[clang] [clang][CIR] Lower SVE svpfalse_b and add tests (PR #228789)

[Andrzej Warzyński via cfe-commits]

================
@@ -0,0 +1,45 @@
+// REQUIRES: aarch64-registered-target
+
+// DEFINE: %{optimize} = opt -passes=mem2reg -S
+
+// RUN: %if cir-enabled %{%clang_cc1_cg_arm64_sve                        -fclangir -emit-cir -disable-O0-optnone -o - %s                | FileCheck %s --check-prefixes=C,CIR %}
+// RUN: %if cir-enabled %{%clang_cc1_cg_arm64_sve -DSVE_OVERLOADED_FORMS -fclangir -emit-cir -disable-O0-optnone -o - %s                | FileCheck %s --check-prefixes=C,CIR %}
+
+// RUN: %if cir-enabled %{%clang_cc1_cg_arm64_sve                        -fclangir -emit-llvm -disable-O0-optnone -o -        %s | %{optimize} | FileCheck %s --check-prefixes=C,LLVM %}
+// RUN: %if cir-enabled %{%clang_cc1_cg_arm64_sve -DSVE_OVERLOADED_FORMS -fclangir -emit-llvm -disable-O0-optnone -o -        %s | %{optimize} | FileCheck %s --check-prefixes=C,LLVM %}
+// RUN: %if cir-enabled %{%clang_cc1_cg_arm64_sve -DSVE_OVERLOADED_FORMS -fclangir -emit-llvm -disable-O0-optnone -o - -x c++ %s | %{optimize} | FileCheck %s --check-prefixes=CPP,LLVM %}
+
+// RUN:                   %clang_cc1_cg_arm64_sve                                  -emit-llvm -disable-O0-optnone -o -        %s | %{optimize} | FileCheck %s --check-prefixes=C,LLVM
+// RUN:                   %clang_cc1_cg_arm64_sve -DSVE_OVERLOADED_FORMS           -emit-llvm -disable-O0-optnone -o -        %s | %{optimize} | FileCheck %s --check-prefixes=C,LLVM
+// RUN:                   %clang_cc1_cg_arm64_sve -DSVE_OVERLOADED_FORMS           -emit-llvm -disable-O0-optnone -o - -x c++ %s | %{optimize} | FileCheck %s --check-prefixes=CPP,LLVM
+
+//=============================================================================
+// NOTES
+//
+// Tests for SVE PFALSE intrinsics (svpfalse_b only; svpfalse_c is SVE2.1)
+//=============================================================================
+
+#include <arm_sve.h>
+
+#if defined __ARM_FEATURE_SME
+#define MODE_ATTR __arm_streaming
+#else
+#define MODE_ATTR
+#endif
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+// C-LABEL: @test_svpfalse_b(
+// CPP-LABEL: @_Z15test_svpfalse_bv(
+svbool_t test_svpfalse_b(void) MODE_ATTR
+{
+// CIR:     cir.const #cir.zero : !cir.vector<[16] x !cir.int<u, 1>>
----------------
banach-space wrote:

We didn't do that for the NEON tests and I wouldn't do it here. 

Ultimately, if the LLVM output is correct than it's highly unlikely that there's something incorrect in CIR output and hence minimal checks like this should be fine.

No objections to adding checks for the return value, but I wouldn't consider it as a requirement.

https://github.com/llvm/llvm-project/pull/228789