From: Joerg Sonnenberger <joerg@bec.de>
Date: Sat, 27 Jul 2019 18:57:59 +0000 (+0000)
Subject: Stricter check for the memory access.
X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=791951bd32ac9af46a8c01d841912059ce8cc8cb;p=platform%2Fupstream%2Fllvm.git

Stricter check for the memory access.

The current pattern would trigger for scheduling changes of the
post-load computation, since those are commutable with the inline asm.
Avoid this by explicitly check the order of load vs asm block.

llvm-svn: 367180
---

diff --git a/llvm/test/CodeGen/X86/inlineasm-sched-bug.ll b/llvm/test/CodeGen/X86/inlineasm-sched-bug.ll
index 25bf5e0..b893496 100644
--- a/llvm/test/CodeGen/X86/inlineasm-sched-bug.ll
+++ b/llvm/test/CodeGen/X86/inlineasm-sched-bug.ll
@@ -1,7 +1,9 @@
 ; PR13504
 ; RUN: llc -mtriple=i686-- -mcpu=atom < %s | FileCheck %s
+; Check that treemap is read before the asm statement.
+; CHECK: movl 8(%{{esp|ebp}})
 ; CHECK: bsfl
-; CHECK-NOT: movl
+; CHECK-NOT: movl 8(%{{esp|ebp}})
 
 define i32 @foo(i32 %treemap) nounwind uwtable {
 entry: