Enable support for precise scheduling of the instruction selection

DAG. Disable using "-disable-sched-cycles".

For ARM, this enables a framework for modeling the cpu pipeline and
counting stalls. It also activates several heuristics to drive
scheduling based on the model. Scheduling is inherently imprecise at
this stage, and until spilling is improved it may defeat attempts to
schedule. However, this framework provides greater control over
tuning codegen.

Although the flag is not target-specific, it should have very little
affect on the default scheduler used by x86. The only two changes that
affect x86 are:
- scheduling a high-latency operation bumps the current cycle so independent
  operations can have their latency covered. i.e. two independent 4
  cycle operations can produce results in 4 cycles, not 8 cycles.
- Two operations with equal register pressure impact and no
  latency-based stalls on their uses will be prioritized by depth before height
  (height is irrelevant if no stalls occur in the schedule below this point).


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@123971 91177308-0d34-0410-b5e6-96231b3b80d8

3 files changed, 12 insertions, 7 deletions

diff --git a/lib/CodeGen/SelectionDAG/ScheduleDAGRRList.cpp b/lib/CodeGen/SelectionDAG/ScheduleDAGRRList.cpp
index e43bcfd..ad83580 100644
--- a/lib/CodeGen/SelectionDAG/ScheduleDAGRRList.cpp
+++ b/lib/CodeGen/SelectionDAG/ScheduleDAGRRList.cpp
@@ -67,7 +67,7 @@ static RegisterScheduler
                       createILPListDAGScheduler);
 
 static cl::opt<bool> DisableSchedCycles(
-  "disable-sched-cycles", cl::Hidden, cl::init(true),
+  "disable-sched-cycles", cl::Hidden, cl::init(false),
   cl::desc("Disable cycle-level precision during preRA scheduling"));
 
 namespace {
diff --git a/test/CodeGen/ARM/vcgt.ll b/test/CodeGen/ARM/vcgt.ll
index 7663da3..c3c4cb3 100644
--- a/test/CodeGen/ARM/vcgt.ll
+++ b/test/CodeGen/ARM/vcgt.ll
@@ -161,9 +161,9 @@ define <4 x i32> @vacgtQf32(<4 x float>* %A, <4 x float>* %B) nounwind {
 ; rdar://7923010
 define <4 x i32> @vcgt_zext(<4 x float>* %A, <4 x float>* %B) nounwind {
 ;CHECK: vcgt_zext:
+;CHECK: vmov.i32 q10, #0x1
 ;CHECK: vcgt.f32 q8
-;CHECK: vmov.i32 q9, #0x1
-;CHECK: vand q8, q8, q9
+;CHECK: vand q8, q8, q10
 	%tmp1 = load <4 x float>* %A
 	%tmp2 = load <4 x float>* %B
 	%tmp3 = fcmp ogt <4 x float> %tmp1, %tmp2
diff --git a/test/CodeGen/Thumb2/2010-03-15-AsmCCClobber.ll b/test/CodeGen/Thumb2/2010-03-15-AsmCCClobber.ll
index 7ee1986..458569e 100644
--- a/test/CodeGen/Thumb2/2010-03-15-AsmCCClobber.ll
+++ b/test/CodeGen/Thumb2/2010-03-15-AsmCCClobber.ll
@@ -1,4 +1,7 @@
-; RUN: llc < %s -mtriple=thumbv7-apple-darwin -mcpu=cortex-a8 | FileCheck %s
+; RUN: llc < %s -mtriple=thumbv7-apple-darwin -mcpu=cortex-a8 \
+; RUN:   -pre-RA-sched=source | FileCheck -check-prefix=SOURCE %s
+; RUN: llc < %s -mtriple=thumbv7-apple-darwin -mcpu=cortex-a8 \
+; RUN:   -pre-RA-sched=list-hybrid | FileCheck -check-prefix=HYBRID %s
 ; Radar 7459078
 target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32-n32"
 
@@ -10,9 +13,11 @@ target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-
 %s5 = type { i32 }
 
 ; Make sure the cmp is not scheduled before the InlineAsm that clobbers cc.
-; CHECK: InlineAsm End
-; CHECK: cmp
-; CHECK: beq
+; SOURCE: InlineAsm End
+; SOURCE: cmp
+; SOURCE: beq
+; HYBRID: InlineAsm End
+; HYBRID: cbz
 define void @test(%s1* %this, i32 %format, i32 %w, i32 %h, i32 %levels, i32* %s, i8* %data, i32* nocapture %rowbytes, void (i8*, i8*)* %release, i8* %info) nounwind {
 entry:
   %tmp1 = getelementptr inbounds %s1* %this, i32 0, i32 0, i32 0, i32 1, i32 0, i32 0