Skip to content

[clang][llvm][aarch64] Add aarch64_sme_in_streaming_mode intrinsic #120265

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
Jan 7, 2025
Merged

[clang][llvm][aarch64] Add aarch64_sme_in_streaming_mode intrinsic #120265

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
898c30b
[clang][llvm][aarch64] Add aarch64_sme_in_streaming_mode intrinsic
NickGuy-Arm Dec 13, 2024
20701ac
Remove redundant __arm_in_streaming_mode declaration
NickGuy-Arm Dec 18, 2024
453324a
Replaced auto with actual types
NickGuy-Arm Dec 18, 2024
0eca79f
Address comments
NickGuy-Arm Dec 18, 2024
afaf447
Address nits and update test
NickGuy-Arm Jan 6, 2025
ba489f4
Format
NickGuy-Arm Jan 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 15 additions & 0 deletions clang/lib/CodeGen/CGBuiltin.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11285,6 +11285,21 @@ Value *CodeGenFunction::EmitAArch64SMEBuiltinExpr(unsigned BuiltinID,
if (Builtin->LLVMIntrinsic == 0)
return nullptr;

if (BuiltinID == SME::BI__builtin_sme___arm_in_streaming_mode) {
// If we already know the streaming mode, don't bother with the intrinsic
// and emit a constant instead
Comment on lines +11289 to +11290
Copy link
Member

@MacDue MacDue Dec 18, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(optional suggestion): Rather than doing this while emitting builtins you could always emit the intrinsic, then handle Intrinsic::aarch64_sme_in_streaming_mode in llvm/lib/Analysis/ConstantFolding.cpp (other target specific intrinsics are already handled there).

auto FD = cast<FunctionDecl>(CurFuncDecl);
Copy link
Collaborator

@sdesmalen-arm sdesmalen-arm Dec 18, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
auto FD = cast<FunctionDecl>(CurFuncDecl);
const auto *FD = cast<FunctionDecl>(CurFuncDecl);

if (const Type *Ty = FD->getType().getTypePtrOrNull())
if (const auto *FPT = Ty->getAs<FunctionProtoType>()) {
Copy link
Collaborator

@sdesmalen-arm sdesmalen-arm Dec 18, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
if (const Type *Ty = FD->getType().getTypePtrOrNull())
if (const auto *FPT = Ty->getAs<FunctionProtoType>()) {
if (const auto *FPT = F->getType()->getAs<FunctionProtoType>()) {

unsigned SMEAttrs = FPT->getAArch64SMEAttributes();
if (!(SMEAttrs & FunctionType::SME_PStateSMCompatibleMask)) {
bool IsStreamingMode =
Copy link
Collaborator

@sdesmalen-arm sdesmalen-arm Dec 18, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit:

Suggested change
bool IsStreamingMode =
bool IsStreaming =

SMEAttrs & FunctionType::SME_PStateSMEnabledMask;
return ConstantInt::getBool(Builder.getContext(), IsStreamingMode);
}
}
}

// Predicates must match the main datatype.
for (unsigned i = 0, e = Ops.size(); i != e; ++i)
if (auto PredTy = dyn_cast<llvm::VectorType>(Ops[i]->getType()))
Expand Down
55 changes: 40 additions & 15 deletions clang/test/CodeGen/AArch64/sme-intrinsics/acle_sme_state_funs.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,28 +6,53 @@

#include <arm_sme.h>

// CHECK-LABEL: @test_in_streaming_mode(
// CHECK-LABEL: @test_in_streaming_mode_streaming_compatible(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call i1 @llvm.aarch64.sme.in.streaming.mode()
// CHECK-NEXT: ret i1 [[TMP0]]
//
// CPP-CHECK-LABEL: @_Z22test_in_streaming_modev(
// CPP-CHECK-LABEL: @_Z43test_in_streaming_mode_streaming_compatiblev(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call i1 @llvm.aarch64.sme.in.streaming.mode()
// CPP-CHECK-NEXT: ret i1 [[TMP0]]
//
bool test_in_streaming_mode(void) __arm_streaming_compatible {
bool test_in_streaming_mode_streaming_compatible(void) __arm_streaming_compatible {
return __arm_in_streaming_mode();
}

// CHECK-LABEL: @test_in_streaming_mode_streaming(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret i1 true
//
// CPP-CHECK-LABEL: @_Z32test_in_streaming_mode_streamingv(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: ret i1 true
//
bool test_in_streaming_mode_streaming(void) __arm_streaming {
//
return __arm_in_streaming_mode();
}

// CHECK-LABEL: @test_in_streaming_mode_non_streaming(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret i1 false
//
// CPP-CHECK-LABEL: @_Z36test_in_streaming_mode_non_streamingv(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: ret i1 false
//
bool test_in_streaming_mode_non_streaming(void) {
return __arm_in_streaming_mode();
}

// CHECK-LABEL: @test_za_disable(
// CHECK-NEXT: entry:
// CHECK-NEXT: tail call void @__arm_za_disable() #[[ATTR5:[0-9]+]]
// CHECK-NEXT: tail call void @__arm_za_disable() #[[ATTR7:[0-9]+]]
// CHECK-NEXT: ret void
//
// CPP-CHECK-LABEL: @_Z15test_za_disablev(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: tail call void @__arm_za_disable() #[[ATTR5:[0-9]+]]
// CPP-CHECK-NEXT: tail call void @__arm_za_disable() #[[ATTR7:[0-9]+]]
// CPP-CHECK-NEXT: ret void
//
void test_za_disable(void) __arm_streaming_compatible {
Expand All @@ -36,14 +61,14 @@ void test_za_disable(void) __arm_streaming_compatible {

// CHECK-LABEL: @test_has_sme(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[TMP0:%.*]] = tail call aarch64_sme_preservemost_from_x2 { i64, i64 } @__arm_sme_state() #[[ATTR5]]
// CHECK-NEXT: [[TMP0:%.*]] = tail call aarch64_sme_preservemost_from_x2 { i64, i64 } @__arm_sme_state() #[[ATTR7]]
// CHECK-NEXT: [[TMP1:%.*]] = extractvalue { i64, i64 } [[TMP0]], 0
// CHECK-NEXT: [[TOBOOL_I:%.*]] = icmp slt i64 [[TMP1]], 0
// CHECK-NEXT: ret i1 [[TOBOOL_I]]
//
// CPP-CHECK-LABEL: @_Z12test_has_smev(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call aarch64_sme_preservemost_from_x2 { i64, i64 } @__arm_sme_state() #[[ATTR5]]
// CPP-CHECK-NEXT: [[TMP0:%.*]] = tail call aarch64_sme_preservemost_from_x2 { i64, i64 } @__arm_sme_state() #[[ATTR7]]
// CPP-CHECK-NEXT: [[TMP1:%.*]] = extractvalue { i64, i64 } [[TMP0]], 0
// CPP-CHECK-NEXT: [[TOBOOL_I:%.*]] = icmp slt i64 [[TMP1]], 0
// CPP-CHECK-NEXT: ret i1 [[TOBOOL_I]]
Expand All @@ -66,12 +91,12 @@ void test_svundef_za(void) __arm_streaming_compatible __arm_out("za") {

// CHECK-LABEL: @test_sc_memcpy(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memcpy(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memcpy(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CHECK-NEXT: ret ptr [[CALL]]
//
// CPP-CHECK-LABEL: @_Z14test_sc_memcpyPvPKvm(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memcpy(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memcpy(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CPP-CHECK-NEXT: ret ptr [[CALL]]
//
void *test_sc_memcpy(void *dest, const void *src, size_t n) __arm_streaming_compatible {
Expand All @@ -80,12 +105,12 @@ void *test_sc_memcpy(void *dest, const void *src, size_t n) __arm_streaming_comp

// CHECK-LABEL: @test_sc_memmove(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memmove(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memmove(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CHECK-NEXT: ret ptr [[CALL]]
//
// CPP-CHECK-LABEL: @_Z15test_sc_memmovePvPKvm(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memmove(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memmove(ptr noundef [[DEST:%.*]], ptr noundef [[SRC:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CPP-CHECK-NEXT: ret ptr [[CALL]]
//
void *test_sc_memmove(void *dest, const void *src, size_t n) __arm_streaming_compatible {
Expand All @@ -94,12 +119,12 @@ void *test_sc_memmove(void *dest, const void *src, size_t n) __arm_streaming_com

// CHECK-LABEL: @test_sc_memset(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memset(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memset(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CHECK-NEXT: ret ptr [[CALL]]
//
// CPP-CHECK-LABEL: @_Z14test_sc_memsetPvim(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memset(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memset(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CPP-CHECK-NEXT: ret ptr [[CALL]]
//
void *test_sc_memset(void *s, int c, size_t n) __arm_streaming_compatible {
Expand All @@ -108,12 +133,12 @@ void *test_sc_memset(void *s, int c, size_t n) __arm_streaming_compatible {

// CHECK-LABEL: @test_sc_memchr(
// CHECK-NEXT: entry:
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memchr(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memchr(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CHECK-NEXT: ret ptr [[CALL]]
//
// CPP-CHECK-LABEL: @_Z14test_sc_memchrPvim(
// CPP-CHECK-NEXT: entry:
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memchr(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR5]]
// CPP-CHECK-NEXT: [[CALL:%.*]] = tail call ptr @__arm_sc_memchr(ptr noundef [[S:%.*]], i32 noundef [[C:%.*]], i64 noundef [[N:%.*]]) #[[ATTR7]]
// CPP-CHECK-NEXT: ret ptr [[CALL]]
//
void *test_sc_memchr(void *s, int c, size_t n) __arm_streaming_compatible {
Expand Down
28 changes: 1 addition & 27 deletions llvm/test/CodeGen/AArch64/sme-intrinsics-state.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
; RUN: llc -mtriple=aarch64-linux-gnu -mattr=+sme -verify-machineinstrs < %s | FileCheck %s


define i1 @streaming_mode_st_compatible() #0 {
define i1 @streaming_mode_streaming_compatible() #0 {
; CHECK-LABEL: streaming_mode_st_compatible:
; CHECK: // %bb.0:
; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill
Expand All @@ -14,31 +14,5 @@ define i1 @streaming_mode_st_compatible() #0 {
ret i1 %mode
}

define i1 @streaming_mode_st_enabled() #1 {
; CHECK-LABEL: streaming_mode_st_enabled:
; CHECK: // %bb.0:
; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and w0, w0, #0x1
; CHECK-NEXT: ldr x30, [sp], #16 // 8-byte Folded Reload
; CHECK-NEXT: ret
%mode = tail call noundef i1 @llvm.aarch64.sme.in.streaming.mode()
ret i1 %mode
}

define i1 @streaming_mode_st_disabled() #2 {
; CHECK-LABEL: streaming_mode_st_disabled:
; CHECK: // %bb.0:
; CHECK-NEXT: str x30, [sp, #-16]! // 8-byte Folded Spill
; CHECK-NEXT: bl __arm_sme_state
; CHECK-NEXT: and w0, w0, #0x1
; CHECK-NEXT: ldr x30, [sp], #16 // 8-byte Folded Reload
; CHECK-NEXT: ret
%mode = tail call noundef i1 @llvm.aarch64.sme.in.streaming.mode()
ret i1 %mode
}


attributes #0 = {nounwind memory(none) "aarch64_pstate_sm_compatible"}
attributes #1 = {nounwind memory(none) "aarch64_pstate_sm_enabled"}
attributes #2 = {nounwind memory(none)}
Loading