From a4b2f5df5eabb86b54b9c4257660305d477a9aca Mon Sep 17 00:00:00 2001
From: Geoff Berry <gberry@codeaurora.org>
Date: Tue, 26 Sep 2017 21:40:41 +0000
Subject: [PATCH] [AArch64][Falkor] Fix correctness bug in falkor prefetcher
 fix pass and correct some opcode tag computations.

Summary:
This addresses a correctness bug for LD[1234]*_POST opcodes that have
the prefetcher fix applied to them: the base register was not being
written back from the temp after being incremented, so it would appear
to never be incremented.

Also, fix some opcode tag computations based on some updated HW details
to get better tag avoidance and thus better prefetcher performance.

Reviewers: mcrosier

Subscribers: aemerson, rengolin, javed.absar, kristof.beyls

Differential Revision: https://reviews.llvm.org/D38256

llvm-svn: 314251
---
 llvm/lib/Target/AArch64/AArch64FalkorHWPFFix.cpp | 101 +++++----
 llvm/test/CodeGen/AArch64/falkor-hwpf-fix.mir    | 277 ++++++++++++++++++++++-
 2 files changed, 318 insertions(+), 60 deletions(-)

diff --git a/llvm/lib/Target/AArch64/AArch64FalkorHWPFFix.cpp b/llvm/lib/Target/AArch64/AArch64FalkorHWPFFix.cpp
index 8ecfcb6..7bf0c94 100644
--- a/llvm/lib/Target/AArch64/AArch64FalkorHWPFFix.cpp
+++ b/llvm/lib/Target/AArch64/AArch64FalkorHWPFFix.cpp
@@ -240,27 +240,27 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
   default:
     return None;
 
+  case AArch64::LD1i64:
+  case AArch64::LD2i64:
+    DestRegIdx = 0;
+    BaseRegIdx = 3;
+    OffsetIdx = -1;
+    IsPrePost = false;
+    break;
+
   case AArch64::LD1i8:
   case AArch64::LD1i16:
   case AArch64::LD1i32:
-  case AArch64::LD1i64:
   case AArch64::LD2i8:
   case AArch64::LD2i16:
   case AArch64::LD2i32:
-  case AArch64::LD2i64:
   case AArch64::LD3i8:
   case AArch64::LD3i16:
   case AArch64::LD3i32:
+  case AArch64::LD3i64:
   case AArch64::LD4i8:
   case AArch64::LD4i16:
   case AArch64::LD4i32:
-    DestRegIdx = 0;
-    BaseRegIdx = 3;
-    OffsetIdx = -1;
-    IsPrePost = false;
-    break;
-
-  case AArch64::LD3i64:
   case AArch64::LD4i64:
     DestRegIdx = -1;
     BaseRegIdx = 3;
@@ -284,23 +284,16 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
   case AArch64::LD1Rv4s:
   case AArch64::LD1Rv8h:
   case AArch64::LD1Rv16b:
-  case AArch64::LD1Twov1d:
-  case AArch64::LD1Twov2s:
-  case AArch64::LD1Twov4h:
-  case AArch64::LD1Twov8b:
-  case AArch64::LD2Twov2s:
-  case AArch64::LD2Twov4s:
-  case AArch64::LD2Twov8b:
-  case AArch64::LD2Rv1d:
-  case AArch64::LD2Rv2s:
-  case AArch64::LD2Rv4s:
-  case AArch64::LD2Rv8b:
     DestRegIdx = 0;
     BaseRegIdx = 1;
     OffsetIdx = -1;
     IsPrePost = false;
     break;
 
+  case AArch64::LD1Twov1d:
+  case AArch64::LD1Twov2s:
+  case AArch64::LD1Twov4h:
+  case AArch64::LD1Twov8b:
   case AArch64::LD1Twov2d:
   case AArch64::LD1Twov4s:
   case AArch64::LD1Twov8h:
@@ -321,10 +314,17 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
   case AArch64::LD1Fourv4s:
   case AArch64::LD1Fourv8h:
   case AArch64::LD1Fourv16b:
+  case AArch64::LD2Twov2s:
+  case AArch64::LD2Twov4s:
+  case AArch64::LD2Twov8b:
   case AArch64::LD2Twov2d:
   case AArch64::LD2Twov4h:
   case AArch64::LD2Twov8h:
   case AArch64::LD2Twov16b:
+  case AArch64::LD2Rv1d:
+  case AArch64::LD2Rv2s:
+  case AArch64::LD2Rv4s:
+  case AArch64::LD2Rv8b:
   case AArch64::LD2Rv2d:
   case AArch64::LD2Rv4h:
   case AArch64::LD2Rv8h:
@@ -365,32 +365,32 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
     IsPrePost = false;
     break;
 
+  case AArch64::LD1i64_POST:
+  case AArch64::LD2i64_POST:
+    DestRegIdx = 1;
+    BaseRegIdx = 4;
+    OffsetIdx = 5;
+    IsPrePost = true;
+    break;
+
   case AArch64::LD1i8_POST:
   case AArch64::LD1i16_POST:
   case AArch64::LD1i32_POST:
-  case AArch64::LD1i64_POST:
   case AArch64::LD2i8_POST:
   case AArch64::LD2i16_POST:
   case AArch64::LD2i32_POST:
-  case AArch64::LD2i64_POST:
   case AArch64::LD3i8_POST:
   case AArch64::LD3i16_POST:
   case AArch64::LD3i32_POST:
+  case AArch64::LD3i64_POST:
   case AArch64::LD4i8_POST:
   case AArch64::LD4i16_POST:
   case AArch64::LD4i32_POST:
-    DestRegIdx = 1;
-    BaseRegIdx = 4;
-    OffsetIdx = 5;
-    IsPrePost = false;
-    break;
-
-  case AArch64::LD3i64_POST:
   case AArch64::LD4i64_POST:
     DestRegIdx = -1;
     BaseRegIdx = 4;
     OffsetIdx = 5;
-    IsPrePost = false;
+    IsPrePost = true;
     break;
 
   case AArch64::LD1Onev1d_POST:
@@ -409,23 +409,16 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
   case AArch64::LD1Rv4s_POST:
   case AArch64::LD1Rv8h_POST:
   case AArch64::LD1Rv16b_POST:
-  case AArch64::LD1Twov1d_POST:
-  case AArch64::LD1Twov2s_POST:
-  case AArch64::LD1Twov4h_POST:
-  case AArch64::LD1Twov8b_POST:
-  case AArch64::LD2Twov2s_POST:
-  case AArch64::LD2Twov4s_POST:
-  case AArch64::LD2Twov8b_POST:
-  case AArch64::LD2Rv1d_POST:
-  case AArch64::LD2Rv2s_POST:
-  case AArch64::LD2Rv4s_POST:
-  case AArch64::LD2Rv8b_POST:
     DestRegIdx = 1;
     BaseRegIdx = 2;
     OffsetIdx = 3;
-    IsPrePost = false;
+    IsPrePost = true;
     break;
 
+  case AArch64::LD1Twov1d_POST:
+  case AArch64::LD1Twov2s_POST:
+  case AArch64::LD1Twov4h_POST:
+  case AArch64::LD1Twov8b_POST:
   case AArch64::LD1Twov2d_POST:
   case AArch64::LD1Twov4s_POST:
   case AArch64::LD1Twov8h_POST:
@@ -446,10 +439,17 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
   case AArch64::LD1Fourv4s_POST:
   case AArch64::LD1Fourv8h_POST:
   case AArch64::LD1Fourv16b_POST:
+  case AArch64::LD2Twov2s_POST:
+  case AArch64::LD2Twov4s_POST:
+  case AArch64::LD2Twov8b_POST:
   case AArch64::LD2Twov2d_POST:
   case AArch64::LD2Twov4h_POST:
   case AArch64::LD2Twov8h_POST:
   case AArch64::LD2Twov16b_POST:
+  case AArch64::LD2Rv1d_POST:
+  case AArch64::LD2Rv2s_POST:
+  case AArch64::LD2Rv4s_POST:
+  case AArch64::LD2Rv8b_POST:
   case AArch64::LD2Rv2d_POST:
   case AArch64::LD2Rv4h_POST:
   case AArch64::LD2Rv8h_POST:
@@ -487,7 +487,7 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
     DestRegIdx = -1;
     BaseRegIdx = 2;
     OffsetIdx = 3;
-    IsPrePost = false;
+    IsPrePost = true;
     break;
 
   case AArch64::LDRBBroW:
@@ -592,16 +592,19 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
     IsPrePost = true;
     break;
 
+  case AArch64::LDNPDi:
+  case AArch64::LDNPQi:
+  case AArch64::LDNPSi:
   case AArch64::LDPQi:
+  case AArch64::LDPDi:
+  case AArch64::LDPSi:
     DestRegIdx = -1;
     BaseRegIdx = 2;
     OffsetIdx = 3;
     IsPrePost = false;
     break;
 
-  case AArch64::LDPDi:
   case AArch64::LDPSWi:
-  case AArch64::LDPSi:
   case AArch64::LDPWi:
   case AArch64::LDPXi:
     DestRegIdx = 0;
@@ -612,18 +615,18 @@ static Optional<LoadInfo> getLoadInfo(const MachineInstr &MI) {
 
   case AArch64::LDPQpost:
   case AArch64::LDPQpre:
+  case AArch64::LDPDpost:
+  case AArch64::LDPDpre:
+  case AArch64::LDPSpost:
+  case AArch64::LDPSpre:
     DestRegIdx = -1;
     BaseRegIdx = 3;
     OffsetIdx = 4;
     IsPrePost = true;
     break;
 
-  case AArch64::LDPDpost:
-  case AArch64::LDPDpre:
   case AArch64::LDPSWpost:
   case AArch64::LDPSWpre:
-  case AArch64::LDPSpost:
-  case AArch64::LDPSpre:
   case AArch64::LDPWpost:
   case AArch64::LDPWpre:
   case AArch64::LDPXpost:
diff --git a/llvm/test/CodeGen/AArch64/falkor-hwpf-fix.mir b/llvm/test/CodeGen/AArch64/falkor-hwpf-fix.mir
index 54c8b16..298e8a0 100644
--- a/llvm/test/CodeGen/AArch64/falkor-hwpf-fix.mir
+++ b/llvm/test/CodeGen/AArch64/falkor-hwpf-fix.mir
@@ -1,12 +1,7 @@
 # RUN: llc -mtriple=aarch64-linux-gnu -mcpu=falkor -run-pass falkor-hwpf-fix-late -o - %s | FileCheck %s
---- |
-  @g = external global i32
-
-  define void @hwpf1() { ret void }
-  define void @hwpf2() { ret void }
-...
 ---
-# Verify that the tag collision between the loads is resolved.
+# Verify that the tag collision between the loads is resolved for various load opcodes.
+
 # CHECK-LABEL: name: hwpf1
 # CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
 # CHECK: LDRWui %[[BASE]], 0
@@ -17,7 +12,7 @@ body: |
   bb.0:
     liveins: %w0, %x1
 
-    %w2 = LDRWui %x1, 0 :: ("aarch64-strided-access" load 4 from @g)
+    %w2 = LDRWui %x1, 0 :: ("aarch64-strided-access" load 4)
     %w2 = LDRWui %x1, 1
 
     %w0 = SUBWri %w0, 1, 0
@@ -28,19 +23,147 @@ body: |
     RET_ReallyLR
 ...
 ---
-# Verify that the tag collision between the loads is resolved and written back for post increment addressing.
 # CHECK-LABEL: name: hwpf2
 # CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1i64 %q2, 0, %[[BASE]]
+# CHECK: LDRWui %x1, 0
+name:            hwpf2
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %q2
+
+    %q2 = LD1i64 %q2, 0, %x1 :: ("aarch64-strided-access" load 4)
+    %w2 = LDRWui %x1, 0
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpf3
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1i8 %q2, 0, %[[BASE]]
+# CHECK: LDRWui %x1, 0
+name:            hwpf3
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %q2
+
+    %q2 = LD1i8 %q2, 0, %x1 :: ("aarch64-strided-access" load 4)
+    %w0 = LDRWui %x1, 0
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpf4
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1Onev1d %[[BASE]]
+# CHECK: LDRWui %x1, 0
+name:            hwpf4
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1
+
+    %d2 = LD1Onev1d %x1 :: ("aarch64-strided-access" load 4)
+    %w2 = LDRWui %x1, 0
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpf5
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1Twov1d %[[BASE]]
+# CHECK: LDRWui %x1, 0
+name:            hwpf5
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1
+
+    %d2_d3 = LD1Twov1d %x1 :: ("aarch64-strided-access" load 4)
+    %w0 = LDRWui %x1, 0
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpf6
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LDPQi %[[BASE]]
+# CHECK: LDRWui %x1, 3
+name:            hwpf6
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1
+
+    %q2, %q3 = LDPQi %x1, 3 :: ("aarch64-strided-access" load 4)
+    %w0 = LDRWui %x1, 3
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpf7
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LDPXi %[[BASE]]
+# CHECK: LDRWui %x1, 2
+name:            hwpf7
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1
+
+    %x2, %x3 = LDPXi %x1, 3 :: ("aarch64-strided-access" load 4)
+    %w2 = LDRWui %x1, 2
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# Verify that the tag collision between the loads is resolved and written back
+# for post increment addressing for various load opcodes.
+
+# CHECK-LABEL: name: hwpfinc1
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
 # CHECK: LDRWpost %[[BASE]], 0
 # CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
 # CHECK: LDRWui %x1, 1
-name:            hwpf2
+name:            hwpfinc1
 tracksRegLiveness: true
 body: |
   bb.0:
     liveins: %w0, %x1
 
-    %x1, %w2 = LDRWpost %x1, 0 :: ("aarch64-strided-access" load 4 from @g)
+    %x1, %w2 = LDRWpost %x1, 0 :: ("aarch64-strided-access" load 4)
     %w2 = LDRWui %x1, 1
 
     %w0 = SUBWri %w0, 1, 0
@@ -50,3 +173,135 @@ body: |
   bb.1:
     RET_ReallyLR
 ...
+---
+# CHECK-LABEL: name: hwpfinc2
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1i64_POST %q2, 0, %[[BASE]]
+# CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
+# CHECK: LDRWui %x1, 1
+name:            hwpfinc2
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %q2
+
+    %x1, %q2 = LD1i64_POST %q2, 0, %x1, %x1 :: ("aarch64-strided-access" load 4)
+    %w2 = LDRWui %x1, 132
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpfinc3
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1i8_POST %q2, 0, %[[BASE]]
+# CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
+# CHECK: LDRWui %x1, 132
+name:            hwpfinc3
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %q2
+
+    %x1, %q2 = LD1i8_POST %q2, 0, %x1, %x1 :: ("aarch64-strided-access" load 4)
+    %w0 = LDRWui %x1, 132
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpfinc4
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD1Rv1d_POST %[[BASE]]
+# CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
+# CHECK: LDRWui %x1, 252
+name:            hwpfinc4
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %q2
+
+    %x1, %d2 = LD1Rv1d_POST %x1, %xzr :: ("aarch64-strided-access" load 4)
+    %w2 = LDRWui %x1, 252
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpfinc5
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LD3Threev2s_POST %[[BASE]]
+# CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
+# CHECK: LDRWroX %x17, %x0
+name:            hwpfinc5
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %x17, %q2
+
+    %x1, %d2_d3_d4 = LD3Threev2s_POST %x1, %x0 :: ("aarch64-strided-access" load 4)
+    %w0 = LDRWroX %x17, %x0, 0, 0
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpfinc6
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LDPDpost %[[BASE]]
+# CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
+# CHECK: LDRWui %x17, 2
+name:            hwpfinc6
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %x17, %q2
+
+    %x1, %d2, %d3 = LDPDpost %x1, 3 :: ("aarch64-strided-access" load 4)
+    %w16 = LDRWui %x17, 2
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
+---
+# CHECK-LABEL: name: hwpfinc7
+# CHECK: %[[BASE:[a-z0-9]+]] = ORRXrs %xzr, %x1, 0
+# CHECK: LDPXpost %[[BASE]]
+# CHECK: %x1 = ORRXrs %xzr, %[[BASE]], 0
+# CHECK: LDRWui %x17, 2
+name:            hwpfinc7
+tracksRegLiveness: true
+body: |
+  bb.0:
+    liveins: %w0, %x1, %x17, %q2
+
+    %x1, %x2, %x3 = LDPXpost %x1, 3 :: ("aarch64-strided-access" load 4)
+    %w18 = LDRWui %x17, 2
+
+    %w0 = SUBWri %w0, 1, 0
+    %wzr = SUBSWri %w0, 0, 0, implicit-def %nzcv
+    Bcc 9, %bb.0, implicit %nzcv
+
+  bb.1:
+    RET_ReallyLR
+...
-- 
2.7.4