lib/multi3.S since gcc doing the same "optimization" there. Adding asm code doesn't look right to me. I'd rather push gcc folks to avoid such codegen. tree: https://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git master head: 7c5556decd0a629e9ee02e93653f75ba7b7da03c commit: 638f5b90d46016372a8e3e0a434f199cc5e12b8c [488/665] bpf: reduce verifier memory consumption config: mips-64r6el_defconfig (attached as .config) compiler: mips64el-linux-gnuabi64-gcc (Debian 7.2.0-11) 7.2.0 reproduce: wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross git checkout 638f5b90d46016372a8e3e0a434f199cc5e12b8c # save the attached .config to linux build tree make.cross ARCH=mips All errors (new ones prefixed by >>): kernel/bpf/verifier.o: In function `realloc_verifier_state.isra.19': >> verifier.c:(.text+0x36fc): undefined reference to `__multi3' crypto/scompress.o: In function `.L82': scompress.c:(.text+0x55c): undefined reference to `__multi3' lib/mpi/generic_mpih-mul1.o: In function `.L2': generic_mpih-mul1.c:(.text+0x60): undefined reference to `__multi3' lib/mpi/generic_mpih-mul2.o: In function `.L2': generic_mpih-mul2.c:(.text+0x5c): undefined reference to `__multi3' lib/mpi/generic_mpih-mul3.o: In function `.L2': generic_mpih-mul3.c:(.text+0x5c): undefined reference to `__multi3' lib/mpi/mpih-div.o:mpih-div.c:(.text+0x1b8): more undefined references to `__multi3' follow --- 0-DAY kernel test infrastructureOpen Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation .config.gz Description: application/gzip ___ kbuild-all mailing list kbuild-...@lists.01.org https://lists.01.org/mailman/listinfo/kbuild-all
