commit 8626b741716390a0119ffeb88b5b9fcf08e13010
author Alexandre Rames <alexandre.rames@linaro.org> Wed Nov 25 16:28:08 2015 +0000
committer Alexandre Rames <alexandre.rames@linaro.org> Wed Nov 25 16:28:08 2015 +0000
tree 28d261dbb8fa3018cba8a5d829319604508ea0a1
parent 0c32fdeaeda2a1e388e280da12662d1d18c834a2

ARM64: Use the shifter operands.

This introduces architecture-specific instruction simplification.
On ARM64 we try to merge shifts and sign-extension operations into
arithmetic and logical instructions.

For example for the Java code

    int res = a + (b << 5);

we would generate

    lsl w3, w2, #5
    add w0, w1, w3

and we now generate

    add w0, w1, w2, lsl #5

Change-Id: Ic03bdff44a1c12e21ddff1b0513bd32a730742b7

compiler/optimizing/nodes.h

diff --git a/compiler/optimizing/nodes.h b/compiler/optimizing/nodes.h
index 1bd626f..b9df92f 100644
--- a/compiler/optimizing/nodes.h
+++ b/compiler/optimizing/nodes.h
@@ -1101,6 +1101,7 @@
 #define FOR_EACH_CONCRETE_INSTRUCTION_ARM64(M)
 #else
 #define FOR_EACH_CONCRETE_INSTRUCTION_ARM64(M)                          \
+  M(Arm64DataProcWithShifterOp, Instruction)                            \
   M(Arm64IntermediateAddress, Instruction)                              \
   M(Arm64MultiplyAccumulate, Instruction)
 #endif