commit 6a0b657a1875b4fbb020b806169e2f73fcb2578b
author Artem Serov <artem.serov@linaro.org> Fri Jul 26 20:38:37 2019 +0100
committer Vladimir Marko <vmarko@google.com> Fri Aug 02 13:31:43 2019 +0000
tree 955bb0e3413e18f2b13b7fee7fa3e6e48a214597
parent 61f071630083775fe64d177455a056daa7071eca

ART: ARM64: Optimize frame size for SIMD graphs.

For SIMD graphs allocate 64 bit instead of 128 bit on stack for
each FP register to be preserved by the callee in the frame entry
as ABI suggests (currently 64-bit registers are preserved but
more space on stack is allocated).

Note: slow paths still require spilling full 128-bit Q-Registers
for SIMD graphs due to register allocator restrictions.

Test: test-art-target.
Change-Id: Ie0b12e4b769158445f3d0f4562c70d4fb0ea7744

compiler/optimizing/code_generator_x86_64.cc

diff --git a/compiler/optimizing/code_generator_x86_64.cc b/compiler/optimizing/code_generator_x86_64.cc
index 48a8320..a75c745 100644
--- a/compiler/optimizing/code_generator_x86_64.cc
+++ b/compiler/optimizing/code_generator_x86_64.cc
@@ -1249,7 +1249,7 @@
   } else {
     __ movsd(Address(CpuRegister(RSP), stack_index), XmmRegister(reg_id));
   }
-  return GetFloatingPointSpillSlotSize();
+  return GetSlowPathFPWidth();
 }
 
 size_t CodeGeneratorX86_64::RestoreFloatingPointRegister(size_t stack_index, uint32_t reg_id) {
@@ -1258,7 +1258,7 @@
   } else {
     __ movsd(XmmRegister(reg_id), Address(CpuRegister(RSP), stack_index));
   }
-  return GetFloatingPointSpillSlotSize();
+  return GetSlowPathFPWidth();
 }
 
 void CodeGeneratorX86_64::InvokeRuntime(QuickEntrypointEnum entrypoint,
@@ -1377,7 +1377,7 @@
   __ subq(CpuRegister(RSP), Immediate(adjust));
   __ cfi().AdjustCFAOffset(adjust);
   uint32_t xmm_spill_location = GetFpuSpillStart();
-  size_t xmm_spill_slot_size = GetFloatingPointSpillSlotSize();
+  size_t xmm_spill_slot_size = GetCalleePreservedFPWidth();
 
   for (int i = arraysize(kFpuCalleeSaves) - 1; i >= 0; --i) {
     if (allocated_registers_.ContainsFloatingPointRegister(kFpuCalleeSaves[i])) {
@@ -1405,7 +1405,7 @@
   __ cfi().RememberState();
   if (!HasEmptyFrame()) {
     uint32_t xmm_spill_location = GetFpuSpillStart();
-    size_t xmm_spill_slot_size = GetFloatingPointSpillSlotSize();
+    size_t xmm_spill_slot_size = GetCalleePreservedFPWidth();
     for (size_t i = 0; i < arraysize(kFpuCalleeSaves); ++i) {
       if (allocated_registers_.ContainsFloatingPointRegister(kFpuCalleeSaves[i])) {
         int offset = xmm_spill_location + (xmm_spill_slot_size * i);