[TargetMachine] Don't imply dso_local on function declarations in Rel…

…oc::Static model for ELF/wasm

clang/lib/CodeGen/CodeGenModule sets dso_local on applicable function declarations,
we don't need to duplicate the work in TargetMachine:shouldAssumeDSOLocal.
(Actually the long-term goal (started by r324535) is to drop TargetMachine::shouldAssumeDSOLocal.)

By not implying dso_local, we will respect dso_local/dso_preemptable specifiers
set by the frontend. This allows the proposed -fno-direct-access-external-data
option to work with -fno-pic and prevent a canonical PLT entry (SHN_UNDEF with non-zero st_value)
when taking the address of a function symbol.

This patch should be NFC in terms of the Clang emitted assembly because the case
we don't set dso_local is a case Clang sets dso_local. However, some tests don't
set dso_local on some function declarations and expose some differences. Most
tests have been fixed to be more robust in the previous commit.

llvm/lib/Target/TargetMachine.cpp

@@ -101,10 +101,9 @@ bool TargetMachine::shouldAssumeDSOLocal(const Module &M,
   // dso_preemptable.  At this point in time, the various IR producers
   // have not been transitioned to always produce a dso_local when it
   // is possible to do so.
-  // In the case of ExternalSymbolSDNode, GV is null and there is nowhere to put
-  // dso_local. Returning false for those will produce worse code in some
-  // architectures. For example, on x86 the caller has to set ebx before calling
-  // a plt.
+  // In the case of ExternalSymbolSDNode, GV is null and we should just return
+  // false. However, COFF currently relies on this to be true
+  //
   // As a result we still have some logic in here to improve the quality of the
   // generated code.
   // FIXME: Add a module level metadata for whether intrinsics should be assumed
@@ -191,12 +190,8 @@ bool TargetMachine::shouldAssumeDSOLocal(const Module &M,
     // drop the if block entirely and respect dso_local/dso_preemptable
     // specifiers set by the frontend.
     if (RM == Reloc::Static) {
-      // We currently respect dso_local/dso_preemptable specifiers for
-      // variables.
-      if (F)
-        return true;
       // TODO Remove the special case for x86-32.
-      if (Arch == Triple::x86 && !GV->isThreadLocal())
+      if (Arch == Triple::x86 && !F && !GV->isThreadLocal())
         return true;
     }
   } else if (TT.isOSBinFormatELF()) {

llvm/lib/Target/X86/X86Subtarget.cpp

@@ -166,6 +166,10 @@ unsigned char X86Subtarget::classifyGlobalReference(const GlobalValue *GV,
     return X86II::MO_DARWIN_NONLAZY_PIC_BASE;
   }
 
+  // 32-bit ELF references GlobalAddress directly in static relocation model.
+  // We cannot use MO_GOT because EBX may not be set up.
+  if (TM.getRelocationModel() == Reloc::Static)
+    return X86II::MO_NO_FLAG;
   return X86II::MO_GOT;
 }
 

llvm/test/CodeGen/AArch64/tiny_model.ll

@@ -370,12 +370,12 @@ declare void @func(...)
 define i8* @externfuncaddr() {
 ; CHECK-LABEL: externfuncaddr:
 ; CHECK:       // %bb.0: // %entry
-; CHECK-NEXT:    adr x0, func
+; CHECK-NEXT:    ldr x0, :got:func
 ; CHECK-NEXT:    ret
 ;
 ; CHECK-GLOBISEL-LABEL: externfuncaddr:
 ; CHECK-GLOBISEL:       // %bb.0: // %entry
-; CHECK-GLOBISEL-NEXT:    adr x0, func
+; CHECK-GLOBISEL-NEXT:    ldr x0, :got:func
 ; CHECK-GLOBISEL-NEXT:    ret
 ;
 ; CHECK-PIC-LABEL: externfuncaddr:

llvm/test/CodeGen/ARM/fast-isel-call.ll

@@ -109,8 +109,9 @@ entry:
 ; ARM-LONG-MACHO: {{(movt [[R1]], :upper16:L_bar\$non_lazy_ptr)?}}
 ; ARM-LONG-MACHO: ldr [[R:r[0-9]+]], {{\[}}[[R1]]]
 
-; ARM-LONG-ELF: movw [[R:l?r[0-9]*]], :lower16:bar
-; ARM-LONG-ELF: {{(movt [[R]], :upper16:L_bar\$non_lazy_ptr)?}}
+; ARM-LONG-ELF: movw [[R1:r[0-9]*]], :lower16:bar
+; ARM-LONG-ELF: movt [[R1]], :upper16:bar
+; ARM-LONG-ELF: ldr [[R:r[0-9]+]], {{\[}}[[R1]]]
 
 ; ARM-LONG: blx [[R]]
 ; THUMB: @t10

llvm/test/CodeGen/RISCV/mir-target-flags.ll

@@ -35,7 +35,7 @@ define i32 @caller(i32 %a) nounwind {
 ; RV32-SMALL:      target-flags(riscv-tprel-hi) @t_le
 ; RV32-SMALL-NEXT: target-flags(riscv-tprel-add) @t_le
 ; RV32-SMALL-NEXT: target-flags(riscv-tprel-lo) @t_le
-; RV32-SMALL:      target-flags(riscv-call) @callee
+; RV32-SMALL:      target-flags(riscv-plt) @callee
 ;
 ; RV32-MED-LABEL: name: caller
 ; RV32-MED:      target-flags(riscv-got-hi) @g_e

llvm/test/CodeGen/X86/abi-isel.ll

@@ -10676,26 +10676,26 @@ define void @lcallee() nounwind {
 ; LINUX-64-STATIC-LABEL: lcallee:
 ; LINUX-64-STATIC:       # %bb.0: # %entry
 ; LINUX-64-STATIC-NEXT:    pushq %rax
-; LINUX-64-STATIC-NEXT:    callq x
-; LINUX-64-STATIC-NEXT:    callq x
-; LINUX-64-STATIC-NEXT:    callq x
-; LINUX-64-STATIC-NEXT:    callq x
-; LINUX-64-STATIC-NEXT:    callq x
-; LINUX-64-STATIC-NEXT:    callq x
-; LINUX-64-STATIC-NEXT:    callq x
+; LINUX-64-STATIC-NEXT:    callq x@PLT
+; LINUX-64-STATIC-NEXT:    callq x@PLT
+; LINUX-64-STATIC-NEXT:    callq x@PLT
+; LINUX-64-STATIC-NEXT:    callq x@PLT
+; LINUX-64-STATIC-NEXT:    callq x@PLT
+; LINUX-64-STATIC-NEXT:    callq x@PLT
+; LINUX-64-STATIC-NEXT:    callq x@PLT
 ; LINUX-64-STATIC-NEXT:    popq %rax
 ; LINUX-64-STATIC-NEXT:    retq
 ;
 ; LINUX-32-STATIC-LABEL: lcallee:
 ; LINUX-32-STATIC:       # %bb.0: # %entry
 ; LINUX-32-STATIC-NEXT:    subl $12, %esp
-; LINUX-32-STATIC-NEXT:    calll x
-; LINUX-32-STATIC-NEXT:    calll x
-; LINUX-32-STATIC-NEXT:    calll x
-; LINUX-32-STATIC-NEXT:    calll x
-; LINUX-32-STATIC-NEXT:    calll x
-; LINUX-32-STATIC-NEXT:    calll x
-; LINUX-32-STATIC-NEXT:    calll x
+; LINUX-32-STATIC-NEXT:    calll x@PLT
+; LINUX-32-STATIC-NEXT:    calll x@PLT
+; LINUX-32-STATIC-NEXT:    calll x@PLT
+; LINUX-32-STATIC-NEXT:    calll x@PLT
+; LINUX-32-STATIC-NEXT:    calll x@PLT
+; LINUX-32-STATIC-NEXT:    calll x@PLT
+; LINUX-32-STATIC-NEXT:    calll x@PLT
 ; LINUX-32-STATIC-NEXT:    addl $12, %esp
 ; LINUX-32-STATIC-NEXT:    retl
 ;
@@ -10827,26 +10827,26 @@ define internal void @dcallee() nounwind {
 ; LINUX-64-STATIC-LABEL: dcallee:
 ; LINUX-64-STATIC:       # %bb.0: # %entry
 ; LINUX-64-STATIC-NEXT:    pushq %rax
-; LINUX-64-STATIC-NEXT:    callq y
-; LINUX-64-STATIC-NEXT:    callq y
-; LINUX-64-STATIC-NEXT:    callq y
-; LINUX-64-STATIC-NEXT:    callq y
-; LINUX-64-STATIC-NEXT:    callq y
-; LINUX-64-STATIC-NEXT:    callq y
-; LINUX-64-STATIC-NEXT:    callq y
+; LINUX-64-STATIC-NEXT:    callq y@PLT
+; LINUX-64-STATIC-NEXT:    callq y@PLT
+; LINUX-64-STATIC-NEXT:    callq y@PLT
+; LINUX-64-STATIC-NEXT:    callq y@PLT
+; LINUX-64-STATIC-NEXT:    callq y@PLT
+; LINUX-64-STATIC-NEXT:    callq y@PLT
+; LINUX-64-STATIC-NEXT:    callq y@PLT
 ; LINUX-64-STATIC-NEXT:    popq %rax
 ; LINUX-64-STATIC-NEXT:    retq
 ;
 ; LINUX-32-STATIC-LABEL: dcallee:
 ; LINUX-32-STATIC:       # %bb.0: # %entry
 ; LINUX-32-STATIC-NEXT:    subl $12, %esp
-; LINUX-32-STATIC-NEXT:    calll y
-; LINUX-32-STATIC-NEXT:    calll y
-; LINUX-32-STATIC-NEXT:    calll y
-; LINUX-32-STATIC-NEXT:    calll y
-; LINUX-32-STATIC-NEXT:    calll y
-; LINUX-32-STATIC-NEXT:    calll y
-; LINUX-32-STATIC-NEXT:    calll y
+; LINUX-32-STATIC-NEXT:    calll y@PLT
+; LINUX-32-STATIC-NEXT:    calll y@PLT
+; LINUX-32-STATIC-NEXT:    calll y@PLT
+; LINUX-32-STATIC-NEXT:    calll y@PLT
+; LINUX-32-STATIC-NEXT:    calll y@PLT
+; LINUX-32-STATIC-NEXT:    calll y@PLT
+; LINUX-32-STATIC-NEXT:    calll y@PLT
 ; LINUX-32-STATIC-NEXT:    addl $12, %esp
 ; LINUX-32-STATIC-NEXT:    retl
 ;
@@ -10977,7 +10977,7 @@ declare void @y()
 define void ()* @address() nounwind {
 ; LINUX-64-STATIC-LABEL: address:
 ; LINUX-64-STATIC:       # %bb.0: # %entry
-; LINUX-64-STATIC-NEXT:    movl $callee, %eax
+; LINUX-64-STATIC-NEXT:    movq callee@{{.*}}(%rip), %rax
 ; LINUX-64-STATIC-NEXT:    retq
 ;
 ; LINUX-32-STATIC-LABEL: address:
@@ -11169,16 +11169,16 @@ define void @caller() nounwind {
 ; LINUX-64-STATIC-LABEL: caller:
 ; LINUX-64-STATIC:       # %bb.0: # %entry
 ; LINUX-64-STATIC-NEXT:    pushq %rax
-; LINUX-64-STATIC-NEXT:    callq callee
-; LINUX-64-STATIC-NEXT:    callq callee
+; LINUX-64-STATIC-NEXT:    callq callee@PLT
+; LINUX-64-STATIC-NEXT:    callq callee@PLT
 ; LINUX-64-STATIC-NEXT:    popq %rax
 ; LINUX-64-STATIC-NEXT:    retq
 ;
 ; LINUX-32-STATIC-LABEL: caller:
 ; LINUX-32-STATIC:       # %bb.0: # %entry
 ; LINUX-32-STATIC-NEXT:    subl $12, %esp
-; LINUX-32-STATIC-NEXT:    calll callee
-; LINUX-32-STATIC-NEXT:    calll callee
+; LINUX-32-STATIC-NEXT:    calll callee@PLT
+; LINUX-32-STATIC-NEXT:    calll callee@PLT
 ; LINUX-32-STATIC-NEXT:    addl $12, %esp
 ; LINUX-32-STATIC-NEXT:    retl
 ;
@@ -11451,14 +11451,14 @@ define void @tailcaller() nounwind {
 ; LINUX-64-STATIC-LABEL: tailcaller:
 ; LINUX-64-STATIC:       # %bb.0: # %entry
 ; LINUX-64-STATIC-NEXT:    pushq %rax
-; LINUX-64-STATIC-NEXT:    callq callee
+; LINUX-64-STATIC-NEXT:    callq callee@PLT
 ; LINUX-64-STATIC-NEXT:    popq %rax
 ; LINUX-64-STATIC-NEXT:    retq
 ;
 ; LINUX-32-STATIC-LABEL: tailcaller:
 ; LINUX-32-STATIC:       # %bb.0: # %entry
 ; LINUX-32-STATIC-NEXT:    subl $12, %esp
-; LINUX-32-STATIC-NEXT:    calll callee
+; LINUX-32-STATIC-NEXT:    calll callee@PLT
 ; LINUX-32-STATIC-NEXT:    addl $12, %esp
 ; LINUX-32-STATIC-NEXT:    retl
 ;

llvm/test/CodeGen/X86/absolute-cmp.ll

@@ -20,8 +20,8 @@ define void @foo8(i64 %val) {
 ; NOPIC-NEXT:  # %bb.1: # %t
 ; NOPIC-NEXT:    pushq %rax # encoding: [0x50]
 ; NOPIC-NEXT:    .cfi_def_cfa_offset 16
-; NOPIC-NEXT:    callq f # encoding: [0xe8,A,A,A,A]
-; NOPIC-NEXT:    # fixup A - offset: 1, value: f-4, kind: reloc_branch_4byte_pcrel
+; NOPIC-NEXT:    callq f@PLT # encoding: [0xe8,A,A,A,A]
+; NOPIC-NEXT:    # fixup A - offset: 1, value: f@PLT-4, kind: FK_PCRel_4
 ; NOPIC-NEXT:    popq %rax # encoding: [0x58]
 ; NOPIC-NEXT:    .cfi_def_cfa_offset 8
 ; NOPIC-NEXT:  .LBB0_2: # %f
@@ -63,8 +63,8 @@ define void @foo32(i64 %val) {
 ; NOPIC-NEXT:  # %bb.1: # %t
 ; NOPIC-NEXT:    pushq %rax # encoding: [0x50]
 ; NOPIC-NEXT:    .cfi_def_cfa_offset 16
-; NOPIC-NEXT:    callq f # encoding: [0xe8,A,A,A,A]
-; NOPIC-NEXT:    # fixup A - offset: 1, value: f-4, kind: reloc_branch_4byte_pcrel
+; NOPIC-NEXT:    callq f@PLT # encoding: [0xe8,A,A,A,A]
+; NOPIC-NEXT:    # fixup A - offset: 1, value: f@PLT-4, kind: FK_PCRel_4
 ; NOPIC-NEXT:    popq %rax # encoding: [0x58]
 ; NOPIC-NEXT:    .cfi_def_cfa_offset 8
 ; NOPIC-NEXT:  .LBB1_2: # %f

llvm/test/CodeGen/X86/absolute-constant.ll

@@ -14,7 +14,7 @@ define void @bar(i8* %x) {
 ; CHECK-NEXT:    je .LBB0_1
 ; CHECK-NEXT:  # %bb.2: # %if.then
 ; CHECK-NEXT:    xorl %eax, %eax
-; CHECK-NEXT:    jmp xf # TAILCALL
+; CHECK-NEXT:    jmp xf@PLT # TAILCALL
 ; CHECK-NEXT:  .LBB0_1: # %if.end
 ; CHECK-NEXT:    retq
 ;

llvm/test/CodeGen/X86/code-model-elf.ll

@@ -341,12 +341,12 @@ define dso_local void ()* @lea_global_fn() #0 {
 define dso_local void ()* @lea_extern_fn() #0 {
 ; SMALL-STATIC-LABEL: lea_extern_fn:
 ; SMALL-STATIC:       # %bb.0:
-; SMALL-STATIC-NEXT:    movl $extern_fn, %eax
+; SMALL-STATIC-NEXT:    movq extern_fn@{{.*}}(%rip), %rax
 ; SMALL-STATIC-NEXT:    retq
 ;
 ; MEDIUM-STATIC-LABEL: lea_extern_fn:
 ; MEDIUM-STATIC:       # %bb.0:
-; MEDIUM-STATIC-NEXT:    movabsq $extern_fn, %rax
+; MEDIUM-STATIC-NEXT:    movq extern_fn@{{.*}}(%rip), %rax
 ; MEDIUM-STATIC-NEXT:    retq
 ;
 ; LARGE-STATIC-LABEL: lea_extern_fn:

llvm/test/CodeGen/X86/conditional-tailcall-pgso.ll

@@ -11,24 +11,24 @@ define void @f(i32 %x, i32 %y) !prof !14 {
 ; CHECK32:       # %bb.0: # %entry
 ; CHECK32-NEXT:    movl {{[0-9]+}}(%esp), %eax # encoding: [0x8b,0x44,0x24,0x04]
 ; CHECK32-NEXT:    cmpl {{[0-9]+}}(%esp), %eax # encoding: [0x3b,0x44,0x24,0x08]
-; CHECK32-NEXT:    jne bar # TAILCALL
+; CHECK32-NEXT:    jne bar@PLT # TAILCALL
 ; CHECK32-NEXT:    # encoding: [0x75,A]
-; CHECK32-NEXT:    # fixup A - offset: 1, value: bar-1, kind: FK_PCRel_1
+; CHECK32-NEXT:    # fixup A - offset: 1, value: bar@PLT-1, kind: FK_PCRel_1
 ; CHECK32-NEXT:  # %bb.1: # %bb1
-; CHECK32-NEXT:    jmp foo # TAILCALL
+; CHECK32-NEXT:    jmp foo@PLT # TAILCALL
 ; CHECK32-NEXT:    # encoding: [0xeb,A]
-; CHECK32-NEXT:    # fixup A - offset: 1, value: foo-1, kind: FK_PCRel_1
+; CHECK32-NEXT:    # fixup A - offset: 1, value: foo@PLT-1, kind: FK_PCRel_1
 ;
 ; CHECK64-LABEL: f:
 ; CHECK64:       # %bb.0: # %entry
 ; CHECK64-NEXT:    cmpl %esi, %edi # encoding: [0x39,0xf7]
-; CHECK64-NEXT:    jne bar # TAILCALL
+; CHECK64-NEXT:    jne bar@PLT # TAILCALL
 ; CHECK64-NEXT:    # encoding: [0x75,A]
-; CHECK64-NEXT:    # fixup A - offset: 1, value: bar-1, kind: FK_PCRel_1
+; CHECK64-NEXT:    # fixup A - offset: 1, value: bar@PLT-1, kind: FK_PCRel_1
 ; CHECK64-NEXT:  # %bb.1: # %bb1
-; CHECK64-NEXT:    jmp foo # TAILCALL
+; CHECK64-NEXT:    jmp foo@PLT # TAILCALL
 ; CHECK64-NEXT:    # encoding: [0xeb,A]
-; CHECK64-NEXT:    # fixup A - offset: 1, value: foo-1, kind: FK_PCRel_1
+; CHECK64-NEXT:    # fixup A - offset: 1, value: foo@PLT-1, kind: FK_PCRel_1
 ;
 ; WIN64-LABEL: f:
 ; WIN64:       # %bb.0: # %entry
@@ -68,16 +68,16 @@ define void @f_non_leaf(i32 %x, i32 %y) !prof !14 {
 ; CHECK32-NEXT:  # %bb.1: # %bb1
 ; CHECK32-NEXT:    popl %ebx # encoding: [0x5b]
 ; CHECK32-NEXT:    .cfi_def_cfa_offset 4
-; CHECK32-NEXT:    jmp foo # TAILCALL
+; CHECK32-NEXT:    jmp foo@PLT # TAILCALL
 ; CHECK32-NEXT:    # encoding: [0xeb,A]
-; CHECK32-NEXT:    # fixup A - offset: 1, value: foo-1, kind: FK_PCRel_1
+; CHECK32-NEXT:    # fixup A - offset: 1, value: foo@PLT-1, kind: FK_PCRel_1
 ; CHECK32-NEXT:  .LBB1_2: # %bb2
 ; CHECK32-NEXT:    .cfi_def_cfa_offset 8
 ; CHECK32-NEXT:    popl %ebx # encoding: [0x5b]
 ; CHECK32-NEXT:    .cfi_def_cfa_offset 4
-; CHECK32-NEXT:    jmp bar # TAILCALL
+; CHECK32-NEXT:    jmp bar@PLT # TAILCALL
 ; CHECK32-NEXT:    # encoding: [0xeb,A]
-; CHECK32-NEXT:    # fixup A - offset: 1, value: bar-1, kind: FK_PCRel_1
+; CHECK32-NEXT:    # fixup A - offset: 1, value: bar@PLT-1, kind: FK_PCRel_1
 ;
 ; CHECK64-LABEL: f_non_leaf:
 ; CHECK64:       # %bb.0: # %entry
@@ -92,16 +92,16 @@ define void @f_non_leaf(i32 %x, i32 %y) !prof !14 {
 ; CHECK64-NEXT:  # %bb.1: # %bb1
 ; CHECK64-NEXT:    popq %rbx # encoding: [0x5b]
 ; CHECK64-NEXT:    .cfi_def_cfa_offset 8
-; CHECK64-NEXT:    jmp foo # TAILCALL
+; CHECK64-NEXT:    jmp foo@PLT # TAILCALL
 ; CHECK64-NEXT:    # encoding: [0xeb,A]
-; CHECK64-NEXT:    # fixup A - offset: 1, value: foo-1, kind: FK_PCRel_1
+; CHECK64-NEXT:    # fixup A - offset: 1, value: foo@PLT-1, kind: FK_PCRel_1
 ; CHECK64-NEXT:  .LBB1_2: # %bb2
 ; CHECK64-NEXT:    .cfi_def_cfa_offset 16
 ; CHECK64-NEXT:    popq %rbx # encoding: [0x5b]
 ; CHECK64-NEXT:    .cfi_def_cfa_offset 8
-; CHECK64-NEXT:    jmp bar # TAILCALL
+; CHECK64-NEXT:    jmp bar@PLT # TAILCALL
 ; CHECK64-NEXT:    # encoding: [0xeb,A]
-; CHECK64-NEXT:    # fixup A - offset: 1, value: bar-1, kind: FK_PCRel_1
+; CHECK64-NEXT:    # fixup A - offset: 1, value: bar@PLT-1, kind: FK_PCRel_1
 ;
 ; WIN64-LABEL: f_non_leaf:
 ; WIN64:       # %bb.0: # %entry
@@ -152,9 +152,9 @@ define x86_thiscallcc zeroext i1 @BlockPlacementTest(i8* %this, i32 %x) !prof !1
 ; CHECK32-NEXT:  # %bb.1: # %land.rhs
 ; CHECK32-NEXT:    movb $1, %al # encoding: [0xb0,0x01]
 ; CHECK32-NEXT:    testb $44, %dl # encoding: [0xf6,0xc2,0x2c]
-; CHECK32-NEXT:    je baz # TAILCALL
+; CHECK32-NEXT:    je baz@PLT # TAILCALL
 ; CHECK32-NEXT:    # encoding: [0x74,A]
-; CHECK32-NEXT:    # fixup A - offset: 1, value: baz-1, kind: FK_PCRel_1
+; CHECK32-NEXT:    # fixup A - offset: 1, value: baz@PLT-1, kind: FK_PCRel_1
 ; CHECK32-NEXT:  .LBB2_2: # %land.end
 ; CHECK32-NEXT:    # kill: def $al killed $al killed $eax
 ; CHECK32-NEXT:    retl $4 # encoding: [0xc2,0x04,0x00]
@@ -171,9 +171,9 @@ define x86_thiscallcc zeroext i1 @BlockPlacementTest(i8* %this, i32 %x) !prof !1
 ; CHECK64-NEXT:  # %bb.1: # %land.rhs
 ; CHECK64-NEXT:    movb $1, %al # encoding: [0xb0,0x01]
 ; CHECK64-NEXT:    testb $44, %sil # encoding: [0x40,0xf6,0xc6,0x2c]
-; CHECK64-NEXT:    je baz # TAILCALL
+; CHECK64-NEXT:    je baz@PLT # TAILCALL
 ; CHECK64-NEXT:    # encoding: [0x74,A]
-; CHECK64-NEXT:    # fixup A - offset: 1, value: baz-1, kind: FK_PCRel_1
+; CHECK64-NEXT:    # fixup A - offset: 1, value: baz@PLT-1, kind: FK_PCRel_1
 ; CHECK64-NEXT:  .LBB2_2: # %land.end
 ; CHECK64-NEXT:    # kill: def $al killed $al killed $eax
 ; CHECK64-NEXT:    retq # encoding: [0xc3]