[llvm] [IR] Return correct memory effects for `convergencectrl` (PR #129874)

[Yingwei Zheng via llvm-commits]

================
@@ -7,14 +7,14 @@ define i32 @nested(i32 %src) #0 {
 ; CHECK-SAME: i32 [[A0:%.*]]) #[[ATTR0:[0-9]+]] {
 ; CHECK-NEXT:  [[BB15160:.*:]]
 ; CHECK-NEXT:    [[T1:%.*]] = call token @llvm.experimental.convergence.entry()
-; CHECK-NEXT:    %"vl15001llvm.experimental.convergence.anchor()" = call token @llvm.experimental.convergence.anchor()
-; CHECK-NEXT:    %"op68297llvm.amdgcn.readfirstlane.i32([[A0]], vl15001llvm.experimental.convergence.anchor())" = call i32 @llvm.amdgcn.readfirstlane.i32(i32 [[A0]]) [ "convergencectrl"(token %"vl15001llvm.experimental.convergence.anchor()") ]
-; CHECK-NEXT:    ret i32 undef
+; CHECK-NEXT:    %"vl77672llvm.experimental.convergence.anchor()" = call token @llvm.experimental.convergence.anchor()
+; CHECK-NEXT:    %"op68297(vl77672)" = call i32 @llvm.amdgcn.readfirstlane.i32(i32 [[A0]]) [ "convergencectrl"(token %"vl77672llvm.experimental.convergence.anchor()") ]
+; CHECK-NEXT:    ret i32 %"op68297(vl77672)"
 ;
   %t1 = call token @llvm.experimental.convergence.entry()
   %t2 = call token @llvm.experimental.convergence.anchor()
   %r2 = call i32 @llvm.amdgcn.readfirstlane(i32 %src) [ "convergencectrl"(token %t2) ]
-  ret i32 undef
+  ret i32 %r2
----------------
dtcxzyw wrote:

IRNormalizer does not set name for dead values.


https://github.com/llvm/llvm-project/pull/129874