1 .. SPDX-License-Identifier: GPL-2.0
3 =====================================
4 Using Propeller with the Linux kernel
5 =====================================
7 This enables Propeller build support for the kernel when using Clang
8 compiler. Propeller is a profile-guided optimization (PGO) method used
9 to optimize binary executables. Like AutoFDO, it utilizes hardware
10 sampling to gather information about the frequency of execution of
11 different code paths within a binary. Unlike AutoFDO, this information
12 is then used right before linking phase to optimize (among others)
13 block layout within and across functions.
15 A few important notes about adopting Propeller optimization:
17 #. Although it can be used as a standalone optimization step, it is
18 strongly recommended to apply Propeller on top of AutoFDO,
19 AutoFDO+ThinLTO or Instrument FDO. The rest of this document
20 assumes this paradigm.
22 #. Propeller uses another round of profiling on top of
23 AutoFDO/AutoFDO+ThinLTO/iFDO. The whole build process involves
24 "build-afdo - train-afdo - build-propeller - train-propeller -
27 #. Propeller requires LLVM 19 release or later for Clang/Clang++
28 and the linker(ld.lld).
30 #. In addition to LLVM toolchain, Propeller requires a profiling
31 conversion tool: https://github.com/google/autofdo with a release
32 after v0.30.1: https://github.com/google/autofdo/releases/tag/v0.30.1.
34 The Propeller optimization process involves the following steps:
36 #. Initial building: Build the AutoFDO or AutoFDO+ThinLTO binary as
37 you would normally do, but with a set of compile-time / link-time
38 flags, so that a special metadata section is created within the
39 kernel binary. The special section is only intend to be used by the
40 profiling tool, it is not part of the runtime image, nor does it
41 change kernel run time text sections.
43 #. Profiling: The above kernel is then run with a representative
44 workload to gather execution frequency data. This data is collected
45 using hardware sampling, via perf. Propeller is most effective on
46 platforms supporting advanced PMU features like LBR on Intel
47 machines. This step is the same as profiling the kernel for AutoFDO
48 (the exact perf parameters can be different).
50 #. Propeller profile generation: Perf output file is converted to a
51 pair of Propeller profiles via an offline tool.
53 #. Optimized build: Build the AutoFDO or AutoFDO+ThinLTO optimized
54 binary as you would normally do, but with a compile-time /
55 link-time flag to pick up the Propeller compile time and link time
56 profiles. This build step uses 3 profiles - the AutoFDO profile,
57 the Propeller compile-time profile and the Propeller link-time
60 #. Deployment: The optimized kernel binary is deployed and used
61 in production environments, providing improved performance
67 Configure the kernel with::
69 CONFIG_AUTOFDO_CLANG=y
70 CONFIG_PROPELLER_CLANG=y
75 The default CONFIG_PROPELLER_CLANG setting covers kernel space objects
76 for Propeller builds. One can, however, enable or disable Propeller build
77 for individual files and directories by adding a line similar to the
78 following to the respective kernel Makefile:
80 - For enabling a single file (e.g. foo.o)::
82 PROPELLER_PROFILE_foo.o := y
84 - For enabling all files in one directory::
86 PROPELLER_PROFILE := y
88 - For disabling one file::
90 PROPELLER_PROFILE_foo.o := n
92 - For disabling all files in one directory::
94 PROPELLER__PROFILE := n
100 Here is an example workflow for building an AutoFDO+Propeller kernel:
102 1) Assuming an AutoFDO profile is already collected following
103 instructions in the AutoFDO document, build the kernel on the host
104 machine, with AutoFDO and Propeller build configs ::
106 CONFIG_AUTOFDO_CLANG=y
107 CONFIG_PROPELLER_CLANG=y
111 $ make LLVM=1 CLANG_AUTOFDO_PROFILE=<autofdo-profile-name>
113 2) Install the kernel on the test machine.
115 3) Run the load tests. The '-c' option in perf specifies the sample
116 event period. We suggest using a suitable prime number, like 500009,
119 - For Intel platforms::
121 $ perf record -e BR_INST_RETIRED.NEAR_TAKEN:k -a -N -b -c <count> -o <perf_file> -- <loadtest>
123 - For AMD platforms::
125 $ perf record --pfm-event RETIRED_TAKEN_BRANCH_INSTRUCTIONS:k -a -N -b -c <count> -o <perf_file> -- <loadtest>
127 Note you can repeat the above steps to collect multiple <perf_file>s.
129 4) (Optional) Download the raw perf file(s) to the host machine.
131 5) Use the create_llvm_prof tool (https://github.com/google/autofdo) to
132 generate Propeller profile. ::
134 $ create_llvm_prof --binary=<vmlinux> --profile=<perf_file>
135 --format=propeller --propeller_output_module_name
136 --out=<propeller_profile_prefix>_cc_profile.txt
137 --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
139 "<propeller_profile_prefix>" can be something like "/home/user/dir/any_string".
141 This command generates a pair of Propeller profiles:
142 "<propeller_profile_prefix>_cc_profile.txt" and
143 "<propeller_profile_prefix>_ld_profile.txt".
145 If there are more than 1 perf_file collected in the previous step,
146 you can create a temp list file "<perf_file_list>" with each line
147 containing one perf file name and run::
149 $ create_llvm_prof --binary=<vmlinux> --profile=@<perf_file_list>
150 --format=propeller --propeller_output_module_name
151 --out=<propeller_profile_prefix>_cc_profile.txt
152 --propeller_symorder=<propeller_profile_prefix>_ld_profile.txt
154 6) Rebuild the kernel using the AutoFDO and Propeller
157 CONFIG_AUTOFDO_CLANG=y
158 CONFIG_PROPELLER_CLANG=y
162 $ make LLVM=1 CLANG_AUTOFDO_PROFILE=<profile_file> CLANG_PROPELLER_PROFILE_PREFIX=<propeller_profile_prefix>