https://bugs.llvm.org/show_bug.cgi?id=43540
Bug ID: 43540
Summary: Miscompile with opt -licm with memory-ssa enabled
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: enhancement
Priority: P
Component: Loop Optimizer
Assignee: unassignedb...@nondot.org
Reporter: mikael.hol...@ericsson.com
CC: llvm-bugs@lists.llvm.org
Created attachment 22622
--> https://bugs.llvm.org/attachment.cgi?id=22622&action=edit
bbi-33452.ll reproducer
With
opt -licm -S -o - bbi-33452.ll
the store to @v_2 is erroneously hoisted out of the loop, so it's done before
the memcpy reading from @v_2.
The loop is only run one round, so the hoisting looks wrong to me.
Input:
define void @foo() {
entry:
br label %for.cond
for.cond: ; preds = %for.body, %entry
%0 = phi i16 [ %inc, %for.body ], [ 0, %entry ]
%cmp = icmp slt i16 %0, 1
br i1 %cmp, label %for.body, label %for.end
for.body: ; preds = %for.cond
call void @llvm.memcpy.p0i8.p0i8.i64(i8* @v_1, i8 * @v_2, i64 1, i1 false)
store i8 1, i8 * @v_2, align 1
%inc = add nsw i16 %0, 1
br label %for.cond
for.end: ; preds = %for.cond
ret void
}
Output:
@v_1 = global i8 0, align 1
@v_2 = global i8 0, align 1
define void @foo() {
entry:
store i8 1, i8* @v_2, align 1
br label %for.cond
for.cond: ; preds = %for.body, %entry
%0 = phi i16 [ %inc, %for.body ], [ 0, %entry ]
%cmp = icmp slt i16 %0, 1
br i1 %cmp, label %for.body, label %for.end
for.body: ; preds = %for.cond
call void @llvm.memcpy.p0i8.p0i8.i64(i8* @v_1, i8* @v_2, i64 1, i1 false)
%inc = add nsw i16 %0, 1
br label %for.cond
for.end: ; preds = %for.cond
ret void
}
; Function Attrs: argmemonly nounwind willreturn
declare void @llvm.memcpy.p0i8.p0i8.i64(i8* noalias nocapture writeonly, i8*
noalias nocapture readonly, i64, i1 immarg) #0
attributes #0 = { argmemonly nounwind willreturn }
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs