[AArch64] Add new subtarget feature to fuse AES crypto operations

This feature enables the fusion of such operations on Cortex A57, as
recommended in its Software Optimisation Guide, section 4.13, and on Exynos
M1.

Differential revision: https://reviews.llvm.org/D28491

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293738 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/AArch64/AArch64.td b/lib/Target/AArch64/AArch64.td
index cb6d8fc96688da34db7196789558115a6153d6d0..1f21da75681aaeb224b358a704740dff520aeb9c 100644 (file)
--- a/lib/Target/AArch64/AArch64.td
+++ b/lib/Target/AArch64/AArch64.td
@@ -103,6 +103,10 @@ def FeatureArithmeticCbzFusion : SubtargetFeature<
     "arith-cbz-fusion", "HasArithmeticCbzFusion", "true",
     "CPU fuses arithmetic + cbz/cbnz operations">;
 
+def FeatureFuseAES : SubtargetFeature<
+    "fuse-aes", "HasFuseAES", "true",
+    "CPU fuses AES crypto operations">;
+
 def FeatureDisableLatencySchedHeuristic : SubtargetFeature<
     "disable-latency-sched-heuristic", "DisableLatencySchedHeuristic", "true",
     "Disable latency scheduling heuristic">;
@@ -184,6 +188,7 @@ def ProcA57     : SubtargetFeature<"a57", "ARMProcFamily", "CortexA57",
                                    FeatureCrypto,
                                    FeatureCustomCheapAsMoveHandling,
                                    FeatureFPARMv8,
+                                   FeatureFuseAES,
                                    FeatureNEON,
                                    FeaturePerfMon,
                                    FeaturePostRAScheduler,
@@ -230,6 +235,7 @@ def ProcExynosM1 : SubtargetFeature<"exynosm1", "ARMProcFamily", "ExynosM1",
                                      FeatureCrypto,
                                      FeatureCustomCheapAsMoveHandling,
                                      FeatureFPARMv8,
+                                     FeatureFuseAES,
                                      FeatureNEON,
                                      FeaturePerfMon,
                                      FeaturePostRAScheduler,

diff --git a/lib/Target/AArch64/AArch64MacroFusion.cpp b/lib/Target/AArch64/AArch64MacroFusion.cpp
index c1adc862fedf83d716c4b1ebe738d00a010c96a6..f6d693262c363e0ee1b21d9aa14e02c3118a8757 100644 (file)
--- a/lib/Target/AArch64/AArch64MacroFusion.cpp
+++ b/lib/Target/AArch64/AArch64MacroFusion.cpp
@@ -116,6 +116,19 @@ static bool shouldScheduleAdjacent(const AArch64InstrInfo &TII,
         return true;
       }
 
+  if (ST.hasFuseAES())
+    // Fuse AES crypto operations.
+    switch(FirstOpcode) {
+    // AES encode.
+    case AArch64::AESErr:
+      return SecondOpcode == AArch64::AESMCrr ||
+             SecondOpcode == AArch64::INSTRUCTION_LIST_END;
+    // AES decode.
+    case AArch64::AESDrr:
+      return SecondOpcode == AArch64::AESIMCrr ||
+             SecondOpcode == AArch64::INSTRUCTION_LIST_END;
+    }
+
   return false;
 }
 

diff --git a/lib/Target/AArch64/AArch64SchedA57.td b/lib/Target/AArch64/AArch64SchedA57.td
index 99c48d0146e473959b7f252cb92a77e925d1f71a..303398ea0b7f3b524b9419a3aecf45c34412642f 100644 (file)
--- a/lib/Target/AArch64/AArch64SchedA57.td
+++ b/lib/Target/AArch64/AArch64SchedA57.td
@@ -162,7 +162,9 @@ def : InstRW<[A57Write_2cyc_1M],    (instregex "BFM")>;
 // Cryptography Extensions
 // -----------------------------------------------------------------------------
 
-def : InstRW<[A57Write_3cyc_1W], (instregex "^AES")>;
+def A57ReadAES  : SchedReadAdvance<3, [A57Write_3cyc_1W]>;
+def : InstRW<[A57Write_3cyc_1W], (instregex "^AES[DE]")>;
+def : InstRW<[A57Write_3cyc_1W, A57ReadAES], (instregex "^AESI?MC")>;
 def : InstRW<[A57Write_6cyc_2V], (instregex "^SHA1SU0")>;
 def : InstRW<[A57Write_3cyc_1W], (instregex "^SHA1(H|SU1)")>;
 def : InstRW<[A57Write_6cyc_2W], (instregex "^SHA1[CMP]")>;

diff --git a/lib/Target/AArch64/AArch64SchedM1.td b/lib/Target/AArch64/AArch64SchedM1.td
index 14d6891253facf6856eb19d2f5fd5594cc611ab9..3fbbc0be682d739c4cf60a633565519fd5d8d1c0 100644 (file)
--- a/lib/Target/AArch64/AArch64SchedM1.td
+++ b/lib/Target/AArch64/AArch64SchedM1.td
@@ -366,7 +366,8 @@ def : InstRW<[M1WriteNALU1],  (instregex "^ZIP[12]v")>;
 // Cryptography instructions.
 def M1WriteAES : SchedWriteRes<[M1UnitNCRYPT]> { let Latency = 1; }
 def M1ReadAES  : SchedReadAdvance<1, [M1WriteAES]>;
-def : InstRW<[M1WriteAES, M1ReadAES], (instregex "^AES")>;
+def : InstRW<[M1WriteAES], (instregex "^AES[DE]")>;
+def : InstRW<[M1WriteAES, M1ReadAES], (instregex "^AESI?MC")>;
 
 def : InstRW<[M1WriteNCRYPT1], (instregex "^PMUL")>;
 def : InstRW<[M1WriteNCRYPT1], (instregex "^SHA1(H|SU)")>;

diff --git a/lib/Target/AArch64/AArch64Subtarget.h b/lib/Target/AArch64/AArch64Subtarget.h
index 48bcdbc82d09cbce3b4b0263664686e4bc2d7f39..26eac3c63493fb2f598926f6f0ba6b017b47cb6b 100644 (file)
--- a/lib/Target/AArch64/AArch64Subtarget.h
+++ b/lib/Target/AArch64/AArch64Subtarget.h
@@ -84,6 +84,7 @@ protected:
   bool UseAlternateSExtLoadCVTF32Pattern = false;
   bool HasArithmeticBccFusion = false;
   bool HasArithmeticCbzFusion = false;
+  bool HasFuseAES = false;
   bool DisableLatencySchedHeuristic = false;
   bool UseRSqrt = false;
   uint8_t MaxInterleaveFactor = 2;
@@ -197,6 +198,7 @@ public:
   }
   bool hasArithmeticBccFusion() const { return HasArithmeticBccFusion; }
   bool hasArithmeticCbzFusion() const { return HasArithmeticCbzFusion; }
+  bool hasFuseAES() const { return HasFuseAES; }
   bool useRSqrt() const { return UseRSqrt; }
   unsigned getMaxInterleaveFactor() const { return MaxInterleaveFactor; }
   unsigned getVectorInsertExtractBaseCost() const {