blob: e36619216d4d0a19434595715649084e0bdddcbd [file] [log] [blame]
Tom Rini10e47792018-05-06 17:58:06 -04001# SPDX-License-Identifier: GPL-2.0+
Simon Glassc05694f2013-04-03 11:07:16 +00002# Copyright (c) 2013 The Chromium OS Authors.
Simon Glassc05694f2013-04-03 11:07:16 +00003
Simon Glass3ec03d42014-08-09 15:32:58 -06004(Please read 'How to change from MAKEALL' if you are used to that tool)
5
Simon Glassd7c8a7d2016-07-27 20:33:06 -06006Quick-start
7===========
8
9If you just want to quickly set up buildman so you can build something (for
10example Raspberry Pi 2):
11
12 cd /path/to/u-boot
13 PATH=$PATH:`pwd`/tools/buildman
14 buildman --fetch-arch arm
15 buildman -k rpi_2
16 ls ../current/rpi_2
17 # u-boot.bin is the output image
18
19
Simon Glassc05694f2013-04-03 11:07:16 +000020What is this?
21=============
22
23This tool handles building U-Boot to check that you have not broken it
24with your patch series. It can build each individual commit and report
25which boards fail on which commits, and which errors come up. It aims
26to make full use of multi-processor machines.
27
28A key feature of buildman is its output summary, which allows warnings,
29errors or image size increases in a particular commit or board to be
30quickly identified and the offending commit pinpointed. This can be a big
31help for anyone working with >10 patches at a time.
32
33
34Caveats
35=======
36
Simon Glassc05694f2013-04-03 11:07:16 +000037Buildman can be stopped and restarted, in which case it will continue
38where it left off. This should happen cleanly and without side-effects.
39If not, it is a bug, for which a patch would be welcome.
40
41Buildman gets so tied up in its work that it can ignore the outside world.
42You may need to press Ctrl-C several times to quit it. Also it will print
Simon Glass7b94ab52016-07-27 20:32:58 -060043out various exceptions when stopped. You may have to kill it since the
44Ctrl-C handling is somewhat broken.
Simon Glassc05694f2013-04-03 11:07:16 +000045
46
47Theory of Operation
48===================
49
50(please read this section in full twice or you will be perpetually confused)
51
52Buildman is a builder. It is not make, although it runs make. It does not
53produce any useful output on the terminal while building, except for
Simon Glass78e418e2014-08-09 15:33:03 -060054progress information (except with -v, see below). All the output (errors,
Dirk Behme5f01c7a2014-12-23 07:41:26 +010055warnings and binaries if you ask for them) is stored in output
Simon Glass78e418e2014-08-09 15:33:03 -060056directories, which you can look at while the build is progressing, or when
57it is finished.
Simon Glassc05694f2013-04-03 11:07:16 +000058
Simon Glass7b94ab52016-07-27 20:32:58 -060059Buildman is designed to build entire git branches, i.e. muliple commits. It
60can be run repeatedly on the same branch. In this case it will automatically
61rebuild commits which have changed (and remove its old results for that
62commit). It is possible to build a branch for one board, then later build it
63for another board. If you want buildman to re-build a commit it has already
64built (e.g. because of a toolchain update), use the -f flag.
65
Simon Glassc05694f2013-04-03 11:07:16 +000066Buildman produces a concise summary of which boards succeeded and failed.
67It shows which commit introduced which board failure using a simple
68red/green colour coding. Full error information can be requested, in which
69case it is de-duped and displayed against the commit that introduced the
70error. An example workflow is below.
71
72Buildman stores image size information and can report changes in image size
73from commit to commit. An example of this is below.
74
75Buildman starts multiple threads, and each thread builds for one board at
76a time. A thread starts at the first commit, configures the source for your
77board and builds it. Then it checks out the next commit and does an
78incremental build. Eventually the thread reaches the last commit and stops.
79If errors or warnings are found along the way, the thread will reconfigure
80after every commit, and your build will be very slow. This is because a
81file that produces just a warning would not normally be rebuilt in an
82incremental build.
83
84Buildman works in an entirely separate place from your U-Boot repository.
85It creates a separate working directory for each thread, and puts the
86output files in the working directory, organised by commit name and board
87name, in a two-level hierarchy.
88
89Buildman is invoked in your U-Boot directory, the one with the .git
90directory. It clones this repository into a copy for each thread, and the
91threads do not affect the state of your git repository. Any checkouts done
92by the thread affect only the working directory for that thread.
93
Simon Glassfa1cfee2014-08-09 15:32:57 -060094Buildman automatically selects the correct tool chain for each board. You
95must supply suitable tool chains, but buildman takes care of selecting the
Simon Glassc05694f2013-04-03 11:07:16 +000096right one.
97
Simon Glass78e418e2014-08-09 15:33:03 -060098Buildman generally builds a branch (with the -b flag), and in this case
99builds the upstream commit as well, for comparison. It cannot build
100individual commits at present, unless (maybe) you point it at an empty
101branch. Put all your commits in a branch, set the branch's upstream to a
102valid value, and all will be well. Otherwise buildman will perform random
103actions. Use -n to check what the random actions might be.
104
Simon Glassd8158f92014-10-16 01:05:56 -0600105If you just want to build the current source tree, leave off the -b flag
106and add -e. This will display results and errors as they happen. You can
107still look at them later using -se. Note that buildman will assume that the
108source has changed, and will build all specified boards in this case.
Simon Glassc05694f2013-04-03 11:07:16 +0000109
110Buildman is optimised for building many commits at once, for many boards.
111On multi-core machines, Buildman is fast because it uses most of the
112available CPU power. When it gets to the end, or if you are building just
113a few commits or boards, it will be pretty slow. As a tip, if you don't
114plan to use your machine for anything else, you can use -T to increase the
115number of threads beyond the default.
116
Simon Glassd9eb9f02018-06-11 23:26:46 -0600117
118Selecting which boards to build
119===============================
120
Stephen Warren08447632013-10-10 10:00:20 -0600121Buildman lets you build all boards, or a subset. Specify the subset by passing
122command-line arguments that list the desired board name, architecture name,
123SOC name, or anything else in the boards.cfg file. Multiple arguments are
124allowed. Each argument will be interpreted as a regular expression, so
125behaviour is a superset of exact or substring matching. Examples are:
126
127* 'tegra20' All boards with a Tegra20 SoC
128* 'tegra' All boards with any Tegra Soc (Tegra20, Tegra30, Tegra114...)
129* '^tegra[23]0$' All boards with either Tegra20 or Tegra30 SoC
130* 'powerpc' All PowerPC boards
Simon Glassc05694f2013-04-03 11:07:16 +0000131
Simon Glassaa40f9a2014-08-09 15:33:08 -0600132While the default is to OR the terms together, you can also make use of
133the '&' operator to limit the selection:
134
135* 'freescale & arm sandbox' All Freescale boards with ARM architecture,
136 plus sandbox
137
Simon Glass924c73a2014-08-28 09:43:41 -0600138You can also use -x to specifically exclude some boards. For example:
139
Baruch Siach752db5b2019-07-11 12:57:41 +0300140 buildman arm -x nvidia,freescale,.*ball$
Simon Glass924c73a2014-08-28 09:43:41 -0600141
142means to build all arm boards except nvidia, freescale and anything ending
143with 'ball'.
144
Simon Glassd9eb9f02018-06-11 23:26:46 -0600145For building specific boards you can use the --boards option, which takes a
146comma-separated list of board target names and be used multiple times on
147the command line:
148
Baruch Siach752db5b2019-07-11 12:57:41 +0300149 buildman --boards sandbox,snow --boards
Simon Glassd9eb9f02018-06-11 23:26:46 -0600150
Dirk Behme5f01c7a2014-12-23 07:41:26 +0100151It is convenient to use the -n option to see what will be built based on
Simon Glass6af145f2017-01-23 05:38:56 -0700152the subset given. Use -v as well to get an actual list of boards.
Simon Glassaa40f9a2014-08-09 15:33:08 -0600153
Simon Glassc05694f2013-04-03 11:07:16 +0000154Buildman does not store intermediate object files. It optionally copies
Simon Glassd9eb9f02018-06-11 23:26:46 -0600155the binary output into a directory when a build is successful (-k). Size
Simon Glassc05694f2013-04-03 11:07:16 +0000156information is always recorded. It needs a fair bit of disk space to work,
157typically 250MB per thread.
158
159
160Setting up
161==========
162
1631. Get the U-Boot source. You probably already have it, but if not these
164steps should get you started with a repo and some commits for testing.
165
166$ cd /path/to/u-boot
167$ git clone git://git.denx.de/u-boot.git .
168$ git checkout -b my-branch origin/master
169$ # Add some commits to the branch, reading for testing
170
Simon Glass3cb033e2014-12-01 17:34:02 -07001712. Create ~/.buildman to tell buildman where to find tool chains (see 'The
172.buildman file' later for details). As an example:
Simon Glassc05694f2013-04-03 11:07:16 +0000173
174# Buildman settings file
175
176[toolchain]
177root: /
178rest: /toolchains/*
179eldk: /opt/eldk-4.2
Simon Glass9a016392014-08-09 15:33:07 -0600180arm: /opt/linaro/gcc-linaro-arm-linux-gnueabihf-4.8-2013.08_linux
181aarch64: /opt/linaro/gcc-linaro-aarch64-none-elf-4.8-2013.10_linux
Simon Glassc05694f2013-04-03 11:07:16 +0000182
183[toolchain-alias]
184x86: i386
185blackfin: bfin
Simon Glassc05694f2013-04-03 11:07:16 +0000186nds32: nds32le
Bin Meng07a50f92016-02-21 21:18:02 -0800187openrisc: or1k
Simon Glassc05694f2013-04-03 11:07:16 +0000188
189
190This selects the available toolchain paths. Add the base directory for
191each of your toolchains here. Buildman will search inside these directories
192and also in any '/usr' and '/usr/bin' subdirectories.
193
194Make sure the tags (here root: rest: and eldk:) are unique.
195
196The toolchain-alias section indicates that the i386 toolchain should be used
197to build x86 commits.
198
Simon Glassf5902732016-03-12 18:50:32 -0700199Note that you can also specific exactly toolchain prefixes if you like:
200
201[toolchain-prefix]
202arm: /opt/arm-eabi-4.6/bin/arm-eabi-
203
204or even:
205
206[toolchain-prefix]
207arm: /opt/arm-eabi-4.6/bin/arm-eabi-gcc
208
209This tells buildman that you want to use this exact toolchain for the arm
210architecture. This will override any toolchains found by searching using the
211[toolchain] settings.
212
213Since the toolchain prefix is an explicit request, buildman will report an
214error if a toolchain is not found with that prefix. The current PATH will be
215searched, so it is possible to use:
216
217[toolchain-prefix]
218arm: arm-none-eabi-
219
220and buildman will find arm-none-eabi-gcc in /usr/bin if you have it installed.
Simon Glassc05694f2013-04-03 11:07:16 +0000221
York Sunfb197a82016-10-04 14:33:51 -0700222[toolchain-wrapper]
223wrapper: ccache
224
225This tells buildman to use a compiler wrapper in front of CROSS_COMPILE. In
226this example, ccache. It doesn't affect the toolchain scan. The wrapper is
227added when CROSS_COMPILE environtal variable is set. The name in this
228section is ignored. If more than one line is provided, only the last one
229is taken.
230
Simon Glass91075712014-12-01 17:34:01 -07002313. Make sure you have the require Python pre-requisites
232
Simon Glass7e803e12014-12-01 17:34:06 -0700233Buildman uses multiprocessing, Queue, shutil, StringIO, ConfigParser and
234urllib2. These should normally be available, but if you get an error like
235this then you will need to obtain those modules:
Simon Glass91075712014-12-01 17:34:01 -0700236
237 ImportError: No module named multiprocessing
238
239
2404. Check the available toolchains
Simon Glassc05694f2013-04-03 11:07:16 +0000241
242Run this check to make sure that you have a toolchain for every architecture.
243
244$ ./tools/buildman/buildman --list-tool-chains
245Scanning for tool chains
Simon Glassf5902732016-03-12 18:50:32 -0700246 - scanning prefix '/opt/gcc-4.6.3-nolibc/x86_64-linux/bin/x86_64-linux-'
247Tool chain test: OK, arch='x86', priority 1
248 - scanning prefix '/opt/arm-eabi-4.6/bin/arm-eabi-'
249Tool chain test: OK, arch='arm', priority 1
250 - scanning path '/toolchains/gcc-4.9.0-nolibc/i386-linux'
251 - looking in '/toolchains/gcc-4.9.0-nolibc/i386-linux/.'
252 - looking in '/toolchains/gcc-4.9.0-nolibc/i386-linux/bin'
253 - found '/toolchains/gcc-4.9.0-nolibc/i386-linux/bin/i386-linux-gcc'
254 - looking in '/toolchains/gcc-4.9.0-nolibc/i386-linux/usr/bin'
255Tool chain test: OK, arch='i386', priority 4
256 - scanning path '/toolchains/gcc-4.9.0-nolibc/aarch64-linux'
257 - looking in '/toolchains/gcc-4.9.0-nolibc/aarch64-linux/.'
258 - looking in '/toolchains/gcc-4.9.0-nolibc/aarch64-linux/bin'
259 - found '/toolchains/gcc-4.9.0-nolibc/aarch64-linux/bin/aarch64-linux-gcc'
260 - looking in '/toolchains/gcc-4.9.0-nolibc/aarch64-linux/usr/bin'
261Tool chain test: OK, arch='aarch64', priority 4
262 - scanning path '/toolchains/gcc-4.9.0-nolibc/microblaze-linux'
263 - looking in '/toolchains/gcc-4.9.0-nolibc/microblaze-linux/.'
264 - looking in '/toolchains/gcc-4.9.0-nolibc/microblaze-linux/bin'
265 - found '/toolchains/gcc-4.9.0-nolibc/microblaze-linux/bin/microblaze-linux-gcc'
266 - looking in '/toolchains/gcc-4.9.0-nolibc/microblaze-linux/usr/bin'
267Tool chain test: OK, arch='microblaze', priority 4
268 - scanning path '/toolchains/gcc-4.9.0-nolibc/mips64-linux'
269 - looking in '/toolchains/gcc-4.9.0-nolibc/mips64-linux/.'
270 - looking in '/toolchains/gcc-4.9.0-nolibc/mips64-linux/bin'
271 - found '/toolchains/gcc-4.9.0-nolibc/mips64-linux/bin/mips64-linux-gcc'
272 - looking in '/toolchains/gcc-4.9.0-nolibc/mips64-linux/usr/bin'
273Tool chain test: OK, arch='mips64', priority 4
274 - scanning path '/toolchains/gcc-4.9.0-nolibc/sparc64-linux'
275 - looking in '/toolchains/gcc-4.9.0-nolibc/sparc64-linux/.'
276 - looking in '/toolchains/gcc-4.9.0-nolibc/sparc64-linux/bin'
277 - found '/toolchains/gcc-4.9.0-nolibc/sparc64-linux/bin/sparc64-linux-gcc'
278 - looking in '/toolchains/gcc-4.9.0-nolibc/sparc64-linux/usr/bin'
279Tool chain test: OK, arch='sparc64', priority 4
280 - scanning path '/toolchains/gcc-4.9.0-nolibc/arm-unknown-linux-gnueabi'
281 - looking in '/toolchains/gcc-4.9.0-nolibc/arm-unknown-linux-gnueabi/.'
282 - looking in '/toolchains/gcc-4.9.0-nolibc/arm-unknown-linux-gnueabi/bin'
283 - found '/toolchains/gcc-4.9.0-nolibc/arm-unknown-linux-gnueabi/bin/arm-unknown-linux-gnueabi-gcc'
284 - looking in '/toolchains/gcc-4.9.0-nolibc/arm-unknown-linux-gnueabi/usr/bin'
285Tool chain test: OK, arch='arm', priority 3
286Toolchain '/toolchains/gcc-4.9.0-nolibc/arm-unknown-linux-gnueabi/bin/arm-unknown-linux-gnueabi-gcc' at priority 3 will be ignored because another toolchain for arch 'arm' has priority 1
287 - scanning path '/toolchains/gcc-4.9.0-nolibc/sparc-linux'
288 - looking in '/toolchains/gcc-4.9.0-nolibc/sparc-linux/.'
289 - looking in '/toolchains/gcc-4.9.0-nolibc/sparc-linux/bin'
290 - found '/toolchains/gcc-4.9.0-nolibc/sparc-linux/bin/sparc-linux-gcc'
291 - looking in '/toolchains/gcc-4.9.0-nolibc/sparc-linux/usr/bin'
292Tool chain test: OK, arch='sparc', priority 4
293 - scanning path '/toolchains/gcc-4.9.0-nolibc/mips-linux'
294 - looking in '/toolchains/gcc-4.9.0-nolibc/mips-linux/.'
295 - looking in '/toolchains/gcc-4.9.0-nolibc/mips-linux/bin'
296 - found '/toolchains/gcc-4.9.0-nolibc/mips-linux/bin/mips-linux-gcc'
297 - looking in '/toolchains/gcc-4.9.0-nolibc/mips-linux/usr/bin'
298Tool chain test: OK, arch='mips', priority 4
299 - scanning path '/toolchains/gcc-4.9.0-nolibc/x86_64-linux'
300 - looking in '/toolchains/gcc-4.9.0-nolibc/x86_64-linux/.'
301 - looking in '/toolchains/gcc-4.9.0-nolibc/x86_64-linux/bin'
302 - found '/toolchains/gcc-4.9.0-nolibc/x86_64-linux/bin/x86_64-linux-gcc'
303 - found '/toolchains/gcc-4.9.0-nolibc/x86_64-linux/bin/x86_64-linux-x86_64-linux-gcc'
304 - looking in '/toolchains/gcc-4.9.0-nolibc/x86_64-linux/usr/bin'
305Tool chain test: OK, arch='x86_64', priority 4
306Tool chain test: OK, arch='x86_64', priority 4
307Toolchain '/toolchains/gcc-4.9.0-nolibc/x86_64-linux/bin/x86_64-linux-x86_64-linux-gcc' at priority 4 will be ignored because another toolchain for arch 'x86_64' has priority 4
308 - scanning path '/toolchains/gcc-4.9.0-nolibc/m68k-linux'
309 - looking in '/toolchains/gcc-4.9.0-nolibc/m68k-linux/.'
310 - looking in '/toolchains/gcc-4.9.0-nolibc/m68k-linux/bin'
311 - found '/toolchains/gcc-4.9.0-nolibc/m68k-linux/bin/m68k-linux-gcc'
312 - looking in '/toolchains/gcc-4.9.0-nolibc/m68k-linux/usr/bin'
313Tool chain test: OK, arch='m68k', priority 4
314 - scanning path '/toolchains/gcc-4.9.0-nolibc/powerpc-linux'
315 - looking in '/toolchains/gcc-4.9.0-nolibc/powerpc-linux/.'
316 - looking in '/toolchains/gcc-4.9.0-nolibc/powerpc-linux/bin'
317 - found '/toolchains/gcc-4.9.0-nolibc/powerpc-linux/bin/powerpc-linux-gcc'
318 - looking in '/toolchains/gcc-4.9.0-nolibc/powerpc-linux/usr/bin'
319Tool chain test: OK, arch='powerpc', priority 4
320 - scanning path '/toolchains/gcc-4.6.3-nolibc/bfin-uclinux'
321 - looking in '/toolchains/gcc-4.6.3-nolibc/bfin-uclinux/.'
322 - looking in '/toolchains/gcc-4.6.3-nolibc/bfin-uclinux/bin'
323 - found '/toolchains/gcc-4.6.3-nolibc/bfin-uclinux/bin/bfin-uclinux-gcc'
324 - looking in '/toolchains/gcc-4.6.3-nolibc/bfin-uclinux/usr/bin'
325Tool chain test: OK, arch='bfin', priority 6
326 - scanning path '/toolchains/gcc-4.6.3-nolibc/sparc-linux'
327 - looking in '/toolchains/gcc-4.6.3-nolibc/sparc-linux/.'
328 - looking in '/toolchains/gcc-4.6.3-nolibc/sparc-linux/bin'
329 - found '/toolchains/gcc-4.6.3-nolibc/sparc-linux/bin/sparc-linux-gcc'
330 - looking in '/toolchains/gcc-4.6.3-nolibc/sparc-linux/usr/bin'
331Tool chain test: OK, arch='sparc', priority 4
332Toolchain '/toolchains/gcc-4.6.3-nolibc/sparc-linux/bin/sparc-linux-gcc' at priority 4 will be ignored because another toolchain for arch 'sparc' has priority 4
333 - scanning path '/toolchains/gcc-4.6.3-nolibc/mips-linux'
334 - looking in '/toolchains/gcc-4.6.3-nolibc/mips-linux/.'
335 - looking in '/toolchains/gcc-4.6.3-nolibc/mips-linux/bin'
336 - found '/toolchains/gcc-4.6.3-nolibc/mips-linux/bin/mips-linux-gcc'
337 - looking in '/toolchains/gcc-4.6.3-nolibc/mips-linux/usr/bin'
338Tool chain test: OK, arch='mips', priority 4
339Toolchain '/toolchains/gcc-4.6.3-nolibc/mips-linux/bin/mips-linux-gcc' at priority 4 will be ignored because another toolchain for arch 'mips' has priority 4
340 - scanning path '/toolchains/gcc-4.6.3-nolibc/m68k-linux'
341 - looking in '/toolchains/gcc-4.6.3-nolibc/m68k-linux/.'
342 - looking in '/toolchains/gcc-4.6.3-nolibc/m68k-linux/bin'
343 - found '/toolchains/gcc-4.6.3-nolibc/m68k-linux/bin/m68k-linux-gcc'
344 - looking in '/toolchains/gcc-4.6.3-nolibc/m68k-linux/usr/bin'
345Tool chain test: OK, arch='m68k', priority 4
346Toolchain '/toolchains/gcc-4.6.3-nolibc/m68k-linux/bin/m68k-linux-gcc' at priority 4 will be ignored because another toolchain for arch 'm68k' has priority 4
347 - scanning path '/toolchains/gcc-4.6.3-nolibc/powerpc-linux'
348 - looking in '/toolchains/gcc-4.6.3-nolibc/powerpc-linux/.'
349 - looking in '/toolchains/gcc-4.6.3-nolibc/powerpc-linux/bin'
350 - found '/toolchains/gcc-4.6.3-nolibc/powerpc-linux/bin/powerpc-linux-gcc'
351 - looking in '/toolchains/gcc-4.6.3-nolibc/powerpc-linux/usr/bin'
352Tool chain test: OK, arch='powerpc', priority 4
353Tool chain test: OK, arch='or32', priority 4
Simon Glassc05694f2013-04-03 11:07:16 +0000354 - scanning path '/'
355 - looking in '/.'
356 - looking in '/bin'
357 - looking in '/usr/bin'
Simon Glassf5902732016-03-12 18:50:32 -0700358 - found '/usr/bin/i586-mingw32msvc-gcc'
Simon Glassc05694f2013-04-03 11:07:16 +0000359 - found '/usr/bin/c89-gcc'
Simon Glassc05694f2013-04-03 11:07:16 +0000360 - found '/usr/bin/x86_64-linux-gnu-gcc'
Simon Glassf5902732016-03-12 18:50:32 -0700361 - found '/usr/bin/gcc'
362 - found '/usr/bin/c99-gcc'
363 - found '/usr/bin/arm-linux-gnueabi-gcc'
364 - found '/usr/bin/aarch64-linux-gnu-gcc'
365 - found '/usr/bin/winegcc'
366 - found '/usr/bin/arm-linux-gnueabihf-gcc'
367Tool chain test: OK, arch='i586', priority 11
368Tool chain test: OK, arch='c89', priority 11
369Tool chain test: OK, arch='x86_64', priority 4
370Toolchain '/usr/bin/x86_64-linux-gnu-gcc' at priority 4 will be ignored because another toolchain for arch 'x86_64' has priority 4
371Tool chain test: OK, arch='sandbox', priority 11
372Tool chain test: OK, arch='c99', priority 11
373Tool chain test: OK, arch='arm', priority 4
374Toolchain '/usr/bin/arm-linux-gnueabi-gcc' at priority 4 will be ignored because another toolchain for arch 'arm' has priority 1
375Tool chain test: OK, arch='aarch64', priority 4
376Toolchain '/usr/bin/aarch64-linux-gnu-gcc' at priority 4 will be ignored because another toolchain for arch 'aarch64' has priority 4
377Tool chain test: OK, arch='sandbox', priority 11
378Toolchain '/usr/bin/winegcc' at priority 11 will be ignored because another toolchain for arch 'sandbox' has priority 11
379Tool chain test: OK, arch='arm', priority 4
380Toolchain '/usr/bin/arm-linux-gnueabihf-gcc' at priority 4 will be ignored because another toolchain for arch 'arm' has priority 1
381List of available toolchains (34):
382aarch64 : /toolchains/gcc-4.9.0-nolibc/aarch64-linux/bin/aarch64-linux-gcc
383alpha : /toolchains/gcc-4.9.0-nolibc/alpha-linux/bin/alpha-linux-gcc
384am33_2.0 : /toolchains/gcc-4.9.0-nolibc/am33_2.0-linux/bin/am33_2.0-linux-gcc
385arm : /opt/arm-eabi-4.6/bin/arm-eabi-gcc
Simon Glassf5902732016-03-12 18:50:32 -0700386bfin : /toolchains/gcc-4.6.3-nolibc/bfin-uclinux/bin/bfin-uclinux-gcc
Simon Glassc05694f2013-04-03 11:07:16 +0000387c89 : /usr/bin/c89-gcc
388c99 : /usr/bin/c99-gcc
Simon Glassf5902732016-03-12 18:50:32 -0700389frv : /toolchains/gcc-4.9.0-nolibc/frv-linux/bin/frv-linux-gcc
390h8300 : /toolchains/gcc-4.9.0-nolibc/h8300-elf/bin/h8300-elf-gcc
391hppa : /toolchains/gcc-4.9.0-nolibc/hppa-linux/bin/hppa-linux-gcc
392hppa64 : /toolchains/gcc-4.9.0-nolibc/hppa64-linux/bin/hppa64-linux-gcc
393i386 : /toolchains/gcc-4.9.0-nolibc/i386-linux/bin/i386-linux-gcc
394i586 : /usr/bin/i586-mingw32msvc-gcc
395ia64 : /toolchains/gcc-4.9.0-nolibc/ia64-linux/bin/ia64-linux-gcc
396m32r : /toolchains/gcc-4.9.0-nolibc/m32r-linux/bin/m32r-linux-gcc
397m68k : /toolchains/gcc-4.9.0-nolibc/m68k-linux/bin/m68k-linux-gcc
398microblaze: /toolchains/gcc-4.9.0-nolibc/microblaze-linux/bin/microblaze-linux-gcc
399mips : /toolchains/gcc-4.9.0-nolibc/mips-linux/bin/mips-linux-gcc
400mips64 : /toolchains/gcc-4.9.0-nolibc/mips64-linux/bin/mips64-linux-gcc
401or32 : /toolchains/gcc-4.5.1-nolibc/or32-linux/bin/or32-linux-gcc
402powerpc : /toolchains/gcc-4.9.0-nolibc/powerpc-linux/bin/powerpc-linux-gcc
403powerpc64 : /toolchains/gcc-4.9.0-nolibc/powerpc64-linux/bin/powerpc64-linux-gcc
404ppc64le : /toolchains/gcc-4.9.0-nolibc/ppc64le-linux/bin/ppc64le-linux-gcc
405s390x : /toolchains/gcc-4.9.0-nolibc/s390x-linux/bin/s390x-linux-gcc
Simon Glassc05694f2013-04-03 11:07:16 +0000406sandbox : /usr/bin/gcc
Simon Glassf5902732016-03-12 18:50:32 -0700407sh4 : /toolchains/gcc-4.6.3-nolibc/sh4-linux/bin/sh4-linux-gcc
408sparc : /toolchains/gcc-4.9.0-nolibc/sparc-linux/bin/sparc-linux-gcc
409sparc64 : /toolchains/gcc-4.9.0-nolibc/sparc64-linux/bin/sparc64-linux-gcc
410tilegx : /toolchains/gcc-4.6.2-nolibc/tilegx-linux/bin/tilegx-linux-gcc
411x86 : /opt/gcc-4.6.3-nolibc/x86_64-linux/bin/x86_64-linux-gcc
412x86_64 : /toolchains/gcc-4.9.0-nolibc/x86_64-linux/bin/x86_64-linux-gcc
Simon Glassc05694f2013-04-03 11:07:16 +0000413
414
415You can see that everything is covered, even some strange ones that won't
416be used (c88 and c99). This is a feature.
417
418
Simon Glass7e803e12014-12-01 17:34:06 -07004195. Install new toolchains if needed
420
421You can download toolchains and update the [toolchain] section of the
422settings file to find them.
423
424To make this easier, buildman can automatically download and install
425toolchains from kernel.org. First list the available architectures:
426
Bin Meng05a50922015-07-16 19:43:46 -0700427$ ./tools/buildman/buildman --fetch-arch list
Simon Glass7e803e12014-12-01 17:34:06 -0700428Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.6.3/
429Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.6.2/
430Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.5.1/
431Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.2.4/
Andy Shevchenko8cb5cdd2017-07-05 16:25:22 +0300432Available architectures: alpha am33_2.0 arm bfin cris crisv32 frv h8300
Simon Glass7e803e12014-12-01 17:34:06 -0700433hppa hppa64 i386 ia64 m32r m68k mips mips64 or32 powerpc powerpc64 s390x sh4
434sparc sparc64 tilegx x86_64 xtensa
435
436Then pick one and download it:
437
Bin Meng05a50922015-07-16 19:43:46 -0700438$ ./tools/buildman/buildman --fetch-arch or32
Simon Glass7e803e12014-12-01 17:34:06 -0700439Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.6.3/
440Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.6.2/
441Checking: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.5.1/
442Downloading: https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.5.1//x86_64-gcc-4.5.1-nolibc_or32-linux.tar.xz
443Unpacking to: /home/sjg/.buildman-toolchains
444Testing
445 - looking in '/home/sjg/.buildman-toolchains/gcc-4.5.1-nolibc/or32-linux/.'
446 - looking in '/home/sjg/.buildman-toolchains/gcc-4.5.1-nolibc/or32-linux/bin'
447 - found '/home/sjg/.buildman-toolchains/gcc-4.5.1-nolibc/or32-linux/bin/or32-linux-gcc'
448Tool chain test: OK
449
Thomas Chou4acc2d42015-11-12 09:29:09 +0800450Or download them all from kernel.org and move them to /toolchains directory,
451
Simon Glass7b94ab52016-07-27 20:32:58 -0600452$ ./tools/buildman/buildman --fetch-arch all
Thomas Chou4acc2d42015-11-12 09:29:09 +0800453$ sudo mkdir -p /toolchains
454$ sudo mv ~/.buildman-toolchains/*/* /toolchains/
455
456For those not available from kernel.org, download from the following links.
457
458arc: https://github.com/foss-for-synopsys-dwc-arc-processors/toolchain/releases/
Alexey Brodkin14c0bde2017-01-26 17:25:59 +0300459 download/arc-2016.09-release/arc_gnu_2016.09_prebuilt_uclibc_le_archs_linux_install.tar.gz
Thomas Chou4acc2d42015-11-12 09:29:09 +0800460blackfin: http://sourceforge.net/projects/adi-toolchain/files/
461 blackfin-toolchain-elf-gcc-4.5-2014R1_45-RC2.x86_64.tar.bz2
462nds32: http://osdk.andestech.com/packages/
463 nds32le-linux-glibc-v1.tgz
464nios2: http://sourcery.mentor.com/public/gnu_toolchain/nios2-linux-gnu/
465 sourceryg++-2015.11-27-nios2-linux-gnu-i686-pc-linux-gnu.tar.bz2
466sh: http://sourcery.mentor.com/public/gnu_toolchain/sh-linux-gnu/
467 renesas-4.4-200-sh-linux-gnu-i686-pc-linux-gnu.tar.bz2
468
Simon Glass7b94ab52016-07-27 20:32:58 -0600469Note openrisc kernel.org toolchain is out of date. Download the latest one from
470http://opencores.org/or1k/OpenRISC_GNU_tool_chain#Prebuilt_versions - eg:
Bin Meng07a50f92016-02-21 21:18:02 -0800471ftp://ocuser:ocuser@openrisc.opencores.org/toolchain/gcc-or1k-elf-4.8.1-x86.tar.bz2.
472
Simon Glass7e803e12014-12-01 17:34:06 -0700473Buildman should now be set up to use your new toolchain.
474
475At the time of writing, U-Boot has these architectures:
476
Andy Shevchenko8cb5cdd2017-07-05 16:25:22 +0300477 arc, arm, blackfin, m68k, microblaze, mips, nds32, nios2, openrisc
Simon Glass7e803e12014-12-01 17:34:06 -0700478 powerpc, sandbox, sh, sparc, x86
479
Michal Simeke0e31f32015-04-20 11:46:24 +0200480Of these, only arc and nds32 are not available at kernel.org..
Simon Glass7e803e12014-12-01 17:34:06 -0700481
482
Simon Glassc05694f2013-04-03 11:07:16 +0000483How to run it
484=============
485
486First do a dry run using the -n flag: (replace <branch> with a real, local
487branch with a valid upstream)
488
489$ ./tools/buildman/buildman -b <branch> -n
490
491If it can't detect the upstream branch, try checking out the branch, and
Simon Glassf204ab12014-12-01 17:33:54 -0700492doing something like 'git branch --set-upstream-to upstream/master'
493or something similar. Buildman will try to guess a suitable upstream branch
494if it can't find one (you will see a message like" Guessing upstream as ...).
Simon Glassc05694f2013-04-03 11:07:16 +0000495
Simon Glassfa1cfee2014-08-09 15:32:57 -0600496As an example:
Simon Glassc05694f2013-04-03 11:07:16 +0000497
498Dry run, so not doing much. But I would do this:
499
500Building 18 commits for 1059 boards (4 threads, 1 job per thread)
501Build directory: ../lcd9b
502 5bb3505 Merge branch 'master' of git://git.denx.de/u-boot-arm
503 c18f1b4 tegra: Use const for pinmux_config_pingroup/table()
504 2f043ae tegra: Add display support to funcmux
505 e349900 tegra: fdt: Add pwm binding and node
506 424a5f0 tegra: fdt: Add LCD definitions for Tegra
507 0636ccf tegra: Add support for PWM
508 a994fe7 tegra: Add SOC support for display/lcd
509 fcd7350 tegra: Add LCD driver
510 4d46e9d tegra: Add LCD support to Nvidia boards
511 991bd48 arm: Add control over cachability of memory regions
512 54e8019 lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment
513 d92aff7 lcd: Add support for flushing LCD fb from dcache after update
514 dbd0677 tegra: Align LCD frame buffer to section boundary
515 0cff9b8 tegra: Support control of cache settings for LCD
516 9c56900 tegra: fdt: Add LCD definitions for Seaboard
517 5cc29db lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console
518 cac5a23 tegra: Enable display/lcd support on Seaboard
519 49ff541 wip
520
521Total boards to build for each commit: 1059
522
523This shows that it will build all 1059 boards, using 4 threads (because
524we have a 4-core CPU). Each thread will run with -j1, meaning that each
525make job will use a single CPU. The list of commits to be built helps you
526confirm that things look about right. Notice that buildman has chosen a
527'base' directory for you, immediately above your source tree.
528
529Buildman works entirely inside the base directory, here ../lcd9b,
530creating a working directory for each thread, and creating output
531directories for each commit and board.
532
533
534Suggested Workflow
535==================
536
537To run the build for real, take off the -n:
538
539$ ./tools/buildman/buildman -b <branch>
540
541Buildman will set up some working directories, and get started. After a
542minute or so it will settle down to a steady pace, with a display like this:
543
544Building 18 commits for 1059 boards (4 threads, 1 job per thread)
545 528 36 124 /19062 1:13:30 : SIMPC8313_SP
546
547This means that it is building 19062 board/commit combinations. So far it
Simon Glassfa1cfee2014-08-09 15:32:57 -0600548has managed to successfully build 528. Another 36 have built with warnings,
Simon Glassc05694f2013-04-03 11:07:16 +0000549and 124 more didn't build at all. Buildman expects to complete the process
Simon Glass7b94ab52016-07-27 20:32:58 -0600550in around an hour and a quarter. Use this time to buy a faster computer.
Simon Glassc05694f2013-04-03 11:07:16 +0000551
552
553To find out how the build went, ask for a summary with -s. You can do this
Dirk Behme5f01c7a2014-12-23 07:41:26 +0100554either before the build completes (presumably in another terminal) or
Simon Glassc05694f2013-04-03 11:07:16 +0000555afterwards. Let's work through an example of how this is used:
556
557$ ./tools/buildman/buildman -b lcd9b -s
558...
55901: Merge branch 'master' of git://git.denx.de/u-boot-arm
560 powerpc: + galaxy5200_LOWBOOT
56102: tegra: Use const for pinmux_config_pingroup/table()
56203: tegra: Add display support to funcmux
56304: tegra: fdt: Add pwm binding and node
56405: tegra: fdt: Add LCD definitions for Tegra
56506: tegra: Add support for PWM
56607: tegra: Add SOC support for display/lcd
56708: tegra: Add LCD driver
56809: tegra: Add LCD support to Nvidia boards
56910: arm: Add control over cachability of memory regions
57011: lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment
57112: lcd: Add support for flushing LCD fb from dcache after update
572 arm: + lubbock
57313: tegra: Align LCD frame buffer to section boundary
57414: tegra: Support control of cache settings for LCD
57515: tegra: fdt: Add LCD definitions for Seaboard
57616: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console
57717: tegra: Enable display/lcd support on Seaboard
57818: wip
579
580This shows which commits have succeeded and which have failed. In this case
581the build is still in progress so many boards are not built yet (use -u to
582see which ones). But still we can see a few failures. The galaxy5200_LOWBOOT
583never builds correctly. This could be a problem with our toolchain, or it
584could be a bug in the upstream. The good news is that we probably don't need
Simon Glass7b94ab52016-07-27 20:32:58 -0600585to blame our commits. The bad news is that our commits are not tested on that
586board.
Simon Glassc05694f2013-04-03 11:07:16 +0000587
588Commit 12 broke lubbock. That's what the '+ lubbock' means. The failure
589is never fixed by a later commit, or you would see lubbock again, in green,
590without the +.
591
592To see the actual error:
593
594$ ./tools/buildman/buildman -b <branch> -se lubbock
595...
59612: lcd: Add support for flushing LCD fb from dcache after update
597 arm: + lubbock
598+common/libcommon.o: In function `lcd_sync':
599+/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range'
600+arm-none-linux-gnueabi-ld: BFD (Sourcery G++ Lite 2010q1-202) 2.19.51.20090709 assertion fail /scratch/julian/2010q1-release-linux-lite/obj/binutils-src-2010q1-202-arm-none-linux-gnueabi-i686-pc-linux-gnu/bfd/elf32-arm.c:12572
601+make: *** [/u-boot/lcd9b/.bm-work/00/build/u-boot] Error 139
60213: tegra: Align LCD frame buffer to section boundary
60314: tegra: Support control of cache settings for LCD
60415: tegra: fdt: Add LCD definitions for Seaboard
60516: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console
606-/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range'
607+/u-boot/lcd9b/.bm-work/00/common/lcd.c:125: undefined reference to `flush_dcache_range'
60817: tegra: Enable display/lcd support on Seaboard
60918: wip
610
611So the problem is in lcd.c, due to missing cache operations. This information
612should be enough to work out what that commit is doing to break these
613boards. (In this case pxa did not have cache operations defined).
614
Simon Glass7b94ab52016-07-27 20:32:58 -0600615If you see error lines marked with '-', that means that the errors were fixed
Simon Glassc05694f2013-04-03 11:07:16 +0000616by that commit. Sometimes commits can be in the wrong order, so that a
617breakage is introduced for a few commits and fixed by later commits. This
618shows up clearly with buildman. You can then reorder the commits and try
619again.
620
Simon Glass7b94ab52016-07-27 20:32:58 -0600621At commit 16, the error moves: you can see that the old error at line 120
Simon Glassc05694f2013-04-03 11:07:16 +0000622is fixed, but there is a new one at line 126. This is probably only because
Dirk Behme5f01c7a2014-12-23 07:41:26 +0100623we added some code and moved the broken line further down the file.
Simon Glassc05694f2013-04-03 11:07:16 +0000624
625If many boards have the same error, then -e will display the error only
Simon Glass3394c9f2014-08-28 09:43:43 -0600626once. This makes the output as concise as possible. To see which boards have
Simon Glass7b94ab52016-07-27 20:32:58 -0600627each error, use -l. So it is safe to omit the board name - you will not get
628lots of repeated output for every board.
Simon Glassc05694f2013-04-03 11:07:16 +0000629
Simon Glass03749d42014-08-28 09:43:44 -0600630Buildman tries to distinguish warnings from errors, and shows warning lines
631separately with a 'w' prefix.
632
Simon Glassc05694f2013-04-03 11:07:16 +0000633The full build output in this case is available in:
634
635../lcd9b/12_of_18_gd92aff7_lcd--Add-support-for/lubbock/
636
637 done: Indicates the build was done, and holds the return code from make.
638 This is 0 for a good build, typically 2 for a failure.
639
640 err: Output from stderr, if any. Errors and warnings appear here.
641
642 log: Output from stdout. Normally there isn't any since buildman runs
Simon Glass54e628e2016-03-12 18:50:33 -0700643 in silent mode. Use -V to force a verbose build (this passes V=1
644 to 'make')
Simon Glassc05694f2013-04-03 11:07:16 +0000645
646 toolchain: Shows information about the toolchain used for the build.
647
648 sizes: Shows image size information.
649
Simon Glass7b94ab52016-07-27 20:32:58 -0600650It is possible to get the build binary output there also. Use the -k option
651for this. In that case you will also see some output files, like:
Simon Glassc05694f2013-04-03 11:07:16 +0000652
653 System.map toolchain u-boot u-boot.bin u-boot.map autoconf.mk
654 (also SPL versions u-boot-spl and u-boot-spl.bin if available)
655
656
657Checking Image Sizes
658====================
659
660A key requirement for U-Boot is that you keep code/data size to a minimum.
661Where a new feature increases this noticeably it should normally be put
Simon Glass7b94ab52016-07-27 20:32:58 -0600662behind a CONFIG flag so that boards can leave it disabled and keep the image
Simon Glassc05694f2013-04-03 11:07:16 +0000663size more or less the same with each new release.
664
665To check the impact of your commits on image size, use -S. For example:
666
667$ ./tools/buildman/buildman -b us-x86 -sS
668Summary of 10 commits for 1066 boards (4 threads, 1 job per thread)
66901: MAKEALL: add support for per architecture toolchains
67002: x86: Add function to get top of usable ram
671 x86: (for 1/3 boards) text -272.0 rodata +41.0
67203: x86: Add basic cache operations
67304: x86: Permit bootstage and timer data to be used prior to relocation
674 x86: (for 1/3 boards) data +16.0
67505: x86: Add an __end symbol to signal the end of the U-Boot binary
676 x86: (for 1/3 boards) text +76.0
67706: x86: Rearrange the output input to remove BSS
678 x86: (for 1/3 boards) bss -2140.0
67907: x86: Support relocation of FDT on start-up
680 x86: + coreboot-x86
68108: x86: Add error checking to x86 relocation code
68209: x86: Adjust link device tree include file
68310: x86: Enable CONFIG_OF_CONTROL on coreboot
684
685
686You can see that image size only changed on x86, which is good because this
687series is not supposed to change any other board. From commit 7 onwards the
688build fails so we don't get code size numbers. The numbers are fractional
689because they are an average of all boards for that architecture. The
690intention is to allow you to quickly find image size problems introduced by
691your commits.
692
693Note that the 'text' region and 'rodata' are split out. You should add the
694two together to get the total read-only size (reported as the first column
695in the output from binutil's 'size' utility).
696
697A useful option is --step which lets you skip some commits. For example
698--step 2 will show the image sizes for only every 2nd commit (so it will
699compare the image sizes of the 1st, 3rd, 5th... commits). You can also use
700--step 0 which will compare only the first and last commits. This is useful
Simon Glass7b94ab52016-07-27 20:32:58 -0600701for an overview of how your entire series affects code size. It will build
702only the upstream commit and your final branch commit.
Simon Glassc05694f2013-04-03 11:07:16 +0000703
704You can also use -d to see a detailed size breakdown for each board. This
705list is sorted in order from largest growth to largest reduction.
706
Simon Glass7b94ab52016-07-27 20:32:58 -0600707It is even possible to go a little further with the -B option (--bloat). This
Simon Glassfa1cfee2014-08-09 15:32:57 -0600708shows where U-Boot has bloated, breaking the size change down to the function
Simon Glassc05694f2013-04-03 11:07:16 +0000709level. Example output is below:
710
711$ ./tools/buildman/buildman -b us-mem4 -sSdB
712...
71319: Roll crc32 into hash infrastructure
714 arm: (for 10/10 boards) all -143.4 bss +1.2 data -4.8 rodata -48.2 text -91.6
715 paz00 : all +23 bss -4 rodata -29 text +56
716 u-boot: add: 1/0, grow: 3/-2 bytes: 168/-104 (64)
717 function old new delta
718 hash_command 80 160 +80
719 crc32_wd_buf - 56 +56
720 ext4fs_read_file 540 568 +28
721 insert_var_value_sub 688 692 +4
722 run_list_real 1996 1992 -4
723 do_mem_crc 168 68 -100
724 trimslice : all -9 bss +16 rodata -29 text +4
725 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12)
726 function old new delta
727 hash_command 80 160 +80
728 crc32_wd_buf - 56 +56
729 ext4fs_iterate_dir 672 668 -4
730 ext4fs_read_file 568 548 -20
731 do_mem_crc 168 68 -100
732 whistler : all -9 bss +16 rodata -29 text +4
733 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12)
734 function old new delta
735 hash_command 80 160 +80
736 crc32_wd_buf - 56 +56
737 ext4fs_iterate_dir 672 668 -4
738 ext4fs_read_file 568 548 -20
739 do_mem_crc 168 68 -100
740 seaboard : all -9 bss -28 rodata -29 text +48
741 u-boot: add: 1/0, grow: 3/-2 bytes: 160/-104 (56)
742 function old new delta
743 hash_command 80 160 +80
744 crc32_wd_buf - 56 +56
745 ext4fs_read_file 548 568 +20
746 run_list_real 1996 2000 +4
747 do_nandboot 760 756 -4
748 do_mem_crc 168 68 -100
Marcel Ziswilercad18b82015-03-26 01:31:54 +0100749 colibri_t20 : all -9 rodata -29 text +20
Simon Glassc05694f2013-04-03 11:07:16 +0000750 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-112 (28)
751 function old new delta
752 hash_command 80 160 +80
753 crc32_wd_buf - 56 +56
754 read_abs_bbt 204 208 +4
755 do_nandboot 760 756 -4
756 ext4fs_read_file 576 568 -8
757 do_mem_crc 168 68 -100
758 ventana : all -37 bss -12 rodata -29 text +4
759 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12)
760 function old new delta
761 hash_command 80 160 +80
762 crc32_wd_buf - 56 +56
763 ext4fs_iterate_dir 672 668 -4
764 ext4fs_read_file 568 548 -20
765 do_mem_crc 168 68 -100
766 harmony : all -37 bss -16 rodata -29 text +8
767 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-124 (16)
768 function old new delta
769 hash_command 80 160 +80
770 crc32_wd_buf - 56 +56
771 nand_write_oob_syndrome 428 432 +4
772 ext4fs_iterate_dir 672 668 -4
773 ext4fs_read_file 568 548 -20
774 do_mem_crc 168 68 -100
775 medcom-wide : all -417 bss +28 data -16 rodata -93 text -336
776 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288)
777 function old new delta
778 crc32_wd_buf - 56 +56
779 do_fat_read_at 2872 2904 +32
780 hash_algo 16 - -16
781 do_mem_crc 168 68 -100
782 hash_command 420 160 -260
783 tec : all -449 bss -4 data -16 rodata -93 text -336
784 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288)
785 function old new delta
786 crc32_wd_buf - 56 +56
787 do_fat_read_at 2872 2904 +32
788 hash_algo 16 - -16
789 do_mem_crc 168 68 -100
790 hash_command 420 160 -260
791 plutux : all -481 bss +16 data -16 rodata -93 text -388
792 u-boot: add: 1/-1, grow: 1/-3 bytes: 68/-408 (-340)
793 function old new delta
794 crc32_wd_buf - 56 +56
795 do_load_serial_bin 1688 1700 +12
796 hash_algo 16 - -16
797 do_fat_read_at 2904 2872 -32
798 do_mem_crc 168 68 -100
799 hash_command 420 160 -260
800 powerpc: (for 5/5 boards) all +37.4 data -3.2 rodata -41.8 text +82.4
801 MPC8610HPCD : all +55 rodata -29 text +84
802 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
803 function old new delta
804 hash_command - 176 +176
805 do_mem_crc 184 88 -96
806 MPC8641HPCN : all +55 rodata -29 text +84
807 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
808 function old new delta
809 hash_command - 176 +176
810 do_mem_crc 184 88 -96
811 MPC8641HPCN_36BIT: all +55 rodata -29 text +84
812 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
813 function old new delta
814 hash_command - 176 +176
815 do_mem_crc 184 88 -96
816 sbc8641d : all +55 rodata -29 text +84
817 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
818 function old new delta
819 hash_command - 176 +176
820 do_mem_crc 184 88 -96
821 xpedite517x : all -33 data -16 rodata -93 text +76
822 u-boot: add: 1/-1, grow: 0/-1 bytes: 176/-112 (64)
823 function old new delta
824 hash_command - 176 +176
825 hash_algo 16 - -16
826 do_mem_crc 184 88 -96
827...
828
829
Simon Glass7b94ab52016-07-27 20:32:58 -0600830This shows that commit 19 has reduced codesize for arm slightly and increased
831it for powerpc. This increase was offset in by reductions in rodata and
832data/bss.
Simon Glassc05694f2013-04-03 11:07:16 +0000833
Dirk Behme5f01c7a2014-12-23 07:41:26 +0100834Shown below the summary lines are the sizes for each board. Below each board
835are the sizes for each function. This information starts with:
Simon Glassc05694f2013-04-03 11:07:16 +0000836
837 add - number of functions added / removed
838 grow - number of functions which grew / shrunk
839 bytes - number of bytes of code added to / removed from all functions,
840 plus the total byte change in brackets
841
842The change seems to be that hash_command() has increased by more than the
843do_mem_crc() function has decreased. The function sizes typically add up to
844roughly the text area size, but note that every read-only section except
845rodata is included in 'text', so the function total does not exactly
846correspond.
847
848It is common when refactoring code for the rodata to decrease as the text size
849increases, and vice versa.
850
851
Simon Glass3cb033e2014-12-01 17:34:02 -0700852The .buildman file
853==================
854
855The .buildman file provides information about the available toolchains and
856also allows build flags to be passed to 'make'. It consists of several
857sections, with the section name in square brackets. Within each section are
858a set of (tag, value) pairs.
859
860'[toolchain]' section
861
862 This lists the available toolchains. The tag here doesn't matter, but
863 make sure it is unique. The value is the path to the toolchain. Buildman
864 will look in that path for a file ending in 'gcc'. It will then execute
865 it to check that it is a C compiler, passing only the --version flag to
866 it. If the return code is 0, buildman assumes that it is a valid C
867 compiler. It uses the first part of the name as the architecture and
868 strips off the last part when setting the CROSS_COMPILE environment
869 variable (parts are delimited with a hyphen).
870
871 For example powerpc-linux-gcc will be noted as a toolchain for 'powerpc'
872 and CROSS_COMPILE will be set to powerpc-linux- when using it.
873
874'[toolchain-alias]' section
875
876 This converts toolchain architecture names to U-Boot names. For example,
877 if an x86 toolchains is called i386-linux-gcc it will not normally be
Simon Glassc1528c12014-12-01 17:34:05 -0700878 used for architecture 'x86'. Adding 'x86: i386 x86_64' to this section
879 will tell buildman that the i386 and x86_64 toolchains can be used for
880 the x86 architecture.
Simon Glass3cb033e2014-12-01 17:34:02 -0700881
882'[make-flags]' section
883
884 U-Boot's build system supports a few flags (such as BUILD_TAG) which
885 affect the build product. These flags can be specified in the buildman
886 settings file. They can also be useful when building U-Boot against other
887 open source software.
888
889 [make-flags]
890 at91-boards=ENABLE_AT91_TEST=1
891 snapper9260=${at91-boards} BUILD_TAG=442
892 snapper9g45=${at91-boards} BUILD_TAG=443
Simon Glasscc246fb2013-09-23 17:35:17 -0600893
Simon Glass3cb033e2014-12-01 17:34:02 -0700894 This will use 'make ENABLE_AT91_TEST=1 BUILD_TAG=442' for snapper9260
895 and 'make ENABLE_AT91_TEST=1 BUILD_TAG=443' for snapper9g45. A special
896 variable ${target} is available to access the target name (snapper9260
897 and snapper9g20 in this case). Variables are resolved recursively. Note
898 that variables can only contain the characters A-Z, a-z, 0-9, hyphen (-)
899 and underscore (_).
Simon Glasscc246fb2013-09-23 17:35:17 -0600900
Simon Glass3cb033e2014-12-01 17:34:02 -0700901 It is expected that any variables added are dealt with in U-Boot's
902 config.mk file and documented in the README.
Simon Glasscc246fb2013-09-23 17:35:17 -0600903
Simon Glass3cb033e2014-12-01 17:34:02 -0700904 Note that you can pass ad-hoc options to the build using environment
905 variables, for example:
Simon Glasscc246fb2013-09-23 17:35:17 -0600906
Simon Glass3cb033e2014-12-01 17:34:02 -0700907 SOME_OPTION=1234 ./tools/buildman/buildman my_board
Simon Glasscc246fb2013-09-23 17:35:17 -0600908
909
Simon Glass78e418e2014-08-09 15:33:03 -0600910Quick Sanity Check
911==================
912
913If you have made changes and want to do a quick sanity check of the
Simon Glassd8158f92014-10-16 01:05:56 -0600914currently checked-out source, run buildman without the -b flag. This will
915build the selected boards and display build status as it runs (i.e. -v is
916enabled automatically). Use -e to see errors/warnings as well.
Simon Glass78e418e2014-08-09 15:33:03 -0600917
918
Simon Glass5eeef462014-12-01 17:33:57 -0700919Building Ranges
920===============
921
922You can build a range of commits by specifying a range instead of a branch
923when using the -b flag. For example:
924
925 upstream/master..us-buildman
926
927will build commits in us-buildman that are not in upstream/master.
928
929
Stephen Warren97c96902016-04-11 10:48:44 -0600930Building Faster
931===============
932
933By default, buildman executes 'make mrproper' prior to building the first
934commit for each board. This causes everything to be built from scratch. If you
935trust the build system's incremental build capabilities, you can pass the -I
936flag to skip the 'make mproper' invocation, which will reduce the amount of
937work 'make' does, and hence speed up the build. This flag will speed up any
938buildman invocation, since it reduces the amount of work done on any build.
939
940One possible application of buildman is as part of a continual edit, build,
941edit, build, ... cycle; repeatedly applying buildman to the same change or
942series of changes while making small incremental modifications to the source
943each time. This provides quick feedback regarding the correctness of recent
944modifications. In this scenario, buildman's default choice of build directory
945causes more build work to be performed than strictly necessary.
946
947By default, each buildman thread uses a single directory for all builds. When a
948thread builds multiple boards, the configuration built in this directory will
949cycle through various different configurations, one per board built by the
950thread. Variations in the configuration will force a rebuild of affected source
951files when a thread switches between boards. Ideally, such buildman-induced
952rebuilds would not happen, thus allowing the build to operate as efficiently as
953the build system and source changes allow. buildman's -P flag may be used to
954enable this; -P causes each board to be built in a separate (board-specific)
955directory, thus avoiding any buildman-induced configuration changes in any
956build directory.
957
958U-Boot's build system embeds information such as a build timestamp into the
959final binary. This information varies each time U-Boot is built. This causes
960various files to be rebuilt even if no source changes are made, which in turn
961requires that the final U-Boot binary be re-linked. This unnecessary work can
962be avoided by turning off the timestamp feature. This can be achieved by
963setting the SOURCE_DATE_EPOCH environment variable to 0.
964
965Combining all of these options together yields the command-line shown below.
966This will provide the quickest possible feedback regarding the current content
967of the source tree, thus allowing rapid tested evolution of the code.
968
969 SOURCE_DATE_EPOCH=0 ./tools/buildman/buildman -I -P tegra
970
971
Simon Glassb50ce022016-11-13 14:25:52 -0700972Checking configuration
973======================
974
975A common requirement when converting CONFIG options to Kconfig is to check
976that the effective configuration has not changed due to the conversion.
977Buildman supports this with the -K option, used after a build. This shows
978differences in effective configuration between one commit and the next.
979
980For example:
981
982 $ buildman -b kc4 -sK
983 ...
984 43: Convert CONFIG_SPL_USBETH_SUPPORT to Kconfig
985 arm:
986 + u-boot.cfg: CONFIG_SPL_ENV_SUPPORT=1 CONFIG_SPL_NET_SUPPORT=1
987 + u-boot-spl.cfg: CONFIG_SPL_MMC_SUPPORT=1 CONFIG_SPL_NAND_SUPPORT=1
988 + all: CONFIG_SPL_ENV_SUPPORT=1 CONFIG_SPL_MMC_SUPPORT=1 CONFIG_SPL_NAND_SUPPORT=1 CONFIG_SPL_NET_SUPPORT=1
989 am335x_evm_usbspl :
990 + u-boot.cfg: CONFIG_SPL_ENV_SUPPORT=1 CONFIG_SPL_NET_SUPPORT=1
991 + u-boot-spl.cfg: CONFIG_SPL_MMC_SUPPORT=1 CONFIG_SPL_NAND_SUPPORT=1
992 + all: CONFIG_SPL_ENV_SUPPORT=1 CONFIG_SPL_MMC_SUPPORT=1 CONFIG_SPL_NAND_SUPPORT=1 CONFIG_SPL_NET_SUPPORT=1
993 44: Convert CONFIG_SPL_USB_HOST_SUPPORT to Kconfig
994 ...
995
996This shows that commit 44 enabled three new options for the board
997am335x_evm_usbspl which were not enabled in commit 43. There is also a
998summary for 'arm' showing all the changes detected for that architecture.
999In this case there is only one board with changes, so 'arm' output is the
1000same as 'am335x_evm_usbspl'/
1001
1002The -K option uses the u-boot.cfg, spl/u-boot-spl.cfg and tpl/u-boot-tpl.cfg
1003files which are produced by a build. If all you want is to check the
1004configuration you can in fact avoid doing a full build, using -D. This tells
1005buildman to configuration U-Boot and create the .cfg files, but not actually
1006build the source. This is 5-10 times faster than doing a full build.
1007
Simon Glasscde5c302016-11-13 14:25:53 -07001008By default buildman considers the follow two configuration methods
1009equivalent:
1010
1011 #define CONFIG_SOME_OPTION
1012
1013 CONFIG_SOME_OPTION=y
1014
1015The former would appear in a header filer and the latter in a defconfig
1016file. The achieve this, buildman considers 'y' to be '1' in configuration
1017variables. This avoids lots of useless output when converting a CONFIG
1018option to Kconfig. To disable this behaviour, use --squash-config-y.
1019
Simon Glassb50ce022016-11-13 14:25:52 -07001020
Alex Kiernan4059e302018-05-31 04:48:34 +00001021Checking the environment
1022========================
1023
1024When converting CONFIG options which manipulate the default environment,
1025a common requirement is to check that the default environment has not
1026changed due to the conversion. Buildman supports this with the -U option,
1027used after a build. This shows differences in the default environment
1028between one commit and the next.
1029
1030For example:
1031
1032$ buildman -b squash brppt1 -sU
1033boards.cfg is up to date. Nothing to do.
1034Summary of 2 commits for 3 boards (3 threads, 3 jobs per thread)
103501: Migrate bootlimit to Kconfig
103602: Squashed commit of the following:
1037 c brppt1_mmc: altbootcmd=mmc dev 1; run mmcboot0; -> mmc dev 1; run mmcboot0
1038 c brppt1_spi: altbootcmd=mmc dev 1; run mmcboot0; -> mmc dev 1; run mmcboot0
1039 + brppt1_nand: altbootcmd=run usbscript
1040 - brppt1_nand: altbootcmd=run usbscript
1041(no errors to report)
1042
1043This shows that commit 2 modified the value of 'altbootcmd' for 'brppt1_mmc'
1044and 'brppt1_spi', removing a trailing semicolon. 'brppt1_nand' gained an a
1045value for 'altbootcmd', but lost one for ' altbootcmd'.
1046
1047The -U option uses the u-boot.env files which are produced by a build.
1048
Simon Glassf77ca5b2019-01-07 16:44:20 -07001049
1050Building with clang
1051===================
1052
1053To build with clang (sandbox only), use the -O option to override the
1054toolchain. For example:
1055
1056 buildman -O clang-7 --board sandbox
1057
1058
Simon Glassc05694f2013-04-03 11:07:16 +00001059Other options
1060=============
1061
1062Buildman has various other command line options. Try --help to see them.
1063
Simon Glassc2f91072014-08-28 09:43:39 -06001064When doing builds, Buildman's return code will reflect the overall result:
1065
1066 0 (success) No errors or warnings found
1067 128 Errors found
1068 129 Warnings found
1069
Simon Glassc05694f2013-04-03 11:07:16 +00001070
Simon Glass3ec03d42014-08-09 15:32:58 -06001071How to change from MAKEALL
1072==========================
1073
1074Buildman includes most of the features of MAKEALL and is generally faster
1075and easier to use. In particular it builds entire branches: if a particular
1076commit introduces an error in a particular board, buildman can easily show
1077you this, even if a later commit fixes that error.
1078
1079The reasons to deprecate MAKEALL are:
1080- We don't want to maintain two build systems
1081- Buildman is typically faster
1082- Buildman has a lot more features
1083
1084But still, many people will be sad to lose MAKEALL. If you are used to
1085MAKEALL, here are a few pointers.
1086
1087First you need to set up your tool chains - see the 'Setting up' section
1088for details. Once you have your required toolchain(s) detected then you are
1089ready to go.
1090
Simon Glass78e418e2014-08-09 15:33:03 -06001091To build the current source tree, run buildman without a -b flag:
1092
1093 ./tools/buildman/buildman <list of things to build>
1094
1095This will build the current source tree for the given boards and display
1096the results and errors.
1097
1098However buildman usually works on entire branches, and for that you must
1099specify a board flag:
Simon Glass3ec03d42014-08-09 15:32:58 -06001100
1101 ./tools/buildman/buildman -b <branch_name> <list of things to build>
1102
1103followed by (afterwards, or perhaps concurrently in another terminal):
1104
1105 ./tools/buildman/buildman -b <branch_name> -s <list of things to build>
1106
1107to see the results of the build. Rather than showing you all the output,
1108buildman just shows a summary, with red indicating that a commit introduced
1109an error and green indicating that a commit fixed an error. Use the -e
Simon Glass3394c9f2014-08-28 09:43:43 -06001110flag to see the full errors and -l to see which boards caused which errors.
Simon Glass3ec03d42014-08-09 15:32:58 -06001111
Simon Glass78e418e2014-08-09 15:33:03 -06001112If you really want to see build results as they happen, use -v when doing a
Simon Glassd8158f92014-10-16 01:05:56 -06001113build (and -e to see the errors/warnings too).
Simon Glass78e418e2014-08-09 15:33:03 -06001114
Simon Glass3ec03d42014-08-09 15:32:58 -06001115You don't need to stick around on that branch while buildman is running. It
1116checks out its own copy of the source code, so you can change branches,
1117add commits, etc. without affecting the build in progress.
1118
1119The <list of things to build> can include board names, architectures or the
1120like. There are no flags to disambiguate since ambiguities are rare. Using
1121the examples from MAKEALL:
1122
1123Examples:
1124 - build all Power Architecture boards:
1125 MAKEALL -a powerpc
1126 MAKEALL --arch powerpc
1127 MAKEALL powerpc
1128 ** buildman -b <branch> powerpc
1129 - build all PowerPC boards manufactured by vendor "esd":
1130 MAKEALL -a powerpc -v esd
1131 ** buildman -b <branch> esd
1132 - build all PowerPC boards manufactured either by "keymile" or "siemens":
1133 MAKEALL -a powerpc -v keymile -v siemens
1134 ** buildman -b <branch> keymile siemens
1135 - build all Freescale boards with MPC83xx CPUs, plus all 4xx boards:
1136 MAKEALL -c mpc83xx -v freescale 4xx
1137 ** buildman -b <branch> mpc83xx freescale 4xx
1138
1139Buildman automatically tries to use all the CPUs in your machine. If you
1140are building a lot of boards it will use one thread for every CPU core
1141it detects in your machine. This is like MAKEALL's BUILD_NBUILDS option.
1142You can use the -T flag to change the number of threads. If you are only
1143building a few boards, buildman will automatically run make with the -j
1144flag to increase the number of concurrent make tasks. It isn't normally
1145that helpful to fiddle with this option, but if you use the BUILD_NCPUS
1146option in MAKEALL then -j is the equivalent in buildman.
1147
1148Buildman puts its output in ../<branch_name> by default but you can change
1149this with the -o option. Buildman normally does out-of-tree builds: use -i
1150to disable that if you really want to. But be careful that once you have
1151used -i you pollute buildman's copies of the source tree, and you will need
1152to remove the build directory (normally ../<branch_name>) to run buildman
1153in normal mode (without -i).
1154
1155Buildman doesn't keep the output result normally, but use the -k option to
1156do this.
1157
1158Please read 'Theory of Operation' a few times as it will make a lot of
1159things clearer.
1160
1161Some options you might like are:
1162
1163 -B shows which functions are growing/shrinking in which commit - great
1164 for finding code bloat.
1165 -S shows image sizes for each commit (just an overall summary)
1166 -u shows boards that you haven't built yet
1167 --step 0 will build just the upstream commit and the last commit of your
1168 branch. This is often a quick sanity check that your branch doesn't
1169 break anything. But note this does not check bisectability!
1170
1171
Simon Glassc05694f2013-04-03 11:07:16 +00001172TODO
1173====
1174
1175This has mostly be written in my spare time as a response to my difficulties
1176in testing large series of patches. Apart from tidying up there is quite a
Simon Glassd8158f92014-10-16 01:05:56 -06001177bit of scope for improvement. Things like better error diffs and easier
Dirk Behme5f01c7a2014-12-23 07:41:26 +01001178access to log files. Also it would be nice if buildman could 'hunt' for
Simon Glassd8158f92014-10-16 01:05:56 -06001179problems, perhaps by building a few boards for each arch, or checking
1180commits for changed files and building only boards which use those files.
Simon Glassc05694f2013-04-03 11:07:16 +00001181
1182
1183Credits
1184=======
1185
1186Thanks to Grant Grundler <grundler@chromium.org> for his ideas for improving
1187the build speed by building all commits for a board instead of the other
1188way around.
1189
1190
Simon Glassc05694f2013-04-03 11:07:16 +00001191Simon Glass
1192sjg@chromium.org
1193Halloween 2012
1194Updated 12-12-12
1195Updated 23-02-13