[Clang][SME2] Add multi-vector zip & unzip builtins (#74841)

Adds the following SME2 builtins:
 - svzip (x2 & x4)
 - svzipq (x2 & x4)
 - svuzp (x2 & x4)
 - svuzpq (x2 & x4)

See https://github.com/ARM-software/acle/pull/217/files

Patch by David Sherwood <[email protected]>

Author: kmclaughlin-arm

clang/include/clang/Basic/arm_sve.td

@@ -2308,6 +2308,22 @@ let TargetGuard = "sme2" in {
   def SVQCVTN_U16_S64_X4 : SInst<"svqcvtn_u16[_{d}_x4]", "b4.d", "l", MergeNone, "aarch64_sve_sqcvtun_x4", [IsStreaming], []>;
 }
 
+//
+// Multi-vector zip/unzip
+//
+
+let TargetGuard = "sme2" in {
+  def SVZIP_X2  : SInst<"svzip[_{d}_x2]",  "22", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_zip_x2",  [IsStreaming], []>;
+  def SVZIPQ_X2 : SInst<"svzipq[_{d}_x2]", "22", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_zipq_x2", [IsStreaming], []>;
+  def SVZIP_X4  : SInst<"svzip[_{d}_x4]",  "44", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_zip_x4",  [IsStreaming], []>;
+  def SVZIPQ_X4 : SInst<"svzipq[_{d}_x4]", "44", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_zipq_x4", [IsStreaming], []>;
+
+  def SVUZP_X2  : SInst<"svuzp[_{d}_x2]",  "22", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_uzp_x2",  [IsStreaming], []>;
+  def SVUZPQ_X2 : SInst<"svuzpq[_{d}_x2]", "22", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_uzpq_x2", [IsStreaming], []>;
+  def SVUZP_X4  : SInst<"svuzp[_{d}_x4]",  "44", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_uzp_x4",  [IsStreaming], []>;
+  def SVUZPQ_X4 : SInst<"svuzpq[_{d}_x4]", "44", "cUcsUsiUilUlbhfd", MergeNone, "aarch64_sve_uzpq_x4", [IsStreaming], []>;
+}
+
 //
 // Multi-vector unpack
 //