build err on arm64 machine, relocation truncated to fit: R_AARCH64_LD64_GOTPAGE_LO15 against symbol

[zhangzhiqiangcs]

Discussed in #3148

^{Originally posted by zhangzhiqiangcs May 2, 2024}
Hi, all, I try to build a static bpftrace binary on my arm64 manchine in a alpine:3.18 container, just followed the docker/Dockerfile.static and static.sh, the build process always filed at final link stage with err

Linking CXX executable bpftrace
 /usr/bin/cmake -E cmake_link_script CMakeFiles/bpftrace.dir/link.txt --verbose=1
/usr/bin/c++ -O3 -DNDEBUG -static-libgcc -static-libstdc++ CMakeFiles/bpftrace.dir/main.cpp.o -o bpftrace  libbpftrace.a resources/libresources.a libruntime.a /usr/lib/libbpf.a /lib/libz.a /usr/lib/libbcc.a /usr/lib/libbcc_bpf.a /usr/lib/libbpf.a /lib/libz.a /usr/lib/libbcc.a /usr/lib/libbcc_bpf.a /usr/lib/libbcc-loader-static.a /usr/lib/liblzma.a /usr/lib/libelf.a aot/libaot.a librequired_resources.a ast/libast.a ../libparser.a ast/libast_defs.a /usr/lib/llvm16/lib/libclang.a /usr/lib/llvm16/lib/libclangExtractAPI.a /usr/lib/llvm16/lib/libclangARCMigrate.a /usr/lib/llvm16/lib/libclangIncludeFixerPlugin.a /usr/lib/llvm16/lib/libclangTooling.a /usr/lib/llvm16/lib/libclangIncludeFixer.a /usr/lib/llvm16/lib/libfindAllSymbols.a /usr/lib/llvm16/lib/libclangIndex.a /usr/lib/llvm16/lib/libclangFormat.a /usr/lib/llvm16/lib/libclangToolingInclusions.a /usr/lib/llvm16/lib/libclangFrontend.a /usr/lib/llvm16/lib/libclangDriver.a /usr/lib/llvm16/lib/libclangParse.a /usr/lib/llvm16/lib/libclangSerialization.a /usr/lib/llvm16/lib/libclangSema.a /usr/lib/llvm16/lib/libclangEdit.a /usr/lib/llvm16/lib/libclangSupport.a /usr/lib/llvm16/lib/libclangTidyPlugin.a /usr/lib/llvm16/lib/libclangTidyAndroidModule.a /usr/lib/llvm16/lib/libclangTidyAbseilModule.a /usr/lib/llvm16/lib/libclangTidyAlteraModule.a /usr/lib/llvm16/lib/libclangTidyBoostModule.a /usr/lib/llvm16/lib/libclangTidyCERTModule.a /usr/lib/llvm16/lib/libclangTidyConcurrencyModule.a /usr/lib/llvm16/lib/libclangTidyDarwinModule.a /usr/lib/llvm16/lib/libclangTidyFuchsiaModule.a /usr/lib/llvm16/lib/libclangTidyHICPPModule.a /usr/lib/llvm16/lib/libclangTidyBugproneModule.a /usr/lib/llvm16/lib/libclangTidyCppCoreGuidelinesModule.a /usr/lib/llvm16/lib/libclangTidyGoogleModule.a /usr/lib/llvm16/lib/libclangTidyLinuxKernelModule.a /usr/lib/llvm16/lib/libclangTidyLLVMModule.a /usr/lib/llvm16/lib/libclangTidyLLVMLibcModule.a /usr/lib/llvm16/lib/libclangTidyMiscModule.a /usr/lib/llvm16/lib/libclangAnalysis.a /usr/lib/llvm16/lib/libclangASTMatchers.a /usr/lib/llvm16/lib/libclangAST.a /usr/lib/llvm16/lib/libclangTidyModernizeModule.a /usr/lib/llvm16/lib/libclangTidyObjCModule.a /usr/lib/llvm16/lib/libclangTidyOpenMPModule.a /usr/lib/llvm16/lib/libclangTidyPerformanceModule.a /usr/lib/llvm16/lib/libclangTidyPortabilityModule.a /usr/lib/llvm16/lib/libclangTidyReadabilityModule.a /usr/lib/llvm16/lib/libclangTidyZirconModule.a /usr/lib/llvm16/lib/libclangTidyMPIModule.a /usr/lib/llvm16/lib/libclangTidyUtils.a /usr/lib/llvm16/lib/libclangTidy.a /usr/lib/llvm16/lib/libclangToolingCore.a /usr/lib/llvm16/lib/libclangRewrite.a /usr/lib/llvm16/lib/libclangLex.a /usr/lib/llvm16/lib/libclangBasic.a /usr/lib/llvm16/lib/libLLVMBPFCodeGen.a /usr/lib/llvm16/lib/libLLVMMCJIT.a /usr/lib/llvm16/lib/libLLVMOrcJIT.a /usr/lib/llvm16/lib/libLLVMExecutionEngine.a /usr/lib/llvm16/lib/libLLVMRuntimeDyld.a /usr/lib/llvm16/lib/libLLVMJITLink.a /usr/lib/llvm16/lib/libLLVMOrcTargetProcess.a /usr/lib/llvm16/lib/libLLVMOrcShared.a /usr/lib/llvm16/lib/libLLVMWindowsDriver.a /usr/lib/llvm16/lib/libLLVMOption.a /usr/lib/llvm16/lib/libLLVMAArch64CodeGen.a /usr/lib/llvm16/lib/libLLVMAArch64AsmParser.a /usr/lib/llvm16/lib/libLLVMAArch64Disassembler.a /usr/lib/llvm16/lib/libLLVMAArch64Desc.a /usr/lib/llvm16/lib/libLLVMAArch64Info.a /usr/lib/llvm16/lib/libLLVMAArch64Utils.a /usr/lib/llvm16/lib/libLLVMAMDGPUCodeGen.a /usr/lib/llvm16/lib/libLLVMPasses.a /usr/lib/llvm16/lib/libLLVMCoroutines.a /usr/lib/llvm16/lib/libLLVMIRPrinter.a /usr/lib/llvm16/lib/libLLVMMIRParser.a /usr/lib/llvm16/lib/libLLVMAMDGPUAsmParser.a /usr/lib/llvm16/lib/libLLVMAMDGPUDisassembler.a /usr/lib/llvm16/lib/libLLVMAMDGPUDesc.a /usr/lib/llvm16/lib/libLLVMAMDGPUInfo.a /usr/lib/llvm16/lib/libLLVMAMDGPUUtils.a /usr/lib/llvm16/lib/libLLVMARMCodeGen.a /usr/lib/llvm16/lib/libLLVMARMAsmParser.a /usr/lib/llvm16/lib/libLLVMARMDisassembler.a /usr/lib/llvm16/lib/libLLVMARMDesc.a /usr/lib/llvm16/lib/libLLVMARMInfo.a /usr/lib/llvm16/lib/libLLVMARMUtils.a /usr/lib/llvm16/lib/libLLVMAVRCodeGen.a /usr/lib/llvm16/lib/libLLVMAVRAsmParser.a /usr/lib/llvm16/lib/libLLVMAVRDesc.a /usr/lib/llvm16/lib/libLLVMAVRDisassembler.a /usr/lib/llvm16/lib/libLLVMAVRInfo.a /usr/lib/llvm16/lib/libLLVMBPFAsmParser.a /usr/lib/llvm16/lib/libLLVMBPFDesc.a /usr/lib/llvm16/lib/libLLVMBPFDisassembler.a /usr/lib/llvm16/lib/libLLVMBPFInfo.a /usr/lib/llvm16/lib/libLLVMHexagonCodeGen.a /usr/lib/llvm16/lib/libLLVMHexagonAsmParser.a /usr/lib/llvm16/lib/libLLVMHexagonDisassembler.a /usr/lib/llvm16/lib/libLLVMHexagonDesc.a /usr/lib/llvm16/lib/libLLVMHexagonInfo.a /usr/lib/llvm16/lib/libLLVMLanaiCodeGen.a /usr/lib/llvm16/lib/libLLVMLanaiAsmParser.a /usr/lib/llvm16/lib/libLLVMLanaiDisassembler.a /usr/lib/llvm16/lib/libLLVMLanaiDesc.a /usr/lib/llvm16/lib/libLLVMLanaiInfo.a /usr/lib/llvm16/lib/libLLVMLoongArchCodeGen.a /usr/lib/llvm16/lib/libLLVMLoongArchAsmParser.a /usr/lib/llvm16/lib/libLLVMLoongArchDisassembler.a /usr/lib/llvm16/lib/libLLVMLoongArchDesc.a /usr/lib/llvm16/lib/libLLVMLoongArchInfo.a /usr/lib/llvm16/lib/libLLVMMipsCodeGen.a /usr/lib/llvm16/lib/libLLVMMipsAsmParser.a /usr/lib/llvm16/lib/libLLVMMipsDesc.a /usr/lib/llvm16/lib/libLLVMMipsDisassembler.a /usr/lib/llvm16/lib/libLLVMMipsInfo.a /usr/lib/llvm16/lib/libLLVMMSP430CodeGen.a /usr/lib/llvm16/lib/libLLVMMSP430AsmParser.a /usr/lib/llvm16/lib/libLLVMMSP430Desc.a /usr/lib/llvm16/lib/libLLVMMSP430Disassembler.a /usr/lib/llvm16/lib/libLLVMMSP430Info.a /usr/lib/llvm16/lib/libLLVMNVPTXCodeGen.a /usr/lib/llvm16/lib/libLLVMNVPTXDesc.a /usr/lib/llvm16/lib/libLLVMNVPTXInfo.a /usr/lib/llvm16/lib/libLLVMPowerPCCodeGen.a /usr/lib/llvm16/lib/libLLVMPowerPCAsmParser.a /usr/lib/llvm16/lib/libLLVMPowerPCDesc.a /usr/lib/llvm16/lib/libLLVMPowerPCDisassembler.a /usr/lib/llvm16/lib/libLLVMPowerPCInfo.a /usr/lib/llvm16/lib/libLLVMRISCVCodeGen.a /usr/lib/llvm16/lib/libLLVMipo.a /usr/lib/llvm16/lib/libLLVMFrontendOpenMP.a /usr/lib/llvm16/lib/libLLVMLinker.a /usr/lib/llvm16/lib/libLLVMVectorize.a /usr/lib/llvm16/lib/libLLVMRISCVAsmParser.a /usr/lib/llvm16/lib/libLLVMRISCVDisassembler.a /usr/lib/llvm16/lib/libLLVMRISCVDesc.a /usr/lib/llvm16/lib/libLLVMRISCVInfo.a /usr/lib/llvm16/lib/libLLVMSparcCodeGen.a /usr/lib/llvm16/lib/libLLVMSparcAsmParser.a /usr/lib/llvm16/lib/libLLVMSparcDesc.a /usr/lib/llvm16/lib/libLLVMSparcDisassembler.a /usr/lib/llvm16/lib/libLLVMSparcInfo.a /usr/lib/llvm16/lib/libLLVMSystemZCodeGen.a /usr/lib/llvm16/lib/libLLVMSystemZAsmParser.a /usr/lib/llvm16/lib/libLLVMSystemZDisassembler.a /usr/lib/llvm16/lib/libLLVMSystemZDesc.a /usr/lib/llvm16/lib/libLLVMSystemZInfo.a /usr/lib/llvm16/lib/libLLVMVECodeGen.a /usr/lib/llvm16/lib/libLLVMVEAsmParser.a /usr/lib/llvm16/lib/libLLVMVEDesc.a /usr/lib/llvm16/lib/libLLVMVEDisassembler.a /usr/lib/llvm16/lib/libLLVMVEInfo.a /usr/lib/llvm16/lib/libLLVMWebAssemblyCodeGen.a /usr/lib/llvm16/lib/libLLVMWebAssemblyAsmParser.a /usr/lib/llvm16/lib/libLLVMWebAssemblyDisassembler.a /usr/lib/llvm16/lib/libLLVMWebAssemblyDesc.a /usr/lib/llvm16/lib/libLLVMWebAssemblyInfo.a /usr/lib/llvm16/lib/libLLVMWebAssemblyUtils.a /usr/lib/llvm16/lib/libLLVMX86CodeGen.a /usr/lib/llvm16/lib/libLLVMInstrumentation.a /usr/lib/llvm16/lib/libLLVMGlobalISel.a /usr/lib/llvm16/lib/libLLVMCFGuard.a /usr/lib/llvm16/lib/libLLVMX86AsmParser.a /usr/lib/llvm16/lib/libLLVMX86Desc.a /usr/lib/llvm16/lib/libLLVMX86Disassembler.a /usr/lib/llvm16/lib/libLLVMX86Info.a /usr/lib/llvm16/lib/libLLVMXCoreCodeGen.a /usr/lib/llvm16/lib/libLLVMAsmPrinter.a /usr/lib/llvm16/lib/libLLVMSelectionDAG.a /usr/lib/llvm16/lib/libLLVMCodeGen.a /usr/lib/llvm16/lib/libLLVMScalarOpts.a /usr/lib/llvm16/lib/libLLVMAggressiveInstCombine.a /usr/lib/llvm16/lib/libLLVMInstCombine.a /usr/lib/llvm16/lib/libLLVMBitWriter.a /usr/lib/llvm16/lib/libLLVMObjCARCOpts.a /usr/lib/llvm16/lib/libLLVMTarget.a /usr/lib/llvm16/lib/libLLVMTransformUtils.a /usr/lib/llvm16/lib/libLLVMAnalysis.a /usr/lib/llvm16/lib/libLLVMProfileData.a /usr/lib/llvm16/lib/libLLVMSymbolize.a /usr/lib/llvm16/lib/libLLVMDebugInfoDWARF.a /usr/lib/llvm16/lib/libLLVMDebugInfoPDB.a /usr/lib/llvm16/lib/libLLVMObject.a /usr/lib/llvm16/lib/libLLVMIRReader.a /usr/lib/llvm16/lib/libLLVMAsmParser.a /usr/lib/llvm16/lib/libLLVMBitReader.a /usr/lib/llvm16/lib/libLLVMMCParser.a /usr/lib/llvm16/lib/libLLVMTextAPI.a /usr/lib/llvm16/lib/libLLVMDebugInfoMSF.a /usr/lib/llvm16/lib/libLLVMCore.a /usr/lib/llvm16/lib/libLLVMRemarks.a /usr/lib/llvm16/lib/libLLVMBitstreamReader.a /usr/lib/llvm16/lib/libLLVMXCoreDesc.a /usr/lib/llvm16/lib/libLLVMXCoreDisassembler.a /usr/lib/llvm16/lib/libLLVMMCDisassembler.a /usr/lib/llvm16/lib/libLLVMXCoreInfo.a /usr/lib/llvm16/lib/libLLVMMC.a /usr/lib/llvm16/lib/libLLVMBinaryFormat.a /usr/lib/llvm16/lib/libLLVMTargetParser.a /usr/lib/llvm16/lib/libLLVMDebugInfoCodeView.a /usr/lib/llvm16/lib/libLLVMSupport.a -ldl -lrt -lm -Wl,-Bdynamic /lib/libz.a /usr/lib/libzstd.a /usr/lib/llvm16/lib/libLLVMDemangle.a arch/libarch.a cxxdemangler/libcxxdemangler_llvm.a -Wl,-Bstatic /usr/lib/libpcap.a -Wl,-Bdynamic -lrt -lpthread -ldl -lm -Wl,-Bstatic -lz 
/usr/lib/libpcap.a(nametoaddr.o): in function `pcap_nametoeproto':
nametoaddr.c:(.text+0x410): relocation truncated to fit: R_AARCH64_LD64_GOTPAGE_LO15 against symbol `eproto_db' defined in .data.rel.local section in /usr/lib/libpcap.a(nametoaddr.o)
/usr/lib/gcc/aarch64-alpine-linux-musl/12.2.1/../../../../aarch64-alpine-linux-musl/bin/ld: nametoaddr.c:(.text+0x410): warning: too many GOT entries for -fpic, please recompile with -fPIC
collect2: error: ld returned 1 exit status
make[2]: *** [src/CMakeFiles/bpftrace.dir/build.make:323: src/bpftrace] Error 1
make[2]: Leaving directory '/go/src/go.megvii-inc.com/bpftrace/build-static'
make[1]: *** [CMakeFiles/Makefile2:574: src/CMakeFiles/bpftrace.dir/all] Error 2
make[1]: Leaving directory '/go/src/go.megvii-inc.com/bpftrace/build-static'
make: *** [Makefile:139: all] Error 2
```</div>

[zhangzhiqiangcs]

Find that the limitation about maximums GOT entries is architecture specific.
In GCC doc, it says: there no limit on x86_64 but 28k on AArch64.
https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html#Code-Gen-Options

-fpic
Generate position-independent code (PIC) suitable for use in a shared library, if supported for the target machine. Such code accesses all constant addresses through a global offset table (GOT). The dynamic loader resolves the GOT entries when the program starts (the dynamic loader is not part of GCC; it is part of the operating system). If the GOT size for the linked executable exceeds a machine-specific maximum size, you get an error message from the linker indicating that -fpic does not work; in that case, recompile with -fPIC instead. (These maximums are 8k on the SPARC, 28k on AArch64 and 32k on the m68k and RS/6000. The x86 has no such limit.)

Position-independent code requires special support, and therefore works only on certain machines. For the x86, GCC supports PIC for System V but not for the Sun 386i. Code generated for the IBM RS/6000 is always position-independent.
When this flag is set, the macros pic and PIC are defined to 1.

-fPIC
If supported for the target machine, emit position-independent code, suitable for dynamic linking and avoiding any limit on the size of the global offset table. This option makes a difference on AArch64, m68k, PowerPC and SPARC.

Position-independent code requires special support, and therefore works only on certain machines.