From f60513f71d3e433491e2bfc51cf764de78b848f8 Mon Sep 17 00:00:00 2001 From: Chris Lattner Date: Mon, 17 Jan 2005 06:25:59 +0000 Subject: [PATCH] New testcase for a problem that occurred in 132.ijpeg git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@19630 91177308-0d34-0410-b5e6-96231b3b80d8 --- test/CodeGen/X86/2005-01-17-CycleInDAG.ll | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) create mode 100644 test/CodeGen/X86/2005-01-17-CycleInDAG.ll diff --git a/test/CodeGen/X86/2005-01-17-CycleInDAG.ll b/test/CodeGen/X86/2005-01-17-CycleInDAG.ll new file mode 100644 index 00000000000..fdfed6cf41d --- /dev/null +++ b/test/CodeGen/X86/2005-01-17-CycleInDAG.ll @@ -0,0 +1,16 @@ +; This testcase was distilled from 132.ijpeg. Bsaically we cannot fold the +; load into the sub instruction here as it induces a cycle in the dag, which +; is invalid code (there is no correct way to order the instruction). Check +; that we do not fold the load into the sub. + +; RUN: llvm-as < %s | llc -march=x86 -disable-pattern-isel=0 | not grep 'sub.*GLOBAL' + +%GLOBAL = external global int + +int %test(int* %P1, int* %P2, int* %P3) { + %L = load int* %GLOBAL + store int 12, int* %P2 + %Y = load int* %P3 + %Z = sub int %Y, %L + ret int %Z +}