PPC: Implement base pointer and stack realignment

This builds on some frame-lowering code that has existed since 2005 (r24224)
but was disabled in 2008 (r48188) because it needed base pointer support to
function correctly. This implementation follows the strategy suggested by Dale
Johannesen in r48188 where the following comment was added:

  This does not currently work, because the delta between old and new stack
  pointers is added to offsets that reference incoming parameters after the
  prolog is generated, and the code that does that doesn't handle a variable
  delta.  You don't want to do that anyway; a better approach is to reserve
  another register that retains to the incoming stack pointer, and reference
  parameters relative to that.

And now we do exactly that. If we don't need a frame pointer, then we use r31
as a base pointer. If we do need a frame pointer, then we use r30 as a base
pointer. The base pointer retains the value of the stack pointer before it was
decremented in the prologue. We then use the base pointer to resolve all
negative frame indicies. The basic scheme follows that for base pointers in the
X86 backend.

We use a base pointer when we need to dynamically realign the incoming stack
pointer. This currently applies only to static objects (dynamic allocas with
large alignments, and base-pointer support in SjLj lowering will come in future
commits).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@186478 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 151 insertions, 0 deletions

diff --git a/test/CodeGen/PowerPC/stack-realign.ll b/test/CodeGen/PowerPC/stack-realign.ll
new file mode 100644
index 0000000000..7bd28f6a7c
--- /dev/null
+++ b/test/CodeGen/PowerPC/stack-realign.ll
@@ -0,0 +1,151 @@
+; RUN: llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 -disable-fp-elim < %s | FileCheck -check-prefix=CHECK-FP %s
+target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
+target triple = "powerpc64-unknown-linux-gnu"
+
+%struct.s = type { i32, i32 }
+
+declare void @bar(i32*)
+
+define void @goo(%struct.s* byval nocapture readonly %a) {
+entry:
+  %x = alloca [2 x i32], align 32
+  %a1 = getelementptr inbounds %struct.s* %a, i64 0, i32 0
+  %0 = load i32* %a1, align 4, !tbaa !0
+  %arrayidx = getelementptr inbounds [2 x i32]* %x, i64 0, i64 0
+  store i32 %0, i32* %arrayidx, align 32, !tbaa !0
+  %b = getelementptr inbounds %struct.s* %a, i64 0, i32 1
+  %1 = load i32* %b, align 4, !tbaa !0
+  %arrayidx2 = getelementptr inbounds [2 x i32]* %x, i64 0, i64 1
+  store i32 %1, i32* %arrayidx2, align 4, !tbaa !0
+  call void @bar(i32* %arrayidx)
+  ret void
+}
+
+; CHECK-LABEL: @goo
+
+; CHECK-DAG: mflr 0
+; CHECK-DAG: rldicl [[REG:[0-9]+]], 1, 0, 59
+; CHECK-DAG: std 31, -8(1)
+; CHECK-DAG: mr 31, 1
+; CHECK-DAG: std 0, 16(1)
+; CHECK-DAG: subfic 0, [[REG]], -160
+; CHECK: stdux 1, 1, 0
+
+; CHECK: .cfi_offset r31, -8
+; CHECK: .cfi_offset lr, 16
+
+; CHECK: std 3, 48(31)
+
+; CHECK: ld 1, 0(1)
+; CHECK-DAG: ld 0, 16(1)
+; CHECK-DAG: ld 31, -8(1)
+; CHECK-DAG: mtlr 0
+; CHECK: blr
+
+; CHECK-FP-LABEL: @goo
+
+; CHECK-FP-DAG: mflr 0
+; CHECK-FP-DAG: rldicl [[REG:[0-9]+]], 1, 0, 59
+; CHECK-FP-DAG: std 31, -8(1)
+; CHECK-FP-DAG: std 30, -16(1)
+; CHECK-FP-DAG: mr 30, 1
+; CHECK-FP-DAG: std 0, 16(1)
+; CHECK-FP-DAG: subfic 0, [[REG]], -160
+; CHECK-FP: stdux 1, 1, 0
+
+; CHECK-FP: .cfi_offset r31, -8
+; CHECK-FP: .cfi_offset r30, -16
+; CHECK-FP: .cfi_offset lr, 16
+
+; CHECK-FP: mr 31, 1
+
+; CHECK-FP: std 3, 48(30)
+
+; CHECK-FP: ld 1, 0(1)
+; CHECK-FP-DAG: ld 0, 16(1)
+; CHECK-FP-DAG: ld 31, -8(1)
+; CHECK-FP-DAG: ld 30, -16(1)
+; CHECK-FP-DAG: mtlr 0
+; CHECK-FP: blr
+
+; The large-frame-size case.
+define void @hoo(%struct.s* byval nocapture readonly %a) {
+entry:
+  %x = alloca [200000 x i32], align 32
+  %a1 = getelementptr inbounds %struct.s* %a, i64 0, i32 0
+  %0 = load i32* %a1, align 4, !tbaa !0
+  %arrayidx = getelementptr inbounds [200000 x i32]* %x, i64 0, i64 0
+  store i32 %0, i32* %arrayidx, align 32, !tbaa !0
+  %b = getelementptr inbounds %struct.s* %a, i64 0, i32 1
+  %1 = load i32* %b, align 4, !tbaa !0
+  %arrayidx2 = getelementptr inbounds [200000 x i32]* %x, i64 0, i64 1
+  store i32 %1, i32* %arrayidx2, align 4, !tbaa !0
+  call void @bar(i32* %arrayidx)
+  ret void
+}
+
+; CHECK-LABEL: @hoo
+
+; CHECK-DAG: lis [[REG1:[0-9]+]], -13
+; CHECK-DAG: rldicl [[REG3:[0-9]+]], 1, 0, 59
+; CHECK-DAG: mflr 0
+; CHECK-DAG: ori [[REG2:[0-9]+]], [[REG1]], 51808
+; CHECK-DAG: std 31, -8(1)
+; CHECK-DAG: mr 31, 1
+; CHECK-DAG: std 0, 16(1)
+; CHECK-DAG: subfc 0, [[REG3]], [[REG2]]
+; CHECK: stdux 1, 1, 0
+
+; CHECK: blr
+
+; Make sure that the FP save area is still allocated correctly relative to
+; where r30 is saved.
+define void @loo(%struct.s* byval nocapture readonly %a) {
+entry:
+  %x = alloca [2 x i32], align 32
+  %a1 = getelementptr inbounds %struct.s* %a, i64 0, i32 0
+  %0 = load i32* %a1, align 4, !tbaa !0
+  %arrayidx = getelementptr inbounds [2 x i32]* %x, i64 0, i64 0
+  store i32 %0, i32* %arrayidx, align 32, !tbaa !0
+  %b = getelementptr inbounds %struct.s* %a, i64 0, i32 1
+  %1 = load i32* %b, align 4, !tbaa !0
+  %arrayidx2 = getelementptr inbounds [2 x i32]* %x, i64 0, i64 1
+  store i32 %1, i32* %arrayidx2, align 4, !tbaa !0
+  call void @bar(i32* %arrayidx)
+  call void asm sideeffect "", "~{f30}"() nounwind
+  ret void
+}
+
+; CHECK-LABEL: @loo
+
+; CHECK-DAG: mflr 0
+; CHECK-DAG: rldicl [[REG:[0-9]+]], 1, 0, 59
+; CHECK-DAG: std 31, -24(1)
+; CHECK-DAG: mr 31, 1
+; CHECK-DAG: std 0, 16(1)
+; CHECK-DAG: subfic 0, [[REG]], -160
+; CHECK: stdux 1, 1, 0
+
+; CHECK: stfd 30, -16(31)
+
+; CHECK: blr
+
+; CHECK-FP-LABEL: @loo
+
+; CHECK-FP-DAG: mflr 0
+; CHECK-FP-DAG: rldicl [[REG:[0-9]+]], 1, 0, 59
+; CHECK-FP-DAG: std 31, -24(1)
+; CHECK-FP-DAG: std 30, -32(1)
+; CHECK-FP-DAG: mr 30, 1
+; CHECK-FP-DAG: std 0, 16(1)
+; CHECK-FP-DAG: subfic 0, [[REG]], -192
+; CHECK-FP: stdux 1, 1, 0
+
+; CHECK-FP: stfd 30, -16(30)
+
+; CHECK-FP: blr
+
+!0 = metadata !{metadata !"int", metadata !1}
+!1 = metadata !{metadata !"omnipotent char", metadata !2}
+!2 = metadata !{metadata !"Simple C/C++ TBAA"}