1 # Copyright (c) 2013 The Chromium OS Authors.
3 # SPDX-License-Identifier: GPL-2.0+
6 (Please read 'How to change from MAKEALL' if you are used to that tool)
11 This tool handles building U-Boot to check that you have not broken it
12 with your patch series. It can build each individual commit and report
13 which boards fail on which commits, and which errors come up. It aims
14 to make full use of multi-processor machines.
16 A key feature of buildman is its output summary, which allows warnings,
17 errors or image size increases in a particular commit or board to be
18 quickly identified and the offending commit pinpointed. This can be a big
19 help for anyone working with >10 patches at a time.
25 Buildman is still in its infancy. It is already a very useful tool, but
26 expect to find problems and send patches.
28 Buildman can be stopped and restarted, in which case it will continue
29 where it left off. This should happen cleanly and without side-effects.
30 If not, it is a bug, for which a patch would be welcome.
32 Buildman gets so tied up in its work that it can ignore the outside world.
33 You may need to press Ctrl-C several times to quit it. Also it will print
34 out various exceptions when stopped.
40 (please read this section in full twice or you will be perpetually confused)
42 Buildman is a builder. It is not make, although it runs make. It does not
43 produce any useful output on the terminal while building, except for
44 progress information (except with -v, see below). All the output (errors,
45 warnings and binaries if you are ask for them) is stored in output
46 directories, which you can look at while the build is progressing, or when
49 Buildman produces a concise summary of which boards succeeded and failed.
50 It shows which commit introduced which board failure using a simple
51 red/green colour coding. Full error information can be requested, in which
52 case it is de-duped and displayed against the commit that introduced the
53 error. An example workflow is below.
55 Buildman stores image size information and can report changes in image size
56 from commit to commit. An example of this is below.
58 Buildman starts multiple threads, and each thread builds for one board at
59 a time. A thread starts at the first commit, configures the source for your
60 board and builds it. Then it checks out the next commit and does an
61 incremental build. Eventually the thread reaches the last commit and stops.
62 If errors or warnings are found along the way, the thread will reconfigure
63 after every commit, and your build will be very slow. This is because a
64 file that produces just a warning would not normally be rebuilt in an
67 Buildman works in an entirely separate place from your U-Boot repository.
68 It creates a separate working directory for each thread, and puts the
69 output files in the working directory, organised by commit name and board
70 name, in a two-level hierarchy.
72 Buildman is invoked in your U-Boot directory, the one with the .git
73 directory. It clones this repository into a copy for each thread, and the
74 threads do not affect the state of your git repository. Any checkouts done
75 by the thread affect only the working directory for that thread.
77 Buildman automatically selects the correct tool chain for each board. You
78 must supply suitable tool chains, but buildman takes care of selecting the
81 Buildman generally builds a branch (with the -b flag), and in this case
82 builds the upstream commit as well, for comparison. It cannot build
83 individual commits at present, unless (maybe) you point it at an empty
84 branch. Put all your commits in a branch, set the branch's upstream to a
85 valid value, and all will be well. Otherwise buildman will perform random
86 actions. Use -n to check what the random actions might be.
88 If you just want to build the current source tree, leave off the -b flag.
89 This will display results and errors as they happen. You can still look
90 at them later using -s. Note that buildman will assume that the source
91 has changed, and will build all specified boards in this case.
93 Buildman is optimised for building many commits at once, for many boards.
94 On multi-core machines, Buildman is fast because it uses most of the
95 available CPU power. When it gets to the end, or if you are building just
96 a few commits or boards, it will be pretty slow. As a tip, if you don't
97 plan to use your machine for anything else, you can use -T to increase the
98 number of threads beyond the default.
100 Buildman lets you build all boards, or a subset. Specify the subset by passing
101 command-line arguments that list the desired board name, architecture name,
102 SOC name, or anything else in the boards.cfg file. Multiple arguments are
103 allowed. Each argument will be interpreted as a regular expression, so
104 behaviour is a superset of exact or substring matching. Examples are:
106 * 'tegra20' All boards with a Tegra20 SoC
107 * 'tegra' All boards with any Tegra Soc (Tegra20, Tegra30, Tegra114...)
108 * '^tegra[23]0$' All boards with either Tegra20 or Tegra30 SoC
109 * 'powerpc' All PowerPC boards
111 While the default is to OR the terms together, you can also make use of
112 the '&' operator to limit the selection:
114 * 'freescale & arm sandbox' All Freescale boards with ARM architecture,
117 It is convenient to use the -n option to see whaat will be built based on
120 Buildman does not store intermediate object files. It optionally copies
121 the binary output into a directory when a build is successful. Size
122 information is always recorded. It needs a fair bit of disk space to work,
123 typically 250MB per thread.
129 1. Get the U-Boot source. You probably already have it, but if not these
130 steps should get you started with a repo and some commits for testing.
133 $ git clone git://git.denx.de/u-boot.git .
134 $ git checkout -b my-branch origin/master
135 $ # Add some commits to the branch, reading for testing
137 2. Create ~/.buildman to tell buildman where to find tool chains. As an
140 # Buildman settings file
146 arm: /opt/linaro/gcc-linaro-arm-linux-gnueabihf-4.8-2013.08_linux
147 aarch64: /opt/linaro/gcc-linaro-aarch64-none-elf-4.8-2013.10_linux
157 This selects the available toolchain paths. Add the base directory for
158 each of your toolchains here. Buildman will search inside these directories
159 and also in any '/usr' and '/usr/bin' subdirectories.
161 Make sure the tags (here root: rest: and eldk:) are unique.
163 The toolchain-alias section indicates that the i386 toolchain should be used
164 to build x86 commits.
167 2. Check the available toolchains
169 Run this check to make sure that you have a toolchain for every architecture.
171 $ ./tools/buildman/buildman --list-tool-chains
172 Scanning for tool chains
176 - looking in '/usr/bin'
177 - found '/usr/bin/gcc'
179 - found '/usr/bin/c89-gcc'
181 - found '/usr/bin/c99-gcc'
183 - found '/usr/bin/x86_64-linux-gnu-gcc'
185 - scanning path '/toolchains/powerpc-linux'
186 - looking in '/toolchains/powerpc-linux/.'
187 - looking in '/toolchains/powerpc-linux/bin'
188 - found '/toolchains/powerpc-linux/bin/powerpc-linux-gcc'
190 - looking in '/toolchains/powerpc-linux/usr/bin'
191 - scanning path '/toolchains/nds32le-linux-glibc-v1f'
192 - looking in '/toolchains/nds32le-linux-glibc-v1f/.'
193 - looking in '/toolchains/nds32le-linux-glibc-v1f/bin'
194 - found '/toolchains/nds32le-linux-glibc-v1f/bin/nds32le-linux-gcc'
196 - looking in '/toolchains/nds32le-linux-glibc-v1f/usr/bin'
197 - scanning path '/toolchains/nios2'
198 - looking in '/toolchains/nios2/.'
199 - looking in '/toolchains/nios2/bin'
200 - found '/toolchains/nios2/bin/nios2-linux-gcc'
202 - found '/toolchains/nios2/bin/nios2-linux-uclibc-gcc'
204 - looking in '/toolchains/nios2/usr/bin'
205 - found '/toolchains/nios2/usr/bin/nios2-linux-gcc'
207 - found '/toolchains/nios2/usr/bin/nios2-linux-uclibc-gcc'
209 - scanning path '/toolchains/microblaze-unknown-linux-gnu'
210 - looking in '/toolchains/microblaze-unknown-linux-gnu/.'
211 - looking in '/toolchains/microblaze-unknown-linux-gnu/bin'
212 - found '/toolchains/microblaze-unknown-linux-gnu/bin/microblaze-unknown-linux-gnu-gcc'
214 - found '/toolchains/microblaze-unknown-linux-gnu/bin/mb-linux-gcc'
216 - looking in '/toolchains/microblaze-unknown-linux-gnu/usr/bin'
217 - scanning path '/toolchains/mips-linux'
218 - looking in '/toolchains/mips-linux/.'
219 - looking in '/toolchains/mips-linux/bin'
220 - found '/toolchains/mips-linux/bin/mips-linux-gcc'
222 - looking in '/toolchains/mips-linux/usr/bin'
223 - scanning path '/toolchains/old'
224 - looking in '/toolchains/old/.'
225 - looking in '/toolchains/old/bin'
226 - looking in '/toolchains/old/usr/bin'
227 - scanning path '/toolchains/i386-linux'
228 - looking in '/toolchains/i386-linux/.'
229 - looking in '/toolchains/i386-linux/bin'
230 - found '/toolchains/i386-linux/bin/i386-linux-gcc'
232 - looking in '/toolchains/i386-linux/usr/bin'
233 - scanning path '/toolchains/bfin-uclinux'
234 - looking in '/toolchains/bfin-uclinux/.'
235 - looking in '/toolchains/bfin-uclinux/bin'
236 - found '/toolchains/bfin-uclinux/bin/bfin-uclinux-gcc'
238 - looking in '/toolchains/bfin-uclinux/usr/bin'
239 - scanning path '/toolchains/sparc-elf'
240 - looking in '/toolchains/sparc-elf/.'
241 - looking in '/toolchains/sparc-elf/bin'
242 - found '/toolchains/sparc-elf/bin/sparc-elf-gcc'
244 - looking in '/toolchains/sparc-elf/usr/bin'
245 - scanning path '/toolchains/arm-2010q1'
246 - looking in '/toolchains/arm-2010q1/.'
247 - looking in '/toolchains/arm-2010q1/bin'
248 - found '/toolchains/arm-2010q1/bin/arm-none-linux-gnueabi-gcc'
250 - looking in '/toolchains/arm-2010q1/usr/bin'
251 - scanning path '/toolchains/from'
252 - looking in '/toolchains/from/.'
253 - looking in '/toolchains/from/bin'
254 - looking in '/toolchains/from/usr/bin'
255 - scanning path '/toolchains/sh4-gentoo-linux-gnu'
256 - looking in '/toolchains/sh4-gentoo-linux-gnu/.'
257 - looking in '/toolchains/sh4-gentoo-linux-gnu/bin'
258 - found '/toolchains/sh4-gentoo-linux-gnu/bin/sh4-gentoo-linux-gnu-gcc'
260 - looking in '/toolchains/sh4-gentoo-linux-gnu/usr/bin'
261 - scanning path '/toolchains/avr32-linux'
262 - looking in '/toolchains/avr32-linux/.'
263 - looking in '/toolchains/avr32-linux/bin'
264 - found '/toolchains/avr32-linux/bin/avr32-gcc'
266 - looking in '/toolchains/avr32-linux/usr/bin'
267 - scanning path '/toolchains/m68k-linux'
268 - looking in '/toolchains/m68k-linux/.'
269 - looking in '/toolchains/m68k-linux/bin'
270 - found '/toolchains/m68k-linux/bin/m68k-linux-gcc'
272 - looking in '/toolchains/m68k-linux/usr/bin'
273 List of available toolchains (17):
274 arm : /toolchains/arm-2010q1/bin/arm-none-linux-gnueabi-gcc
275 avr32 : /toolchains/avr32-linux/bin/avr32-gcc
276 bfin : /toolchains/bfin-uclinux/bin/bfin-uclinux-gcc
277 c89 : /usr/bin/c89-gcc
278 c99 : /usr/bin/c99-gcc
279 i386 : /toolchains/i386-linux/bin/i386-linux-gcc
280 m68k : /toolchains/m68k-linux/bin/m68k-linux-gcc
281 mb : /toolchains/microblaze-unknown-linux-gnu/bin/mb-linux-gcc
282 microblaze: /toolchains/microblaze-unknown-linux-gnu/bin/microblaze-unknown-linux-gnu-gcc
283 mips : /toolchains/mips-linux/bin/mips-linux-gcc
284 nds32le : /toolchains/nds32le-linux-glibc-v1f/bin/nds32le-linux-gcc
285 nios2 : /toolchains/nios2/bin/nios2-linux-gcc
286 powerpc : /toolchains/powerpc-linux/bin/powerpc-linux-gcc
287 sandbox : /usr/bin/gcc
288 sh4 : /toolchains/sh4-gentoo-linux-gnu/bin/sh4-gentoo-linux-gnu-gcc
289 sparc : /toolchains/sparc-elf/bin/sparc-elf-gcc
290 x86_64 : /usr/bin/x86_64-linux-gnu-gcc
293 You can see that everything is covered, even some strange ones that won't
294 be used (c88 and c99). This is a feature.
300 First do a dry run using the -n flag: (replace <branch> with a real, local
301 branch with a valid upstream)
303 $ ./tools/buildman/buildman -b <branch> -n
305 If it can't detect the upstream branch, try checking out the branch, and
306 doing something like 'git branch --set-upstream <branch> upstream/master'
307 or something similar.
311 Dry run, so not doing much. But I would do this:
313 Building 18 commits for 1059 boards (4 threads, 1 job per thread)
314 Build directory: ../lcd9b
315 5bb3505 Merge branch 'master' of git://git.denx.de/u-boot-arm
316 c18f1b4 tegra: Use const for pinmux_config_pingroup/table()
317 2f043ae tegra: Add display support to funcmux
318 e349900 tegra: fdt: Add pwm binding and node
319 424a5f0 tegra: fdt: Add LCD definitions for Tegra
320 0636ccf tegra: Add support for PWM
321 a994fe7 tegra: Add SOC support for display/lcd
322 fcd7350 tegra: Add LCD driver
323 4d46e9d tegra: Add LCD support to Nvidia boards
324 991bd48 arm: Add control over cachability of memory regions
325 54e8019 lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment
326 d92aff7 lcd: Add support for flushing LCD fb from dcache after update
327 dbd0677 tegra: Align LCD frame buffer to section boundary
328 0cff9b8 tegra: Support control of cache settings for LCD
329 9c56900 tegra: fdt: Add LCD definitions for Seaboard
330 5cc29db lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console
331 cac5a23 tegra: Enable display/lcd support on Seaboard
334 Total boards to build for each commit: 1059
336 This shows that it will build all 1059 boards, using 4 threads (because
337 we have a 4-core CPU). Each thread will run with -j1, meaning that each
338 make job will use a single CPU. The list of commits to be built helps you
339 confirm that things look about right. Notice that buildman has chosen a
340 'base' directory for you, immediately above your source tree.
342 Buildman works entirely inside the base directory, here ../lcd9b,
343 creating a working directory for each thread, and creating output
344 directories for each commit and board.
350 To run the build for real, take off the -n:
352 $ ./tools/buildman/buildman -b <branch>
354 Buildman will set up some working directories, and get started. After a
355 minute or so it will settle down to a steady pace, with a display like this:
357 Building 18 commits for 1059 boards (4 threads, 1 job per thread)
358 528 36 124 /19062 1:13:30 : SIMPC8313_SP
360 This means that it is building 19062 board/commit combinations. So far it
361 has managed to successfully build 528. Another 36 have built with warnings,
362 and 124 more didn't build at all. Buildman expects to complete the process
363 in an hour and 15 minutes. Use this time to buy a faster computer.
366 To find out how the build went, ask for a summary with -s. You can do this
367 either before the build completes (presumably in another terminal) or or
368 afterwards. Let's work through an example of how this is used:
370 $ ./tools/buildman/buildman -b lcd9b -s
372 01: Merge branch 'master' of git://git.denx.de/u-boot-arm
373 powerpc: + galaxy5200_LOWBOOT
374 02: tegra: Use const for pinmux_config_pingroup/table()
375 03: tegra: Add display support to funcmux
376 04: tegra: fdt: Add pwm binding and node
377 05: tegra: fdt: Add LCD definitions for Tegra
378 06: tegra: Add support for PWM
379 07: tegra: Add SOC support for display/lcd
380 08: tegra: Add LCD driver
381 09: tegra: Add LCD support to Nvidia boards
382 10: arm: Add control over cachability of memory regions
383 11: lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment
384 12: lcd: Add support for flushing LCD fb from dcache after update
386 13: tegra: Align LCD frame buffer to section boundary
387 14: tegra: Support control of cache settings for LCD
388 15: tegra: fdt: Add LCD definitions for Seaboard
389 16: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console
390 17: tegra: Enable display/lcd support on Seaboard
393 This shows which commits have succeeded and which have failed. In this case
394 the build is still in progress so many boards are not built yet (use -u to
395 see which ones). But still we can see a few failures. The galaxy5200_LOWBOOT
396 never builds correctly. This could be a problem with our toolchain, or it
397 could be a bug in the upstream. The good news is that we probably don't need
398 to blame our commits. The bad news is it isn't tested on that board.
400 Commit 12 broke lubbock. That's what the '+ lubbock' means. The failure
401 is never fixed by a later commit, or you would see lubbock again, in green,
404 To see the actual error:
406 $ ./tools/buildman/buildman -b <branch> -se lubbock
408 12: lcd: Add support for flushing LCD fb from dcache after update
410 +common/libcommon.o: In function `lcd_sync':
411 +/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range'
412 +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
413 +make: *** [/u-boot/lcd9b/.bm-work/00/build/u-boot] Error 139
414 13: tegra: Align LCD frame buffer to section boundary
415 14: tegra: Support control of cache settings for LCD
416 15: tegra: fdt: Add LCD definitions for Seaboard
417 16: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console
418 -/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range'
419 +/u-boot/lcd9b/.bm-work/00/common/lcd.c:125: undefined reference to `flush_dcache_range'
420 17: tegra: Enable display/lcd support on Seaboard
423 So the problem is in lcd.c, due to missing cache operations. This information
424 should be enough to work out what that commit is doing to break these
425 boards. (In this case pxa did not have cache operations defined).
427 If you see error lines marked with - that means that the errors were fixed
428 by that commit. Sometimes commits can be in the wrong order, so that a
429 breakage is introduced for a few commits and fixed by later commits. This
430 shows up clearly with buildman. You can then reorder the commits and try
433 At commit 16, the error moves - you can see that the old error at line 120
434 is fixed, but there is a new one at line 126. This is probably only because
435 we added some code and moved the broken line father down the file.
437 If many boards have the same error, then -e will display the error only
438 once. This makes the output as concise as possible.
440 The full build output in this case is available in:
442 ../lcd9b/12_of_18_gd92aff7_lcd--Add-support-for/lubbock/
444 done: Indicates the build was done, and holds the return code from make.
445 This is 0 for a good build, typically 2 for a failure.
447 err: Output from stderr, if any. Errors and warnings appear here.
449 log: Output from stdout. Normally there isn't any since buildman runs
450 in silent mode for now.
452 toolchain: Shows information about the toolchain used for the build.
454 sizes: Shows image size information.
456 It is possible to get the build output there also. Use the -k option for
457 this. In that case you will also see some output files, like:
459 System.map toolchain u-boot u-boot.bin u-boot.map autoconf.mk
460 (also SPL versions u-boot-spl and u-boot-spl.bin if available)
466 A key requirement for U-Boot is that you keep code/data size to a minimum.
467 Where a new feature increases this noticeably it should normally be put
468 behind a CONFIG flag so that boards can leave it off and keep the image
469 size more or less the same with each new release.
471 To check the impact of your commits on image size, use -S. For example:
473 $ ./tools/buildman/buildman -b us-x86 -sS
474 Summary of 10 commits for 1066 boards (4 threads, 1 job per thread)
475 01: MAKEALL: add support for per architecture toolchains
476 02: x86: Add function to get top of usable ram
477 x86: (for 1/3 boards) text -272.0 rodata +41.0
478 03: x86: Add basic cache operations
479 04: x86: Permit bootstage and timer data to be used prior to relocation
480 x86: (for 1/3 boards) data +16.0
481 05: x86: Add an __end symbol to signal the end of the U-Boot binary
482 x86: (for 1/3 boards) text +76.0
483 06: x86: Rearrange the output input to remove BSS
484 x86: (for 1/3 boards) bss -2140.0
485 07: x86: Support relocation of FDT on start-up
487 08: x86: Add error checking to x86 relocation code
488 09: x86: Adjust link device tree include file
489 10: x86: Enable CONFIG_OF_CONTROL on coreboot
492 You can see that image size only changed on x86, which is good because this
493 series is not supposed to change any other board. From commit 7 onwards the
494 build fails so we don't get code size numbers. The numbers are fractional
495 because they are an average of all boards for that architecture. The
496 intention is to allow you to quickly find image size problems introduced by
499 Note that the 'text' region and 'rodata' are split out. You should add the
500 two together to get the total read-only size (reported as the first column
501 in the output from binutil's 'size' utility).
503 A useful option is --step which lets you skip some commits. For example
504 --step 2 will show the image sizes for only every 2nd commit (so it will
505 compare the image sizes of the 1st, 3rd, 5th... commits). You can also use
506 --step 0 which will compare only the first and last commits. This is useful
507 for an overview of how your entire series affects code size.
509 You can also use -d to see a detailed size breakdown for each board. This
510 list is sorted in order from largest growth to largest reduction.
512 It is possible to go a little further with the -B option (--bloat). This
513 shows where U-Boot has bloated, breaking the size change down to the function
514 level. Example output is below:
516 $ ./tools/buildman/buildman -b us-mem4 -sSdB
518 19: Roll crc32 into hash infrastructure
519 arm: (for 10/10 boards) all -143.4 bss +1.2 data -4.8 rodata -48.2 text -91.6
520 paz00 : all +23 bss -4 rodata -29 text +56
521 u-boot: add: 1/0, grow: 3/-2 bytes: 168/-104 (64)
522 function old new delta
523 hash_command 80 160 +80
524 crc32_wd_buf - 56 +56
525 ext4fs_read_file 540 568 +28
526 insert_var_value_sub 688 692 +4
527 run_list_real 1996 1992 -4
528 do_mem_crc 168 68 -100
529 trimslice : all -9 bss +16 rodata -29 text +4
530 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12)
531 function old new delta
532 hash_command 80 160 +80
533 crc32_wd_buf - 56 +56
534 ext4fs_iterate_dir 672 668 -4
535 ext4fs_read_file 568 548 -20
536 do_mem_crc 168 68 -100
537 whistler : all -9 bss +16 rodata -29 text +4
538 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12)
539 function old new delta
540 hash_command 80 160 +80
541 crc32_wd_buf - 56 +56
542 ext4fs_iterate_dir 672 668 -4
543 ext4fs_read_file 568 548 -20
544 do_mem_crc 168 68 -100
545 seaboard : all -9 bss -28 rodata -29 text +48
546 u-boot: add: 1/0, grow: 3/-2 bytes: 160/-104 (56)
547 function old new delta
548 hash_command 80 160 +80
549 crc32_wd_buf - 56 +56
550 ext4fs_read_file 548 568 +20
551 run_list_real 1996 2000 +4
552 do_nandboot 760 756 -4
553 do_mem_crc 168 68 -100
554 colibri_t20_iris: all -9 rodata -29 text +20
555 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-112 (28)
556 function old new delta
557 hash_command 80 160 +80
558 crc32_wd_buf - 56 +56
559 read_abs_bbt 204 208 +4
560 do_nandboot 760 756 -4
561 ext4fs_read_file 576 568 -8
562 do_mem_crc 168 68 -100
563 ventana : all -37 bss -12 rodata -29 text +4
564 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12)
565 function old new delta
566 hash_command 80 160 +80
567 crc32_wd_buf - 56 +56
568 ext4fs_iterate_dir 672 668 -4
569 ext4fs_read_file 568 548 -20
570 do_mem_crc 168 68 -100
571 harmony : all -37 bss -16 rodata -29 text +8
572 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-124 (16)
573 function old new delta
574 hash_command 80 160 +80
575 crc32_wd_buf - 56 +56
576 nand_write_oob_syndrome 428 432 +4
577 ext4fs_iterate_dir 672 668 -4
578 ext4fs_read_file 568 548 -20
579 do_mem_crc 168 68 -100
580 medcom-wide : all -417 bss +28 data -16 rodata -93 text -336
581 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288)
582 function old new delta
583 crc32_wd_buf - 56 +56
584 do_fat_read_at 2872 2904 +32
586 do_mem_crc 168 68 -100
587 hash_command 420 160 -260
588 tec : all -449 bss -4 data -16 rodata -93 text -336
589 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288)
590 function old new delta
591 crc32_wd_buf - 56 +56
592 do_fat_read_at 2872 2904 +32
594 do_mem_crc 168 68 -100
595 hash_command 420 160 -260
596 plutux : all -481 bss +16 data -16 rodata -93 text -388
597 u-boot: add: 1/-1, grow: 1/-3 bytes: 68/-408 (-340)
598 function old new delta
599 crc32_wd_buf - 56 +56
600 do_load_serial_bin 1688 1700 +12
602 do_fat_read_at 2904 2872 -32
603 do_mem_crc 168 68 -100
604 hash_command 420 160 -260
605 powerpc: (for 5/5 boards) all +37.4 data -3.2 rodata -41.8 text +82.4
606 MPC8610HPCD : all +55 rodata -29 text +84
607 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
608 function old new delta
609 hash_command - 176 +176
610 do_mem_crc 184 88 -96
611 MPC8641HPCN : all +55 rodata -29 text +84
612 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
613 function old new delta
614 hash_command - 176 +176
615 do_mem_crc 184 88 -96
616 MPC8641HPCN_36BIT: all +55 rodata -29 text +84
617 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
618 function old new delta
619 hash_command - 176 +176
620 do_mem_crc 184 88 -96
621 sbc8641d : all +55 rodata -29 text +84
622 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80)
623 function old new delta
624 hash_command - 176 +176
625 do_mem_crc 184 88 -96
626 xpedite517x : all -33 data -16 rodata -93 text +76
627 u-boot: add: 1/-1, grow: 0/-1 bytes: 176/-112 (64)
628 function old new delta
629 hash_command - 176 +176
631 do_mem_crc 184 88 -96
635 This shows that commit 19 has increased text size for arm (although only one
636 board was built) and by 96 bytes for powerpc. This increase was offset in both
637 cases by reductions in rodata and data/bss.
639 Shown below the summary lines is the sizes for each board. Below each board
640 is the sizes for each function. This information starts with:
642 add - number of functions added / removed
643 grow - number of functions which grew / shrunk
644 bytes - number of bytes of code added to / removed from all functions,
645 plus the total byte change in brackets
647 The change seems to be that hash_command() has increased by more than the
648 do_mem_crc() function has decreased. The function sizes typically add up to
649 roughly the text area size, but note that every read-only section except
650 rodata is included in 'text', so the function total does not exactly
653 It is common when refactoring code for the rodata to decrease as the text size
654 increases, and vice versa.
657 Providing 'make' flags
658 ======================
660 U-Boot's build system supports a few flags (such as BUILD_TAG) which affect
661 the build product. These flags can be specified in the buildman settings
662 file. They can also be useful when building U-Boot against other open source
666 at91-boards=ENABLE_AT91_TEST=1
667 snapper9260=${at91-boards} BUILD_TAG=442
668 snapper9g45=${at91-boards} BUILD_TAG=443
670 This will use 'make ENABLE_AT91_TEST=1 BUILD_TAG=442' for snapper9260
671 and 'make ENABLE_AT91_TEST=1 BUILD_TAG=443' for snapper9g45. A special
672 variable ${target} is available to access the target name (snapper9260 and
673 snapper9g20 in this case). Variables are resolved recursively.
675 It is expected that any variables added are dealt with in U-Boot's
676 config.mk file and documented in the README.
682 If you have made changes and want to do a quick sanity check of the
683 currently-checked-out source, run buildman without the -b flag. This will
684 build the selected boards and display build status and errors as it runs
685 (i.e. -v amd -e are enabled automatically).
691 Buildman has various other command line options. Try --help to see them.
694 How to change from MAKEALL
695 ==========================
697 Buildman includes most of the features of MAKEALL and is generally faster
698 and easier to use. In particular it builds entire branches: if a particular
699 commit introduces an error in a particular board, buildman can easily show
700 you this, even if a later commit fixes that error.
702 The reasons to deprecate MAKEALL are:
703 - We don't want to maintain two build systems
704 - Buildman is typically faster
705 - Buildman has a lot more features
707 But still, many people will be sad to lose MAKEALL. If you are used to
708 MAKEALL, here are a few pointers.
710 First you need to set up your tool chains - see the 'Setting up' section
711 for details. Once you have your required toolchain(s) detected then you are
714 To build the current source tree, run buildman without a -b flag:
716 ./tools/buildman/buildman <list of things to build>
718 This will build the current source tree for the given boards and display
719 the results and errors.
721 However buildman usually works on entire branches, and for that you must
722 specify a board flag:
724 ./tools/buildman/buildman -b <branch_name> <list of things to build>
726 followed by (afterwards, or perhaps concurrently in another terminal):
728 ./tools/buildman/buildman -b <branch_name> -s <list of things to build>
730 to see the results of the build. Rather than showing you all the output,
731 buildman just shows a summary, with red indicating that a commit introduced
732 an error and green indicating that a commit fixed an error. Use the -e
733 flag to see the full errors.
735 If you really want to see build results as they happen, use -v when doing a
736 build (and -e if you want to see errors as well).
738 You don't need to stick around on that branch while buildman is running. It
739 checks out its own copy of the source code, so you can change branches,
740 add commits, etc. without affecting the build in progress.
742 The <list of things to build> can include board names, architectures or the
743 like. There are no flags to disambiguate since ambiguities are rare. Using
744 the examples from MAKEALL:
747 - build all Power Architecture boards:
749 MAKEALL --arch powerpc
751 ** buildman -b <branch> powerpc
752 - build all PowerPC boards manufactured by vendor "esd":
753 MAKEALL -a powerpc -v esd
754 ** buildman -b <branch> esd
755 - build all PowerPC boards manufactured either by "keymile" or "siemens":
756 MAKEALL -a powerpc -v keymile -v siemens
757 ** buildman -b <branch> keymile siemens
758 - build all Freescale boards with MPC83xx CPUs, plus all 4xx boards:
759 MAKEALL -c mpc83xx -v freescale 4xx
760 ** buildman -b <branch> mpc83xx freescale 4xx
762 Buildman automatically tries to use all the CPUs in your machine. If you
763 are building a lot of boards it will use one thread for every CPU core
764 it detects in your machine. This is like MAKEALL's BUILD_NBUILDS option.
765 You can use the -T flag to change the number of threads. If you are only
766 building a few boards, buildman will automatically run make with the -j
767 flag to increase the number of concurrent make tasks. It isn't normally
768 that helpful to fiddle with this option, but if you use the BUILD_NCPUS
769 option in MAKEALL then -j is the equivalent in buildman.
771 Buildman puts its output in ../<branch_name> by default but you can change
772 this with the -o option. Buildman normally does out-of-tree builds: use -i
773 to disable that if you really want to. But be careful that once you have
774 used -i you pollute buildman's copies of the source tree, and you will need
775 to remove the build directory (normally ../<branch_name>) to run buildman
776 in normal mode (without -i).
778 Buildman doesn't keep the output result normally, but use the -k option to
781 Please read 'Theory of Operation' a few times as it will make a lot of
784 Some options you might like are:
786 -B shows which functions are growing/shrinking in which commit - great
787 for finding code bloat.
788 -S shows image sizes for each commit (just an overall summary)
789 -u shows boards that you haven't built yet
790 --step 0 will build just the upstream commit and the last commit of your
791 branch. This is often a quick sanity check that your branch doesn't
792 break anything. But note this does not check bisectability!
798 This has mostly be written in my spare time as a response to my difficulties
799 in testing large series of patches. Apart from tidying up there is quite a
800 bit of scope for improvement. Things like better error diffs, easier access
801 to log files, error display while building. Also it would be nice it buildman
802 could 'hunt' for problems, perhaps by building a few boards for each arch,
803 or checking commits for changed files and building only boards which use
810 Thanks to Grant Grundler <grundler@chromium.org> for his ideas for improving
811 the build speed by building all commits for a board instead of the other