| Issue |
138208
|
| Summary |
Codegenprepare creates invalid IR
|
| Labels |
new issue
|
| Assignees |
|
| Reporter |
aeubanks
|
```
$ cat a.ll
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
target triple = "x86_64-grtev4-linux-gnu"
declare void @g(ptr)
define void @f(ptr %arg) {
bb:
%getelementptr = getelementptr i8, ptr %arg, i64 -64
%getelementptr1 = getelementptr i8, ptr %arg, i64 -56
call void @g(ptr %getelementptr)
br label %bb3
bb3:
%load = load ptr, ptr %getelementptr, align 8
%load4 = load i32, ptr %getelementptr1, align 8
%load5 = load ptr, ptr %getelementptr, align 8
%add = add i32 1, 0
%icmp = icmp eq i32 %add, 0
br i1 %icmp, label %bb7, label %bb7
bb7:
ret void
}
$ opt -p 'require<profile-summary>,codegenprepare' a.ll -cgpp-huge-func=0 -disable-output
Instruction does not dominate all uses!
%sunkaddr1 = getelementptr i8, ptr %arg, i64 -64
%load = load ptr, ptr %sunkaddr1, align 8
LLVM ERROR: Broken module found, compilation aborted!
```
seems like the GEP we create [here](https://github.com/llvm/llvm-project/blob/a76936f1c01c7cadbce8ea6553af758d0f614b6a/llvm/lib/CodeGen/CodeGenPrepare.cpp#L6059) is getting inserted right before `%load5` to use in that load, but then we end up reusing that one [here](https://github.com/llvm/llvm-project/blob/a76936f1c01c7cadbce8ea6553af758d0f614b6a/llvm/lib/CodeGen/CodeGenPrepare.cpp#L5909 for `%load` in a later iteration
![]()
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
