[lld] r273238 - Refactor X86_64TargetInfo::relaxTlsIeToLe.

Sean Silva via llvm-commits llvm-commits at lists.llvm.org
Mon Jun 20 21:55:13 PDT 2016


I think this broke the bots:
http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/15038/steps/build/logs/stdio



On Mon, Jun 20, 2016 at 8:42 PM, Rui Ueyama via llvm-commits <
llvm-commits at lists.llvm.org> wrote:

> Author: ruiu
> Date: Mon Jun 20 22:42:32 2016
> New Revision: 273238
>
> URL: http://llvm.org/viewvc/llvm-project?rev=273238&view=rev
> Log:
> Refactor X86_64TargetInfo::relaxTlsIeToLe.
>
> This patch is to rewrite the function with a table-lookup-ish approach
> so that it can read as a series of "convert this pattern to this" pattern.
>
> Modified:
>     lld/trunk/ELF/Target.cpp
>     lld/trunk/test/ELF/tls-opt.s
>
> Modified: lld/trunk/ELF/Target.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/lld/trunk/ELF/Target.cpp?rev=273238&r1=273237&r2=273238&view=diff
>
> ==============================================================================
> --- lld/trunk/ELF/Target.cpp (original)
> +++ lld/trunk/ELF/Target.cpp Mon Jun 20 22:42:32 2016
> @@ -677,34 +677,40 @@ void X86_64TargetInfo::relaxTlsGdToIe(ui
>  // R_X86_64_TPOFF32 so that it does not use GOT.
>  void X86_64TargetInfo::relaxTlsIeToLe(uint8_t *Loc, uint32_t Type,
>                                        uint64_t Val) const {
> -  // Ulrich's document section 6.5 says that @gottpoff(%rip) must be
> -  // used in MOVQ or ADDQ instructions only.
> -  // "MOVQ foo at GOTTPOFF(%RIP), %REG" is transformed to "MOVQ $foo, %REG".
> -  // "ADDQ foo at GOTTPOFF(%RIP), %REG" is transformed to "LEAQ foo(%REG),
> %REG"
> -  // (if the register is not RSP/R12) or "ADDQ $foo, %RSP".
> -  // Opcodes info can be found at http://ref.x86asm.net/coder64.html#x48.
> -  uint8_t *Prefix = Loc - 3;
> -  uint8_t *Inst = Loc - 2;
> -  uint8_t *RegSlot = Loc - 1;
> +  uint8_t *Inst = Loc - 3;
>    uint8_t Reg = Loc[-1] >> 3;
> -  bool IsMov = *Inst == 0x8b;
> -  bool RspAdd = !IsMov && Reg == 4;
> +  uint8_t RegSlot = Loc - 1;
>
> -  // r12 and rsp registers requires special handling.
> -  // Problem is that for other registers, for example leaq
> 0xXXXXXXXX(%r11),%r11
> -  // result out is 7 bytes: 4d 8d 9b XX XX XX XX,
> -  // but leaq 0xXXXXXXXX(%r12),%r12 is 8 bytes: 4d 8d a4 24 XX XX XX XX.
> -  // The same true for rsp. So we convert to addq for them, saving 1 byte
> that
> -  // we dont have.
> -  if (RspAdd)
> -    *Inst = 0x81;
> -  else
> -    *Inst = IsMov ? 0xc7 : 0x8d;
> -  if (*Prefix == 0x4c)
> -    *Prefix = (IsMov || RspAdd) ? 0x49 : 0x4d;
> -  *RegSlot = (IsMov || RspAdd) ? (0xc0 | Reg) : (0x80 | Reg | (Reg << 3));
> -  // The original code used a pc relative relocation and so we have to
> -  // compensate for the -4 in had in the addend.
> +  // Note that LEA with RSP or R12 is converted to ADD instead of LEA
> +  // because LEA with these registers needs 4 bytes to encode and thus
> +  // wouldn't fit the space.
> +
> +  if (memcmp(Inst, "\x48\x03\x25", 3) == 0) {
> +    // "addq foo at gottpoff(%rip),%rsp" -> "addq $foo,%rsp"
> +    memcpy(Inst, "\x48\x81\xc4", 3);
> +  } else if (memcmp(Inst, "\x4c\x03\x25", 3) == 0) {
> +    // "addq foo at gottpoff(%rip),%r12" -> "addq $foo,%r12"
> +    memcpy(Inst, "\x49\x81\xc4", 3);
> +  } else if (memcmp(Inst, "\x4c\x03", 2) == 0) {
> +    // "addq foo at gottpoff(%rip),%r[8-15]" -> "leaq
> foo(%r[8-15]),%r[8-15]"
> +    memcpy(Inst, "\x4d\x8d", 2);
> +    *RegSlot = 0x80 | (Reg << 3) | Reg;
> +  } else if (memcmp(Inst, "\x48\x03", 2) == 0) {
> +    // "addq foo at gottpoff(%rip),%reg -> "leaq foo(%reg),%reg"
> +    memcpy(Inst, "\x48\x8d", 2);
> +    *RegSlot = 0x80 | (Reg << 3) | Reg;
> +  } else if (memcmp(Inst, "\x4c\x8b", 2) == 0) {
> +    // "movq foo at gottpoff(%rip),%r[8-15]" -> "movq $foo,%r[8-15]"
> +    memcpy(Inst, "\x49\xc7", 2);
> +    *RegSlot = 0xc0 | Reg;
> +  } else if (memcmp(Inst, "\x48\x8b", 2) == 0) {
> +    // "movq foo at gottpoff(%rip),%reg" -> "movq $foo,%reg"
> +    memcpy(Inst, "\x48\xc7", 2);
> +    *RegSlot = 0xc0 | Reg;
> +  }
> +
> +  // The original code used a PC relative relocation.
> +  // Need to compensate for the -4 it had in the addend.
>    relocateOne(Loc, R_X86_64_TPOFF32, Val + 4);
>  }
>
>
> Modified: lld/trunk/test/ELF/tls-opt.s
> URL:
> http://llvm.org/viewvc/llvm-project/lld/trunk/test/ELF/tls-opt.s?rev=273238&r1=273237&r2=273238&view=diff
>
> ==============================================================================
> --- lld/trunk/test/ELF/tls-opt.s (original)
> +++ lld/trunk/test/ELF/tls-opt.s Mon Jun 20 22:42:32 2016
> @@ -20,11 +20,9 @@
>  // DISASM-NEXT: 1103f: 4d 8d bf fc ff ff ff leaq -4(%r15), %r15
>  // DISASM-NEXT: 11046: 48 81 c4 fc ff ff ff addq $-4, %rsp
>  // DISASM-NEXT: 1104d: 49 81 c4 fc ff ff ff addq $-4, %r12
> -// Corrupred output:
> -// DISASM-NEXT: 11054: 48 8d 80 f8 ff ff ff leaq -8(%rax), %rax
> -// DISASM-NEXT: 1105b: 48 d1 81 c4 fc ff ff rolq -828(%rcx)
> -// DISASM-NEXT: 11062: ff 48 d1 decl -47(%rax)
> -// DISASM-NEXT: 11065: 81 c4 fc ff ff ff addl $4294967292, %esp
> +// DISASM-NEXT: 11054: 48 87 05 f8 ff ff ff    xchgq   %rax, -8(%rip)
> +// DISASM-NEXT: 1105b: 48 d1 24 25 fc ff ff ff         shlq    -4
> +// DISASM-NEXT: 11063: 48 d1 04 25 fc ff ff ff         rolq    -4
>  // LD to LE:
>  // DISASM-NEXT: 1106b: 66 66 66 64 48 8b 04 25 00 00 00 00 movq %fs:0,
> %rax
>  // DISASM-NEXT: 11077: 48 8d 88 f8 ff ff ff                leaq -8(%rax),
> %rcx
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160620/bd66860e/attachment.html>


More information about the llvm-commits mailing list