[llvm-branch-commits] [llvm] [IR] Introduce the `ptrtoaddr` instruction (PR #139357)

[Alexander Richardson via llvm-branch-commits]

================
@@ -7,9 +7,9 @@ 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:    %"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)"
+; CHECK-NEXT:    %"vl14659llvm.experimental.convergence.anchor()" = call token @llvm.experimental.convergence.anchor()
+; CHECK-NEXT:    %"op15516(vl14659)" = call i32 @llvm.amdgcn.readfirstlane.i32(i32 [[A0]]) [ "convergencectrl"(token %"vl14659llvm.experimental.convergence.anchor()") ]
+; CHECK-NEXT:    ret i32 %"op15516(vl14659)"
----------------
arichardson wrote:

I believe this is because all the other instruction kind numbers after the casts changed to the overall hash value is different. I can do the same thing as the C API and allocate the next unused value but e.g. freeze didn't do this either.

https://github.com/llvm/llvm-project/pull/139357