[ELF] Support aarch64_be

[MaskRay]

Summary

This patch adds

• Big-endian values for R_AARCH64_{ABS,PREL}{16,32,64} and R_AARCH64_PLT32
• aarch64elfb & aarch64linuxb BFD emulations
• elf64-bigaarch64 output format (bfdname)

Link: https://github.com/ClangBuiltLinux/linux/issues/1288

[MaskRay]

Tested in an aarch64_be rootfs

ld.lld can link gcc compiled object files and produce working executables and shared objects.

[MaskRay]

Support bitcode file e_machine inference
Test .got.plt

[peter.smith]

Changes LGTM.

I think the PLT32 relocation also needs to be write32 rather than write32le (https://github.com/ARM-software/abi-aa/blob/master/aaelf64/aaelf64.rst#575static-data-relocations) I couldn't immediately spot anything else missing.

lld/ELF/Arch/AArch64.cpp

335

I think PLT32 should be write as it is data (a 32-bit offset).

At the moment it is only used by Fuchsia for relative vtables so I wouldn't expect this to show up in linux.

[MaskRay]

Add PLT32

lld/ELF/Arch/AArch64.cpp

 
 RelType AArch64::getDynRel(RelType type) const {
   if (type == R_AARCH64_ABS64)
     return type;
   return R_AARCH64_NONE;
 }
 
 void AArch64::writeGotPlt(uint8_t *buf, const Symbol &) const {
-  write64le(buf, in.plt->getVA());
+  write64(buf, in.plt->getVA());
 }
 
 void AArch64::writePltHeader(uint8_t *buf) const {
   const uint8_t pltData[] = {
       0xf0, 0x7b, 0xbf, 0xa9, // stp    x16, x30, [sp,#-16]!
       0x10, 0x00, 0x00, 0x90, // adrp   x16, Page(&(.plt.got[2]))
       0x11, 0x02, 0x40, 0xf9, // ldr    x17, [x16, Offset(&(.plt.got[2]))]
       0x10, 0x02, 0x00, 0x91, // add    x16, x16, Offset(&(.plt.got[2]))

 }
 
 void AArch64::relocate(uint8_t *loc, const Relocation &rel,
                        uint64_t val) const {
   switch (rel.type) {
   case R_AARCH64_ABS16:
   case R_AARCH64_PREL16:
     checkIntUInt(loc, val, 16, rel);
-    write16le(loc, val);
+    write16(loc, val);
     break;
   case R_AARCH64_ABS32:
   case R_AARCH64_PREL32:
     checkIntUInt(loc, val, 32, rel);
-    write32le(loc, val);
+    write32(loc, val);
     break;
   case R_AARCH64_PLT32:
     checkInt(loc, val, 32, rel);
     write32le(loc, val);
     break;
   case R_AARCH64_ABS64:
   case R_AARCH64_PREL64:
-    write64le(loc, val);
+    write64(loc, val);
     break;
   case R_AARCH64_ADD_ABS_LO12_NC:
     or32AArch64Imm(loc, val);
     break;
   case R_AARCH64_ADR_GOT_PAGE:
   case R_AARCH64_ADR_PREL_PG_HI21:
   case R_AARCH64_TLSIE_ADR_GOTTPREL_PAGE21:
   case R_AARCH64_TLSDESC_ADR_PAGE21:

lld/ELF/Driver.cpp

if (s.endswith("_fbsd")) {
     s = s.drop_back(5);
     osabi = ELFOSABI_FREEBSD;
   }
 
   std::pair<ELFKind, uint16_t> ret =
       StringSwitch<std::pair<ELFKind, uint16_t>>(s)
           .Cases("aarch64elf", "aarch64linux", "aarch64_elf64_le_vec",
                  {ELF64LEKind, EM_AARCH64})
+          .Cases("aarch64elfb", "aarch64linuxb", {ELF64BEKind, EM_AARCH64})
           .Cases("armelf", "armelf_linux_eabi", {ELF32LEKind, EM_ARM})
           .Case("elf32_x86_64", {ELF32LEKind, EM_X86_64})
           .Cases("elf32btsmip", "elf32btsmipn32", {ELF32BEKind, EM_MIPS})
           .Cases("elf32ltsmip", "elf32ltsmipn32", {ELF32LEKind, EM_MIPS})
           .Case("elf32lriscv", {ELF32LEKind, EM_RISCV})
           .Cases("elf32ppc", "elf32ppclinux", {ELF32BEKind, EM_PPC})
           .Cases("elf32lppc", "elf32lppclinux", {ELF32LEKind, EM_PPC})
           .Case("elf64btsmip", {ELF64BEKind, EM_MIPS})

lld/ELF/ScriptParser.cpp

 static std::pair<ELFKind, uint16_t> parseBfdName(StringRef s) {
   return StringSwitch<std::pair<ELFKind, uint16_t>>(s)
       .Case("elf32-i386", {ELF32LEKind, EM_386})
       .Case("elf32-iamcu", {ELF32LEKind, EM_IAMCU})
       .Case("elf32-littlearm", {ELF32LEKind, EM_ARM})
       .Case("elf32-x86-64", {ELF32LEKind, EM_X86_64})
       .Case("elf64-aarch64", {ELF64LEKind, EM_AARCH64})
       .Case("elf64-littleaarch64", {ELF64LEKind, EM_AARCH64})
+      .Case("elf64-bigaarch64", {ELF64BEKind, EM_AARCH64})
       .Case("elf32-powerpc", {ELF32BEKind, EM_PPC})
       .Case("elf32-powerpcle", {ELF32LEKind, EM_PPC})
       .Case("elf64-powerpc", {ELF64BEKind, EM_PPC64})
       .Case("elf64-powerpcle", {ELF64LEKind, EM_PPC64})
       .Case("elf64-x86-64", {ELF64LEKind, EM_X86_64})
       .Cases("elf32-tradbigmips", "elf32-bigmips", {ELF32BEKind, EM_MIPS})
       .Case("elf32-ntradbigmips", {ELF32BEKind, EM_MIPS})
       .Case("elf32-tradlittlemips", {ELF32LEKind, EM_MIPS})

lld/test/ELF/aarch64-abs16.s

 // REQUIRES: aarch64
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %s -o %t.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs255.s -o %t255.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs256.s -o %t256.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs257.s -o %t257.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %s -o %t.be.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %S/Inputs/abs256.s -o %t256.be.o
 
 .globl _start
 _start:
 .data
   .hword foo + 0xfeff
   .hword foo - 0x8100
 
 // RUN: ld.lld %t.o %t256.o -o %t
 // RUN: llvm-objdump -s --section=.data %t | FileCheck %s --check-prefixes=CHECK,LE
+// RUN: ld.lld %t.be.o %t256.be.o -o %t.be
+// RUN: llvm-objdump -s --section=.data %t.be | FileCheck %s --check-prefixes=CHECK,BE
 
 // CHECK: Contents of section .data:
 // 220158: S = 0x100, A = 0xfeff
 //         S + A = 0xffff
 // 22015c: S = 0x100, A = -0x8100
 //         S + A = 0x8000
 // LE-NEXT: 220158 ffff0080
+// BE-NEXT: 220158 ffff8000
 
 // RUN: not ld.lld %t.o %t255.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW1
 // OVERFLOW1: relocation R_AARCH64_ABS16 out of range: -32769 is not in [-32768, 65535]; references foo
 
 // RUN: not ld.lld %t.o %t257.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW2
 // OVERFLOW2: relocation R_AARCH64_ABS16 out of range: 65536 is not in [-32768, 65535]; references foo

lld/test/ELF/aarch64-abs32.s

 // REQUIRES: aarch64
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %s -o %t.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs255.s -o %t255.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs256.s -o %t256.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs257.s -o %t257.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %s -o %t.be.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %S/Inputs/abs256.s -o %t256.be.o
 
 .globl _start
 _start:
 .data
   .word foo + 0xfffffeff
   .word foo - 0x80000100
 
 // RUN: ld.lld %t.o %t256.o -o %t
 // RUN: llvm-objdump -s --section=.data %t | FileCheck %s --check-prefixes=CHECK,LE
+// RUN: ld.lld %t.be.o %t256.be.o -o %t.be
+// RUN: llvm-objdump -s --section=.data %t.be | FileCheck %s --check-prefixes=CHECK,BE
 
 // CHECK: Contents of section .data:
 // 220158: S = 0x100, A = 0xfffffeff
 //         S + A = 0xffffffff
 // 22015c: S = 0x100, A = -0x80000100
 //         S + A = 0x80000000
 // LE-NEXT: 220158 ffffffff 00000080
+// BE-NEXT: 220158 ffffffff 80000000
 
 // RUN: not ld.lld %t.o %t255.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW1
 // OVERFLOW1: relocation R_AARCH64_ABS32 out of range: -2147483649 is not in [-2147483648, 4294967295]; references foo
 
 // RUN: not ld.lld %t.o %t257.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW2
 // OVERFLOW2: relocation R_AARCH64_ABS32 out of range: 4294967296 is not in [-2147483648, 4294967295]; references foo

lld/test/ELF/aarch64-data-relocs.s

 // REQUIRES: aarch64
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %s -o %t.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs256.s -o %t256.o
 // RUN: ld.lld %t.o %t256.o -o %t
 // RUN: llvm-objdump -s %t | FileCheck %s --check-prefixes=CHECK,LE
 
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %s -o %t.be.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %S/Inputs/abs256.s -o %t256.be.o
+// RUN: ld.lld %t.be.o %t256.be.o -o %t.be
+// RUN: llvm-objdump -s %t.be | FileCheck %s --check-prefixes=CHECK,BE
+
 .globl _start
 _start:
 .section .R_AARCH64_ABS64, "ax",@progbits
   .xword foo + 0x24
 
 // S = 0x100, A = 0x24
 // S + A = 0x124
 // CHECK: Contents of section .R_AARCH64_ABS64:
 // LE-NEXT: 210120 24010000 00000000
+// BE-NEXT: 210120 00000000 00000124
 
 .section .R_AARCH64_PREL64, "ax",@progbits
   .xword foo - . + 0x24
 
 // S + A - P = 0x100 + 0x24 - 0x210128 = 0xffffffffffdefffc
 // CHECK: Contents of section .R_AARCH64_PREL64:
 // LE-NEXT: 210128 fcffdeff ffffffff
+// BE-NEXT: 210128 ffffffff ffdefffc

lld/test/ELF/aarch64-prel16.s

 // REQUIRES: aarch64
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %s -o %t.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs255.s -o %t255.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs256.s -o %t256.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs257.s -o %t257.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %s -o %t.be.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %S/Inputs/abs256.s -o %t256.be.o
 
 .globl _start
 _start:
 .data
   .hword foo - . + 0x212057
   .hword foo - . + 0x1fa05a
 
 // Note: If this test fails, it probably happens because of
 //       the change of the address of the .data section.
 //       You may found the correct address in the aarch64_abs16.s test,
 //       if it is already fixed. Then, update addends accordingly.
 // RUN: ld.lld -z max-page-size=4096 %t.o %t256.o -o %t
 // RUN: llvm-objdump -s --section=.data %t | FileCheck %s --check-prefixes=CHECK,LE
+// RUN: ld.lld -z max-page-size=4096 %t.be.o %t256.be.o -o %t.be
+// RUN: llvm-objdump -s --section=.data %t.be | FileCheck %s --check-prefixes=CHECK,BE
 
 // CHECK: Contents of section .data:
 // 202158: S = 0x100, A = 0x212157, P = 0x202158
 //         S + A - P = 0xffff
 // 212a5a: S = 0x100, A = 0x1fa05a, P = 0x20215a
 //         S + A - P = 0x8000
 // LE-NEXT: 202158 ffff0080
 
 // RUN: not ld.lld -z max-page-size=4096 %t.o %t255.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW1
 // OVERFLOW1: relocation R_AARCH64_PREL16 out of range: -32769 is not in [-32768, 65535]; references foo
 
 // RUN: not ld.lld -z max-page-size=4096 %t.o %t257.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW2
 // OVERFLOW2: relocation R_AARCH64_PREL16 out of range: 65536 is not in [-32768, 65535]; references foo

lld/test/ELF/aarch64-prel32.s

 // REQUIRES: aarch64
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %s -o %t.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs255.s -o %t255.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs256.s -o %t256.o
 // RUN: llvm-mc -filetype=obj -triple=aarch64 %S/Inputs/abs257.s -o %t257.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %s -o %t.be.o
+// RUN: llvm-mc -filetype=obj -triple=aarch64_be %S/Inputs/abs256.s -o %t256.be.o
 
 .globl _start
 _start:
 .data
   .word foo - . + 0x100202057
   .word foo - . - 0x7fdfdfa4
 
 // Note: If this test fails, it probably happens because of
 //       the change of the address of the .data section.
 //       You may found the correct address in the aarch64_abs32.s test,
 //       if it is already fixed. Then, update addends accordingly.
 // RUN: ld.lld -z max-page-size=4096 %t.o %t256.o -o %t
 // RUN: llvm-objdump -s --section=.data %t | FileCheck %s --check-prefixes=CHECK,LE
+// RUN: ld.lld -z max-page-size=4096 %t.be.o %t256.be.o -o %t.be
+// RUN: llvm-objdump -s --section=.data %t.be | FileCheck %s --check-prefixes=CHECK,BE
 
 // CHECK: Contents of section .data:
 // 202158: S = 0x100, A = 0x100202057, P = 0x202158
 //         S + A - P = 0xffffffff
 // 20215c: S = 0x100, A = -0x7fdfdfa4, P = 0x20215c
 //         S + A - P = 0x80000000
 // LE-NEXT: 202158 ffffffff 00000080
+// BE-NEXT: 202158 ffffffff 80000000
 
 // RUN: not ld.lld -z max-page-size=4096 %t.o %t255.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW1
 // OVERFLOW1: relocation R_AARCH64_PREL32 out of range: -2147483649 is not in [-2147483648, 4294967295]; references foo
 
 // RUN: not ld.lld -z max-page-size=4096 %t.o %t257.o -o /dev/null 2>&1 | FileCheck %s --check-prefix=OVERFLOW2
 // OVERFLOW2: relocation R_AARCH64_PREL32 out of range: 4294967296 is not in [-2147483648, 4294967295]; references foo

lld/test/ELF/emulation-aarch64.s

 # REQUIRES: aarch64
 # RUN: llvm-mc -filetype=obj -triple=aarch64 %s -o %t.o
 # RUN: ld.lld %t.o -o %t
 # RUN: llvm-readobj --file-headers %t | FileCheck --check-prefixes=AARCH64,LE %s
 # RUN: ld.lld -m aarch64linux %t.o -o %t1
 # RUN: llvm-readobj --file-headers %t1 | FileCheck --check-prefixes=AARCH64,LE %s
 # RUN: ld.lld -m aarch64elf %t.o -o %t2
 # RUN: llvm-readobj --file-headers %t2 | FileCheck --check-prefixes=AARCH64,LE %s
 # RUN: echo 'OUTPUT_FORMAT(elf64-littleaarch64)' > %t.script
 # RUN: ld.lld %t.script %t.o -o %t3
 # RUN: llvm-readobj --file-headers %t3 | FileCheck --check-prefixes=AARCH64,LE %s
 # RUN: ld.lld -m aarch64_elf64_le_vec %t.o -o %taosp
 # RUN: llvm-readobj --file-headers %taosp | FileCheck --check-prefixes=AARCH64,LE %s
 
+# RUN: llvm-mc -filetype=obj -triple=aarch64_be %s -o %t.be.o
+# RUN: ld.lld %t.be.o -o %t
+# RUN: llvm-readobj --file-headers %t | FileCheck --check-prefixes=AARCH64,BE %s
+# RUN: ld.lld -m aarch64linuxb %t.be.o -o %t1.be
+# RUN: llvm-readobj --file-headers %t1.be | FileCheck --check-prefixes=AARCH64,BE %s
+# RUN: ld.lld -m aarch64elfb %t.be.o -o %t2.be
+# RUN: llvm-readobj --file-headers %t2.be | FileCheck --check-prefixes=AARCH64,BE %s
+# RUN: echo 'OUTPUT_FORMAT(elf64-bigaarch64)' > %t.script
+# RUN: ld.lld %t.script %t.be.o -o %t3.be
+# RUN: llvm-readobj --file-headers %t3.be | FileCheck --check-prefixes=AARCH64,BE %s
+
 # AARCH64:      ElfHeader {
 # AARCH64-NEXT:   Ident {
 # AARCH64-NEXT:     Magic: (7F 45 4C 46)
 # AARCH64-NEXT:     Class: 64-bit (0x2)
 # LE-NEXT:          DataEncoding: LittleEndian (0x1)
+# BE-NEXT:          DataEncoding: BigEndian (0x2)
 # AARCH64-NEXT:     FileVersion: 1
 # AARCH64-NEXT:     OS/ABI: SystemV (0x0)
 # AARCH64-NEXT:     ABIVersion: 0
 # AARCH64-NEXT:     Unused: (00 00 00 00 00 00 00)
 # AARCH64-NEXT:   }
 # AARCH64-NEXT:   Type: Executable (0x2)
 # AARCH64-NEXT:   Machine: EM_AARCH64 (0xB7)
 # AARCH64-NEXT:   Version: 1