[clang] [AArch64] Enable vscale_range with +sme (PR #124466)

[David Green via cfe-commits]

https://github.com/davemgreen created https://github.com/llvm/llvm-project/pull/124466

If we have +sme but not +sve, we would not set vscale_range on functions. It should be valid to apply it with the same range with just +sme, which can help improve the performance of generated code.

>From 7919c40e6e06d5a5adfbd418f275aad9852de274 Mon Sep 17 00:00:00 2001
From: David Green <david.green at arm.com>
Date: Sun, 26 Jan 2025 13:47:58 +0000
Subject: [PATCH] [AArch64] Enable vscale_range with +sme

If we have +sme but not +sve, we would not set vscale_range on functions. It
should be valid to apply it with the same range with just +sme, which can help
improve the performance of generated code.
---
 clang/lib/Basic/Targets/AArch64.cpp              |  2 +-
 .../AArch64/sme-intrinsics/aarch64-sme-attrs.cpp | 16 ++++++++--------
 .../CodeGen/arm-sve-vector-bits-vscale-range.c   |  1 +
 3 files changed, 10 insertions(+), 9 deletions(-)

diff --git a/clang/lib/Basic/Targets/AArch64.cpp b/clang/lib/Basic/Targets/AArch64.cpp
index 0b899137bbb5c7..e9afe0fb4ee456 100644
--- a/clang/lib/Basic/Targets/AArch64.cpp
+++ b/clang/lib/Basic/Targets/AArch64.cpp
@@ -708,7 +708,7 @@ AArch64TargetInfo::getVScaleRange(const LangOptions &LangOpts) const {
     return std::pair<unsigned, unsigned>(
         LangOpts.VScaleMin ? LangOpts.VScaleMin : 1, LangOpts.VScaleMax);
 
-  if (hasFeature("sve"))
+  if (hasFeature("sve") || hasFeature("sme"))
     return std::pair<unsigned, unsigned>(1, 16);
 
   return std::nullopt;
diff --git a/clang/test/CodeGen/AArch64/sme-intrinsics/aarch64-sme-attrs.cpp b/clang/test/CodeGen/AArch64/sme-intrinsics/aarch64-sme-attrs.cpp
index 54762c8b414124..cabd9d01e46652 100644
--- a/clang/test/CodeGen/AArch64/sme-intrinsics/aarch64-sme-attrs.cpp
+++ b/clang/test/CodeGen/AArch64/sme-intrinsics/aarch64-sme-attrs.cpp
@@ -300,19 +300,19 @@ int test_variadic_template() __arm_inout("za") {
               preserves_za_decl);
 }
 
-// CHECK: attributes #[[SM_ENABLED]] = { mustprogress noinline nounwind "aarch64_pstate_sm_enabled" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[SM_ENABLED]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_pstate_sm_enabled" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
 // CHECK: attributes #[[NORMAL_DECL]] = { "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
 // CHECK: attributes #[[SM_ENABLED_DECL]] = { "aarch64_pstate_sm_enabled" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[SM_COMPATIBLE]] = { mustprogress noinline nounwind "aarch64_pstate_sm_compatible" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[SM_COMPATIBLE]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_pstate_sm_compatible" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
 // CHECK: attributes #[[SM_COMPATIBLE_DECL]] = { "aarch64_pstate_sm_compatible" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[SM_BODY]] = { mustprogress noinline nounwind "aarch64_pstate_sm_body" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[ZA_SHARED]] = { mustprogress noinline nounwind "aarch64_inout_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[SM_BODY]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_pstate_sm_body" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[ZA_SHARED]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_inout_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
 // CHECK: attributes #[[ZA_SHARED_DECL]] = { "aarch64_inout_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[ZA_PRESERVED]] = { mustprogress noinline nounwind "aarch64_preserves_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[ZA_PRESERVED]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_preserves_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
 // CHECK: attributes #[[ZA_PRESERVED_DECL]] = { "aarch64_preserves_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[ZA_NEW]] = { mustprogress noinline nounwind "aarch64_new_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[ZA_AGNOSTIC]] = { mustprogress noinline nounwind "aarch64_za_state_agnostic" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
-// CHECK: attributes #[[NORMAL_DEF]] = { mustprogress noinline nounwind "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[ZA_NEW]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_new_za" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[ZA_AGNOSTIC]] = { mustprogress noinline nounwind vscale_range(1,16) "aarch64_za_state_agnostic" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
+// CHECK: attributes #[[NORMAL_DEF]] = { mustprogress noinline nounwind vscale_range(1,16) "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+bf16,+sme" }
 // CHECK: attributes #[[SM_ENABLED_CALL]] = { "aarch64_pstate_sm_enabled" }
 // CHECK: attributes #[[SM_COMPATIBLE_CALL]] = { "aarch64_pstate_sm_compatible" }
 // CHECK: attributes #[[SM_BODY_CALL]] = { "aarch64_pstate_sm_body" }
diff --git a/clang/test/CodeGen/arm-sve-vector-bits-vscale-range.c b/clang/test/CodeGen/arm-sve-vector-bits-vscale-range.c
index bd424172a18650..66b67b036ef4ca 100644
--- a/clang/test/CodeGen/arm-sve-vector-bits-vscale-range.c
+++ b/clang/test/CodeGen/arm-sve-vector-bits-vscale-range.c
@@ -13,6 +13,7 @@
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2 -mvscale-min=1 -mvscale-max=0 -emit-llvm -o - %s | FileCheck %s --check-prefix=CHECK-UNBOUNDED
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -mvscale-min=1 -mvscale-max=0 -emit-llvm -o - %s | FileCheck %s --check-prefix=CHECK-UNBOUNDED
 // RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve -emit-llvm -o - %s | FileCheck %s --check-prefix=CHECK-NONE
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme -emit-llvm -o - %s | FileCheck %s --check-prefix=CHECK-NONE
 
 // CHECK-LABEL: @func() #0
 // CHECK: attributes #0 = { {{.*}} vscale_range([[#VBITS]],[[#VBITS]]) {{.*}} }