[BPF] Make -mcpu=v3 as the default (#107008)

Before llvm20, (void)__sync_fetch_and_add(...) always generates locked
xadd insns. In linux kernel upstream discussion [1], it is found that
for arm64 architecture, the original semantics of
(void)__sync_fetch_and_add(...), i.e., __atomic_fetch_add(...), is
preferred in order for jit to emit proper native barrier insns.

In llvm commits [2] and [3], (void)__sync_fetch_and_add(...) will
generate the following insns:
  - for cpu v1/v2: locked xadd insns to keep backward compatibility
  - for cpu v3/v4: __atomic_fetch_add() insns

To ensure proper barrier semantics for (void)__sync_fetch_and_add(...),
cpu v3/v4 is recommended.

This patch enables cpu=v3 as the default cpu version. For users wanting
to use cpu v1, -mcpu=v1 needs to be explicitly added to clang/llc
command line.

  [1]
https://lore.kernel.org/bpf/[email protected]/T/#mb68d67bc8f39e35a0c3db52468b9de59b79f021f
  [2] #101428
  [3] #106494

Author: yonghong-song

clang/lib/Basic/Targets/BPF.cpp

@@ -38,7 +38,10 @@ void BPFTargetInfo::getTargetDefines(const LangOptions &Opts,
 
   Builder.defineMacro("__BPF_FEATURE_ADDR_SPACE_CAST");
 
-  if (CPU.empty() || CPU == "generic" || CPU == "v1") {
+  if (CPU.empty())
+    CPU = "v3";
+
+  if (CPU == "generic" || CPU == "v1") {
     Builder.defineMacro("__BPF_CPU_VERSION__", "1");
     return;
   }

clang/test/Preprocessor/bpf-predefined-macros.c

@@ -1,5 +1,5 @@
-// RUN: %clang -E -target bpfel -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_NO %s
-// RUN: %clang -E -target bpfeb -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_NO %s
+// RUN: %clang -E -target bpfel -mcpu=v1 -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_NO %s
+// RUN: %clang -E -target bpfeb -mcpu=v1 -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_NO %s
 // RUN: %clang -E -target bpfel -mcpu=v1 -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_V1 %s
 // RUN: %clang -E -target bpfel -mcpu=v2 -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_V2 %s
 // RUN: %clang -E -target bpfel -mcpu=v3 -x c -o - %s | FileCheck -check-prefix=CHECK -check-prefix=CPU_V3 %s

llvm/lib/Target/BPF/BPFSubtarget.cpp

@@ -65,6 +65,8 @@ void BPFSubtarget::initializeEnvironment() {
 }
 
 void BPFSubtarget::initSubtargetFeatures(StringRef CPU, StringRef FS) {
+  if (CPU.empty())
+    CPU = "v3";
   if (CPU == "probe")
     CPU = sys::detail::getHostCPUNameForBPF();
   if (CPU == "generic" || CPU == "v1")

llvm/test/CodeGen/BPF/32-bit-subreg-cond-select.ll

@@ -1,4 +1,4 @@
-; RUN: llc -O2 -march=bpfel -mattr=+alu32 < %s | FileCheck %s
+; RUN: llc -O2 -march=bpfel -mcpu=v1 -mattr=+alu32 < %s | FileCheck %s
 ;
 ; unsigned int select_cc_32 (unsigned a, unsigned b, int c, int d)
 ; {

llvm/test/CodeGen/BPF/CORE/field-reloc-bitfield-1-bpfeb.ll

@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 3
-; RUN: opt -O2 -S < %s | llc -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
-; RUN: opt -O2 -S < %s | llc -mattr=+alu32 -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
+; RUN: opt -O2 -S < %s | llc -mcpu=v1 -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
+; RUN: opt -O2 -S < %s | llc -mcpu=v1 -mattr=+alu32 -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
 ; Source code:
 ;   struct s {
 ;     unsigned long long f1;

llvm/test/CodeGen/BPF/CORE/field-reloc-bitfield-1.ll

@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 3
-; RUN: opt -O2 -S < %s | llc -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
-; RUN: opt -O2 -S < %s | llc -mattr=+alu32 -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
+; RUN: opt -O2 -S < %s | llc -mcpu=v1 -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
+; RUN: opt -O2 -S < %s | llc -mcpu=v1 -mattr=+alu32 -filetype=asm | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
 ; Source code:
 ;   struct s {
 ;     unsigned long long f1;

llvm/test/CodeGen/BPF/CORE/field-reloc-bitfield-2-bpfeb.ll

@@ -1,6 +1,6 @@
 ; RUN: opt -O2 %s | llvm-dis > %t1
-; RUN: llc -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EB,CHECK-ALU64 %s
-; RUN: llc -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EB,CHECK-ALU32 %s
+; RUN: llc -mcpu=v1 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EB,CHECK-ALU64 %s
+; RUN: llc -mcpu=v1 -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EB,CHECK-ALU32 %s
 ; Source code:
 ;   struct s {
 ;     char f1;

llvm/test/CodeGen/BPF/CORE/field-reloc-bitfield-2.ll

@@ -1,6 +1,6 @@
 ; RUN: opt -O2 %s | llvm-dis > %t1
-; RUN: llc -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EL,CHECK-ALU64 %s
-; RUN: llc -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EL,CHECK-ALU32 %s
+; RUN: llc -mcpu=v1 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EL,CHECK-ALU64 %s
+; RUN: llc -mcpu=v1 -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-EL,CHECK-ALU32 %s
 ; Source code:
 ;   struct s {
 ;     char f1;

llvm/test/CodeGen/BPF/CORE/intrinsic-fieldinfo-byte-size-1.ll

@@ -1,6 +1,6 @@
 ; RUN: opt -O2 %s | llvm-dis > %t1
-; RUN: llc -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
-; RUN: llc -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
+; RUN: llc -mcpu=v1 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
+; RUN: llc -mcpu=v1 -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
 ; Source code:
 ;   typedef struct s1 { int a1:7; int a2:4; int a3:5; int a4:16;} __s1;
 ;   union u1 { int b1; __s1 b2; };

llvm/test/CodeGen/BPF/CORE/intrinsic-fieldinfo-byte-size-2.ll

@@ -1,6 +1,6 @@
 ; RUN: opt -O2 %s | llvm-dis > %t1
-; RUN: llc -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
-; RUN: llc -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
+; RUN: llc -mcpu=v1 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU64 %s
+; RUN: llc -mcpu=v1 -mattr=+alu32 -filetype=asm -o - %t1 | FileCheck -check-prefixes=CHECK,CHECK-ALU32 %s
 ; Source code:
 ;   typedef struct s1 { int a1; char a2; } __s1;
 ;   union u1 { int b1; __s1 b2; };