1# Copyright (c) 2013 The Chromium OS Authors. 2# 3# SPDX-License-Identifier: GPL-2.0+ 4# 5 6(Please read 'How to change from MAKEALL' if you are used to that tool) 7 8What is this? 9============= 10 11This tool handles building U-Boot to check that you have not broken it 12with your patch series. It can build each individual commit and report 13which boards fail on which commits, and which errors come up. It aims 14to make full use of multi-processor machines. 15 16A key feature of buildman is its output summary, which allows warnings, 17errors or image size increases in a particular commit or board to be 18quickly identified and the offending commit pinpointed. This can be a big 19help for anyone working with >10 patches at a time. 20 21 22Caveats 23======= 24 25Buildman is still in its infancy. It is already a very useful tool, but 26expect to find problems and send patches. 27 28Buildman can be stopped and restarted, in which case it will continue 29where it left off. This should happen cleanly and without side-effects. 30If not, it is a bug, for which a patch would be welcome. 31 32Buildman gets so tied up in its work that it can ignore the outside world. 33You may need to press Ctrl-C several times to quit it. Also it will print 34out various exceptions when stopped. 35 36 37Theory of Operation 38=================== 39 40(please read this section in full twice or you will be perpetually confused) 41 42Buildman is a builder. It is not make, although it runs make. It does not 43produce any useful output on the terminal while building, except for 44progress information. All the output (errors, warnings and binaries if you 45are ask for them) is stored in output directories, which you can look at 46while the build is progressing, or when it is finished. 47 48Buildman produces a concise summary of which boards succeeded and failed. 49It shows which commit introduced which board failure using a simple 50red/green colour coding. Full error information can be requested, in which 51case it is de-duped and displayed against the commit that introduced the 52error. An example workflow is below. 53 54Buildman stores image size information and can report changes in image size 55from commit to commit. An example of this is below. 56 57Buildman starts multiple threads, and each thread builds for one board at 58a time. A thread starts at the first commit, configures the source for your 59board and builds it. Then it checks out the next commit and does an 60incremental build. Eventually the thread reaches the last commit and stops. 61If errors or warnings are found along the way, the thread will reconfigure 62after every commit, and your build will be very slow. This is because a 63file that produces just a warning would not normally be rebuilt in an 64incremental build. 65 66Buildman works in an entirely separate place from your U-Boot repository. 67It creates a separate working directory for each thread, and puts the 68output files in the working directory, organised by commit name and board 69name, in a two-level hierarchy. 70 71Buildman is invoked in your U-Boot directory, the one with the .git 72directory. It clones this repository into a copy for each thread, and the 73threads do not affect the state of your git repository. Any checkouts done 74by the thread affect only the working directory for that thread. 75 76Buildman automatically selects the correct tool chain for each board. You 77must supply suitable tool chains, but buildman takes care of selecting the 78right one. 79 80Buildman always builds a branch, and always builds the upstream commit as 81well, for comparison. It cannot build individual commits at present, unless 82(maybe) you point it at an empty branch. Put all your commits in a branch, 83set the branch's upstream to a valid value, and all will be well. Otherwise 84buildman will perform random actions. Use -n to check what the random 85actions might be. 86 87Buildman is optimised for building many commits at once, for many boards. 88On multi-core machines, Buildman is fast because it uses most of the 89available CPU power. When it gets to the end, or if you are building just 90a few commits or boards, it will be pretty slow. As a tip, if you don't 91plan to use your machine for anything else, you can use -T to increase the 92number of threads beyond the default. 93 94Buildman lets you build all boards, or a subset. Specify the subset by passing 95command-line arguments that list the desired board name, architecture name, 96SOC name, or anything else in the boards.cfg file. Multiple arguments are 97allowed. Each argument will be interpreted as a regular expression, so 98behaviour is a superset of exact or substring matching. Examples are: 99 100* 'tegra20' All boards with a Tegra20 SoC 101* 'tegra' All boards with any Tegra Soc (Tegra20, Tegra30, Tegra114...) 102* '^tegra[23]0$' All boards with either Tegra20 or Tegra30 SoC 103* 'powerpc' All PowerPC boards 104 105Buildman does not store intermediate object files. It optionally copies 106the binary output into a directory when a build is successful. Size 107information is always recorded. It needs a fair bit of disk space to work, 108typically 250MB per thread. 109 110 111Setting up 112========== 113 1141. Get the U-Boot source. You probably already have it, but if not these 115steps should get you started with a repo and some commits for testing. 116 117$ cd /path/to/u-boot 118$ git clone git://git.denx.de/u-boot.git . 119$ git checkout -b my-branch origin/master 120$ # Add some commits to the branch, reading for testing 121 1222. Create ~/.buildman to tell buildman where to find tool chains. As an 123example: 124 125# Buildman settings file 126 127[toolchain] 128root: / 129rest: /toolchains/* 130eldk: /opt/eldk-4.2 131 132[toolchain-alias] 133x86: i386 134blackfin: bfin 135sh: sh4 136nds32: nds32le 137openrisc: or32 138 139 140This selects the available toolchain paths. Add the base directory for 141each of your toolchains here. Buildman will search inside these directories 142and also in any '/usr' and '/usr/bin' subdirectories. 143 144Make sure the tags (here root: rest: and eldk:) are unique. 145 146The toolchain-alias section indicates that the i386 toolchain should be used 147to build x86 commits. 148 149 1502. Check the available toolchains 151 152Run this check to make sure that you have a toolchain for every architecture. 153 154$ ./tools/buildman/buildman --list-tool-chains 155Scanning for tool chains 156 - scanning path '/' 157 - looking in '/.' 158 - looking in '/bin' 159 - looking in '/usr/bin' 160 - found '/usr/bin/gcc' 161Tool chain test: OK 162 - found '/usr/bin/c89-gcc' 163Tool chain test: OK 164 - found '/usr/bin/c99-gcc' 165Tool chain test: OK 166 - found '/usr/bin/x86_64-linux-gnu-gcc' 167Tool chain test: OK 168 - scanning path '/toolchains/powerpc-linux' 169 - looking in '/toolchains/powerpc-linux/.' 170 - looking in '/toolchains/powerpc-linux/bin' 171 - found '/toolchains/powerpc-linux/bin/powerpc-linux-gcc' 172Tool chain test: OK 173 - looking in '/toolchains/powerpc-linux/usr/bin' 174 - scanning path '/toolchains/nds32le-linux-glibc-v1f' 175 - looking in '/toolchains/nds32le-linux-glibc-v1f/.' 176 - looking in '/toolchains/nds32le-linux-glibc-v1f/bin' 177 - found '/toolchains/nds32le-linux-glibc-v1f/bin/nds32le-linux-gcc' 178Tool chain test: OK 179 - looking in '/toolchains/nds32le-linux-glibc-v1f/usr/bin' 180 - scanning path '/toolchains/nios2' 181 - looking in '/toolchains/nios2/.' 182 - looking in '/toolchains/nios2/bin' 183 - found '/toolchains/nios2/bin/nios2-linux-gcc' 184Tool chain test: OK 185 - found '/toolchains/nios2/bin/nios2-linux-uclibc-gcc' 186Tool chain test: OK 187 - looking in '/toolchains/nios2/usr/bin' 188 - found '/toolchains/nios2/usr/bin/nios2-linux-gcc' 189Tool chain test: OK 190 - found '/toolchains/nios2/usr/bin/nios2-linux-uclibc-gcc' 191Tool chain test: OK 192 - scanning path '/toolchains/microblaze-unknown-linux-gnu' 193 - looking in '/toolchains/microblaze-unknown-linux-gnu/.' 194 - looking in '/toolchains/microblaze-unknown-linux-gnu/bin' 195 - found '/toolchains/microblaze-unknown-linux-gnu/bin/microblaze-unknown-linux-gnu-gcc' 196Tool chain test: OK 197 - found '/toolchains/microblaze-unknown-linux-gnu/bin/mb-linux-gcc' 198Tool chain test: OK 199 - looking in '/toolchains/microblaze-unknown-linux-gnu/usr/bin' 200 - scanning path '/toolchains/mips-linux' 201 - looking in '/toolchains/mips-linux/.' 202 - looking in '/toolchains/mips-linux/bin' 203 - found '/toolchains/mips-linux/bin/mips-linux-gcc' 204Tool chain test: OK 205 - looking in '/toolchains/mips-linux/usr/bin' 206 - scanning path '/toolchains/old' 207 - looking in '/toolchains/old/.' 208 - looking in '/toolchains/old/bin' 209 - looking in '/toolchains/old/usr/bin' 210 - scanning path '/toolchains/i386-linux' 211 - looking in '/toolchains/i386-linux/.' 212 - looking in '/toolchains/i386-linux/bin' 213 - found '/toolchains/i386-linux/bin/i386-linux-gcc' 214Tool chain test: OK 215 - looking in '/toolchains/i386-linux/usr/bin' 216 - scanning path '/toolchains/bfin-uclinux' 217 - looking in '/toolchains/bfin-uclinux/.' 218 - looking in '/toolchains/bfin-uclinux/bin' 219 - found '/toolchains/bfin-uclinux/bin/bfin-uclinux-gcc' 220Tool chain test: OK 221 - looking in '/toolchains/bfin-uclinux/usr/bin' 222 - scanning path '/toolchains/sparc-elf' 223 - looking in '/toolchains/sparc-elf/.' 224 - looking in '/toolchains/sparc-elf/bin' 225 - found '/toolchains/sparc-elf/bin/sparc-elf-gcc' 226Tool chain test: OK 227 - looking in '/toolchains/sparc-elf/usr/bin' 228 - scanning path '/toolchains/arm-2010q1' 229 - looking in '/toolchains/arm-2010q1/.' 230 - looking in '/toolchains/arm-2010q1/bin' 231 - found '/toolchains/arm-2010q1/bin/arm-none-linux-gnueabi-gcc' 232Tool chain test: OK 233 - looking in '/toolchains/arm-2010q1/usr/bin' 234 - scanning path '/toolchains/from' 235 - looking in '/toolchains/from/.' 236 - looking in '/toolchains/from/bin' 237 - looking in '/toolchains/from/usr/bin' 238 - scanning path '/toolchains/sh4-gentoo-linux-gnu' 239 - looking in '/toolchains/sh4-gentoo-linux-gnu/.' 240 - looking in '/toolchains/sh4-gentoo-linux-gnu/bin' 241 - found '/toolchains/sh4-gentoo-linux-gnu/bin/sh4-gentoo-linux-gnu-gcc' 242Tool chain test: OK 243 - looking in '/toolchains/sh4-gentoo-linux-gnu/usr/bin' 244 - scanning path '/toolchains/avr32-linux' 245 - looking in '/toolchains/avr32-linux/.' 246 - looking in '/toolchains/avr32-linux/bin' 247 - found '/toolchains/avr32-linux/bin/avr32-gcc' 248Tool chain test: OK 249 - looking in '/toolchains/avr32-linux/usr/bin' 250 - scanning path '/toolchains/m68k-linux' 251 - looking in '/toolchains/m68k-linux/.' 252 - looking in '/toolchains/m68k-linux/bin' 253 - found '/toolchains/m68k-linux/bin/m68k-linux-gcc' 254Tool chain test: OK 255 - looking in '/toolchains/m68k-linux/usr/bin' 256List of available toolchains (17): 257arm : /toolchains/arm-2010q1/bin/arm-none-linux-gnueabi-gcc 258avr32 : /toolchains/avr32-linux/bin/avr32-gcc 259bfin : /toolchains/bfin-uclinux/bin/bfin-uclinux-gcc 260c89 : /usr/bin/c89-gcc 261c99 : /usr/bin/c99-gcc 262i386 : /toolchains/i386-linux/bin/i386-linux-gcc 263m68k : /toolchains/m68k-linux/bin/m68k-linux-gcc 264mb : /toolchains/microblaze-unknown-linux-gnu/bin/mb-linux-gcc 265microblaze: /toolchains/microblaze-unknown-linux-gnu/bin/microblaze-unknown-linux-gnu-gcc 266mips : /toolchains/mips-linux/bin/mips-linux-gcc 267nds32le : /toolchains/nds32le-linux-glibc-v1f/bin/nds32le-linux-gcc 268nios2 : /toolchains/nios2/bin/nios2-linux-gcc 269powerpc : /toolchains/powerpc-linux/bin/powerpc-linux-gcc 270sandbox : /usr/bin/gcc 271sh4 : /toolchains/sh4-gentoo-linux-gnu/bin/sh4-gentoo-linux-gnu-gcc 272sparc : /toolchains/sparc-elf/bin/sparc-elf-gcc 273x86_64 : /usr/bin/x86_64-linux-gnu-gcc 274 275 276You can see that everything is covered, even some strange ones that won't 277be used (c88 and c99). This is a feature. 278 279 280How to run it 281============= 282 283First do a dry run using the -n flag: (replace <branch> with a real, local 284branch with a valid upstream) 285 286$ ./tools/buildman/buildman -b <branch> -n 287 288If it can't detect the upstream branch, try checking out the branch, and 289doing something like 'git branch --set-upstream <branch> upstream/master' 290or something similar. 291 292As an example: 293 294Dry run, so not doing much. But I would do this: 295 296Building 18 commits for 1059 boards (4 threads, 1 job per thread) 297Build directory: ../lcd9b 298 5bb3505 Merge branch 'master' of git://git.denx.de/u-boot-arm 299 c18f1b4 tegra: Use const for pinmux_config_pingroup/table() 300 2f043ae tegra: Add display support to funcmux 301 e349900 tegra: fdt: Add pwm binding and node 302 424a5f0 tegra: fdt: Add LCD definitions for Tegra 303 0636ccf tegra: Add support for PWM 304 a994fe7 tegra: Add SOC support for display/lcd 305 fcd7350 tegra: Add LCD driver 306 4d46e9d tegra: Add LCD support to Nvidia boards 307 991bd48 arm: Add control over cachability of memory regions 308 54e8019 lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment 309 d92aff7 lcd: Add support for flushing LCD fb from dcache after update 310 dbd0677 tegra: Align LCD frame buffer to section boundary 311 0cff9b8 tegra: Support control of cache settings for LCD 312 9c56900 tegra: fdt: Add LCD definitions for Seaboard 313 5cc29db lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console 314 cac5a23 tegra: Enable display/lcd support on Seaboard 315 49ff541 wip 316 317Total boards to build for each commit: 1059 318 319This shows that it will build all 1059 boards, using 4 threads (because 320we have a 4-core CPU). Each thread will run with -j1, meaning that each 321make job will use a single CPU. The list of commits to be built helps you 322confirm that things look about right. Notice that buildman has chosen a 323'base' directory for you, immediately above your source tree. 324 325Buildman works entirely inside the base directory, here ../lcd9b, 326creating a working directory for each thread, and creating output 327directories for each commit and board. 328 329 330Suggested Workflow 331================== 332 333To run the build for real, take off the -n: 334 335$ ./tools/buildman/buildman -b <branch> 336 337Buildman will set up some working directories, and get started. After a 338minute or so it will settle down to a steady pace, with a display like this: 339 340Building 18 commits for 1059 boards (4 threads, 1 job per thread) 341 528 36 124 /19062 1:13:30 : SIMPC8313_SP 342 343This means that it is building 19062 board/commit combinations. So far it 344has managed to successfully build 528. Another 36 have built with warnings, 345and 124 more didn't build at all. Buildman expects to complete the process 346in an hour and 15 minutes. Use this time to buy a faster computer. 347 348 349To find out how the build went, ask for a summary with -s. You can do this 350either before the build completes (presumably in another terminal) or or 351afterwards. Let's work through an example of how this is used: 352 353$ ./tools/buildman/buildman -b lcd9b -s 354... 35501: Merge branch 'master' of git://git.denx.de/u-boot-arm 356 powerpc: + galaxy5200_LOWBOOT 35702: tegra: Use const for pinmux_config_pingroup/table() 35803: tegra: Add display support to funcmux 35904: tegra: fdt: Add pwm binding and node 36005: tegra: fdt: Add LCD definitions for Tegra 36106: tegra: Add support for PWM 36207: tegra: Add SOC support for display/lcd 36308: tegra: Add LCD driver 36409: tegra: Add LCD support to Nvidia boards 36510: arm: Add control over cachability of memory regions 36611: lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment 36712: lcd: Add support for flushing LCD fb from dcache after update 368 arm: + lubbock 36913: tegra: Align LCD frame buffer to section boundary 37014: tegra: Support control of cache settings for LCD 37115: tegra: fdt: Add LCD definitions for Seaboard 37216: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console 37317: tegra: Enable display/lcd support on Seaboard 37418: wip 375 376This shows which commits have succeeded and which have failed. In this case 377the build is still in progress so many boards are not built yet (use -u to 378see which ones). But still we can see a few failures. The galaxy5200_LOWBOOT 379never builds correctly. This could be a problem with our toolchain, or it 380could be a bug in the upstream. The good news is that we probably don't need 381to blame our commits. The bad news is it isn't tested on that board. 382 383Commit 12 broke lubbock. That's what the '+ lubbock' means. The failure 384is never fixed by a later commit, or you would see lubbock again, in green, 385without the +. 386 387To see the actual error: 388 389$ ./tools/buildman/buildman -b <branch> -se lubbock 390... 39112: lcd: Add support for flushing LCD fb from dcache after update 392 arm: + lubbock 393+common/libcommon.o: In function `lcd_sync': 394+/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range' 395+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 396+make: *** [/u-boot/lcd9b/.bm-work/00/build/u-boot] Error 139 39713: tegra: Align LCD frame buffer to section boundary 39814: tegra: Support control of cache settings for LCD 39915: tegra: fdt: Add LCD definitions for Seaboard 40016: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console 401-/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range' 402+/u-boot/lcd9b/.bm-work/00/common/lcd.c:125: undefined reference to `flush_dcache_range' 40317: tegra: Enable display/lcd support on Seaboard 40418: wip 405 406So the problem is in lcd.c, due to missing cache operations. This information 407should be enough to work out what that commit is doing to break these 408boards. (In this case pxa did not have cache operations defined). 409 410If you see error lines marked with - that means that the errors were fixed 411by that commit. Sometimes commits can be in the wrong order, so that a 412breakage is introduced for a few commits and fixed by later commits. This 413shows up clearly with buildman. You can then reorder the commits and try 414again. 415 416At commit 16, the error moves - you can see that the old error at line 120 417is fixed, but there is a new one at line 126. This is probably only because 418we added some code and moved the broken line father down the file. 419 420If many boards have the same error, then -e will display the error only 421once. This makes the output as concise as possible. 422 423The full build output in this case is available in: 424 425../lcd9b/12_of_18_gd92aff7_lcd--Add-support-for/lubbock/ 426 427 done: Indicates the build was done, and holds the return code from make. 428 This is 0 for a good build, typically 2 for a failure. 429 430 err: Output from stderr, if any. Errors and warnings appear here. 431 432 log: Output from stdout. Normally there isn't any since buildman runs 433 in silent mode for now. 434 435 toolchain: Shows information about the toolchain used for the build. 436 437 sizes: Shows image size information. 438 439It is possible to get the build output there also. Use the -k option for 440this. In that case you will also see some output files, like: 441 442 System.map toolchain u-boot u-boot.bin u-boot.map autoconf.mk 443 (also SPL versions u-boot-spl and u-boot-spl.bin if available) 444 445 446Checking Image Sizes 447==================== 448 449A key requirement for U-Boot is that you keep code/data size to a minimum. 450Where a new feature increases this noticeably it should normally be put 451behind a CONFIG flag so that boards can leave it off and keep the image 452size more or less the same with each new release. 453 454To check the impact of your commits on image size, use -S. For example: 455 456$ ./tools/buildman/buildman -b us-x86 -sS 457Summary of 10 commits for 1066 boards (4 threads, 1 job per thread) 45801: MAKEALL: add support for per architecture toolchains 45902: x86: Add function to get top of usable ram 460 x86: (for 1/3 boards) text -272.0 rodata +41.0 46103: x86: Add basic cache operations 46204: x86: Permit bootstage and timer data to be used prior to relocation 463 x86: (for 1/3 boards) data +16.0 46405: x86: Add an __end symbol to signal the end of the U-Boot binary 465 x86: (for 1/3 boards) text +76.0 46606: x86: Rearrange the output input to remove BSS 467 x86: (for 1/3 boards) bss -2140.0 46807: x86: Support relocation of FDT on start-up 469 x86: + coreboot-x86 47008: x86: Add error checking to x86 relocation code 47109: x86: Adjust link device tree include file 47210: x86: Enable CONFIG_OF_CONTROL on coreboot 473 474 475You can see that image size only changed on x86, which is good because this 476series is not supposed to change any other board. From commit 7 onwards the 477build fails so we don't get code size numbers. The numbers are fractional 478because they are an average of all boards for that architecture. The 479intention is to allow you to quickly find image size problems introduced by 480your commits. 481 482Note that the 'text' region and 'rodata' are split out. You should add the 483two together to get the total read-only size (reported as the first column 484in the output from binutil's 'size' utility). 485 486A useful option is --step which lets you skip some commits. For example 487--step 2 will show the image sizes for only every 2nd commit (so it will 488compare the image sizes of the 1st, 3rd, 5th... commits). You can also use 489--step 0 which will compare only the first and last commits. This is useful 490for an overview of how your entire series affects code size. 491 492You can also use -d to see a detailed size breakdown for each board. This 493list is sorted in order from largest growth to largest reduction. 494 495It is possible to go a little further with the -B option (--bloat). This 496shows where U-Boot has bloated, breaking the size change down to the function 497level. Example output is below: 498 499$ ./tools/buildman/buildman -b us-mem4 -sSdB 500... 50119: Roll crc32 into hash infrastructure 502 arm: (for 10/10 boards) all -143.4 bss +1.2 data -4.8 rodata -48.2 text -91.6 503 paz00 : all +23 bss -4 rodata -29 text +56 504 u-boot: add: 1/0, grow: 3/-2 bytes: 168/-104 (64) 505 function old new delta 506 hash_command 80 160 +80 507 crc32_wd_buf - 56 +56 508 ext4fs_read_file 540 568 +28 509 insert_var_value_sub 688 692 +4 510 run_list_real 1996 1992 -4 511 do_mem_crc 168 68 -100 512 trimslice : all -9 bss +16 rodata -29 text +4 513 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12) 514 function old new delta 515 hash_command 80 160 +80 516 crc32_wd_buf - 56 +56 517 ext4fs_iterate_dir 672 668 -4 518 ext4fs_read_file 568 548 -20 519 do_mem_crc 168 68 -100 520 whistler : all -9 bss +16 rodata -29 text +4 521 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12) 522 function old new delta 523 hash_command 80 160 +80 524 crc32_wd_buf - 56 +56 525 ext4fs_iterate_dir 672 668 -4 526 ext4fs_read_file 568 548 -20 527 do_mem_crc 168 68 -100 528 seaboard : all -9 bss -28 rodata -29 text +48 529 u-boot: add: 1/0, grow: 3/-2 bytes: 160/-104 (56) 530 function old new delta 531 hash_command 80 160 +80 532 crc32_wd_buf - 56 +56 533 ext4fs_read_file 548 568 +20 534 run_list_real 1996 2000 +4 535 do_nandboot 760 756 -4 536 do_mem_crc 168 68 -100 537 colibri_t20_iris: all -9 rodata -29 text +20 538 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-112 (28) 539 function old new delta 540 hash_command 80 160 +80 541 crc32_wd_buf - 56 +56 542 read_abs_bbt 204 208 +4 543 do_nandboot 760 756 -4 544 ext4fs_read_file 576 568 -8 545 do_mem_crc 168 68 -100 546 ventana : all -37 bss -12 rodata -29 text +4 547 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12) 548 function old new delta 549 hash_command 80 160 +80 550 crc32_wd_buf - 56 +56 551 ext4fs_iterate_dir 672 668 -4 552 ext4fs_read_file 568 548 -20 553 do_mem_crc 168 68 -100 554 harmony : all -37 bss -16 rodata -29 text +8 555 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-124 (16) 556 function old new delta 557 hash_command 80 160 +80 558 crc32_wd_buf - 56 +56 559 nand_write_oob_syndrome 428 432 +4 560 ext4fs_iterate_dir 672 668 -4 561 ext4fs_read_file 568 548 -20 562 do_mem_crc 168 68 -100 563 medcom-wide : all -417 bss +28 data -16 rodata -93 text -336 564 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288) 565 function old new delta 566 crc32_wd_buf - 56 +56 567 do_fat_read_at 2872 2904 +32 568 hash_algo 16 - -16 569 do_mem_crc 168 68 -100 570 hash_command 420 160 -260 571 tec : all -449 bss -4 data -16 rodata -93 text -336 572 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288) 573 function old new delta 574 crc32_wd_buf - 56 +56 575 do_fat_read_at 2872 2904 +32 576 hash_algo 16 - -16 577 do_mem_crc 168 68 -100 578 hash_command 420 160 -260 579 plutux : all -481 bss +16 data -16 rodata -93 text -388 580 u-boot: add: 1/-1, grow: 1/-3 bytes: 68/-408 (-340) 581 function old new delta 582 crc32_wd_buf - 56 +56 583 do_load_serial_bin 1688 1700 +12 584 hash_algo 16 - -16 585 do_fat_read_at 2904 2872 -32 586 do_mem_crc 168 68 -100 587 hash_command 420 160 -260 588 powerpc: (for 5/5 boards) all +37.4 data -3.2 rodata -41.8 text +82.4 589 MPC8610HPCD : all +55 rodata -29 text +84 590 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 591 function old new delta 592 hash_command - 176 +176 593 do_mem_crc 184 88 -96 594 MPC8641HPCN : all +55 rodata -29 text +84 595 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 596 function old new delta 597 hash_command - 176 +176 598 do_mem_crc 184 88 -96 599 MPC8641HPCN_36BIT: all +55 rodata -29 text +84 600 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 601 function old new delta 602 hash_command - 176 +176 603 do_mem_crc 184 88 -96 604 sbc8641d : all +55 rodata -29 text +84 605 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 606 function old new delta 607 hash_command - 176 +176 608 do_mem_crc 184 88 -96 609 xpedite517x : all -33 data -16 rodata -93 text +76 610 u-boot: add: 1/-1, grow: 0/-1 bytes: 176/-112 (64) 611 function old new delta 612 hash_command - 176 +176 613 hash_algo 16 - -16 614 do_mem_crc 184 88 -96 615... 616 617 618This shows that commit 19 has increased text size for arm (although only one 619board was built) and by 96 bytes for powerpc. This increase was offset in both 620cases by reductions in rodata and data/bss. 621 622Shown below the summary lines is the sizes for each board. Below each board 623is the sizes for each function. This information starts with: 624 625 add - number of functions added / removed 626 grow - number of functions which grew / shrunk 627 bytes - number of bytes of code added to / removed from all functions, 628 plus the total byte change in brackets 629 630The change seems to be that hash_command() has increased by more than the 631do_mem_crc() function has decreased. The function sizes typically add up to 632roughly the text area size, but note that every read-only section except 633rodata is included in 'text', so the function total does not exactly 634correspond. 635 636It is common when refactoring code for the rodata to decrease as the text size 637increases, and vice versa. 638 639 640Providing 'make' flags 641====================== 642 643U-Boot's build system supports a few flags (such as BUILD_TAG) which affect 644the build product. These flags can be specified in the buildman settings 645file. They can also be useful when building U-Boot against other open source 646software. 647 648[make-flags] 649at91-boards=ENABLE_AT91_TEST=1 650snapper9260=${at91-boards} BUILD_TAG=442 651snapper9g45=${at91-boards} BUILD_TAG=443 652 653This will use 'make ENABLE_AT91_TEST=1 BUILD_TAG=442' for snapper9260 654and 'make ENABLE_AT91_TEST=1 BUILD_TAG=443' for snapper9g45. A special 655variable ${target} is available to access the target name (snapper9260 and 656snapper9g20 in this case). Variables are resolved recursively. 657 658It is expected that any variables added are dealt with in U-Boot's 659config.mk file and documented in the README. 660 661 662Other options 663============= 664 665Buildman has various other command line options. Try --help to see them. 666 667 668How to change from MAKEALL 669========================== 670 671Buildman includes most of the features of MAKEALL and is generally faster 672and easier to use. In particular it builds entire branches: if a particular 673commit introduces an error in a particular board, buildman can easily show 674you this, even if a later commit fixes that error. 675 676The reasons to deprecate MAKEALL are: 677- We don't want to maintain two build systems 678- Buildman is typically faster 679- Buildman has a lot more features 680 681But still, many people will be sad to lose MAKEALL. If you are used to 682MAKEALL, here are a few pointers. 683 684First you need to set up your tool chains - see the 'Setting up' section 685for details. Once you have your required toolchain(s) detected then you are 686ready to go. 687 688Buildman works on entire branches, so the normal use is: 689 690 ./tools/buildman/buildman -b <branch_name> <list of things to build> 691 692followed by (afterwards, or perhaps concurrently in another terminal): 693 694 ./tools/buildman/buildman -b <branch_name> -s <list of things to build> 695 696to see the results of the build. Rather than showing you all the output, 697buildman just shows a summary, with red indicating that a commit introduced 698an error and green indicating that a commit fixed an error. Use the -e 699flag to see the full errors. 700 701You don't need to stick around on that branch while buildman is running. It 702checks out its own copy of the source code, so you can change branches, 703add commits, etc. without affecting the build in progress. 704 705The <list of things to build> can include board names, architectures or the 706like. There are no flags to disambiguate since ambiguities are rare. Using 707the examples from MAKEALL: 708 709Examples: 710 - build all Power Architecture boards: 711 MAKEALL -a powerpc 712 MAKEALL --arch powerpc 713 MAKEALL powerpc 714 ** buildman -b <branch> powerpc 715 - build all PowerPC boards manufactured by vendor "esd": 716 MAKEALL -a powerpc -v esd 717 ** buildman -b <branch> esd 718 - build all PowerPC boards manufactured either by "keymile" or "siemens": 719 MAKEALL -a powerpc -v keymile -v siemens 720 ** buildman -b <branch> keymile siemens 721 - build all Freescale boards with MPC83xx CPUs, plus all 4xx boards: 722 MAKEALL -c mpc83xx -v freescale 4xx 723 ** buildman -b <branch> mpc83xx freescale 4xx 724 725Buildman automatically tries to use all the CPUs in your machine. If you 726are building a lot of boards it will use one thread for every CPU core 727it detects in your machine. This is like MAKEALL's BUILD_NBUILDS option. 728You can use the -T flag to change the number of threads. If you are only 729building a few boards, buildman will automatically run make with the -j 730flag to increase the number of concurrent make tasks. It isn't normally 731that helpful to fiddle with this option, but if you use the BUILD_NCPUS 732option in MAKEALL then -j is the equivalent in buildman. 733 734Buildman puts its output in ../<branch_name> by default but you can change 735this with the -o option. Buildman normally does out-of-tree builds: use -i 736to disable that if you really want to. But be careful that once you have 737used -i you pollute buildman's copies of the source tree, and you will need 738to remove the build directory (normally ../<branch_name>) to run buildman 739in normal mode (without -i). 740 741Buildman doesn't keep the output result normally, but use the -k option to 742do this. 743 744Please read 'Theory of Operation' a few times as it will make a lot of 745things clearer. 746 747Some options you might like are: 748 749 -B shows which functions are growing/shrinking in which commit - great 750 for finding code bloat. 751 -S shows image sizes for each commit (just an overall summary) 752 -u shows boards that you haven't built yet 753 --step 0 will build just the upstream commit and the last commit of your 754 branch. This is often a quick sanity check that your branch doesn't 755 break anything. But note this does not check bisectability! 756 757 758TODO 759==== 760 761This has mostly be written in my spare time as a response to my difficulties 762in testing large series of patches. Apart from tidying up there is quite a 763bit of scope for improvement. Things like better error diffs, easier access 764to log files, error display while building. Also it would be nice it buildman 765could 'hunt' for problems, perhaps by building a few boards for each arch, 766or checking commits for changed files and building only boards which use 767those files. 768 769 770Credits 771======= 772 773Thanks to Grant Grundler <grundler@chromium.org> for his ideas for improving 774the build speed by building all commits for a board instead of the other 775way around. 776 777 778Simon Glass 779sjg@chromium.org 780Halloween 2012 781Updated 12-12-12 782Updated 23-02-13 783