[PPC64] Add TLS local dynamic to local exec relaxation

[syzaara]

Summary

This patch adds the target call back relaxTlsLdToLe to support TLS relaxation from local dynamic to local exec model.

[ruiu]

lld/ELF/Arch/PPC64.cpp

178–189 ↗

(On Diff #152744)

I'm little confused by this part of the comment because these instructions are not really related to this function. Maybe just explaining the first four instructions is better?

198 ↗

(On Diff #152744)

What is this?

200 ↗

(On Diff #152744)

nit: you don't need {} after case if you don't define any variable inside it, as {} is added to introduce a new local variable scope.

225 ↗

(On Diff #152744)

ditto

[sfertile]

lld/ELF/Arch/PPC64.cpp

185 ↗

(On Diff #152939)

I think this comment should be // addis r3, r13, 0

lld/ELF/Relocations.cpp

210 ↗

(On Diff #152939)

At this point haven't we already returned if the RelExpr is a R_TLSLD_HINT? (likewise for R_TLSLD_GOT_FROM_END and R_TLSLD_PC).

lld/ELF/Relocations.h

47 ↗

(On Diff #152939)

This should be put in sorted order.

[ruiu]

LGTM

lld/trunk/ELF/Arch/PPC64.cpp

public:
   void relocateOne(uint8_t *Loc, RelType Type, uint64_t Val) const override;
   void writeGotHeader(uint8_t *Buf) const override;
   bool needsThunk(RelExpr Expr, RelType Type, const InputFile *File,
                   uint64_t BranchAddr, const Symbol &S) const override;
   RelExpr adjustRelaxExpr(RelType Type, const uint8_t *Data,
                           RelExpr Expr) const override;
   void relaxTlsGdToIe(uint8_t *Loc, RelType Type, uint64_t Val) const override;
   void relaxTlsGdToLe(uint8_t *Loc, RelType Type, uint64_t Val) const override;
+  void relaxTlsLdToLe(uint8_t *Loc, RelType Type, uint64_t Val) const override;
 };
 } // namespace
 
 // Relocation masks following the #lo(value), #hi(value), #ha(value),
 // #higher(value), #highera(value), #highest(value), and #highesta(value)
 // macros defined in section 4.5.1. Relocation Types of the PPC-elf64abi
 // document.
 static uint16_t lo(uint64_t V) { return V; }
case R_PPC64_TLSGD:
     write32(Loc + 4, 0x38630000); // addi r3, r3
     relocateOne(Loc + 4 + EndianOffset, R_PPC64_TPREL16_LO, Val);
     break;
   default:
     llvm_unreachable("unsupported relocation for TLS GD to LE relaxation");
   }
 }
 
+
+void PPC64::relaxTlsLdToLe(uint8_t *Loc, RelType Type, uint64_t Val) const {
+  // Reference: 3.7.4.3 of the 64-bit ELF V2 abi supplement.
+  // The local dynamic code sequence for a global `x` will look like:
+  // Instruction                    Relocation                Symbol
+  // addis r3, r2, x@got@tlsld@ha   R_PPC64_GOT_TLSLD16_HA      x
+  // addi  r3, r3, x@got@tlsld@l    R_PPC64_GOT_TLSLD16_LO      x
+  // bl __tls_get_addr(x@tlsgd)     R_PPC64_TLSLD               x
+  //                                R_PPC64_REL24               __tls_get_addr
+  // nop                            None                       None
+
+  // Relaxing to local exec entails converting:
+  // addis r3, r2, x@got@tlsld@ha   into      nop
+  // addi  r3, r3, x@got@tlsld@l    into      addis r3, r13, 0
+  // bl __tls_get_addr(x@tlsgd)     into      nop
+  // nop                            into      addi r3, r3, 4096
+
+  uint32_t EndianOffset = Config->EKind == ELF64BEKind ? 2U : 0U;
+  switch (Type) {
+  case R_PPC64_GOT_TLSLD16_HA:
+    write32(Loc - EndianOffset, 0x60000000); // nop
+    break;
+  case R_PPC64_GOT_TLSLD16_LO:
+    write32(Loc - EndianOffset, 0x3c6d0000); // addis r3, r13, 0
+    break;
+  case R_PPC64_TLSLD:
+    write32(Loc, 0x60000000);     // nop
+    write32(Loc + 4, 0x38631000); // addi r3, r3, 4096
+    break;
+  case R_PPC64_DTPREL16:
+  case R_PPC64_DTPREL16_HA:
+  case R_PPC64_DTPREL16_HI:
+  case R_PPC64_DTPREL16_DS:
+  case R_PPC64_DTPREL16_LO:
+  case R_PPC64_DTPREL16_LO_DS:
+  case R_PPC64_GOT_DTPREL16_HA:
+  case R_PPC64_GOT_DTPREL16_LO_DS:
+  case R_PPC64_GOT_DTPREL16_DS:
+  case R_PPC64_GOT_DTPREL16_HI:
+    relocateOne(Loc, Type, Val);
+    break;
+  default:
+    llvm_unreachable("unsupported relocation for TLS LD to LE relaxation");
+  }
+}
+
 RelExpr PPC64::getRelExpr(RelType Type, const Symbol &S,
                           const uint8_t *Loc) const {
   switch (Type) {
   case R_PPC64_TOC16:
   case R_PPC64_TOC16_DS:
   case R_PPC64_TOC16_HA:
   case R_PPC64_TOC16_HI:
   case R_PPC64_TOC16_LO:
RelExpr PPC64::getRelExpr(RelType Type, const Symbol &S,
   case R_PPC64_DTPREL16_HIGHESTA:
   case R_PPC64_DTPREL16_LO:
   case R_PPC64_DTPREL16_LO_DS:
   case R_PPC64_DTPREL64:
     return R_ABS;
   case R_PPC64_TLSGD:
     return R_TLSDESC_CALL;
   case R_PPC64_TLSLD:
+    return R_TLSLD_HINT;
   case R_PPC64_TLS:
     return R_HINT;
   default:
     return R_ABS;
   }
 }
 
 void PPC64::writeGotHeader(uint8_t *Buf) const {
bool PPC64::needsThunk(RelExpr Expr, RelType Type, const InputFile *File,
   // a call stub.
   return Type == R_PPC64_REL24 && S.isInPlt();
 }
 
 RelExpr PPC64::adjustRelaxExpr(RelType Type, const uint8_t *Data,
                                RelExpr Expr) const {
   if (Expr == R_RELAX_TLS_GD_TO_IE)
     return R_RELAX_TLS_GD_TO_IE_GOT_OFF;
+  if (Expr == R_RELAX_TLS_LD_TO_LE)
+    return R_RELAX_TLS_LD_TO_LE_ABS;
   return Expr;
 }
 
 // Reference: 3.7.4.1 of the 64-bit ELF V2 abi supplement.
 // The general dynamic code sequence for a global `x` uses 4 instructions.
 // Instruction                    Relocation                Symbol
 // addis r3, r2, x@got@tlsgd@ha   R_PPC64_GOT_TLSGD16_HA      x
 // addi  r3, r3, x@got@tlsgd@l    R_PPC64_GOT_TLSGD16_LO      x

lld/trunk/ELF/InputSection.cpp

 }
 
 static uint64_t getRelocTargetVA(const InputFile *File, RelType Type, int64_t A,
                                  uint64_t P, const Symbol &Sym, RelExpr Expr) {
   switch (Expr) {
   case R_INVALID:
     return 0;
   case R_ABS:
+  case R_RELAX_TLS_LD_TO_LE_ABS:
   case R_RELAX_GOT_PC_NOPIC:
     return Sym.getVA(A);
   case R_ADDEND:
     return A;
   case R_ARM_SBREL:
     return Sym.getVA(A) - getARMStaticBase(Sym);
   case R_GOT:
   case R_RELAX_TLS_GD_TO_IE_ABS:
static uint64_t getRelocTargetVA(const InputFile *File, RelType Type, int64_t A,
   case R_RELAX_TLS_GD_TO_IE_PAGE_PC:
     return getAArch64Page(Sym.getGotVA() + A) - getAArch64Page(P);
   case R_GOT_PC:
   case R_RELAX_TLS_GD_TO_IE:
     return Sym.getGotVA() + A - P;
   case R_HINT:
   case R_NONE:
   case R_TLSDESC_CALL:
+  case R_TLSLD_HINT:
     llvm_unreachable("cannot relocate hint relocs");
   case R_MIPS_GOTREL:
     return Sym.getVA(A) - InX::MipsGot->getGp(File);
   case R_MIPS_GOT_GP:
     return InX::MipsGot->getGp(File) + A;
   case R_MIPS_GOT_GP_PC: {
     // R_MIPS_LO16 expression has R_MIPS_GOT_GP_PC type iif the target
     // is _gp_disp symbol. In that case we should use the following
for (const Relocation &Rel : Relocations) {
     case R_RELAX_GOT_PC:
     case R_RELAX_GOT_PC_NOPIC:
       Target->relaxGot(BufLoc, TargetVA);
       break;
     case R_RELAX_TLS_IE_TO_LE:
       Target->relaxTlsIeToLe(BufLoc, Type, TargetVA);
       break;
     case R_RELAX_TLS_LD_TO_LE:
+    case R_RELAX_TLS_LD_TO_LE_ABS:
       Target->relaxTlsLdToLe(BufLoc, Type, TargetVA);
       break;
     case R_RELAX_TLS_GD_TO_LE:
     case R_RELAX_TLS_GD_TO_LE_NEG:
       Target->relaxTlsGdToLe(BufLoc, Type, TargetVA);
       break;
     case R_RELAX_TLS_GD_TO_IE:
     case R_RELAX_TLS_GD_TO_IE_ABS:

lld/trunk/ELF/Relocations.h

enum RelExpr {
   R_RELAX_TLS_GD_TO_IE_ABS,
   R_RELAX_TLS_GD_TO_IE_END,
   R_RELAX_TLS_GD_TO_IE_GOT_OFF,
   R_RELAX_TLS_GD_TO_IE_PAGE_PC,
   R_RELAX_TLS_GD_TO_LE,
   R_RELAX_TLS_GD_TO_LE_NEG,
   R_RELAX_TLS_IE_TO_LE,
   R_RELAX_TLS_LD_TO_LE,
+  R_RELAX_TLS_LD_TO_LE_ABS,
   R_SIZE,
   R_TLS,
   R_TLSDESC,
   R_TLSDESC_CALL,
   R_TLSDESC_PAGE,
   R_TLSGD_GOT,
   R_TLSGD_GOT_FROM_END,
   R_TLSGD_PC,
   R_TLSLD_GOT,
   R_TLSLD_GOT_FROM_END,
   R_TLSLD_GOT_OFF,
+  R_TLSLD_HINT,
   R_TLSLD_PC,
 };
 
 // Build a bitmask with one bit set for each RelExpr.
 //
 // Constexpr function arguments can't be used in static asserts, so we
 // use template arguments to build the mask.
 // But function template partial specializations don't exist (needed

lld/trunk/ELF/Relocations.cpp

if (InX::Got->addDynTlsEntry(Sym)) {
       InX::RelaDyn->addReloc(
           {Target->TlsDescRel, InX::Got, Off, !Sym.IsPreemptible, &Sym, 0});
     }
     if (Expr != R_TLSDESC_CALL)
       C.Relocations.push_back({Expr, Type, Offset, Addend, &Sym});
     return 1;
   }
 
-  if (isRelExprOneOf<R_TLSLD_GOT, R_TLSLD_GOT_FROM_END, R_TLSLD_PC>(Expr)) {
+  if (isRelExprOneOf<R_TLSLD_GOT, R_TLSLD_GOT_FROM_END, R_TLSLD_PC,
+                     R_TLSLD_HINT>(Expr)) {
     // Local-Dynamic relocs can be relaxed to Local-Exec.
     if (!Config->Shared) {
       C.Relocations.push_back(
-          {R_RELAX_TLS_LD_TO_LE, Type, Offset, Addend, &Sym});
-      return 2;
+          {Target->adjustRelaxExpr(Type, nullptr, R_RELAX_TLS_LD_TO_LE), Type,
+           Offset, Addend, &Sym});
+      return Target->TlsGdRelaxSkip;
     }
+    if (Expr == R_TLSLD_HINT)
+      return 1;
     if (InX::Got->addTlsIndex())
       InX::RelaDyn->addReloc(Target->TlsModuleIndexRel, InX::Got,
                              InX::Got->getTlsIndexOff(), nullptr);
     C.Relocations.push_back({Expr, Type, Offset, Addend, &Sym});
     return 1;
   }
 
   // Local-Dynamic relocs can be relaxed to Local-Exec.
-  if (isRelExprOneOf<R_ABS, R_TLSLD_GOT_FROM_END, R_TLSLD_PC>(Expr) &&
-      !Config->Shared) {
-    C.Relocations.push_back({R_RELAX_TLS_LD_TO_LE, Type, Offset, Addend, &Sym});
+  if (Expr == R_ABS && !Config->Shared) {
+    C.Relocations.push_back(
+        {Target->adjustRelaxExpr(Type, nullptr, R_RELAX_TLS_LD_TO_LE), Type,
+         Offset, Addend, &Sym});
     return 1;
   }
 
   // Local-Dynamic sequence where offset of tls variable relative to dynamic
   // thread pointer is stored in the got.
   if (Expr == R_TLSLD_GOT_OFF) {
     // Local-Dynamic relocs can be relaxed to local-exec
     if (!Config->Shared) {

 static bool isStaticLinkTimeConstant(RelExpr E, RelType Type, const Symbol &Sym,
                                      InputSectionBase &S, uint64_t RelOff) {
   // These expressions always compute a constant
   if (isRelExprOneOf<
           R_GOT_FROM_END, R_GOT_OFF, R_TLSLD_GOT_OFF, R_MIPS_GOT_LOCAL_PAGE,
           R_MIPS_GOTREL, R_MIPS_GOT_OFF, R_MIPS_GOT_OFF32, R_MIPS_GOT_GP_PC,
           R_MIPS_TLSGD, R_GOT_PAGE_PC, R_GOT_PC, R_GOTONLY_PC,
           R_GOTONLY_PC_FROM_END, R_PLT_PC, R_TLSGD_GOT, R_TLSGD_GOT_FROM_END,
-          R_TLSGD_PC, R_PPC_CALL_PLT, R_TLSDESC_CALL, R_TLSDESC_PAGE, R_HINT>(
-          E))
+          R_TLSGD_PC, R_PPC_CALL_PLT, R_TLSDESC_CALL, R_TLSDESC_PAGE, R_HINT,
+          R_TLSLD_HINT>(E))
     return true;
 
   // These never do, except if the entire file is position dependent or if
   // only the low bits are used.
   if (E == R_GOT || E == R_PLT || E == R_TLSDESC)
     return Target->usesOnlyLowPageBits(Type) || !Config->Pic;
 
   if (Sym.IsPreemptible)

lld/trunk/test/ELF/ppc64-tls-ld-le.s

+// REQUIRES: ppc
+
+// RUN: llvm-mc -filetype=obj -triple=powerpc64le-unknown-linux %s -o %t.o
+// RUN: llvm-readelf -relocations --wide %t.o | FileCheck --check-prefix=InputRelocs %s
+// RUN: ld.lld  %t.o -o %t
+// RUN: llvm-objdump -D %t | FileCheck --check-prefix=Dis %s
+// RUN: llvm-readelf -relocations --wide %t | FileCheck --check-prefix=OutputRelocs %s
+
+// RUN: llvm-mc -filetype=obj -triple=powerpc64-unknown-linux %s -o %t.o
+// RUN: llvm-readelf -relocations --wide %t.o | FileCheck --check-prefix=InputRelocs %s
+// RUN: ld.lld  %t.o -o %t
+// RUN: llvm-objdump -D %t | FileCheck --check-prefix=Dis %s
+// RUN: llvm-readelf -relocations --wide %t | FileCheck --check-prefix=OutputRelocs %s
+
+  .text
+  .abiversion 2
+  .globl  _start                    # -- Begin function _start
+  .p2align  4
+  .type _start,@function
+_start:                                   # @_start
+.Lfunc_begin0:
+.Lfunc_gep0:
+  addis 2, 12, .TOC.-.Lfunc_gep0@ha
+  addi 2, 2, .TOC.-.Lfunc_gep0@l
+.Lfunc_lep0:
+  .localentry _start, .Lfunc_lep0-.Lfunc_gep0
+# %bb.0:                                # %entry
+  mflr 0
+  std 31, -8(1)
+  std 0, 16(1)
+  stdu 1, -64(1)
+  mr 31, 1
+  std 30, 48(31)                  # 8-byte Folded Spill
+  li 3, 0
+  stw 3, 44(31)
+  addis 3, 2, a@got@tlsld@ha
+  addi 3, 3, a@got@tlsld@l
+  bl __tls_get_addr(a@tlsld)
+  nop
+  addis 3, 3, a@dtprel@ha
+  addi 3, 3, a@dtprel@l
+  lwz 30, 0(3)
+  extsw 3, 30
+  ld 30, 48(31)                   # 8-byte Folded Reload
+  addi 1, 1, 64
+  ld 0, 16(1)
+  ld 31, -8(1)
+  mtlr 0
+  blr
+  .long 0
+  .quad 0
+.Lfunc_end0:
+  .size _start, .Lfunc_end0-.Lfunc_begin0
+                                        # -- End function
+.globl __tls_get_addr
+.type __tls_get_addr,@function
+__tls_get_addr:
+  .type a,@object               # @a
+  .section  .tdata,"awT",@progbits
+  .p2align  2
+a:
+  .long 2                       # 0x2
+  .size a, 4
+
+// Verify that the input has local-dynamic tls relocation types
+// InputRelocs:  Relocation section '.rela.text'
+// InputRelocs: R_PPC64_GOT_TLSLD16_HA  {{0+}}  a + 0
+// InputRelocs: R_PPC64_GOT_TLSLD16_LO  {{0+}}  a + 0
+// InputRelocs: R_PPC64_TLSLD           {{0+}}  a + 0
+
+// Verify that the local-dynamic sequence is relaxed to local exec.
+// Dis: _start:
+// Dis: nop
+// Dis: addis 3, 13, 0
+// Dis: nop
+// Dis: addi 3, 3, 4096
+
+// #ha(a@dtprel) --> (0x0 -0x8000 + 0x8000) >> 16 = 0
+// #lo(a@dtprel) --> (0x0 -0x8000) = -0x8000 = -32768
+// Dis: addis 3, 3, 0
+// Dis: addi 3, 3, -32768
+
+// Verify that no local-dynamic relocations exist for the dynamic linker.
+// OutputRelocs-NOT: R_PPC64_DTPMOD64