Refactor the tablegen DAGISelEmitter code for outputing calls to

getTargetNode and SelectNodeTo to reduce duplication, and to
make some of the getTargetNode code available to SelectNodeTo.
Use SelectNodeTo instead of getTargetNode in several new
interesting cases, as it mutates nodes in place instead of
creating new ones.

This triggers some scheduling behavior differences due to nodes
being presented to the scheduler in a different order. Some of the
arbitrary scheduling decisions it makes are now arbitrarily made
differently. This is visible in CodeGen/PowerPC/LargeAbsoluteAddr.ll,
where a trivial scheduling difference led to a trivial register
allocation difference.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@53203 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 2 insertions, 2 deletions

diff --git a/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll b/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll
index 1705379..f3ded0e 100644
--- a/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll
+++ b/test/CodeGen/PowerPC/LargeAbsoluteAddr.ll
@@ -1,7 +1,7 @@
 ; RUN: llvm-as < %s | llc -march=ppc32 -mtriple=powerpc-apple-darwin | \
-; RUN:   grep {stw r3, 32751}
+; RUN:   grep {stw r2, 32751}
 ; RUN: llvm-as < %s | llc -march=ppc64 -mtriple=powerpc-apple-darwin | \
-; RUN:   grep {stw r3, 32751}
+; RUN:   grep {stw r2, 32751}
 ; RUN: llvm-as < %s | llc -march=ppc64 -mtriple=powerpc-apple-darwin | \
 ; RUN:   grep {std r2, 9024}