* [PATCH v4 1/2] kconfig: Remove the architecture specific config for AutoFDO
2026-06-04 19:56 [PATCH v4 0/2] kconfig: Remove the architecture specific config for AutoFDO and Propeller xur
@ 2026-06-04 19:56 ` xur
2026-06-04 19:56 ` [PATCH v4 2/2] kconfig: Remove the architecture specific config for Propeller xur
1 sibling, 0 replies; 3+ messages in thread
From: xur @ 2026-06-04 19:56 UTC (permalink / raw)
To: Yabin Cui, Will Deacon, Han Shen, Thomas Gleixner, Ingo Molnar,
Borislav Petkov, Dave Hansen, H. Peter Anvin, Kees Cook,
Nathan Chancellor, Nicolas Schier, Linus Walleij, Arnd Bergmann,
Mathieu Desnoyers, Rong Xu, Miguel Ojeda, Peter Zijlstra,
Jinjie Ruan, Lukas Bulwahn, linux-kernel, Juergen Gross,
Helge Deller, Ryan Roberts, Marc Zyngier, Ard Biesheuvel,
Vincent Donnefort, Alice Ryhl
Cc: x86, linux-arm-kernel
From: Rong Xu <xur@google•com>
The CONFIG_AUTOFDO_CLANG option currently depends on
ARCH_SUPPORTS_AUTOFDO_CLANG, but this dependency seems unnecessary.
Remove ARCH_SUPPORTS_AUTOFDO_CLANG and allow users to control AutoFDO
builds solely through CONFIG_AUTOFDO_CLANG. This simplifies the kconfig
and avoids potential confusion.
Expand the AutoFDO documentation to include instructions for arm64.
Contributor acknowledgments:
* SPE instructions: Daniel Hoekwater <hoekwater@google•com>
* ETM instructions: Yabin Cui <yabinc@google•com>
Signed-off-by: Rong Xu <xur@google•com>
Suggested-by: Will Deacon <will@kernel•org>
Tested-by: Yabin Cui <yabinc@google•com>
Reviewed-by: Kees Cook <kees@kernel•org>
---
Documentation/dev-tools/autofdo.rst | 41 +++++++++++++++++++++++++++++
arch/Kconfig | 4 ---
arch/x86/Kconfig | 1 -
3 files changed, 41 insertions(+), 5 deletions(-)
diff --git a/Documentation/dev-tools/autofdo.rst b/Documentation/dev-tools/autofdo.rst
index bcf06e7d6ffa..ae03c4dfedc1 100644
--- a/Documentation/dev-tools/autofdo.rst
+++ b/Documentation/dev-tools/autofdo.rst
@@ -61,6 +61,9 @@ process consists of the following steps:
the AutoFDO profile via offline tools.
The support requires a Clang compiler LLVM 17 or later.
+Current supported architectures include x86/x86_64 (via LBR) and
+arm64 (via SPE or ETM).
+
Preparation
===========
@@ -141,6 +144,35 @@ Here is an example workflow for AutoFDO kernel:
$ perf record --pfm-events RETIRED_TAKEN_BRANCH_INSTRUCTIONS:k -a -N -b -c <count> -o <perf_file> -- <loadtest>
+ - For arm64 with SPE:
+
+ There are a few kernel features that must be enabled to collect SPE profiles on Arm.
+ Below is a list of the required features:
+
+ - CONFIG_ARM_SPE_PMU=y
+ - CONFIG_PID_IN_CONTEXTIDR=y
+ - kpti=off
+
+ Use the following command to generate SPE perf data file::
+
+ $ perf record -e ' arm_spe_0/branch_filter=1,load_filter=0,store_filter=0/' -a -c <count> -N --no-switch-events -o <perf_file> -- <loadtest>
+
+ - For arm64 with ETM trace:
+
+ Follow the instructions in `Linaro OpenCSD document
+ <https://github.com/Linaro/OpenCSD/blob/master/decoder/tests/auto-fdo/autofdo.md>`_
+ to record ETM traces for AutoFDO::
+
+ $ perf record -e cs_etm/@tmc_etr0/k -a -o <etm_perf_file> -- <loadtest>
+ $ perf inject -i <etm_perf_file> -o <perf_file> --itrace=i500009il
+
+ For ARM platforms running Android, follow the instructions in `Android simpleperf
+ document <https://android.googlesource.com/kernel/common/+/refs/heads/android-mainline/gki/aarch64/afdo>`_
+ to record ETM traces for AutoFDO::
+
+ $ simpleperf record -e cs-etm:k -a -o <etm_perf_file> -- <loadtest>
+ $ simpleperf inject -i <etm_perf_file> -o <text_perf_file> --symdir <vmlinux_dir>
+
4) (Optional) Download the raw perf file to the host machine.
5) To generate an AutoFDO profile, two offline tools are available:
@@ -162,6 +194,15 @@ Here is an example workflow for AutoFDO kernel:
$ llvm-profdata merge -o <profile_file> <profile_1> <profile_2> ... <profile_n>
+ For arm64 SPE, use the following command::
+
+ $ create_llvm_prof --binary=<vmlinux> --profile=<perf_file> --profiler=perf_spe --format=extbinary --out=<profile_file>
+
+ For arm64 ETM, use the following command::
+
+ $ create_llvm_prof --binary=<vmlinux> --profile=<text_perf_file> --profiler=text -format=extbinary -out=<profile_file>
+
+
6) Rebuild the kernel using the AutoFDO profile file with the same config as step 1,
(Note CONFIG_AUTOFDO_CLANG needs to be enabled)::
diff --git a/arch/Kconfig b/arch/Kconfig
index 0848932d1c8e..5e878924939a 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -879,12 +879,8 @@ config LTO_CLANG_THIN_DIST
module-specific compiler options, and simplifies debugging.
endchoice
-config ARCH_SUPPORTS_AUTOFDO_CLANG
- bool
-
config AUTOFDO_CLANG
bool "Enable Clang's AutoFDO build (EXPERIMENTAL)"
- depends on ARCH_SUPPORTS_AUTOFDO_CLANG
depends on CC_IS_CLANG
help
This option enables Clang’s AutoFDO build. When
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index f3f7cb01d69d..10bf3984102e 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -130,7 +130,6 @@ config X86
select ARCH_SUPPORTS_LTO_CLANG
select ARCH_SUPPORTS_LTO_CLANG_THIN
select ARCH_SUPPORTS_RT
- select ARCH_SUPPORTS_AUTOFDO_CLANG
select ARCH_SUPPORTS_PROPELLER_CLANG if X86_64
select ARCH_USE_BUILTIN_BSWAP
select ARCH_USE_CMPXCHG_LOCKREF if X86_CX8
--
2.54.0.1032.g2f8565e1d1-goog
^ permalink raw reply related [flat|nested] 3+ messages in thread* [PATCH v4 2/2] kconfig: Remove the architecture specific config for Propeller
2026-06-04 19:56 [PATCH v4 0/2] kconfig: Remove the architecture specific config for AutoFDO and Propeller xur
2026-06-04 19:56 ` [PATCH v4 1/2] kconfig: Remove the architecture specific config for AutoFDO xur
@ 2026-06-04 19:56 ` xur
1 sibling, 0 replies; 3+ messages in thread
From: xur @ 2026-06-04 19:56 UTC (permalink / raw)
To: Yabin Cui, Will Deacon, Han Shen, Thomas Gleixner, Ingo Molnar,
Borislav Petkov, Dave Hansen, H. Peter Anvin, Kees Cook,
Nathan Chancellor, Nicolas Schier, Linus Walleij, Arnd Bergmann,
Mathieu Desnoyers, Rong Xu, Miguel Ojeda, Peter Zijlstra,
Jinjie Ruan, Lukas Bulwahn, linux-kernel, Juergen Gross,
Helge Deller, Ryan Roberts, Marc Zyngier, Ard Biesheuvel,
Vincent Donnefort, Alice Ryhl
Cc: x86, linux-arm-kernel
From: Rong Xu <xur@google•com>
The CONFIG_PROPELLER_CLANG option currently depends on
ARCH_SUPPORTS_PROPELLER_CLANG, but this dependency seems unnecessary.
Remove ARCH_SUPPORTS_PROPELLER_CLANG and allow users to control
Propeller builds solely through CONFIG_PROPELLER_CLANG. This simplifies
the kconfig and avoids potential confusion.
Move the .llvm_bb_addr_map sections grouping to
include/asm-generic/vmlinux.lds.h.
The Propeller documentation has been updated to reflect the most
recent tool location and now includes instructions for arm64.
Contributor Acknowledgments:
* SPE instructions: Daniel Hoekwater <hoekwater@google•com>
Signed-off-by: Rong Xu <xur@google•com>
Suggested-by: Will Deacon <will@kernel•org>
Suggested-by: Nathan Chancellor <nathan@kernel•org>
Tested-by: Yabin Cui <yabinc@google•com>
Reviewed-by: Kees Cook <kees@kernel•org>
---
Documentation/dev-tools/propeller.rst | 49 ++++++++++++++++++++-------
arch/Kconfig | 5 +--
arch/arm64/kernel/vmlinux.lds.S | 1 +
arch/x86/Kconfig | 1 -
arch/x86/kernel/vmlinux.lds.S | 5 +--
include/asm-generic/vmlinux.lds.h | 6 ++++
6 files changed, 46 insertions(+), 21 deletions(-)
diff --git a/Documentation/dev-tools/propeller.rst b/Documentation/dev-tools/propeller.rst
index 92195958e3db..e927319941c9 100644
--- a/Documentation/dev-tools/propeller.rst
+++ b/Documentation/dev-tools/propeller.rst
@@ -28,8 +28,10 @@ A few important notes about adopting Propeller optimization:
and the linker(ld.lld).
#. In addition to LLVM toolchain, Propeller requires a profiling
- conversion tool: https://github.com/google/autofdo with a release
- after v0.30.1: https://github.com/google/autofdo/releases/tag/v0.30.1.
+ conversion tool: https://github.com/google/llvm-propeller.
+
+Current supported architectures include x86/X86_64 (via LBR),
+and arm64 (via SPE).
The Propeller optimization process involves the following steps:
@@ -124,17 +126,30 @@ Here is an example workflow for building an AutoFDO+Propeller kernel:
$ perf record --pfm-event RETIRED_TAKEN_BRANCH_INSTRUCTIONS:k -a -N -b -c <count> -o <perf_file> -- <loadtest>
- Note you can repeat the above steps to collect multiple <perf_file>s.
+ - For arm64 with SPE::
+ There are a few kernel features that must be enabled to collect SPE profiles on Arm.
+ Below is a list of the required features:
+
+ - CONFIG_ARM_SPE_PMU=y
+ - CONFIG_PID_IN_CONTEXTIDR=y
+ - kpti=off
+
+ Use the following command to generate SPE perf data file::
+
+ $ perf record -e 'arm_spe_0/branch_filter=1,load_filter=0,store_filter=0/' -a -N -c <count> --no-switch-events -o <perf_file> -- <loadtest>
+
+ Note you can repeat the above steps to collect multiple <perf_file>s.
4) (Optional) Download the raw perf file(s) to the host machine.
-5) Use the create_llvm_prof tool (https://github.com/google/autofdo) to
+5) Use the generate_propeller_profiles tool (https://github.com/google/llvm-propeller) to
generate Propeller profile. ::
- $ create_llvm_prof --binary=<vmlinux> --profile=<perf_file>
- --format=propeller --propeller_output_module_name
- --out=<propeller_profile_prefix>_cc_profile.txt
- --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
+ $ generate_propeller_profiles \
+ --binary=<vmlinux> --profile=<perf_file> \
+ --format=propeller --propeller_output_module_name \
+ --out=<propeller_profile_prefix>_cc_profile.txt \
+ --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
"<propeller_profile_prefix>" can be something like "/home/user/dir/any_string".
@@ -146,10 +161,20 @@ Here is an example workflow for building an AutoFDO+Propeller kernel:
you can create a temp list file "<perf_file_list>" with each line
containing one perf file name and run::
- $ create_llvm_prof --binary=<vmlinux> --profile=@<perf_file_list>
- --format=propeller --propeller_output_module_name
- --out=<propeller_profile_prefix>_cc_profile.txt
- --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
+ $ generate_propeller_profiles \
+ --binary=<vmlinux> --profile=@<perf_file_list> \
+ --format=propeller --propeller_output_module_name \
+ --out=<propeller_profile_prefix>_cc_profile.txt \
+ --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
+
+ For arm64 SPE, add the option '--profiler=perf_spe', like::
+
+ $ generate_propeller_profiles \
+ --binary=<vmlinux> --profile=<perf_file> \
+ --profiler=perf_spe \
+ --format=propeller --propeller_output_module_name \
+ --out=<propeller_profile_prefix>_cc_profile.txt \
+ --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
6) Rebuild the kernel using the AutoFDO and Propeller
profiles. ::
diff --git a/arch/Kconfig b/arch/Kconfig
index 5e878924939a..99c2017eb515 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -895,13 +895,10 @@ config AUTOFDO_CLANG
If unsure, say N.
-config ARCH_SUPPORTS_PROPELLER_CLANG
- bool
-
config PROPELLER_CLANG
bool "Enable Clang's Propeller build"
- depends on ARCH_SUPPORTS_PROPELLER_CLANG
depends on CC_IS_CLANG && CLANG_VERSION >= 190000
+ depends on $(cc-option,-fbasic-block-sections=list=/dev/null)
help
This option enables Clang’s Propeller build. When the Propeller
profiles is specified in variable CLANG_PROPELLER_PROFILE_PREFIX
diff --git a/arch/arm64/kernel/vmlinux.lds.S b/arch/arm64/kernel/vmlinux.lds.S
index e1ac876200a3..8aaf404980a7 100644
--- a/arch/arm64/kernel/vmlinux.lds.S
+++ b/arch/arm64/kernel/vmlinux.lds.S
@@ -368,6 +368,7 @@ SECTIONS
STABS_DEBUG
DWARF_DEBUG
+ PROPELLER_DATA
MODINFO
ELF_DETAILS
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 10bf3984102e..b875d2f27e48 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -130,7 +130,6 @@ config X86
select ARCH_SUPPORTS_LTO_CLANG
select ARCH_SUPPORTS_LTO_CLANG_THIN
select ARCH_SUPPORTS_RT
- select ARCH_SUPPORTS_PROPELLER_CLANG if X86_64
select ARCH_USE_BUILTIN_BSWAP
select ARCH_USE_CMPXCHG_LOCKREF if X86_CX8
select ARCH_USE_MEMTEST
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index 4711a35e706c..74e336d7f9dd 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -423,10 +423,7 @@ SECTIONS
STABS_DEBUG
DWARF_DEBUG
-#ifdef CONFIG_PROPELLER_CLANG
- .llvm_bb_addr_map : { *(.llvm_bb_addr_map) }
-#endif
-
+ PROPELLER_DATA
MODINFO
ELF_DETAILS
diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h
index 60c8c22fd3e4..5659f4b5a125 100644
--- a/include/asm-generic/vmlinux.lds.h
+++ b/include/asm-generic/vmlinux.lds.h
@@ -1011,6 +1011,12 @@
#define PERCPU_DECRYPTED_SECTION
#endif
+#ifdef CONFIG_PROPELLER_CLANG
+#define PROPELLER_DATA \
+ .llvm_bb_addr_map : { *(.llvm_bb_addr_map) }
+#else
+#define PROPELLER_DATA
+#endif
/*
* Default discarded sections.
--
2.54.0.1032.g2f8565e1d1-goog
^ permalink raw reply related [flat|nested] 3+ messages in thread