[Clang][SME2] Add builtins for moving multi-vectors to/from ZA

clang/include/clang/Basic/arm_sme.td

@@ -299,6 +299,44 @@ multiclass ZAAddSub<string n_suffix> {
 defm SVADD : ZAAddSub<"add">;
 defm SVSUB : ZAAddSub<"sub">;
 
+// SME2 - MOVA
+
+//
+// Single, 2 and 4 vector-group read/write intrinsics.
+//
+
+multiclass ZAWrite_VG<string n, string t, string i, list<ImmCheck> checks> {
+  def NAME # _VG2_H : Inst<"svwrite_hor_" # n # "[_{d}]_vg2",   "vim2", t, MergeNone, i # "_hor_vg2", [IsSharedZA, IsStreaming], checks>;
+  def NAME # _VG2_V : Inst<"svwrite_ver_" # n # "[_{d}]_vg2",   "vim2", t, MergeNone, i # "_ver_vg2", [IsSharedZA, IsStreaming], checks>;
+  def NAME # _VG4_H : Inst<"svwrite_hor_" # n # "[_{d}]_vg4",   "vim4", t, MergeNone, i # "_hor_vg4", [IsSharedZA, IsStreaming], checks>;
+  def NAME # _VG4_V : Inst<"svwrite_ver_" # n # "[_{d}]_vg4",   "vim4", t, MergeNone, i # "_ver_vg4", [IsSharedZA, IsStreaming], checks>;
+  def NAME # _VG1x2 : Inst<"svwrite_"     # n # "[_{d}]_vg1x2", "vm2",  t, MergeNone, i # "_vg1x2",   [IsSharedZA, IsStreaming], []>;
+  def NAME # _VG1x4 : Inst<"svwrite_"     # n # "[_{d}]_vg1x4", "vm4",  t, MergeNone, i # "_vg1x4",   [IsSharedZA, IsStreaming], []>;
+}
+
+let TargetGuard = "sme2" in {
+  defm SVWRITE_ZA8  : ZAWrite_VG<"za8",  "cUc",   "aarch64_sme_write", [ImmCheck<0, ImmCheck0_0>]>;
+  defm SVWRITE_ZA16 : ZAWrite_VG<"za16", "sUshb", "aarch64_sme_write", [ImmCheck<0, ImmCheck0_1>]>;
+  defm SVWRITE_ZA32 : ZAWrite_VG<"za32", "iUif",  "aarch64_sme_write", [ImmCheck<0, ImmCheck0_3>]>;
+  defm SVWRITE_ZA64 : ZAWrite_VG<"za64", "lUld",  "aarch64_sme_write", [ImmCheck<0, ImmCheck0_7>]>;
+}
+
+multiclass ZARead_VG<string n, string t, string i, list<ImmCheck> checks> {
+  def NAME # _VG2_H : Inst<"svread_hor_" # n # "_{d}_vg2",   "2im", t, MergeNone, i # "_hor_vg2", [IsSharedZA, IsPreservesZA, IsStreaming], checks>;
+  def NAME # _VG2_V : Inst<"svread_ver_" # n # "_{d}_vg2",   "2im", t, MergeNone, i # "_ver_vg2", [IsSharedZA, IsPreservesZA, IsStreaming], checks>;
+  def NAME # _VG4_H : Inst<"svread_hor_" # n # "_{d}_vg4",   "4im", t, MergeNone, i # "_hor_vg4", [IsSharedZA, IsPreservesZA, IsStreaming], checks>;
+  def NAME # _VG4_V : Inst<"svread_ver_" # n # "_{d}_vg4",   "4im", t, MergeNone, i # "_ver_vg4", [IsSharedZA, IsPreservesZA, IsStreaming], checks>;
+  def NAME # _VG1x2 : Inst<"svread_"     # n # "_{d}_vg1x2", "2m",  t, MergeNone, i # "_vg1x2",   [IsSharedZA, IsPreservesZA, IsStreaming], []>;
+  def NAME # _VG1x4 : Inst<"svread_"     # n # "_{d}_vg1x4", "4m",  t, MergeNone, i # "_vg1x4",   [IsSharedZA, IsPreservesZA, IsStreaming], []>;
+}
+
+let TargetGuard = "sme2" in {
+  defm SVREAD_ZA8  : ZARead_VG<"za8",  "cUc",   "aarch64_sme_read", [ImmCheck<0, ImmCheck0_0>]>;
+  defm SVREAD_ZA16 : ZARead_VG<"za16", "sUshb", "aarch64_sme_read", [ImmCheck<0, ImmCheck0_1>]>;
+  defm SVREAD_ZA32 : ZARead_VG<"za32", "iUif",  "aarch64_sme_read", [ImmCheck<0, ImmCheck0_3>]>;
+  defm SVREAD_ZA64 : ZARead_VG<"za64", "lUld",  "aarch64_sme_read", [ImmCheck<0, ImmCheck0_7>]>;
+}
+
 //
 // Outer product and accumulate/subtract
 //

clang/test/Sema/aarch64-sme2-intrinsics/acle_sme2_imm.cpp

@@ -5,6 +5,57 @@
 
 #include <arm_sme_draft_spec_subject_to_change.h>
 
+void test_multivector_read(uint32_t base) __arm_streaming __arm_shared_za __arm_preserves_za {
+
+  // Test Tile Range
+  svread_hor_za8_u8_vg2(1, base); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+  svread_ver_za8_u8_vg2(1, base); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+  svread_hor_za8_u8_vg4(1, base); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+  svread_ver_za8_u8_vg4(1, base); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+
+  svread_hor_za16_u16_vg2(2, base); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+  svread_ver_za16_u16_vg2(2, base); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+  svread_hor_za16_u16_vg4(2, base); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+  svread_ver_za16_u16_vg4(2, base); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+
+  svread_hor_za32_u32_vg2(4, base); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+  svread_ver_za32_u32_vg2(4, base); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+  svread_hor_za32_u32_vg4(4, base); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+  svread_ver_za32_u32_vg4(4, base); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+
+  svread_hor_za64_u64_vg2(8, base); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+  svread_ver_za64_u64_vg2(8, base); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+  svread_hor_za64_u64_vg4(8, base); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+  svread_ver_za64_u64_vg4(8, base); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+}
+
+void test_multivector_write(uint32_t base, svuint8x2_t v8x2, svuint8x4_t v8x4,
+                            svuint16x2_t v16x2, svuint16x4_t v16x4,
+                            svuint32x2_t v32x2, svuint32x4_t v32x4,
+                            svuint64x2_t v64x2, svuint64x4_t v64x4) __arm_streaming __arm_shared_za {
+
+  // Test Tile Range
+  svwrite_hor_za8_u8_vg2(1, base, v8x2); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+  svwrite_ver_za8_u8_vg2(1, base, v8x2); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+  svwrite_hor_za8_u8_vg4(1, base, v8x4); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+  svwrite_ver_za8_u8_vg4(1, base, v8x4); // expected-error {{argument value 1 is outside the valid range [0, 0]}}
+
+  svwrite_hor_za16_u16_vg2(2, base, v16x2); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+  svwrite_ver_za16_u16_vg2(2, base, v16x2); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+  svwrite_hor_za16_u16_vg4(2, base, v16x4); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+  svwrite_ver_za16_u16_vg4(2, base, v16x4); // expected-error {{argument value 2 is outside the valid range [0, 1]}}
+
+  svwrite_hor_za32_u32_vg2(4, base, v32x2); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+  svwrite_ver_za32_u32_vg2(4, base, v32x2); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+  svwrite_hor_za32_u32_vg4(4, base, v32x4); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+  svwrite_ver_za32_u32_vg4(4, base, v32x4); // expected-error {{argument value 4 is outside the valid range [0, 3]}}
+
+  svwrite_hor_za64_u64_vg2(8, base, v64x2); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+  svwrite_ver_za64_u64_vg2(8, base, v64x2); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+  svwrite_hor_za64_u64_vg4(8, base, v64x4); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+  svwrite_ver_za64_u64_vg4(8, base, v64x4); // expected-error {{argument value 8 is outside the valid range [0, 7]}}
+}
+
 void test_outer_product(svbool_t pred, svint16_t s16, svuint16_t u16, svint32_t s32, svuint32_t u32) __arm_streaming __arm_shared_za {
   // Test Tile Range
   svmopa_za32_u16_m(4, pred, pred, u16, u16); // expected-error {{argument value 4 is outside the valid range [0, 3]}}

llvm/lib/Target/AArch64/SMEInstrFormats.td

@@ -3686,8 +3686,14 @@ multiclass sme2_mova_vec_to_array_vg2_multi<string mnemonic, SDPatternOperator i
 
   def NAME # _PSEUDO : sme2_move_to_za_pseudo<NAME, sme_elm_idx0_7, ZZ_d_mul_r, SMEMatrixArray>;
 
-  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv2i64, sme_elm_idx0_7, tileslice16>;
-  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv2f64, sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv16i8,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv8i16,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv8f16,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv8bf16, sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv4i32,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv4f32,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv2i64,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x2_Multi_Pat<NAME, intrinsic, nxv2f64,  sme_elm_idx0_7, tileslice16>;
 
   defm : sme2_mova_vec_to_tile_or_array_aliases<0, !cast<Instruction>(NAME),
                                                 MatrixOp8,
@@ -3779,8 +3785,14 @@ multiclass sme2_mova_vec_to_array_vg4_multi<string mnemonic, SDPatternOperator i
 
   def NAME # _PSEUDO : sme2_move_to_za_pseudo<NAME, sme_elm_idx0_7, ZZZZ_d_mul_r, SMEMatrixArray>;
 
-  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv2i64, sme_elm_idx0_7, tileslice16>;
-  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv2f64, sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv16i8,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv8i16,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv8f16,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv8bf16, sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv4i32,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv4f32,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv2i64,  sme_elm_idx0_7, tileslice16>;
+  def : SME2_ZA_VG1x4_Multi_Pat<NAME, intrinsic, nxv2f64,  sme_elm_idx0_7, tileslice16>;
 
   defm : sme2_mova_vec_to_tile_or_array_aliases<0, !cast<Instruction>(NAME),
                                                 MatrixOp8,