[SKX] Enabling SKX target and AVX512BW, AVX512DQ, AVX512VL features.

Enabling HasAVX512{DQ,BW,VL} predicates.
Adding VK2, VK4, VK32, VK64 masked register classes.
Adding new types (v64i8, v32i16) to VR512.
Extending calling conventions for new types (v64i8, v32i16)

Patch by Zinovy Nis <zinovy.y.nis@intel.com>
Reviewed by Elena Demikhovsky <elena.demikhovsky@intel.com>


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213545 91177308-0d34-0410-b5e6-96231b3b80d8

10 files changed, 91 insertions, 23 deletions

diff --git a/lib/Target/X86/AsmParser/X86AsmParser.cpp b/lib/Target/X86/AsmParser/X86AsmParser.cpp
index a259c963203..a11a238fc97 100644
--- a/lib/Target/X86/AsmParser/X86AsmParser.cpp
+++ b/lib/Target/X86/AsmParser/X86AsmParser.cpp
@@ -1666,6 +1666,8 @@ bool X86AsmParser::HandleAVX512Operand(OperandVector &Operands,
         // Recognize only reasonable suffixes.
         const char *BroadcastPrimitive =
           StringSwitch<const char*>(getLexer().getTok().getIdentifier())
+            .Case("to2",  "{1to2}")
+            .Case("to4",  "{1to4}")
             .Case("to8",  "{1to8}")
             .Case("to16", "{1to16}")
             .Default(nullptr);
diff --git a/lib/Target/X86/Disassembler/X86DisassemblerDecoderCommon.h b/lib/Target/X86/Disassembler/X86DisassemblerDecoderCommon.h
index 4baaf1e6873..13a7b557b44 100644
--- a/lib/Target/X86/Disassembler/X86DisassemblerDecoderCommon.h
+++ b/lib/Target/X86/Disassembler/X86DisassemblerDecoderCommon.h
@@ -265,7 +265,7 @@ enum attributeBits {
   ENUM_ENTRY(IC_EVEX_L2_W_KZ,        3,  "requires EVEX_KZ, L2 and W")               \
   ENUM_ENTRY(IC_EVEX_L2_W_XS_KZ,     4,  "requires EVEX_KZ, L2, W and XS prefix")    \
   ENUM_ENTRY(IC_EVEX_L2_W_XD_KZ,     4,  "requires EVEX_KZ, L2, W and XD prefix")    \
-  ENUM_ENTRY(IC_EVEX_L2_W_OPSIZE_KZ, 4,  "requires EVEX_KZ, L2, W and OpSize")     
+  ENUM_ENTRY(IC_EVEX_L2_W_OPSIZE_KZ, 4,  "requires EVEX_KZ, L2, W and OpSize")
 
 #define ENUM_ENTRY(n, r, d) n,
 enum InstructionContext {
@@ -453,8 +453,12 @@ enum OperandEncoding {
   ENUM_ENTRY(TYPE_XMM256,     "32-byte")                                       \
   ENUM_ENTRY(TYPE_XMM512,     "64-byte")                                       \
   ENUM_ENTRY(TYPE_VK1,        "1-bit")                                         \
+  ENUM_ENTRY(TYPE_VK2,        "2-bit")                                         \
+  ENUM_ENTRY(TYPE_VK4,        "4-bit")                                         \
   ENUM_ENTRY(TYPE_VK8,        "8-bit")                                         \
   ENUM_ENTRY(TYPE_VK16,       "16-bit")                                        \
+  ENUM_ENTRY(TYPE_VK32,       "32-bit")                                        \
+  ENUM_ENTRY(TYPE_VK64,       "64-bit")                                        \
   ENUM_ENTRY(TYPE_XMM0,       "Implicit use of XMM0")                          \
   ENUM_ENTRY(TYPE_SEGMENTREG, "Segment register operand")                      \
   ENUM_ENTRY(TYPE_DEBUGREG,   "Debug register operand")                        \
diff --git a/lib/Target/X86/X86.td b/lib/Target/X86/X86.td
index 93f516a151a..cd32a0f2423 100644
--- a/lib/Target/X86/X86.td
+++ b/lib/Target/X86/X86.td
@@ -104,7 +104,15 @@ def FeatureCDI      : SubtargetFeature<"avx512cd", "HasCDI", "true",
 def FeaturePFI      : SubtargetFeature<"avx512pf", "HasPFI", "true",
                       "Enable AVX-512 PreFetch Instructions",
                                       [FeatureAVX512]>;
-
+def FeatureDQI     : SubtargetFeature<"avx512dq", "HasDQI", "true",
+                      "Enable AVX-512 Doubleword and Quadword Instructions",
+                                      [FeatureAVX512]>;
+def FeatureBWI     : SubtargetFeature<"avx512bw", "HasBWI", "true",
+                      "Enable AVX-512 Byte and Word Instructions",
+                                      [FeatureAVX512]>;
+def FeatureVLX     : SubtargetFeature<"avx512vl", "HasVLX", "true",
+                      "Enable AVX-512 Vector Length eXtensions",
+                                      [FeatureAVX512]>;
 def FeaturePCLMUL  : SubtargetFeature<"pclmul", "HasPCLMUL", "true",
                          "Enable packed carry-less multiplication instructions",
                                [FeatureSSE2]>;
@@ -276,6 +284,17 @@ def : ProcessorModel<"knl", HaswellModel,
                       FeatureBMI2, FeatureFMA, FeatureRTM, FeatureHLE,
                       FeatureSlowIncDec]>;
 
+// SKX
+// FIXME: define SKX model
+def : ProcessorModel<"skx", HaswellModel,
+                     [FeatureAVX512, FeatureCDI,
+                      FeatureDQI, FeatureBWI, FeatureVLX,
+                      FeatureCMPXCHG16B, FeatureFastUAMem, FeaturePOPCNT,
+                      FeatureAES, FeaturePCLMUL, FeatureRDRAND, FeatureF16C,
+                      FeatureFSGSBase, FeatureMOVBE, FeatureLZCNT, FeatureBMI,
+                      FeatureBMI2, FeatureFMA, FeatureRTM, FeatureHLE,
+                      FeatureSlowIncDec]>;
+
 def : Proc<"k6",              [FeatureMMX]>;
 def : Proc<"k6-2",            [Feature3DNow]>;
 def : Proc<"k6-3",            [Feature3DNow]>;
diff --git a/lib/Target/X86/X86CallingConv.td b/lib/Target/X86/X86CallingConv.td
index 0824d4ed660..86c01bd6464 100644
--- a/lib/Target/X86/X86CallingConv.td
+++ b/lib/Target/X86/X86CallingConv.td
@@ -52,7 +52,7 @@ def RetCC_X86Common : CallingConv<[
   // 512-bit vectors are returned in ZMM0 and ZMM1, when they fit. ZMM2 and ZMM3
   // can only be used by ABI non-compliant code. This vector type is only
   // supported while using the AVX-512 target feature.
-  CCIfType<[v16i32, v8i64, v16f32, v8f64],
+  CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64],
             CCAssignToReg<[ZMM0,ZMM1,ZMM2,ZMM3]>>,
 
   // MMX vector types are always returned in MM0. If the target doesn't have
@@ -252,7 +252,7 @@ def CC_X86_64_C : CallingConv<[
                                          YMM4, YMM5, YMM6, YMM7]>>>>,
 
   // The first 8 512-bit vector arguments are passed in ZMM registers.
-  CCIfNotVarArg<CCIfType<[v16i32, v8i64, v16f32, v8f64],
+  CCIfNotVarArg<CCIfType<[v64i8, v32i16, v16i32, v8i64, v16f32, v8f64],
             CCIfSubtarget<"hasAVX512()",
             CCAssignToReg<[ZMM0, ZMM1, ZMM2, ZMM3, ZMM4, ZMM5, ZMM6, ZMM7]>>>>,
 
diff --git a/lib/Target/X86/X86InstrAVX512.td b/lib/Target/X86/X86InstrAVX512.td
index 41e900ed11a..d2894088b80 100644
--- a/lib/Target/X86/X86InstrAVX512.td
+++ b/lib/Target/X86/X86InstrAVX512.td
@@ -1,19 +1,36 @@
 // Bitcasts between 512-bit vector types. Return the original type since
 // no instruction is needed for the conversion
 let Predicates = [HasAVX512] in {
-  def : Pat<(v8f64  (bitconvert (v16f32 VR512:$src))), (v8f64 VR512:$src)>;
-  def : Pat<(v8f64  (bitconvert (v16i32 VR512:$src))), (v8f64 VR512:$src)>;
   def : Pat<(v8f64  (bitconvert (v8i64 VR512:$src))),  (v8f64 VR512:$src)>;
-  def : Pat<(v16f32 (bitconvert (v16i32 VR512:$src))), (v16f32 VR512:$src)>;
+  def : Pat<(v8f64  (bitconvert (v16i32 VR512:$src))), (v8f64 VR512:$src)>;
+  def : Pat<(v8f64  (bitconvert (v32i16 VR512:$src))),  (v8f64 VR512:$src)>;
+  def : Pat<(v8f64  (bitconvert (v64i8 VR512:$src))), (v8f64 VR512:$src)>;
+  def : Pat<(v8f64  (bitconvert (v16f32 VR512:$src))), (v8f64 VR512:$src)>;
   def : Pat<(v16f32 (bitconvert (v8i64 VR512:$src))),  (v16f32 VR512:$src)>;
+  def : Pat<(v16f32 (bitconvert (v16i32 VR512:$src))), (v16f32 VR512:$src)>;
+  def : Pat<(v16f32 (bitconvert (v32i16 VR512:$src))), (v16f32 VR512:$src)>;
+  def : Pat<(v16f32 (bitconvert (v64i8 VR512:$src))), (v16f32 VR512:$src)>;
   def : Pat<(v16f32 (bitconvert (v8f64 VR512:$src))),  (v16f32 VR512:$src)>;
-  def : Pat<(v8i64  (bitconvert (v16f32 VR512:$src))), (v8i64 VR512:$src)>;
   def : Pat<(v8i64  (bitconvert (v16i32 VR512:$src))), (v8i64 VR512:$src)>;
+  def : Pat<(v8i64  (bitconvert (v32i16 VR512:$src))), (v8i64 VR512:$src)>;
+  def : Pat<(v8i64  (bitconvert (v64i8 VR512:$src))), (v8i64 VR512:$src)>;
   def : Pat<(v8i64  (bitconvert (v8f64 VR512:$src))),  (v8i64 VR512:$src)>;
-  def : Pat<(v16i32 (bitconvert (v16f32 VR512:$src))), (v16i32 VR512:$src)>;
-  def : Pat<(v16i32 (bitconvert (v8i64 VR512:$src))),  (v16i32 VR512:$src)>;
+  def : Pat<(v8i64  (bitconvert (v16f32 VR512:$src))), (v8i64 VR512:$src)>;
+  def : Pat<(v16i32 (bitconvert (v8i64 VR512:$src))), (v16i32 VR512:$src)>;
+  def : Pat<(v16i32 (bitconvert (v32i16 VR512:$src))),  (v16i32 VR512:$src)>;
+  def : Pat<(v16i32 (bitconvert (v64i8 VR512:$src))),  (v16i32 VR512:$src)>;
   def : Pat<(v16i32 (bitconvert (v8f64 VR512:$src))),  (v16i32 VR512:$src)>;
-  def : Pat<(v8f64  (bitconvert (v8i64 VR512:$src))),  (v8f64 VR512:$src)>;
+  def : Pat<(v32i16 (bitconvert (v8i64 VR512:$src))), (v32i16 VR512:$src)>;
+  def : Pat<(v32i16 (bitconvert (v16i32 VR512:$src))),  (v32i16 VR512:$src)>;
+  def : Pat<(v32i16 (bitconvert (v64i8 VR512:$src))),  (v32i16 VR512:$src)>;
+  def : Pat<(v32i16 (bitconvert (v8f64 VR512:$src))),  (v32i16 VR512:$src)>;
+  def : Pat<(v32i16 (bitconvert (v16f32 VR512:$src))), (v32i16 VR512:$src)>;
+  def : Pat<(v32i16 (bitconvert (v16f32 VR512:$src))), (v32i16 VR512:$src)>;
+  def : Pat<(v64i8  (bitconvert (v8i64 VR512:$src))), (v64i8 VR512:$src)>;
+  def : Pat<(v64i8  (bitconvert (v16i32 VR512:$src))), (v64i8 VR512:$src)>;
+  def : Pat<(v64i8  (bitconvert (v32i16 VR512:$src))), (v64i8 VR512:$src)>;
+  def : Pat<(v64i8  (bitconvert (v8f64 VR512:$src))),  (v64i8 VR512:$src)>;
+  def : Pat<(v64i8  (bitconvert (v16f32 VR512:$src))), (v64i8 VR512:$src)>;
 
   def : Pat<(v2i64 (bitconvert (v4i32 VR128X:$src))), (v2i64 VR128X:$src)>;
   def : Pat<(v2i64 (bitconvert (v8i16 VR128X:$src))), (v2i64 VR128X:$src)>;
@@ -135,7 +152,6 @@ def VINSERTI32x4rm : AVX512AIi8<0x38, MRMSrcMem, (outs VR512:$dst),
           (ins VR512:$src1, i128mem:$src2, i8imm:$src3),
           "vinserti32x4\t{$src3, $src2, $src1, $dst|$dst, $src1, $src2, $src3}",
           []>, EVEX_4V, EVEX_V512, EVEX_CD8<32, CD8VT4>;
-
 }
 
 let hasSideEffects = 0 in {
diff --git a/lib/Target/X86/X86InstrFormats.td b/lib/Target/X86/X86InstrFormats.td
index ab6d2885fb3..8ef5f901c18 100644
--- a/lib/Target/X86/X86InstrFormats.td
+++ b/lib/Target/X86/X86InstrFormats.td
@@ -184,6 +184,8 @@ class EVEX_KZ : EVEX_K { bit hasEVEX_Z = 1; }
 class EVEX_B { bit hasEVEX_B = 1; }
 class EVEX_RC { bit hasEVEX_RC = 1; }
 class EVEX_V512 { bit hasEVEX_L2 = 1; bit hasVEX_L = 0; }
+class EVEX_V256 { bit hasEVEX_L2 = 0; bit hasVEX_L = 1; }
+class EVEX_V128 { bit hasEVEX_L2 = 0; bit hasVEX_L = 0; }
 
 // Specify AVX512 8-bit compressed displacement encoding based on the vector
 // element size in bits (8, 16, 32, 64) and the CDisp8 form.
diff --git a/lib/Target/X86/X86InstrInfo.td b/lib/Target/X86/X86InstrInfo.td
index e7b532c6af8..0f872a676c2 100644
--- a/lib/Target/X86/X86InstrInfo.td
+++ b/lib/Target/X86/X86InstrInfo.td
@@ -719,10 +719,14 @@ def HasAVX512    : Predicate<"Subtarget->hasAVX512()">,
                      AssemblerPredicate<"FeatureAVX512", "AVX-512 ISA">;
 def UseAVX       : Predicate<"Subtarget->hasAVX() && !Subtarget->hasAVX512()">;
 def UseAVX2      : Predicate<"Subtarget->hasAVX2() && !Subtarget->hasAVX512()">;
-def NoAVX512       : Predicate<"!Subtarget->hasAVX512()">;
+def NoAVX512     : Predicate<"!Subtarget->hasAVX512()">;
 def HasCDI       : Predicate<"Subtarget->hasCDI()">;
 def HasPFI       : Predicate<"Subtarget->hasPFI()">;
 def HasERI       : Predicate<"Subtarget->hasERI()">;
+def HasDQI       : Predicate<"Subtarget->hasDQI()">;
+def HasBWI       : Predicate<"Subtarget->hasBWI()">;
+def HasVLX       : Predicate<"Subtarget->hasVLX()">,
+                     AssemblerPredicate<"FeatureVLX", "AVX-512 VLX ISA">;
 
 def HasPOPCNT    : Predicate<"Subtarget->hasPOPCNT()">;
 def HasAES       : Predicate<"Subtarget->hasAES()">;
diff --git a/lib/Target/X86/X86RegisterInfo.td b/lib/Target/X86/X86RegisterInfo.td
index 33c402b69a4..0da98637496 100644
--- a/lib/Target/X86/X86RegisterInfo.td
+++ b/lib/Target/X86/X86RegisterInfo.td
@@ -449,7 +449,7 @@ def FPCCR : RegisterClass<"X86", [i16], 16, (add FPSW)> {
 }
 
 // AVX-512 vector/mask registers.
-def VR512 : RegisterClass<"X86", [v16f32, v8f64, v16i32, v8i64], 512,
+def VR512 : RegisterClass<"X86", [v16f32, v8f64, v64i8, v32i16, v16i32, v8i64], 512,
     (sequence "ZMM%u", 0, 31)>;
 
 // Scalar AVX-512 floating point registers.
@@ -463,13 +463,19 @@ def VR128X : RegisterClass<"X86", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64],
 def VR256X : RegisterClass<"X86", [v32i8, v16i16, v8i32, v4i64, v8f32, v4f64],
                           256, (sequence "YMM%u", 0, 31)>;
 
-// The size of the all masked registers is 16 bit because we have only one
-// KMOVW istruction that can store this register in memory, and it writes 2 bytes
-def VK1     : RegisterClass<"X86", [i1],    16, (sequence "K%u", 0, 7)>;
-def VK8     : RegisterClass<"X86", [v8i1],  16, (add VK1)> {let Size = 16;}
+// Mask registers
+def VK1     : RegisterClass<"X86", [i1],    16, (sequence "K%u", 0, 7)> {let Size = 16;}
+def VK2     : RegisterClass<"X86", [v2i1],  16, (add VK1)> {let Size = 16;}
+def VK4     : RegisterClass<"X86", [v4i1],  16, (add VK2)> {let Size = 16;}
+def VK8     : RegisterClass<"X86", [v8i1],  16, (add VK4)> {let Size = 16;}
 def VK16    : RegisterClass<"X86", [v16i1], 16, (add VK8)> {let Size = 16;}
+def VK32    : RegisterClass<"X86", [v32i1], 32, (add VK16)> {let Size = 32;}
+def VK64    : RegisterClass<"X86", [v64i1], 64, (add VK32)> {let Size = 64;}
 
 def VK1WM   : RegisterClass<"X86", [i1],    16, (sub VK1, K0)> {let Size = 16;}
+def VK2WM   : RegisterClass<"X86", [v2i1],  16, (sub VK2, K0)> {let Size = 16;}
+def VK4WM   : RegisterClass<"X86", [v4i1],  16, (sub VK4, K0)> {let Size = 16;}
 def VK8WM   : RegisterClass<"X86", [v8i1],  16, (sub VK8, K0)> {let Size = 16;}
-def VK16WM  : RegisterClass<"X86", [v16i1], 16, (add VK8WM)>;
-
+def VK16WM  : RegisterClass<"X86", [v16i1], 16, (add VK8WM)>   {let Size = 16;}
+def VK32WM  : RegisterClass<"X86", [v32i1], 32, (add VK16WM)> {let Size = 32;}
+def VK64WM  : RegisterClass<"X86", [v64i1], 64, (add VK32WM)> {let Size = 64;}
diff --git a/lib/Target/X86/X86Subtarget.cpp b/lib/Target/X86/X86Subtarget.cpp
index d1b71c48a92..41551a1d677 100644
--- a/lib/Target/X86/X86Subtarget.cpp
+++ b/lib/Target/X86/X86Subtarget.cpp
@@ -272,6 +272,9 @@ void X86Subtarget::initializeEnvironment() {
   HasERI = false;
   HasCDI = false;
   HasPFI = false;
+  HasDQI = false;
+  HasBWI = false;
+  HasVLX = false;
   HasADX = false;
   HasSHA = false;
   HasPRFCHW = false;
diff --git a/lib/Target/X86/X86Subtarget.h b/lib/Target/X86/X86Subtarget.h
index 30c3b08a890..5f5df5e0818 100644
--- a/lib/Target/X86/X86Subtarget.h
+++ b/lib/Target/X86/X86Subtarget.h
@@ -189,13 +189,22 @@ protected:
 
   /// Processor has AVX-512 PreFetch Instructions
   bool HasPFI;
-  
+
   /// Processor has AVX-512 Exponential and Reciprocal Instructions
   bool HasERI;
-  
+
   /// Processor has AVX-512 Conflict Detection Instructions
   bool HasCDI;
-  
+
+  /// Processor has AVX-512 Doubleword and Quadword instructions
+  bool HasDQI;
+
+  /// Processor has AVX-512 Byte and Word instructions
+  bool HasBWI;
+
+  /// Processor has AVX-512 Vector Length eXtenstions
+  bool HasVLX;
+
   /// stackAlignment - The minimum alignment known to hold of the stack frame on
   /// entry to the function and which must be maintained by every function.
   unsigned stackAlignment;
@@ -349,6 +358,9 @@ public:
   bool hasCDI() const { return HasCDI; }
   bool hasPFI() const { return HasPFI; }
   bool hasERI() const { return HasERI; }
+  bool hasDQI() const { return HasDQI; }
+  bool hasBWI() const { return HasBWI; }
+  bool hasVLX() const { return HasVLX; }
 
   bool isAtom() const { return X86ProcFamily == IntelAtom; }
   bool isSLM() const { return X86ProcFamily == IntelSLM; }