[X86] Correctly select registers to pop into for x86_64

[mkuper]

Summary

When trying to replace an add to esp with pops, we need to choose dead registers to pop into.
Registers clobbered by the call and not imp-def'd by it should be safe. Except that it's not enough to check the register itself isn't defined, we also need to make sure no overlapping registers are defined either.

This fixes PR26711.

[aaboud]

LGTM

[rnk]

lgtm

test/CodeGen/X86/pop-stack-cleanup.ll

82

Add 'LINUX64: retq' so that if more tests are added later the -NOT line will be bound inside this function.

[hans]

Thanks for tracking this down!

lib/Target/X86/X86FrameLowering.cpp

for (auto Candidate : RegClass) {
     // Poor man's liveness:
     // Since we're immediately after a call, any register that is clobbered
     // by the call and not defined by it can be considered dead.
     if (!RegMask.clobbersPhysReg(Candidate))
       continue;
 
     bool IsDef = false;
     for (const MachineOperand &MO : Prev->implicit_operands()) {
-      if (MO.isReg() && MO.isDef() && MO.getReg() == Candidate) {
+      if (MO.isReg() && MO.isDef() && 
+          TRI->isSuperOrSubRegisterEq(MO.getReg(), Candidate)) {
         IsDef = true;
         break;
       }
     }
 
     if (IsDef)
       continue;
 

test/CodeGen/X86/pop-stack-cleanup.ll

 ; RUN: llc < %s -mtriple=i686-windows | FileCheck %s -check-prefix=CHECK 
 ; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s -check-prefix=LINUX64
 
 declare void @param1(i32 %a)
 declare i32 @param2_ret(i32 %a, i32 %b)
 declare i64 @param2_ret64(i32 %a, i32 %b)
 declare void @param2(i32 %a, i32 %b)
 declare void @param3(i32 %a, i32 %b, i32 %c)
 declare void @param8(i64, i64, i64, i64, i64, i64, i64, i64)
+declare i32 @param8_ret(i64, i64, i64, i64, i64, i64, i64, i64)
 
 
 define void @test() minsize nounwind {
 ; CHECK-LABEL: test:
 ; CHECK: calll _param1
 ; CHECK-NEXT: popl %eax
 ; CHECK: calll _param2
 ; CHECK-NEXT: popl %eax

 ; LINUX64: callq param8
 ; LINUX64-NEXT: popq %rax
 ; LINUX64-NEXT: popq %rcx
 
   %a = alloca i64, i32 %size, align 8
   call void @param8(i64 1, i64 2, i64 3, i64 4, i64 5, i64 6, i64 7, i64 8)
   ret void
 }
+
+define i32 @test_linux64_i32(i32 %size) minsize nounwind {
+; LINUX64-LABEL: test_linux64_i32:
+; LINUX64: callq param8
+; LINUX64-NOT: popq %rax
+  %a = alloca i64, i32 %size, align 8
+  %r = call i32 @param8_ret(i64 1, i64 2, i64 3, i64 4, i64 5, i64 6, i64 7, i64 8)
+  ret i32 %r
+}