[LLVM][AARCH64]Replace +sme2p1+smef16f16 by +smef16f16

llvm/include/llvm/TargetParser/AArch64TargetParser.h

@@ -268,7 +268,7 @@ inline constexpr ExtensionInfo Extensions[] = {
     {"sha3", AArch64::AEK_SHA3, "+sha3", "-sha3", FEAT_SHA3, "+sha3,+sha2,+fp-armv8,+neon", 140},
     {"simd", AArch64::AEK_SIMD, "+neon", "-neon", FEAT_SIMD, "+fp-armv8,+neon", 100},
     {"sm4", AArch64::AEK_SM4, "+sm4", "-sm4", FEAT_SM4, "+sm4,+fp-armv8,+neon", 106},
-    {"sme-f16f16", AArch64::AEK_SMEF16F16, "+sme-f16f16", "-sme-f16f16", FEAT_INIT, "", 0},
+    {"sme-f16f16", AArch64::AEK_SMEF16F16, "+sme-f16f16", "-sme-f16f16", FEAT_INIT, "+sme2,+sme-f16f16", 0},
     {"sme-f64f64", AArch64::AEK_SMEF64F64, "+sme-f64f64", "-sme-f64f64", FEAT_SME_F64, "+sme,+sme-f64f64,+bf16", 560},
     {"sme-i16i64", AArch64::AEK_SMEI16I64, "+sme-i16i64", "-sme-i16i64", FEAT_SME_I64, "+sme,+sme-i16i64,+bf16", 570},
     {"sme", AArch64::AEK_SME, "+sme", "-sme", FEAT_SME, "+sme,+bf16", 430},
@@ -302,7 +302,7 @@ inline constexpr ExtensionInfo Extensions[] = {
     {"ssve-fp8dot4", AArch64::AEK_SSVE_FP8DOT4, "+ssve-fp8dot4", "-ssve-fp8dot4", FEAT_INIT, "+sme2", 0},
     {"lut", AArch64::AEK_LUT, "+lut", "-lut", FEAT_INIT, "", 0},
     {"sme-lutv2", AArch64::AEK_SME_LUTv2, "+sme-lutv2", "-sme-lutv2", FEAT_INIT, "", 0},
-    {"sme-f8f16", AArch64::AEK_SMEF8F16, "+sme-f8f16", "-sme-f8f16", FEAT_INIT, "+sme2,+fp8", 0},
+    {"sme-f8f16", AArch64::AEK_SMEF8F16, "+sme-f8f16", "-sme-f8f16", FEAT_INIT, "+fp8,+sme2", 0},
     {"sme-f8f32", AArch64::AEK_SMEF8F32, "+sme-f8f32", "-sme-f8f32", FEAT_INIT, "+sme2,+fp8", 0},
     {"sme-fa64",  AArch64::AEK_SMEFA64,  "+sme-fa64", "-sme-fa64",  FEAT_INIT, "", 0},
     {"cpa", AArch64::AEK_CPA, "+cpa", "-cpa", FEAT_INIT, "", 0},

llvm/lib/Target/AArch64/AArch64.td

@@ -76,7 +76,7 @@ def SME2p1Unsupported : AArch64Unsupported;
 
 def SME2Unsupported : AArch64Unsupported {
   let F = !listconcat([HasSME2, HasSVE2orSME2, HasSVE2p1_or_HasSME2, HasSSVE_FP8FMA,
-                      HasSMEF8F16, HasSMEF8F32],
+                      HasSMEF8F16, HasSMEF8F32, HasSMEF16F16orSMEF8F16],
                       SME2p1Unsupported.F);
 }
 

llvm/lib/Target/AArch64/AArch64Features.td

@@ -505,15 +505,15 @@ def FeatureSMEF64F64 : SubtargetFeature<"sme-f64f64", "HasSMEF64F64", "true",
 def FeatureSMEI16I64 : SubtargetFeature<"sme-i16i64", "HasSMEI16I64", "true",
   "Enable Scalable Matrix Extension (SME) I16I64 instructions (FEAT_SME_I16I64)", [FeatureSME]>;
 
-def FeatureSMEF16F16 : SubtargetFeature<"sme-f16f16", "HasSMEF16F16", "true",
-  "Enable SME2.1 non-widening Float16 instructions (FEAT_SME_F16F16)", []>;
-
 def FeatureSMEFA64 : SubtargetFeature<"sme-fa64", "HasSMEFA64", "true",
   "Enable the full A64 instruction set in streaming SVE mode (FEAT_SME_FA64)", [FeatureSME, FeatureSVE2]>;
 
 def FeatureSME2 : SubtargetFeature<"sme2", "HasSME2", "true",
   "Enable Scalable Matrix Extension 2 (SME2) instructions", [FeatureSME]>;
 
+def FeatureSMEF16F16 : SubtargetFeature<"sme-f16f16", "HasSMEF16F16", "true",
+  "Enable SME non-widening Float16 instructions (FEAT_SME_F16F16)", [FeatureSME2]>;
+
 def FeatureSME2p1 : SubtargetFeature<"sme2p1", "HasSME2p1", "true",
   "Enable Scalable Matrix Extension 2.1 (FEAT_SME2p1) instructions", [FeatureSME2]>;
 

llvm/lib/Target/AArch64/AArch64InstrInfo.td

@@ -230,6 +230,12 @@ def HasSVE2p1_or_HasSME2
 def HasSVE2p1_or_HasSME2p1
     : Predicate<"Subtarget->hasSVE2p1() || Subtarget->hasSME2p1()">,
                  AssemblerPredicateWithAll<(any_of FeatureSME2p1, FeatureSVE2p1), "sme2p1 or sve2p1">;
+
+def HasSMEF16F16orSMEF8F16
+    : Predicate<"Subtarget->hasSMEF16F16() || Subtarget->hasSMEF8F16()">,
+                AssemblerPredicateWithAll<(any_of FeatureSMEF16F16, FeatureSMEF8F16),
+                "sme-f16f16 or sme-f8f16">;
+
 // A subset of NEON instructions are legal in Streaming SVE execution mode,
 // they should be enabled if either has been specified.
 def HasNEONorSME

llvm/lib/Target/AArch64/AArch64SMEInstrInfo.td

@@ -792,12 +792,14 @@ defm LUTI4_S_2ZTZI : sme2p1_luti4_vector_vg2_index<"luti4">;
 defm LUTI4_S_4ZTZI : sme2p1_luti4_vector_vg4_index<"luti4">;
 }
 
-let Predicates = [HasSME2p1, HasSMEF16F16] in {
+let Predicates = [HasSMEF16F16orSMEF8F16] in {
 defm FADD_VG2_M2Z_H : sme2_multivec_accum_add_sub_vg2<"fadd", 0b0100, MatrixOp16, ZZ_h_mul_r, nxv8f16, null_frag>;
 defm FADD_VG4_M4Z_H : sme2_multivec_accum_add_sub_vg4<"fadd", 0b0100, MatrixOp16, ZZZZ_h_mul_r, nxv8f16, null_frag>;
 defm FSUB_VG2_M2Z_H : sme2_multivec_accum_add_sub_vg2<"fsub", 0b0101, MatrixOp16, ZZ_h_mul_r, nxv8f16, null_frag>;
 defm FSUB_VG4_M4Z_H : sme2_multivec_accum_add_sub_vg4<"fsub", 0b0101, MatrixOp16, ZZZZ_h_mul_r, nxv8f16, null_frag>;
+}
 
+let Predicates = [HasSMEF16F16] in {
 defm FMLA_VG2_M2ZZI_H : sme2p1_multi_vec_array_vg2_index_16b<"fmla", 0b00, 0b100, ZZ_h_mul_r, ZPR4b16>;
 defm FMLA_VG4_M4ZZI_H : sme2p1_multi_vec_array_vg4_index_16b<"fmla", 0b000, ZZZZ_h_mul_r, ZPR4b16>;
 defm FMLA_VG2_M2ZZ_H :  sme2_dot_mla_add_sub_array_vg24_single<"fmla", 0b0011100, MatrixOp16, ZZ_h, ZPR4b16>;

llvm/test/MC/AArch64/SME2p1/fadd-diagnostics.s

@@ -1,4 +1,4 @@
-// RUN: not llvm-mc -triple=aarch64 -show-encoding -mattr=+sme2p1,+sme-f16f16 2>&1 < %s | FileCheck %s
+// RUN: not llvm-mc -triple=aarch64 -show-encoding -mattr=+sme-f16f16 2>&1 < %s | FileCheck %s
 
 // --------------------------------------------------------------------------//
 // Out of range index offset