[PATCH 0/8] perf build: Support building with Clang

[PATCH 0/8] perf build: Support building with Clang

[Leo Yan]

This series adds support for building perf with Clang, following the
discussion in [1]. Because the Android NDK uses Clang by default,
enabling Clang builds naturally covers Android as well.

The series is organized as follows:

  - Patches 1–2: Fix two bugs found with Clang.
  - Patches 3–6: Address Clang build warnings. Because these warnings
    do not break the build, no Fixes tag is added to avoid backporting.
  - Patches 7–8: Enable Clang in the Makefile and update the
    documentation.

Testing:

  - Clang 15.0.7 on Ubuntu 22.04.5: native and cross-compiling (aarch64)
  - Clang 18.1.3 on Ubuntu 24.04.2: native and cross-compiling (aarch64)
  - Android NDK r27d (latest LTS): cross-compiling (aarch64)

[1] https://lore.kernel.org/linux-perf-users/20240715143342.52236-1-leo.yan@arm.com/

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
Leo Yan (8):
      tools build: Align warning options with perf
      perf python: split Clang options when invoking Popen
      bpftool: Conditionally add -Wformat-signedness flag
      perf test coresight: Dismiss clang warning for memcpy thread
      perf test coresight: Dismiss clang warning for thread loop
      perf test coresight: Dismiss clang warning for unroll loop thread
      perf build: Support build with clang
      perf docs: Document building with Clang

 tools/bpf/bpftool/Makefile                         | 12 +++-
 tools/build/feature/Makefile                       |  4 +-
 tools/perf/Documentation/Build.txt                 | 18 +++++
 tools/perf/Documentation/android.txt               | 82 +++-------------------
 tools/perf/Makefile.config                         | 32 ++++++++-
 .../shell/coresight/memcpy_thread/memcpy_thread.c  |  2 +
 .../shell/coresight/thread_loop/thread_loop.c      |  4 +-
 .../unroll_loop_thread/unroll_loop_thread.c        |  4 +-
 tools/perf/util/setup.py                           |  5 +-
 9 files changed, 81 insertions(+), 82 deletions(-)
---
base-commit: c17dda8013495d8132c976cbf349be9949d0fbd1
change-id: 20250915-perf_build_android_ndk-500b53bea895

Best regards,
-- 
Leo Yan <leo.yan@arm.com>

[PATCH 1/8] tools build: Align warning options with perf

[Leo Yan]

The feature test programs are built without enabling '-Wall -Werror'
options. As a result, a feature may appear to be available, but later
building in perf can fail with stricter checks.

Make the feature test program use the same warning options as perf.

Fixes: 1925459b4d92 ("tools build: Fix feature Makefile issues with 'O='")
Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/build/feature/Makefile | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/build/feature/Makefile b/tools/build/feature/Makefile
index b41a42818d8ac232ade78ecb41363d26ce2a9471..bd615a708a0aa89ddbe87401f04bd736e384a9c4 100644
--- a/tools/build/feature/Makefile
+++ b/tools/build/feature/Makefile
@@ -316,10 +316,10 @@ $(OUTPUT)test-libcapstone.bin:
 	$(BUILD) # -lcapstone provided by $(FEATURE_CHECK_LDFLAGS-libcapstone)
 
 $(OUTPUT)test-compile-32.bin:
-	$(CC) -m32 -o $@ test-compile.c
+	$(CC) -m32 -Wall -Werror -o $@ test-compile.c
 
 $(OUTPUT)test-compile-x32.bin:
-	$(CC) -mx32 -o $@ test-compile.c
+	$(CC) -mx32 -Wall -Werror -o $@ test-compile.c
 
 $(OUTPUT)test-zlib.bin:
 	$(BUILD) -lz

-- 
2.34.1

[PATCH 2/8] perf python: split Clang options when invoking Popen

[Leo Yan]

When passing a list to subprocess.Popen, each element maps to one argv
token. Current code bundles multiple Clang flags into a single element,
something like:

  cmd = ['clang',
         '--target=x86_64-linux-gnu -fintegrated-as -Wno-cast-function-type-mismatch',
	 'test-hello.c']

So Clang only sees one long, invalid option instead of separate flags,
as a result, the script cannot capture any log via PIPE.

Fix this by using shlex.split() to separate the string so each option
becomes its own argv element. The fixed list will be:

  cmd = ['clang',
         '--target=x86_64-linux-gnu',
	 '-fintegrated-as',
	 '-Wno-cast-function-type-mismatch',
	 'test-hello.c']

Fixes: 09e6f9f98370 ("perf python: Fix splitting CC into compiler and options")
Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/util/setup.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/tools/perf/util/setup.py b/tools/perf/util/setup.py
index dd289d15acfd62ff058bbaed7e565bb958e3a3c8..9cae2c472f4ad4d9579e8528b8bb0152df6fe20e 100644
--- a/tools/perf/util/setup.py
+++ b/tools/perf/util/setup.py
@@ -1,6 +1,7 @@
 from os import getenv, path
 from subprocess import Popen, PIPE
 from re import sub
+import shlex
 
 cc = getenv("CC")
 assert cc, "Environment variable CC not set"
@@ -22,7 +23,9 @@ assert srctree, "Environment variable srctree, for the Linux sources, not set"
 src_feature_tests  = f'{srctree}/tools/build/feature'
 
 def clang_has_option(option):
-    cc_output = Popen([cc, cc_options + option, path.join(src_feature_tests, "test-hello.c") ], stderr=PIPE).stderr.readlines()
+    cmd = shlex.split(f"{cc} {cc_options} {option}")
+    cmd.append(path.join(src_feature_tests, "test-hello.c"))
+    cc_output = Popen(cmd, stderr=PIPE).stderr.readlines()
     return [o for o in cc_output if ((b"unknown argument" in o) or (b"is not supported" in o) or (b"unknown warning option" in o))] == [ ]
 
 if cc_is_clang:

-- 
2.34.1

[PATCH 3/8] bpftool: Conditionally add -Wformat-signedness flag

[Leo Yan]

clang-18.1.3 on Ubuntu 24.04.2 reports warning:

  warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]

Conditionally add the option only when it is supported by compiler.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/bpf/bpftool/Makefile | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/tools/bpf/bpftool/Makefile b/tools/bpf/bpftool/Makefile
index 9e9a5f006cd2aabe1e89bd83e394455c0d4473e0..948a0cc98b39d3f9afa0de73643eab04e8798ff5 100644
--- a/tools/bpf/bpftool/Makefile
+++ b/tools/bpf/bpftool/Makefile
@@ -64,11 +64,21 @@ $(LIBBPF_BOOTSTRAP)-clean: FORCE | $(LIBBPF_BOOTSTRAP_OUTPUT)
 prefix ?= /usr/local
 bash_compdir ?= /usr/share/bash-completion/completions
 
+try-run = $(shell set -e;		\
+	if ($(1)) >/dev/null 2>&1;	\
+	then echo "$(2)";		\
+	else echo "$(3)";		\
+	fi)
+
+__cc-option = $(call try-run,\
+	$(1) -Werror $(2) -c -x c /dev/null -o /dev/null,$(2),)
+cc-option = $(call __cc-option, $(CC),$(1))
+
 CFLAGS += -O2
 CFLAGS += -W
 CFLAGS += -Wall
 CFLAGS += -Wextra
-CFLAGS += -Wformat-signedness
+CFLAGS += $(call cc-option,-Wformat-signedness)
 CFLAGS += -Wno-unused-parameter
 CFLAGS += -Wno-missing-field-initializers
 CFLAGS += $(filter-out -Wswitch-enum -Wnested-externs,$(EXTRA_WARNINGS))

-- 
2.34.1

[PATCH 4/8] perf test coresight: Dismiss clang warning for memcpy thread

[Leo Yan]

clang-18.1.3 on Ubuntu 24.04.2 reports warning:

  memcpy_thread.c:30:1: warning: non-void function does not return a value in all control paths [-Wreturn-type]
     30 | }
        | ^

Dismiss the warning with returning NULL from the thread function.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c b/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c
index 5f886cd09e6b3a62b5690dade94f1f8cae3279d2..7e879217be30a86431989dbf1f36d2134ef259cc 100644
--- a/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c
+++ b/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c
@@ -27,6 +27,8 @@ static void *thrfn(void *arg)
 	}
 	for (i = 0; i < len; i++)
 		memcpy(dst, src, a->size * 1024);
+
+	return NULL;
 }
 
 static pthread_t new_thr(void *(*fn) (void *arg), void *arg)

-- 
2.34.1

[PATCH 5/8] perf test coresight: Dismiss clang warning for thread loop

[Leo Yan]

clang-18.1.3 on Ubuntu 24.04.2 reports warning:

  thread_loop.c:41:23: warning: value size does not match register size specified by the constraint and modifier [-Wasm-operand-widths]
     41 |                 : /* in */ [i] "r" (i), [len] "r" (len)
        |                                     ^
  thread_loop.c:37:8: note: use constraint modifier "w"
     37 |                 "add %[i], %[i], #1\n"
        |                      ^~~~
        |                      %w[i]
  thread_loop.c:41:23: warning: value size does not match register size specified by the constraint and modifier [-Wasm-operand-widths]
     41 |                 : /* in */ [i] "r" (i), [len] "r" (len)
        |                                     ^
  thread_loop.c:37:14: note: use constraint modifier "w"
     37 |                 "add %[i], %[i], #1\n"
        |                            ^~~~
        |                            %w[i]
  thread_loop.c:41:23: warning: value size does not match register size specified by the constraint and modifier [-Wasm-operand-widths]
     41 |                 : /* in */ [i] "r" (i), [len] "r" (len)
        |                                     ^
  thread_loop.c:38:8: note: use constraint modifier "w"
     38 |                 "cmp %[i], %[len]\n"
        |                      ^~~~
        |                      %w[i]
  thread_loop.c:41:38: warning: value size does not match register size specified by the constraint and modifier [-Wasm-operand-widths]
     41 |                 : /* in */ [i] "r" (i), [len] "r" (len)
        |                                                    ^
  thread_loop.c:38:14: note: use constraint modifier "w"
     38 |                 "cmp %[i], %[len]\n"
        |                            ^~~~~~
        |                            %w[len]

Use the modifier "w" for 32-bit register access.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/tests/shell/coresight/thread_loop/thread_loop.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/perf/tests/shell/coresight/thread_loop/thread_loop.c b/tools/perf/tests/shell/coresight/thread_loop/thread_loop.c
index e05a559253ca9d9366ad321d520349042fb07fca..86f3f548b00631682767665fc5e9d5b8551a3634 100644
--- a/tools/perf/tests/shell/coresight/thread_loop/thread_loop.c
+++ b/tools/perf/tests/shell/coresight/thread_loop/thread_loop.c
@@ -34,8 +34,8 @@ static void *thrfn(void *arg)
 	}
 	asm volatile(
 		"loop:\n"
-		"add %[i], %[i], #1\n"
-		"cmp %[i], %[len]\n"
+		"add %w[i], %w[i], #1\n"
+		"cmp %w[i], %w[len]\n"
 		"blt loop\n"
 		: /* out */
 		: /* in */ [i] "r" (i), [len] "r" (len)

-- 
2.34.1

[PATCH 6/8] perf test coresight: Dismiss clang warning for unroll loop thread

[Leo Yan]

clang-18.1.3 on Ubuntu 24.04.2 reports warning:

  unroll_loop_thread.c:35:25: warning: value size does not match register size specified by the constraint and modifier [-Wasm-operand-widths]
     35 |                         : /* in */ [in] "r" (in)
        |                                              ^
  unroll_loop_thread.c:39:1: warning: non-void function does not return a value [-Wreturn-type]
     39 | }
        | ^

Use the modifier "w" for 32-bit register access and return NULL at the
end of thread function.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 .../tests/shell/coresight/unroll_loop_thread/unroll_loop_thread.c     | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/tools/perf/tests/shell/coresight/unroll_loop_thread/unroll_loop_thread.c b/tools/perf/tests/shell/coresight/unroll_loop_thread/unroll_loop_thread.c
index 0fc7bf1a25af3607b40f091f62176134ddb7f9f6..8f4e1c985ca38ab545a05189432a14a6d888d34c 100644
--- a/tools/perf/tests/shell/coresight/unroll_loop_thread/unroll_loop_thread.c
+++ b/tools/perf/tests/shell/coresight/unroll_loop_thread/unroll_loop_thread.c
@@ -20,7 +20,7 @@ static void *thrfn(void *arg)
 	for (i = 0; i < 10000; i++) {
 		asm volatile (
 // force an unroll of thia add instruction so we can test long runs of code
-#define SNIP1 "add %[in], %[in], #1\n"
+#define SNIP1 "add %w[in], %w[in], #1\n"
 // 10
 #define SNIP2 SNIP1 SNIP1 SNIP1 SNIP1 SNIP1 SNIP1 SNIP1 SNIP1 SNIP1 SNIP1
 // 100
@@ -36,6 +36,8 @@ static void *thrfn(void *arg)
 			: /* clobber */
 		);
 	}
+
+	return NULL;
 }
 
 static pthread_t new_thr(void *(*fn) (void *arg), void *arg)

-- 
2.34.1

[PATCH 7/8] perf build: Support build with clang

[Leo Yan]

Add support for building perf with clang. For cross compilation, the
Makefile dynamically selects target flag for corresponding arch.

This patch has been verified on x86_64 machine with Ubuntu distro, it
can build successfully for native target, and for cross building Arm64
and s390.

Example: native build on x86_64 / Ubuntu machine:

  $ HOSTCC=clang CC=clang CXX=clang++ make -C tools/perf

Example: cross building s390 target on x86_64 / Ubuntu machine:

  # Install x390x cross toolchain and headers
  $ sudo apt-get install gcc-s390x-linux-gnu g++-s390x-linux-gnu \
         libc6-dev-s390x-cross linux-libc-dev-s390x-cross

  # Build with clang
  $ HOSTCC=clang CC=clang CXX=clang++ \
    ARCH=s390 CROSS_COMPILE=s390x-linux-gnu- \
    make -C tools/perf NO_LIBELF=1 NO_LIBTRACEEVENT=1 NO_LIBPYTHON=1

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/Makefile.config | 32 +++++++++++++++++++++++++++++++-
 1 file changed, 31 insertions(+), 1 deletion(-)

diff --git a/tools/perf/Makefile.config b/tools/perf/Makefile.config
index 5a5832ee7b53c604ec0d90c5ca7e2bee5bfa6a17..a2b5c291a364cbf331c16384764520dfe330a04c 100644
--- a/tools/perf/Makefile.config
+++ b/tools/perf/Makefile.config
@@ -23,8 +23,38 @@ HOSTCFLAGS := $(filter-out -Wnested-externs,$(EXTRA_WARNINGS))
 # borrowed from kernel headers depends on it, e.g. put_unaligned_*().
 CFLAGS += -fno-strict-aliasing
 
-# Enabled Wthread-safety analysis for clang builds.
+# Set target flag and options when using clang as compiler.
 ifeq ($(CC_NO_CLANG), 0)
+  CLANG_TARGET_FLAGS_arm	:= arm-linux-gnueabi
+  CLANG_TARGET_FLAGS_arm64	:= aarch64-linux-gnu
+  CLANG_TARGET_FLAGS_m68k	:= m68k-linux-gnu
+  CLANG_TARGET_FLAGS_mips	:= mipsel-linux-gnu
+  CLANG_TARGET_FLAGS_powerpc	:= powerpc64le-linux-gnu
+  CLANG_TARGET_FLAGS_riscv	:= riscv64-linux-gnu
+  CLANG_TARGET_FLAGS_s390	:= s390x-linux-gnu
+  CLANG_TARGET_FLAGS_x86	:= x86_64-linux-gnu
+
+  # Default to host architecture if ARCH is not explicitly given.
+  ifeq ($(ARCH),)
+    CLANG_TARGET_FLAGS := $(shell $(CLANG) -print-target-triple)
+  else
+    CLANG_TARGET_FLAGS := $(CLANG_TARGET_FLAGS_$(ARCH))
+  endif
+
+  ifeq ($(CROSS_COMPILE),)
+    ifeq ($(CLANG_TARGET_FLAGS),)
+      $(error Specify CROSS_COMPILE or add CLANG_TARGET_FLAGS for $(ARCH))
+    else
+      CLANG_FLAGS += --target=$(CLANG_TARGET_FLAGS)
+    endif # CLANG_TARGET_FLAGS
+  else
+    CLANG_FLAGS += --target=$(notdir $(CROSS_COMPILE:%-=%))
+  endif # CROSS_COMPILE
+
+  CC := $(CLANG) $(CLANG_FLAGS) -fintegrated-as
+  CXX := $(CXX) $(CLANG_FLAGS) -fintegrated-as
+
+  # Enabled Wthread-safety analysis for clang builds.
   CFLAGS += -Wthread-safety
 endif
 

-- 
2.34.1

[PATCH 8/8] perf docs: Document building with Clang

[Leo Yan]

Add example commands for building perf with Clang.

Since recent Android NDK releases use Clang as the default compiler, a
separate Android specific document is no longer needed; point to the
general build documentation instead.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/Documentation/Build.txt   | 18 ++++++++
 tools/perf/Documentation/android.txt | 82 ++++--------------------------------
 2 files changed, 26 insertions(+), 74 deletions(-)

diff --git a/tools/perf/Documentation/Build.txt b/tools/perf/Documentation/Build.txt
index 83dc87c662b63ecc17553a15cc15a6b8d6f01d83..3e4104e605ac0d7d30b4408ef413cf1f90b034c1 100644
--- a/tools/perf/Documentation/Build.txt
+++ b/tools/perf/Documentation/Build.txt
@@ -99,3 +99,21 @@ configuration paths for cross building:
 In this case, the variable PKG_CONFIG_SYSROOT_DIR can be used alongside the
 variable PKG_CONFIG_LIBDIR or PKG_CONFIG_PATH to prepend the sysroot path to
 the library paths for cross compilation.
+
+5) Build with clang
+===================
+By default, the makefile uses GCC as compiler. With specifying environment
+variables HOSTCC, CC and CXX, it allows to build perf with clang.
+
+Using clang for native build:
+
+  $ HOSTCC=clang CC=clang CXX=clang++ make -C tools/perf
+
+Using clang for cross compilation:
+
+  $ HOSTCC=clang CC=clang CXX=clang++ \
+    make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- -C tools/perf \
+    NO_LIBELF=1 NO_LIBTRACEEVENT=1 NO_JEVENTS=1
+
+In the example above, due to lack libelf, python and libtraceevent for
+cross comiplation, disable the features accordingly.
diff --git a/tools/perf/Documentation/android.txt b/tools/perf/Documentation/android.txt
index 24a59998fc91e814ad96f658d3481d88d798b60c..e65204cf2921f6bd8a79875784c5b3d5487ce05d 100644
--- a/tools/perf/Documentation/android.txt
+++ b/tools/perf/Documentation/android.txt
@@ -1,78 +1,12 @@
 How to compile perf for Android
-=========================================
+===============================
 
-I. Set the Android NDK environment
-------------------------------------------------
+There have two ways to build perf and run it on Android.
 
-(a). Use the Android NDK
-------------------------------------------------
-1. You need to download and install the Android Native Development Kit (NDK).
-Set the NDK variable to point to the path where you installed the NDK:
-  export NDK=/path/to/android-ndk
+- The first method is to build perf with static linking, please refer to
+  Build.txt, section "4) Cross compilation" for how to build a static
+  perf binary.
 
-2. Set cross-compiling environment variables for NDK toolchain and sysroot.
-For arm:
-  export NDK_TOOLCHAIN=${NDK}/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-
-  export NDK_SYSROOT=${NDK}/platforms/android-24/arch-arm
-For x86:
-  export NDK_TOOLCHAIN=${NDK}/toolchains/x86-4.9/prebuilt/linux-x86_64/bin/i686-linux-android-
-  export NDK_SYSROOT=${NDK}/platforms/android-24/arch-x86
-
-This method is only tested for Android NDK versions Revision 11b and later.
-perf uses some bionic enhancements that are not included in prior NDK versions.
-You can use method (b) described below instead.
-
-(b). Use the Android source tree
------------------------------------------------
-1. Download the master branch of the Android source tree.
-Set the environment for the target you want using:
-  source build/envsetup.sh
-  lunch
-
-2. Build your own NDK sysroot to contain latest bionic changes and set the
-NDK sysroot environment variable.
-  cd ${ANDROID_BUILD_TOP}/ndk
-For arm:
-  ./build/tools/build-ndk-sysroot.sh --abi=arm
-  export NDK_SYSROOT=${ANDROID_BUILD_TOP}/ndk/build/platforms/android-3/arch-arm
-For x86:
-  ./build/tools/build-ndk-sysroot.sh --abi=x86
-  export NDK_SYSROOT=${ANDROID_BUILD_TOP}/ndk/build/platforms/android-3/arch-x86
-
-3. Set the NDK toolchain environment variable.
-For arm:
-  export NDK_TOOLCHAIN=${ANDROID_TOOLCHAIN}/arm-linux-androideabi-
-For x86:
-  export NDK_TOOLCHAIN=${ANDROID_TOOLCHAIN}/i686-linux-android-
-
-II. Compile perf for Android
-------------------------------------------------
-You need to run make with the NDK toolchain and sysroot defined above:
-For arm:
-  make WERROR=0 ARCH=arm CROSS_COMPILE=${NDK_TOOLCHAIN} EXTRA_CFLAGS="-pie --sysroot=${NDK_SYSROOT}"
-For x86:
-  make WERROR=0 ARCH=x86 CROSS_COMPILE=${NDK_TOOLCHAIN} EXTRA_CFLAGS="-pie --sysroot=${NDK_SYSROOT}"
-
-III. Install perf
------------------------------------------------
-You need to connect to your Android device/emulator using adb.
-Install perf using:
-  adb push perf /data/perf
-
-If you also want to use perf-archive you need busybox tools for Android.
-For installing perf-archive, you first need to replace #!/bin/bash with #!/system/bin/sh:
-  sed 's/#!\/bin\/bash/#!\/system\/bin\/sh/g' perf-archive >> /tmp/perf-archive
-  chmod +x /tmp/perf-archive
-  adb push /tmp/perf-archive /data/perf-archive
-
-IV. Environment settings for running perf
-------------------------------------------------
-Some perf features need environment variables to run properly.
-You need to set these before running perf on the target:
-  adb shell
-  # PERF_PAGER=cat
-
-IV. Run perf
-------------------------------------------------
-Run perf on your device/emulator to which you previously connected using adb:
-  # ./data/perf
+- The second method is to download Android NDK, then use the contained
+  clang compiler for building perf. Please refer to Build.txt, section
+  "5) Build with clang" for details.

-- 
2.34.1

Re: [PATCH 3/8] bpftool: Conditionally add -Wformat-signedness flag

[Quentin Monnet]

2025-09-25 11:26 UTC+0100 ~ Leo Yan <leo.yan@arm.com>
> clang-18.1.3 on Ubuntu 24.04.2 reports warning:
> 
>   warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
> 
> Conditionally add the option only when it is supported by compiler.
> 
> Signed-off-by: Leo Yan <leo.yan@arm.com>


Hi, how annoying is this warning? I'm asking because as far as I
understand, the option has been introduced in LLVM 19.1.0 [0] - the
latest being 21.1.0 already - so we won't need this check once distros
have transitioned, and I'm a bit reluctant to add it.

Quentin


[0]
https://github.com/llvm/llvm-project/commit/ea92b1f9d0fc31f1fd97ad04eb0412003a37cb0d

Re: [PATCH 3/8] bpftool: Conditionally add -Wformat-signedness flag

[Leo Yan]

On Thu, Sep 25, 2025 at 11:42:39AM +0100, Quentin Monnet wrote:
> 2025-09-25 11:26 UTC+0100 ~ Leo Yan <leo.yan@arm.com>
> > clang-18.1.3 on Ubuntu 24.04.2 reports warning:
> > 
> >   warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
> > 
> > Conditionally add the option only when it is supported by compiler.
> > 
> > Signed-off-by: Leo Yan <leo.yan@arm.com>
> 
> 
> Hi, how annoying is this warning?

Not too bad, it prints out 5~6 times warnings (see the build log below).
Though the warnings are a bit noisy, they does not break the build.

> I'm asking because as far as I
> understand, the option has been introduced in LLVM 19.1.0 [0] - the
> latest being 21.1.0 already - so we won't need this check once distros
> have transitioned, and I'm a bit reluctant to add it.

I agree the warning likely won't affect distros, as package maintainers
will hide it. But it can annoy _paranoid_ developers in day-to-day
builds ;)

As the warning doesn't block anything, so I am fine with not merging
this patch.

Thanks,
Leo

> [0]
> https://github.com/llvm/llvm-project/commit/ea92b1f9d0fc31f1fd97ad04eb0412003a37cb0d


---8<---

Auto-detecting system features:
...                         clang-bpf-co-re: [ on  ]
...                                    llvm: [ OFF ]
...                                  libcap: [ on  ]
...                                  libbfd: [ OFF ]
...                          libbfd-liberty: [ OFF ]
...                        libbfd-liberty-z: [ OFF ]
...                  disassembler-four-args: [ OFF ]
...                disassembler-init-styled: [ OFF ]
...                             libelf-zstd: [ o
  MKDIR   /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf
  MKDIR   /build/util/bpf_skel/.tmp/bootstrap/libbpf/
  MKDIR   /build/util/bpf_skel/.tmp/bootstrap/
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/hashmap.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/relo_core.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/libbpf_internal.h
warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
1 warning generated.
  GEN     /build/util/bpf_skel/.tmp/bootstrap/libbpf/bpf_helper_defs.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/bpf.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/libbpf.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/libbpf_legacy.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/btf.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/bpf_helpers.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/libbpf_common.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/bpf_tracing.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/bpf_core_read.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/bpf_endian.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/libbpf_version.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/skel_internal.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/usdt.bpf.h
  INSTALL /build/util/bpf_skel/.tmp/bootstrap/libbpf/include/bpf/bpf_helper_defs.h
  INSTALL libbpf_headers
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/libbpf.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/bpf.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/nlattr.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/btf.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/libbpf_errno.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/str_error.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/netlink.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/bpf_prog_linfo.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/libbpf_probes.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/hashmap.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/btf_dump.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/ringbuf.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/strset.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/gen_loader.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/relo_core.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/linker.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/usdt.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/zip.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/elf.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/features.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/btf_iter.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/btf_relocate.o
  LD      /build/util/bpf_skel/.tmp/bootstrap/libbpf/staticobjs/libbpf-in.o
  LINK    /build/util/bpf_skel/.tmp/bootstrap/libbpf/libbpf.a
  CC      /build/util/bpf_skel/.tmp/bootstrap/main.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/common.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/json_writer.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/gen.o
  CC      /build/util/bpf_skel/.tmp/bootstrap/btf.o
warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]warning:
unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
warning: unknown warning option '-Wformat-signedness' [-Wunknown-warning-option]
1 warning generated.
1 warning generated.
1 warning generated.
1 warning generated.
1 warning generated.
  LINK    /build/util/bpf_skel/.tmp/bootstrap/bpftool

Re: [PATCH 4/8] perf test coresight: Dismiss clang warning for memcpy thread

[Ian Rogers]

On Thu, Sep 25, 2025 at 3:26 AM Leo Yan <leo.yan@arm.com> wrote:
>
> clang-18.1.3 on Ubuntu 24.04.2 reports warning:
>
>   memcpy_thread.c:30:1: warning: non-void function does not return a value in all control paths [-Wreturn-type]
>      30 | }
>         | ^
>
> Dismiss the warning with returning NULL from the thread function.
>
> Signed-off-by: Leo Yan <leo.yan@arm.com>

lgtm, should this be moved into being a perf test workload as in
tools/perf/tests/workloads/ ?
There doesn't seem to be anything overly arch specific in the test and
being a workload avoids a cc dependency in the test, as well as
meaning we test with our usual compiler flags, etc.

Thanks,
Ian

> ---
>  tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c b/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c
> index 5f886cd09e6b3a62b5690dade94f1f8cae3279d2..7e879217be30a86431989dbf1f36d2134ef259cc 100644
> --- a/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c
> +++ b/tools/perf/tests/shell/coresight/memcpy_thread/memcpy_thread.c
> @@ -27,6 +27,8 @@ static void *thrfn(void *arg)
>         }
>         for (i = 0; i < len; i++)
>                 memcpy(dst, src, a->size * 1024);
> +
> +       return NULL;
>  }
>
>  static pthread_t new_thr(void *(*fn) (void *arg), void *arg)
>
> --
> 2.34.1
>

Re: [PATCH 8/8] perf docs: Document building with Clang

[Ian Rogers]

On Thu, Sep 25, 2025 at 3:27 AM Leo Yan <leo.yan@arm.com> wrote:
>
> Add example commands for building perf with Clang.
>
> Since recent Android NDK releases use Clang as the default compiler, a
> separate Android specific document is no longer needed; point to the
> general build documentation instead.
>
> Signed-off-by: Leo Yan <leo.yan@arm.com>
> ---
>  tools/perf/Documentation/Build.txt   | 18 ++++++++
>  tools/perf/Documentation/android.txt | 82 ++++--------------------------------
>  2 files changed, 26 insertions(+), 74 deletions(-)
>
> diff --git a/tools/perf/Documentation/Build.txt b/tools/perf/Documentation/Build.txt
> index 83dc87c662b63ecc17553a15cc15a6b8d6f01d83..3e4104e605ac0d7d30b4408ef413cf1f90b034c1 100644
> --- a/tools/perf/Documentation/Build.txt
> +++ b/tools/perf/Documentation/Build.txt
> @@ -99,3 +99,21 @@ configuration paths for cross building:
>  In this case, the variable PKG_CONFIG_SYSROOT_DIR can be used alongside the
>  variable PKG_CONFIG_LIBDIR or PKG_CONFIG_PATH to prepend the sysroot path to
>  the library paths for cross compilation.
> +
> +5) Build with clang
> +===================
> +By default, the makefile uses GCC as compiler. With specifying environment
> +variables HOSTCC, CC and CXX, it allows to build perf with clang.
> +
> +Using clang for native build:
> +
> +  $ HOSTCC=clang CC=clang CXX=clang++ make -C tools/perf
> +
> +Using clang for cross compilation:
> +
> +  $ HOSTCC=clang CC=clang CXX=clang++ \
> +    make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- -C tools/perf \
> +    NO_LIBELF=1 NO_LIBTRACEEVENT=1 NO_JEVENTS=1

The three NO_-s here are going to cripple the build quite a lot, I
wonder if we can list package dependencies to install and failing that
use the NO_-s.

> +
> +In the example above, due to lack libelf, python and libtraceevent for
> +cross comiplation, disable the features accordingly.

nit: s/comiplation/compilation/

> diff --git a/tools/perf/Documentation/android.txt b/tools/perf/Documentation/android.txt
> index 24a59998fc91e814ad96f658d3481d88d798b60c..e65204cf2921f6bd8a79875784c5b3d5487ce05d 100644
> --- a/tools/perf/Documentation/android.txt
> +++ b/tools/perf/Documentation/android.txt
> @@ -1,78 +1,12 @@
>  How to compile perf for Android
> -=========================================
> +===============================
>
> -I. Set the Android NDK environment
> -------------------------------------------------
> +There have two ways to build perf and run it on Android.

nit: s/There have/There are/

Thanks for doing this series!
Ian

>
> -(a). Use the Android NDK
> -------------------------------------------------
> -1. You need to download and install the Android Native Development Kit (NDK).
> -Set the NDK variable to point to the path where you installed the NDK:
> -  export NDK=/path/to/android-ndk
> +- The first method is to build perf with static linking, please refer to
> +  Build.txt, section "4) Cross compilation" for how to build a static
> +  perf binary.
>
> -2. Set cross-compiling environment variables for NDK toolchain and sysroot.
> -For arm:
> -  export NDK_TOOLCHAIN=${NDK}/toolchains/arm-linux-androideabi-4.9/prebuilt/linux-x86_64/bin/arm-linux-androideabi-
> -  export NDK_SYSROOT=${NDK}/platforms/android-24/arch-arm
> -For x86:
> -  export NDK_TOOLCHAIN=${NDK}/toolchains/x86-4.9/prebuilt/linux-x86_64/bin/i686-linux-android-
> -  export NDK_SYSROOT=${NDK}/platforms/android-24/arch-x86
> -
> -This method is only tested for Android NDK versions Revision 11b and later.
> -perf uses some bionic enhancements that are not included in prior NDK versions.
> -You can use method (b) described below instead.
> -
> -(b). Use the Android source tree
> ------------------------------------------------
> -1. Download the master branch of the Android source tree.
> -Set the environment for the target you want using:
> -  source build/envsetup.sh
> -  lunch
> -
> -2. Build your own NDK sysroot to contain latest bionic changes and set the
> -NDK sysroot environment variable.
> -  cd ${ANDROID_BUILD_TOP}/ndk
> -For arm:
> -  ./build/tools/build-ndk-sysroot.sh --abi=arm
> -  export NDK_SYSROOT=${ANDROID_BUILD_TOP}/ndk/build/platforms/android-3/arch-arm
> -For x86:
> -  ./build/tools/build-ndk-sysroot.sh --abi=x86
> -  export NDK_SYSROOT=${ANDROID_BUILD_TOP}/ndk/build/platforms/android-3/arch-x86
> -
> -3. Set the NDK toolchain environment variable.
> -For arm:
> -  export NDK_TOOLCHAIN=${ANDROID_TOOLCHAIN}/arm-linux-androideabi-
> -For x86:
> -  export NDK_TOOLCHAIN=${ANDROID_TOOLCHAIN}/i686-linux-android-
> -
> -II. Compile perf for Android
> -------------------------------------------------
> -You need to run make with the NDK toolchain and sysroot defined above:
> -For arm:
> -  make WERROR=0 ARCH=arm CROSS_COMPILE=${NDK_TOOLCHAIN} EXTRA_CFLAGS="-pie --sysroot=${NDK_SYSROOT}"
> -For x86:
> -  make WERROR=0 ARCH=x86 CROSS_COMPILE=${NDK_TOOLCHAIN} EXTRA_CFLAGS="-pie --sysroot=${NDK_SYSROOT}"
> -
> -III. Install perf
> ------------------------------------------------
> -You need to connect to your Android device/emulator using adb.
> -Install perf using:
> -  adb push perf /data/perf
> -
> -If you also want to use perf-archive you need busybox tools for Android.
> -For installing perf-archive, you first need to replace #!/bin/bash with #!/system/bin/sh:
> -  sed 's/#!\/bin\/bash/#!\/system\/bin\/sh/g' perf-archive >> /tmp/perf-archive
> -  chmod +x /tmp/perf-archive
> -  adb push /tmp/perf-archive /data/perf-archive
> -
> -IV. Environment settings for running perf
> -------------------------------------------------
> -Some perf features need environment variables to run properly.
> -You need to set these before running perf on the target:
> -  adb shell
> -  # PERF_PAGER=cat
> -
> -IV. Run perf
> -------------------------------------------------
> -Run perf on your device/emulator to which you previously connected using adb:
> -  # ./data/perf
> +- The second method is to download Android NDK, then use the contained
> +  clang compiler for building perf. Please refer to Build.txt, section
> +  "5) Build with clang" for details.
>
> --
> 2.34.1
>

Re: [PATCH 8/8] perf docs: Document building with Clang

[Leo Yan]

Hi Ian,

On Thu, Sep 25, 2025 at 12:43:55PM -0700, Ian Rogers wrote:

[...]

> > +5) Build with clang
> > +===================
> > +By default, the makefile uses GCC as compiler. With specifying environment
> > +variables HOSTCC, CC and CXX, it allows to build perf with clang.
> > +
> > +Using clang for native build:
> > +
> > +  $ HOSTCC=clang CC=clang CXX=clang++ make -C tools/perf
> > +
> > +Using clang for cross compilation:
> > +
> > +  $ HOSTCC=clang CC=clang CXX=clang++ \
> > +    make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- -C tools/perf \
> > +    NO_LIBELF=1 NO_LIBTRACEEVENT=1 NO_JEVENTS=1
> 
> The three NO_-s here are going to cripple the build quite a lot, I
> wonder if we can list package dependencies to install and failing that
> use the NO_-s.

In the next version, I will drop NO_ options to avoid confusing.

Missing libraries can occur in both native build and cross build, so
this is not a specific issue for cross build. OTOH, the build logs
already provide clear reminders, here no need to bother readers for
these NO_ options.

> > +
> > +In the example above, due to lack libelf, python and libtraceevent for
> > +cross comiplation, disable the features accordingly.
> 
> nit: s/comiplation/compilation/
> 
> > diff --git a/tools/perf/Documentation/android.txt b/tools/perf/Documentation/android.txt
> > index 24a59998fc91e814ad96f658d3481d88d798b60c..e65204cf2921f6bd8a79875784c5b3d5487ce05d 100644
> > --- a/tools/perf/Documentation/android.txt
> > +++ b/tools/perf/Documentation/android.txt
> > @@ -1,78 +1,12 @@
> >  How to compile perf for Android
> > -=========================================
> > +===============================
> >
> > -I. Set the Android NDK environment
> > -------------------------------------------------
> > +There have two ways to build perf and run it on Android.
> 
> nit: s/There have/There are/

Will do.  Thanks a lot for review!

Leo

Re: [PATCH 4/8] perf test coresight: Dismiss clang warning for memcpy thread

[Leo Yan]

On Thu, Sep 25, 2025 at 12:38:36PM -0700, Ian Rogers wrote:
> On Thu, Sep 25, 2025 at 3:26 AM Leo Yan <leo.yan@arm.com> wrote:
> >
> > clang-18.1.3 on Ubuntu 24.04.2 reports warning:
> >
> >   memcpy_thread.c:30:1: warning: non-void function does not return a value in all control paths [-Wreturn-type]
> >      30 | }
> >         | ^
> >
> > Dismiss the warning with returning NULL from the thread function.
> >
> > Signed-off-by: Leo Yan <leo.yan@arm.com>
> 
> lgtm, should this be moved into being a perf test workload as in
> tools/perf/tests/workloads/ ?

Now we have four programs under shell folder:

  $ cd linux/tools/perf/tests/shell
  $ find ./ -name *.c -o -name *.S
  ./coresight/unroll_loop_thread/unroll_loop_thread.c
  ./coresight/asm_pure_loop/asm_pure_loop.S
  ./coresight/thread_loop/thread_loop.c
  ./coresight/memcpy_thread/memcpy_thread.c
 
If really do this, can move all programs into folders?

  tools/perf/tests/workloads/       - Common programs
  tools/perf/tests/workloads/arm64  - Arch specific programs

Anyway, this should be a separate series to avoid complexity for
enabling Clang build.

Thanks,
Leo