1# Copyright (c) 2013 The Chromium OS Authors. 2# 3# SPDX-License-Identifier: GPL-2.0+ 4# 5 6What is this? 7============= 8 9This tool handles building U-Boot to check that you have not broken it 10with your patch series. It can build each individual commit and report 11which boards fail on which commits, and which errors come up. It aims 12to make full use of multi-processor machines. 13 14A key feature of buildman is its output summary, which allows warnings, 15errors or image size increases in a particular commit or board to be 16quickly identified and the offending commit pinpointed. This can be a big 17help for anyone working with >10 patches at a time. 18 19 20Caveats 21======= 22 23Buildman is still in its infancy. It is already a very useful tool, but 24expect to find problems and send patches. 25 26Buildman can be stopped and restarted, in which case it will continue 27where it left off. This should happen cleanly and without side-effects. 28If not, it is a bug, for which a patch would be welcome. 29 30Buildman gets so tied up in its work that it can ignore the outside world. 31You may need to press Ctrl-C several times to quit it. Also it will print 32out various exceptions when stopped. 33 34 35Theory of Operation 36=================== 37 38(please read this section in full twice or you will be perpetually confused) 39 40Buildman is a builder. It is not make, although it runs make. It does not 41produce any useful output on the terminal while building, except for 42progress information. All the output (errors, warnings and binaries if you 43are ask for them) is stored in output directories, which you can look at 44while the build is progressing, or when it is finished. 45 46Buildman produces a concise summary of which boards succeeded and failed. 47It shows which commit introduced which board failure using a simple 48red/green colour coding. Full error information can be requested, in which 49case it is de-duped and displayed against the commit that introduced the 50error. An example workflow is below. 51 52Buildman stores image size information and can report changes in image size 53from commit to commit. An example of this is below. 54 55Buildman starts multiple threads, and each thread builds for one board at 56a time. A thread starts at the first commit, configures the source for your 57board and builds it. Then it checks out the next commit and does an 58incremental build. Eventually the thread reaches the last commit and stops. 59If errors or warnings are found along the way, the thread will reconfigure 60after every commit, and your build will be very slow. This is because a 61file that produces just a warning would not normally be rebuilt in an 62incremental build. 63 64Buildman works in an entirely separate place from your U-Boot repository. 65It creates a separate working directory for each thread, and puts the 66output files in the working directory, organised by commit name and board 67name, in a two-level hierarchy. 68 69Buildman is invoked in your U-Boot directory, the one with the .git 70directory. It clones this repository into a copy for each thread, and the 71threads do not affect the state of your git repository. Any checkouts done 72by the thread affect only the working directory for that thread. 73 74Buildman automatically selects the correct toolchain for each board. You 75must supply suitable toolchains, but buildman takes care of selecting the 76right one. 77 78Buildman always builds a branch, and always builds the upstream commit as 79well, for comparison. It cannot build individual commits at present, unless 80(maybe) you point it at an empty branch. Put all your commits in a branch, 81set the branch's upstream to a valid value, and all will be well. Otherwise 82buildman will perform random actions. Use -n to check what the random 83actions might be. 84 85Buildman is optimised for building many commits at once, for many boards. 86On multi-core machines, Buildman is fast because it uses most of the 87available CPU power. When it gets to the end, or if you are building just 88a few commits or boards, it will be pretty slow. As a tip, if you don't 89plan to use your machine for anything else, you can use -T to increase the 90number of threads beyond the default. 91 92Buildman lets you build all boards, or a subset. Specify the subset by passing 93command-line arguments that list the desired board name, architecture name, 94SOC name, or anything else in the boards.cfg file. Multiple arguments are 95allowed. Each argument will be interpreted as a regular expression, so 96behaviour is a superset of exact or substring matching. Examples are: 97 98* 'tegra20' All boards with a Tegra20 SoC 99* 'tegra' All boards with any Tegra Soc (Tegra20, Tegra30, Tegra114...) 100* '^tegra[23]0$' All boards with either Tegra20 or Tegra30 SoC 101* 'powerpc' All PowerPC boards 102 103Buildman does not store intermediate object files. It optionally copies 104the binary output into a directory when a build is successful. Size 105information is always recorded. It needs a fair bit of disk space to work, 106typically 250MB per thread. 107 108 109Setting up 110========== 111 1121. Get the U-Boot source. You probably already have it, but if not these 113steps should get you started with a repo and some commits for testing. 114 115$ cd /path/to/u-boot 116$ git clone git://git.denx.de/u-boot.git . 117$ git checkout -b my-branch origin/master 118$ # Add some commits to the branch, reading for testing 119 1202. Create ~/.buildman to tell buildman where to find tool chains. As an 121example: 122 123# Buildman settings file 124 125[toolchain] 126root: / 127rest: /toolchains/* 128eldk: /opt/eldk-4.2 129 130[toolchain-alias] 131x86: i386 132blackfin: bfin 133sh: sh4 134nds32: nds32le 135openrisc: or32 136 137 138This selects the available toolchain paths. Add the base directory for 139each of your toolchains here. Buildman will search inside these directories 140and also in any '/usr' and '/usr/bin' subdirectories. 141 142Make sure the tags (here root: rest: and eldk:) are unique. 143 144The toolchain-alias section indicates that the i386 toolchain should be used 145to build x86 commits. 146 147 1482. Check the available toolchains 149 150Run this check to make sure that you have a toolchain for every architecture. 151 152$ ./tools/buildman/buildman --list-tool-chains 153Scanning for tool chains 154 - scanning path '/' 155 - looking in '/.' 156 - looking in '/bin' 157 - looking in '/usr/bin' 158 - found '/usr/bin/gcc' 159Tool chain test: OK 160 - found '/usr/bin/c89-gcc' 161Tool chain test: OK 162 - found '/usr/bin/c99-gcc' 163Tool chain test: OK 164 - found '/usr/bin/x86_64-linux-gnu-gcc' 165Tool chain test: OK 166 - scanning path '/toolchains/powerpc-linux' 167 - looking in '/toolchains/powerpc-linux/.' 168 - looking in '/toolchains/powerpc-linux/bin' 169 - found '/toolchains/powerpc-linux/bin/powerpc-linux-gcc' 170Tool chain test: OK 171 - looking in '/toolchains/powerpc-linux/usr/bin' 172 - scanning path '/toolchains/nds32le-linux-glibc-v1f' 173 - looking in '/toolchains/nds32le-linux-glibc-v1f/.' 174 - looking in '/toolchains/nds32le-linux-glibc-v1f/bin' 175 - found '/toolchains/nds32le-linux-glibc-v1f/bin/nds32le-linux-gcc' 176Tool chain test: OK 177 - looking in '/toolchains/nds32le-linux-glibc-v1f/usr/bin' 178 - scanning path '/toolchains/nios2' 179 - looking in '/toolchains/nios2/.' 180 - looking in '/toolchains/nios2/bin' 181 - found '/toolchains/nios2/bin/nios2-linux-gcc' 182Tool chain test: OK 183 - found '/toolchains/nios2/bin/nios2-linux-uclibc-gcc' 184Tool chain test: OK 185 - looking in '/toolchains/nios2/usr/bin' 186 - found '/toolchains/nios2/usr/bin/nios2-linux-gcc' 187Tool chain test: OK 188 - found '/toolchains/nios2/usr/bin/nios2-linux-uclibc-gcc' 189Tool chain test: OK 190 - scanning path '/toolchains/microblaze-unknown-linux-gnu' 191 - looking in '/toolchains/microblaze-unknown-linux-gnu/.' 192 - looking in '/toolchains/microblaze-unknown-linux-gnu/bin' 193 - found '/toolchains/microblaze-unknown-linux-gnu/bin/microblaze-unknown-linux-gnu-gcc' 194Tool chain test: OK 195 - found '/toolchains/microblaze-unknown-linux-gnu/bin/mb-linux-gcc' 196Tool chain test: OK 197 - looking in '/toolchains/microblaze-unknown-linux-gnu/usr/bin' 198 - scanning path '/toolchains/mips-linux' 199 - looking in '/toolchains/mips-linux/.' 200 - looking in '/toolchains/mips-linux/bin' 201 - found '/toolchains/mips-linux/bin/mips-linux-gcc' 202Tool chain test: OK 203 - looking in '/toolchains/mips-linux/usr/bin' 204 - scanning path '/toolchains/old' 205 - looking in '/toolchains/old/.' 206 - looking in '/toolchains/old/bin' 207 - looking in '/toolchains/old/usr/bin' 208 - scanning path '/toolchains/i386-linux' 209 - looking in '/toolchains/i386-linux/.' 210 - looking in '/toolchains/i386-linux/bin' 211 - found '/toolchains/i386-linux/bin/i386-linux-gcc' 212Tool chain test: OK 213 - looking in '/toolchains/i386-linux/usr/bin' 214 - scanning path '/toolchains/bfin-uclinux' 215 - looking in '/toolchains/bfin-uclinux/.' 216 - looking in '/toolchains/bfin-uclinux/bin' 217 - found '/toolchains/bfin-uclinux/bin/bfin-uclinux-gcc' 218Tool chain test: OK 219 - looking in '/toolchains/bfin-uclinux/usr/bin' 220 - scanning path '/toolchains/sparc-elf' 221 - looking in '/toolchains/sparc-elf/.' 222 - looking in '/toolchains/sparc-elf/bin' 223 - found '/toolchains/sparc-elf/bin/sparc-elf-gcc' 224Tool chain test: OK 225 - looking in '/toolchains/sparc-elf/usr/bin' 226 - scanning path '/toolchains/arm-2010q1' 227 - looking in '/toolchains/arm-2010q1/.' 228 - looking in '/toolchains/arm-2010q1/bin' 229 - found '/toolchains/arm-2010q1/bin/arm-none-linux-gnueabi-gcc' 230Tool chain test: OK 231 - looking in '/toolchains/arm-2010q1/usr/bin' 232 - scanning path '/toolchains/from' 233 - looking in '/toolchains/from/.' 234 - looking in '/toolchains/from/bin' 235 - looking in '/toolchains/from/usr/bin' 236 - scanning path '/toolchains/sh4-gentoo-linux-gnu' 237 - looking in '/toolchains/sh4-gentoo-linux-gnu/.' 238 - looking in '/toolchains/sh4-gentoo-linux-gnu/bin' 239 - found '/toolchains/sh4-gentoo-linux-gnu/bin/sh4-gentoo-linux-gnu-gcc' 240Tool chain test: OK 241 - looking in '/toolchains/sh4-gentoo-linux-gnu/usr/bin' 242 - scanning path '/toolchains/avr32-linux' 243 - looking in '/toolchains/avr32-linux/.' 244 - looking in '/toolchains/avr32-linux/bin' 245 - found '/toolchains/avr32-linux/bin/avr32-gcc' 246Tool chain test: OK 247 - looking in '/toolchains/avr32-linux/usr/bin' 248 - scanning path '/toolchains/m68k-linux' 249 - looking in '/toolchains/m68k-linux/.' 250 - looking in '/toolchains/m68k-linux/bin' 251 - found '/toolchains/m68k-linux/bin/m68k-linux-gcc' 252Tool chain test: OK 253 - looking in '/toolchains/m68k-linux/usr/bin' 254List of available toolchains (17): 255arm : /toolchains/arm-2010q1/bin/arm-none-linux-gnueabi-gcc 256avr32 : /toolchains/avr32-linux/bin/avr32-gcc 257bfin : /toolchains/bfin-uclinux/bin/bfin-uclinux-gcc 258c89 : /usr/bin/c89-gcc 259c99 : /usr/bin/c99-gcc 260i386 : /toolchains/i386-linux/bin/i386-linux-gcc 261m68k : /toolchains/m68k-linux/bin/m68k-linux-gcc 262mb : /toolchains/microblaze-unknown-linux-gnu/bin/mb-linux-gcc 263microblaze: /toolchains/microblaze-unknown-linux-gnu/bin/microblaze-unknown-linux-gnu-gcc 264mips : /toolchains/mips-linux/bin/mips-linux-gcc 265nds32le : /toolchains/nds32le-linux-glibc-v1f/bin/nds32le-linux-gcc 266nios2 : /toolchains/nios2/bin/nios2-linux-gcc 267powerpc : /toolchains/powerpc-linux/bin/powerpc-linux-gcc 268sandbox : /usr/bin/gcc 269sh4 : /toolchains/sh4-gentoo-linux-gnu/bin/sh4-gentoo-linux-gnu-gcc 270sparc : /toolchains/sparc-elf/bin/sparc-elf-gcc 271x86_64 : /usr/bin/x86_64-linux-gnu-gcc 272 273 274You can see that everything is covered, even some strange ones that won't 275be used (c88 and c99). This is a feature. 276 277 278How to run it 279============= 280 281First do a dry run using the -n flag: (replace <branch> with a real, local 282branch with a valid upstream) 283 284$ ./tools/buildman/buildman -b <branch> -n 285 286If it can't detect the upstream branch, try checking out the branch, and 287doing something like 'git branch --set-upstream <branch> upstream/master' 288or something similar. 289 290As an exmmple: 291 292Dry run, so not doing much. But I would do this: 293 294Building 18 commits for 1059 boards (4 threads, 1 job per thread) 295Build directory: ../lcd9b 296 5bb3505 Merge branch 'master' of git://git.denx.de/u-boot-arm 297 c18f1b4 tegra: Use const for pinmux_config_pingroup/table() 298 2f043ae tegra: Add display support to funcmux 299 e349900 tegra: fdt: Add pwm binding and node 300 424a5f0 tegra: fdt: Add LCD definitions for Tegra 301 0636ccf tegra: Add support for PWM 302 a994fe7 tegra: Add SOC support for display/lcd 303 fcd7350 tegra: Add LCD driver 304 4d46e9d tegra: Add LCD support to Nvidia boards 305 991bd48 arm: Add control over cachability of memory regions 306 54e8019 lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment 307 d92aff7 lcd: Add support for flushing LCD fb from dcache after update 308 dbd0677 tegra: Align LCD frame buffer to section boundary 309 0cff9b8 tegra: Support control of cache settings for LCD 310 9c56900 tegra: fdt: Add LCD definitions for Seaboard 311 5cc29db lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console 312 cac5a23 tegra: Enable display/lcd support on Seaboard 313 49ff541 wip 314 315Total boards to build for each commit: 1059 316 317This shows that it will build all 1059 boards, using 4 threads (because 318we have a 4-core CPU). Each thread will run with -j1, meaning that each 319make job will use a single CPU. The list of commits to be built helps you 320confirm that things look about right. Notice that buildman has chosen a 321'base' directory for you, immediately above your source tree. 322 323Buildman works entirely inside the base directory, here ../lcd9b, 324creating a working directory for each thread, and creating output 325directories for each commit and board. 326 327 328Suggested Workflow 329================== 330 331To run the build for real, take off the -n: 332 333$ ./tools/buildman/buildman -b <branch> 334 335Buildman will set up some working directories, and get started. After a 336minute or so it will settle down to a steady pace, with a display like this: 337 338Building 18 commits for 1059 boards (4 threads, 1 job per thread) 339 528 36 124 /19062 1:13:30 : SIMPC8313_SP 340 341This means that it is building 19062 board/commit combinations. So far it 342has managed to succesfully build 528. Another 36 have built with warnings, 343and 124 more didn't build at all. Buildman expects to complete the process 344in an hour and 15 minutes. Use this time to buy a faster computer. 345 346 347To find out how the build went, ask for a summary with -s. You can do this 348either before the build completes (presumably in another terminal) or or 349afterwards. Let's work through an example of how this is used: 350 351$ ./tools/buildman/buildman -b lcd9b -s 352... 35301: Merge branch 'master' of git://git.denx.de/u-boot-arm 354 powerpc: + galaxy5200_LOWBOOT 35502: tegra: Use const for pinmux_config_pingroup/table() 35603: tegra: Add display support to funcmux 35704: tegra: fdt: Add pwm binding and node 35805: tegra: fdt: Add LCD definitions for Tegra 35906: tegra: Add support for PWM 36007: tegra: Add SOC support for display/lcd 36108: tegra: Add LCD driver 36209: tegra: Add LCD support to Nvidia boards 36310: arm: Add control over cachability of memory regions 36411: lcd: Add CONFIG_LCD_ALIGNMENT to select frame buffer alignment 36512: lcd: Add support for flushing LCD fb from dcache after update 366 arm: + lubbock 36713: tegra: Align LCD frame buffer to section boundary 36814: tegra: Support control of cache settings for LCD 36915: tegra: fdt: Add LCD definitions for Seaboard 37016: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console 37117: tegra: Enable display/lcd support on Seaboard 37218: wip 373 374This shows which commits have succeeded and which have failed. In this case 375the build is still in progress so many boards are not built yet (use -u to 376see which ones). But still we can see a few failures. The galaxy5200_LOWBOOT 377never builds correctly. This could be a problem with our toolchain, or it 378could be a bug in the upstream. The good news is that we probably don't need 379to blame our commits. The bad news is it isn't tested on that board. 380 381Commit 12 broke lubbock. That's what the '+ lubbock' means. The failure 382is never fixed by a later commit, or you would see lubbock again, in green, 383without the +. 384 385To see the actual error: 386 387$ ./tools/buildman/buildman -b <branch> -se lubbock 388... 38912: lcd: Add support for flushing LCD fb from dcache after update 390 arm: + lubbock 391+common/libcommon.o: In function `lcd_sync': 392+/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range' 393+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 394+make: *** [/u-boot/lcd9b/.bm-work/00/build/u-boot] Error 139 39513: tegra: Align LCD frame buffer to section boundary 39614: tegra: Support control of cache settings for LCD 39715: tegra: fdt: Add LCD definitions for Seaboard 39816: lcd: Add CONFIG_CONSOLE_SCROLL_LINES option to speed console 399-/u-boot/lcd9b/.bm-work/00/common/lcd.c:120: undefined reference to `flush_dcache_range' 400+/u-boot/lcd9b/.bm-work/00/common/lcd.c:125: undefined reference to `flush_dcache_range' 40117: tegra: Enable display/lcd support on Seaboard 40218: wip 403 404So the problem is in lcd.c, due to missing cache operations. This information 405should be enough to work out what that commit is doing to break these 406boards. (In this case pxa did not have cache operations defined). 407 408If you see error lines marked with - that means that the errors were fixed 409by that commit. Sometimes commits can be in the wrong order, so that a 410breakage is introduced for a few commits and fixed by later commits. This 411shows up clearly with buildman. You can then reorder the commits and try 412again. 413 414At commit 16, the error moves - you can see that the old error at line 120 415is fixed, but there is a new one at line 126. This is probably only because 416we added some code and moved the broken line futher down the file. 417 418If many boards have the same error, then -e will display the error only 419once. This makes the output as concise as possible. 420 421The full build output in this case is available in: 422 423../lcd9b/12_of_18_gd92aff7_lcd--Add-support-for/lubbock/ 424 425 done: Indicates the build was done, and holds the return code from make. 426 This is 0 for a good build, typically 2 for a failure. 427 428 err: Output from stderr, if any. Errors and warnings appear here. 429 430 log: Output from stdout. Normally there isn't any since buildman runs 431 in silent mode for now. 432 433 toolchain: Shows information about the toolchain used for the build. 434 435 sizes: Shows image size information. 436 437It is possible to get the build output there also. Use the -k option for 438this. In that case you will also see some output files, like: 439 440 System.map toolchain u-boot u-boot.bin u-boot.map autoconf.mk 441 (also SPL versions u-boot-spl and u-boot-spl.bin if available) 442 443 444Checking Image Sizes 445==================== 446 447A key requirement for U-Boot is that you keep code/data size to a minimum. 448Where a new feature increases this noticeably it should normally be put 449behind a CONFIG flag so that boards can leave it off and keep the image 450size more or less the same with each new release. 451 452To check the impact of your commits on image size, use -S. For example: 453 454$ ./tools/buildman/buildman -b us-x86 -sS 455Summary of 10 commits for 1066 boards (4 threads, 1 job per thread) 45601: MAKEALL: add support for per architecture toolchains 45702: x86: Add function to get top of usable ram 458 x86: (for 1/3 boards) text -272.0 rodata +41.0 45903: x86: Add basic cache operations 46004: x86: Permit bootstage and timer data to be used prior to relocation 461 x86: (for 1/3 boards) data +16.0 46205: x86: Add an __end symbol to signal the end of the U-Boot binary 463 x86: (for 1/3 boards) text +76.0 46406: x86: Rearrange the output input to remove BSS 465 x86: (for 1/3 boards) bss -2140.0 46607: x86: Support relocation of FDT on start-up 467 x86: + coreboot-x86 46808: x86: Add error checking to x86 relocation code 46909: x86: Adjust link device tree include file 47010: x86: Enable CONFIG_OF_CONTROL on coreboot 471 472 473You can see that image size only changed on x86, which is good because this 474series is not supposed to change any other board. From commit 7 onwards the 475build fails so we don't get code size numbers. The numbers are fractional 476because they are an average of all boards for that architecture. The 477intention is to allow you to quickly find image size problems introduced by 478your commits. 479 480Note that the 'text' region and 'rodata' are split out. You should add the 481two together to get the total read-only size (reported as the first column 482in the output from binutil's 'size' utility). 483 484A useful option is --step which lets you skip some commits. For example 485--step 2 will show the image sizes for only every 2nd commit (so it will 486compare the image sizes of the 1st, 3rd, 5th... commits). You can also use 487--step 0 which will compare only the first and last commits. This is useful 488for an overview of how your entire series affects code size. 489 490You can also use -d to see a detailed size breakdown for each board. This 491list is sorted in order from largest growth to largest reduction. 492 493It is possible to go a little further with the -B option (--bloat). This 494shows where U-Boot has bloted, breaking the size change down to the function 495level. Example output is below: 496 497$ ./tools/buildman/buildman -b us-mem4 -sSdB 498... 49919: Roll crc32 into hash infrastructure 500 arm: (for 10/10 boards) all -143.4 bss +1.2 data -4.8 rodata -48.2 text -91.6 501 paz00 : all +23 bss -4 rodata -29 text +56 502 u-boot: add: 1/0, grow: 3/-2 bytes: 168/-104 (64) 503 function old new delta 504 hash_command 80 160 +80 505 crc32_wd_buf - 56 +56 506 ext4fs_read_file 540 568 +28 507 insert_var_value_sub 688 692 +4 508 run_list_real 1996 1992 -4 509 do_mem_crc 168 68 -100 510 trimslice : all -9 bss +16 rodata -29 text +4 511 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12) 512 function old new delta 513 hash_command 80 160 +80 514 crc32_wd_buf - 56 +56 515 ext4fs_iterate_dir 672 668 -4 516 ext4fs_read_file 568 548 -20 517 do_mem_crc 168 68 -100 518 whistler : all -9 bss +16 rodata -29 text +4 519 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12) 520 function old new delta 521 hash_command 80 160 +80 522 crc32_wd_buf - 56 +56 523 ext4fs_iterate_dir 672 668 -4 524 ext4fs_read_file 568 548 -20 525 do_mem_crc 168 68 -100 526 seaboard : all -9 bss -28 rodata -29 text +48 527 u-boot: add: 1/0, grow: 3/-2 bytes: 160/-104 (56) 528 function old new delta 529 hash_command 80 160 +80 530 crc32_wd_buf - 56 +56 531 ext4fs_read_file 548 568 +20 532 run_list_real 1996 2000 +4 533 do_nandboot 760 756 -4 534 do_mem_crc 168 68 -100 535 colibri_t20_iris: all -9 rodata -29 text +20 536 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-112 (28) 537 function old new delta 538 hash_command 80 160 +80 539 crc32_wd_buf - 56 +56 540 read_abs_bbt 204 208 +4 541 do_nandboot 760 756 -4 542 ext4fs_read_file 576 568 -8 543 do_mem_crc 168 68 -100 544 ventana : all -37 bss -12 rodata -29 text +4 545 u-boot: add: 1/0, grow: 1/-3 bytes: 136/-124 (12) 546 function old new delta 547 hash_command 80 160 +80 548 crc32_wd_buf - 56 +56 549 ext4fs_iterate_dir 672 668 -4 550 ext4fs_read_file 568 548 -20 551 do_mem_crc 168 68 -100 552 harmony : all -37 bss -16 rodata -29 text +8 553 u-boot: add: 1/0, grow: 2/-3 bytes: 140/-124 (16) 554 function old new delta 555 hash_command 80 160 +80 556 crc32_wd_buf - 56 +56 557 nand_write_oob_syndrome 428 432 +4 558 ext4fs_iterate_dir 672 668 -4 559 ext4fs_read_file 568 548 -20 560 do_mem_crc 168 68 -100 561 medcom-wide : all -417 bss +28 data -16 rodata -93 text -336 562 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288) 563 function old new delta 564 crc32_wd_buf - 56 +56 565 do_fat_read_at 2872 2904 +32 566 hash_algo 16 - -16 567 do_mem_crc 168 68 -100 568 hash_command 420 160 -260 569 tec : all -449 bss -4 data -16 rodata -93 text -336 570 u-boot: add: 1/-1, grow: 1/-2 bytes: 88/-376 (-288) 571 function old new delta 572 crc32_wd_buf - 56 +56 573 do_fat_read_at 2872 2904 +32 574 hash_algo 16 - -16 575 do_mem_crc 168 68 -100 576 hash_command 420 160 -260 577 plutux : all -481 bss +16 data -16 rodata -93 text -388 578 u-boot: add: 1/-1, grow: 1/-3 bytes: 68/-408 (-340) 579 function old new delta 580 crc32_wd_buf - 56 +56 581 do_load_serial_bin 1688 1700 +12 582 hash_algo 16 - -16 583 do_fat_read_at 2904 2872 -32 584 do_mem_crc 168 68 -100 585 hash_command 420 160 -260 586 powerpc: (for 5/5 boards) all +37.4 data -3.2 rodata -41.8 text +82.4 587 MPC8610HPCD : all +55 rodata -29 text +84 588 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 589 function old new delta 590 hash_command - 176 +176 591 do_mem_crc 184 88 -96 592 MPC8641HPCN : all +55 rodata -29 text +84 593 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 594 function old new delta 595 hash_command - 176 +176 596 do_mem_crc 184 88 -96 597 MPC8641HPCN_36BIT: all +55 rodata -29 text +84 598 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 599 function old new delta 600 hash_command - 176 +176 601 do_mem_crc 184 88 -96 602 sbc8641d : all +55 rodata -29 text +84 603 u-boot: add: 1/0, grow: 0/-1 bytes: 176/-96 (80) 604 function old new delta 605 hash_command - 176 +176 606 do_mem_crc 184 88 -96 607 xpedite517x : all -33 data -16 rodata -93 text +76 608 u-boot: add: 1/-1, grow: 0/-1 bytes: 176/-112 (64) 609 function old new delta 610 hash_command - 176 +176 611 hash_algo 16 - -16 612 do_mem_crc 184 88 -96 613... 614 615 616This shows that commit 19 has increased text size for arm (although only one 617board was built) and by 96 bytes for powerpc. This increase was offset in both 618cases by reductions in rodata and data/bss. 619 620Shown below the summary lines is the sizes for each board. Below each board 621is the sizes for each function. This information starts with: 622 623 add - number of functions added / removed 624 grow - number of functions which grew / shrunk 625 bytes - number of bytes of code added to / removed from all functions, 626 plus the total byte change in brackets 627 628The change seems to be that hash_command() has increased by more than the 629do_mem_crc() function has decreased. The function sizes typically add up to 630roughly the text area size, but note that every read-only section except 631rodata is included in 'text', so the function total does not exactly 632correspond. 633 634It is common when refactoring code for the rodata to decrease as the text size 635increases, and vice versa. 636 637 638Providing 'make' flags 639====================== 640 641U-Boot's build system supports a few flags (such as BUILD_TAG) which affect 642the build product. These flags can be specified in the buildman settings 643file. They can also be useful when building U-Boot against other open source 644software. 645 646[make-flags] 647at91-boards=ENABLE_AT91_TEST=1 648snapper9260=${at91-boards} BUILD_TAG=442 649snapper9g45=${at91-boards} BUILD_TAG=443 650 651This will use 'make ENABLE_AT91_TEST=1 BUILD_TAG=442' for snapper9260 652and 'make ENABLE_AT91_TEST=1 BUILD_TAG=443' for snapper9g45. A special 653variable ${target} is available to access the target name (snapper9260 and 654snapper9g20 in this case). Variables are resolved recursively. 655 656It is expected that any variables added are dealt with in U-Boot's 657config.mk file and documented in the README. 658 659 660Other options 661============= 662 663Buildman has various other command line options. Try --help to see them. 664 665 666TODO 667==== 668 669This has mostly be written in my spare time as a response to my difficulties 670in testing large series of patches. Apart from tidying up there is quite a 671bit of scope for improvement. Things like better error diffs, easier access 672to log files, error display while building. Also it would be nice it buildman 673could 'hunt' for problems, perhaps by building a few boards for each arch, 674or checking commits for changed files and building only boards which use 675those files. 676 677 678Credits 679======= 680 681Thanks to Grant Grundler <grundler@chromium.org> for his ideas for improving 682the build speed by building all commits for a board instead of the other 683way around. 684 685 686Simon Glass 687sjg@chromium.org 688Halloween 2012 689Updated 12-12-12 690Updated 23-02-13 691