docs/overview.txt
author Yann E. MORIN
Sun Jun 14 21:44:33 2009 +0200 (2009-06-14)
changeset 1406 515580cad631
parent 1405 8b86a6f004ae
child 1491 ae44a02d67fb
permissions -rw-r--r--
[populate] enhance documentation

This patch adds explanations on how
populate works internally.
     1 File.........: overview.txt
     2 Content......: Overview of how crosstool-NG works.
     3 Copyrigth....: (C) 2007 Yann E. MORIN <yann.morin.1998@anciens.enib.fr>
     4 License......: Creative Commons Attribution Share Alike (CC-by-sa), v2.5
     5 
     6 ____________________
     7                    /
     8 Table Of Content  /
     9 _________________/
    10 
    11 
    12 Introduction
    13 History
    14 Installing crosstool-NG
    15   Install method
    16   The hacker's way
    17   Preparing for packaging
    18   Shell completion
    19   Contributed code
    20 Configuring crosstool-NG
    21   Interesting config options
    22   Re-building an existing toolchain
    23 Running crosstool-NG
    24   Stopping and restarting a build
    25   Testing all toolchains at once
    26   Overriding the number of // jobs
    27 Using the toolchain
    28 Toolchain types
    29 Internals
    30   Makefile front-end
    31   Kconfig parser
    32   Architecture-specific
    33   Adding a new version of a component
    34   Build scripts
    35 
    36 ________________
    37                /
    38 Introduction  /
    39 _____________/
    40 
    41 crosstool-NG aims at building toolchains. Toolchains are an essential component
    42 in a software development project. It will compile, assemble and link the code
    43 that is being developed. Some pieces of the toolchain will eventually end up
    44 in the resulting binary/ies: static libraries are but an example.
    45 
    46 So, a toolchain is a very sensitive piece of software, as any bug in one of the
    47 components, or a poorly configured component, can lead to execution problems,
    48 ranging from poor performance, to applications ending unexpectedly, to
    49 mis-behaving software (which more than often is hard to detect), to hardware
    50 damage, or even to human risks (which is more than regrettable).
    51 
    52 Toolchains are made of different piece of software, each being quite complex
    53 and requiring specially crafted options to build and work seamlessly. This
    54 is usually not that easy, even in the not-so-trivial case of native toolchains.
    55 The work reaches a higher degree of complexity when it comes to cross-
    56 compilation, where it can become quite a nightmare...
    57 
    58 Some cross-toolchains exist on the internet, and can be used for general
    59 development, but they have a number of limitations:
    60   - they can be general purpose, in that they are configured for the majority:
    61     no optimisation for your specific target,
    62   - they can be prepared for a specific target and thus are not easy to use,
    63     nor optimised for, or even supporting your target,
    64   - they often are using aging components (compiler, C library, etc...) not
    65     supporting special features of your shiny new processor;
    66 On the other side, these toolchain offer some advantages:
    67   - they are ready to use and quite easy to install and setup,
    68   - they are proven if used by a wide community.
    69 
    70 But once you want to get all the juice out of your specific hardware, you will
    71 want to build your own toolchain. This is where crosstool-NG comes into play.
    72 
    73 There are also a number of tools that build toolchains for specific needs,
    74 which are not really scalable. Examples are:
    75   - buildroot (buildroot.uclibc.org) whose main purpose is to build root file
    76     systems, hence the name. But once you have your toolchain with buildroot,
    77     part of it is installed in the root-to-be, so if you want to build a whole
    78     new root, you either have to save the existing one as a template and
    79     restore it later, or restart again from scratch. This is not convenient,
    80   - ptxdist (www.pengutronix.de/software/ptxdist), whose purpose is very
    81     similar to buildroot,
    82   - other projects (openembedded.org for example), which are again used to
    83     build root file systems.
    84 
    85 crosstool-NG is really targeted at building toolchains, and only toolchains.
    86 It is then up to you to use it the way you want.
    87 
    88 ___________
    89           /
    90 History  /
    91 ________/
    92 
    93 crosstool was first 'conceived' by Dan Kegel, who offered it to the community
    94 as a set of scripts, a repository of patches, and some pre-configured, general
    95 purpose setup files to be used to configure crosstool. This is available at
    96 http://www.kegel.com/crosstool, and the subversion repository is hosted on
    97 google at http://code.google.com/p/crosstool/.
    98 
    99 I once managed to add support for uClibc-based toolchains, but it did not make
   100 into mainline, mostly because I didn't have time to port the patch forward to
   101 the new versions, due in part to the big effort it was taking.
   102 
   103 So I decided to clean up crosstool in the state it was, re-order the things
   104 in place, add appropriate support for what I needed, that is uClibc support
   105 and a menu-driven configuration, named the new implementation crosstool-NG,
   106 (standing for crosstool Next Generation, as many other comunity projects do,
   107 and as a wink at the TV series "Star Trek: The Next Generation" ;-) ) and
   108 made it available to the community, in case it was of interest to any one.
   109 
   110 ___________________________
   111                           /
   112 Installing crosstool-NG  /
   113 ________________________/
   114 
   115 There are two ways you can use crosstool-NG:
   116  - build and install it, then get rid of the sources like you'd do for most
   117    programs,
   118  - or only build it and run from the source directory.
   119 
   120 The former should be used if you got crosstool-NG from a packaged tarball, see
   121 "Install method", below, while the latter is most useful for developpers that
   122 checked the code out from SVN, and want to submit patches, see "The Hacker's
   123 way", below.
   124 
   125 Install method |
   126 ---------------+
   127 
   128 If you go for the install, then you just follow the classical, but yet easy
   129 ./configure way:
   130   ./configure --prefix=/some/place
   131   make
   132   make install
   133   export PATH="${PATH}:/some/place/bin"
   134 
   135 You can then get rid of crosstool-NG source. Next create a directory to serve
   136 as a working place, cd in there and run:
   137   ct-ng help
   138 
   139 See below for complete usage.
   140 
   141 The Hacker's way |
   142 -----------------+
   143 
   144 If you go the hacker's way, then the usage is a bit different, although very
   145 simple:
   146   ./configure --local
   147   make
   148   make install
   149 
   150 Now, *do not* remove crosstool-NG sources. They are needed to run crosstool-NG!
   151 Stay in the directory holding the sources, and run:
   152   ./ct-ng help
   153 
   154 See below for complete usage.
   155 
   156 Now, provided you checked-out the code, you can send me your interesting changes
   157 by running:
   158   svn diff
   159 
   160 and mailing me the result! :-P
   161 
   162 Preparing for packaging |
   163 ------------------------+
   164 
   165 If you plan on packaging crosstool-NG, you surely don't want to install it
   166 in your root file system. The install procedure of crosstool-NG honors the
   167 DESTDIR variable:
   168 
   169   ./configure --prefix=/usr
   170   make
   171   make DESTDIR=/packaging/place install
   172 
   173 Shell completion |
   174 -----------------+
   175 
   176 crosstool-NG comes with a shell script fragment that defines bash-compatible
   177 completion. That shell fragment is currently not installed automatically, but
   178 this is planned.
   179 
   180 To install the shell script fragment, you have two options:
   181  - install system-wide, most probably by copying ct-ng.comp into
   182    /etc/bash_completion.d/
   183  - install for a single user, by copying ct-ng.comp into ${HOME}/ and
   184    sourcing this file from your ${HOME}/.bashrc
   185 
   186 Contributed code |
   187 -----------------+
   188 
   189 Some people contibuted code that couldn't get merged for various reasons. This
   190 code is available as patches in the contrib/ sub-directory. These patches are
   191 to be applied to the source of crosstool-NG, prior to installing.
   192 
   193 An easy way to use contributed code is to pass the --with-contrib= option to
   194 ./configure. The possible values depend upon which contributions are packaged
   195 with your version, but you can get with it with passing one of those two
   196 special values:
   197   --with-contrib=list
   198     will list all available contributions
   199 
   200   --with-contrib=all
   201     will select all avalaible contributions
   202 
   203 There is no guarantee that a particuliar contribution applies to the current
   204 version of crosstool-ng, or that it will work at all. Use contributions at
   205 your own risk.
   206 
   207 ____________________________
   208                            /
   209 Configuring crosstool-NG  /
   210 _________________________/
   211 
   212 crosstool-NG is configured with a configurator presenting a menu-stuctured set
   213 of options. These options let you specify the way you want your toolchain
   214 built, where you want it installed, what architecture and specific processor it
   215 will support, the version of the components you want to use, etc... The
   216 value for those options are then stored in a configuration file.
   217 
   218 The configurator works the same way you configure your Linux kernel. It is
   219 assumed you now how to handle this.
   220 
   221 To enter the menu, type:
   222   ct-ng menuconfig
   223 
   224 Almost every config item has a help entry. Read them carefully.
   225 
   226 String and number options can refer to environment variables. In such a case,
   227 you must use the shell syntax: ${VAR}. You shall neither single- nor double-
   228 quote the string/number options.
   229 
   230 There are three environment variables that are computed by crosstool-NG, and
   231 that you can use:
   232 
   233 CT_TARGET:
   234   It represents the target tuple you are building for. You can use it for
   235   example in the installation/prefix directory, such as:
   236     /opt/x-tools/${CT_TARGET}
   237 
   238 CT_TOP_DIR:
   239   The top directory where crosstool-NG is running. You shouldn't need it in
   240   most cases. There is one case where you may need it: if you have local
   241   patches and you store them in your running directory, you can refer to them
   242   by using CT_TOP_DIR, such as:
   243     ${CT_TOP_DIR}/patches.myproject
   244 
   245 CT_VERSION:
   246   The version of crosstool-NG you are using. Not much use for you, but it's
   247   there if you need it.
   248 
   249 Interesting config options |
   250 ---------------------------+
   251 
   252 CT_LOCAL_TARBALLS_DIR:
   253   If you already have some tarballs in a direcotry, enter it here. That will
   254   speed up the retrieving phase, where crosstool-NG would otherwise download
   255   those tarballs.
   256 
   257 CT_PREFIX_DIR:
   258   This is where the toolchain will be installed in (and for now, where it
   259   will run from). Common use is to add the target tuple in the directory
   260   path, such as (see above):
   261     /opt/x-tools/${CT_TARGET}
   262 
   263 CT_TARGET_VENDOR:
   264   An identifier for your toolchain, will take place in the vendor part of the
   265   target tuple. It shall *not* contain spaces or dashes. Usually, keep it
   266   to a one-word string, or use underscores to separate words if you need.
   267   Avoid dots, commas, and special characters.
   268 
   269 CT_TARGET_ALIAS:
   270   An alias for the toolchian. It will be used as a prefix to the toolchain
   271   tools. For example, you will have ${CT_TARGET_ALIAS}-gcc
   272 
   273 Also, if you think you don't see enough versions, you can try to enable one of
   274 those:
   275 
   276 CT_OBSOLETE:
   277   Show obsolete versions or tools. Most of the time, you don't want to base
   278   your toolchain on too old a version (of gcc, for example). But at times, it
   279   can come handy to use such an old version for regression tests. Those old
   280   versions are hidden behind CT_OBSOLETE. Those versions (or features) are so
   281   marked because maintaining support for those in crosstool-NG would be too
   282   costly, time-wise, and time is dear.
   283 
   284 CT_EXPERIMENTAL:
   285   Show experimental versions or tools. Again, you might not want to base your
   286   toolchain on too recent tools (eg. gcc) for production. But if you need a
   287   feature present only in a recent version, or a new tool, you can find them
   288   hidden behind CT_EXPERIMENTAL. Those versions (or features) did not (yet)
   289   receive thorough testing in crosstool-NG, and/or are not mature enough to
   290   be blindly trusted.
   291 
   292 Re-building an existing toolchain |
   293 ----------------------------------+
   294 
   295 If you have an existing toolchain, you can re-use the options used to build it
   296 to create a new toolchain. That needs a very little bit of effort on your side
   297 but is quite easy. The options to build a toolchain are saved with the
   298 toolchain, and you can retrieve this configuration by running:
   299   ${CT_TARGET}-config
   300 
   301 This will dump the configuration to stdout, so to rebuild a toolchain with this
   302 configuration, the following is all you need to do:
   303   ${CT_TARGET}-config >.config
   304   ct-ng oldconfig
   305 
   306 Then, you can review and change the configuration by running:
   307   ct-ng menuconfig
   308 
   309 ________________________
   310                        /
   311 Running crosstool-NG  /
   312 _____________________/
   313 
   314 To build the toolchain, simply type:
   315   ct-ng build
   316 
   317 This will use the above configuration to retrieve, extract and patch the
   318 components, build, install and eventually test your newly built toolchain.
   319 
   320 You are then free to add the toolchain /bin directory in your PATH to use
   321 it at will.
   322 
   323 In any case, you can get some terse help. Just type:
   324   ct-ng help
   325 or:
   326   man 1 ct-ng
   327 
   328 Stopping and restarting a build |
   329 --------------------------------+
   330 
   331 If you want to stop the build after a step you are debugging, you can pass the
   332 variable STOP to make:
   333   ct-ng STOP=some_step
   334 
   335 Conversely, if you want to restart a build at a specific step you are
   336 debugging, you can pass the RESTART variable to make:
   337   ct-ng RESTART=some_step
   338 
   339 Alternatively, you can call make with the name of a step to just do that step:
   340   ct-ng libc_headers
   341 is equivalent to:
   342   ct-ng RESTART=libc_headers STOP=libc_headers
   343 
   344 The shortcuts +step_name and step_name+ allow to respectively stop or restart
   345 at that step. Thus:
   346   ct-ng +libc_headers        and:    ct-ng libc_headers+
   347 are equivalent to:
   348   ct-ng STOP=libc_headers    and:    ct-ng RESTART=libc_headers
   349 
   350 To obtain the list of acceptable steps, please call:
   351   ct-ng list-steps
   352 
   353 Note that in order to restart a build, you'll have to say 'Y' to the config
   354 option CT_DEBUG_CT_SAVE_STEPS, and that the previous build effectively went
   355 that far.
   356 
   357 Building all toolchains at once |
   358 --------------------------------+
   359 
   360 You can build all samples; simply call:
   361   ct-ng build-all
   362 
   363 Overriding the number of // jobs |
   364 ---------------------------------+
   365 
   366 If you want to override the number of jobs to run in // (the -j option to
   367 make), you can either re-enter the menuconfig, or simply add it on the command
   368 line, as such:
   369   ct-ng build.4
   370 
   371 which tells crosstool-NG to override the number of // jobs to 4.
   372 
   373 You can see the actions that support overriding the number of // jobs in
   374 the help menu. Those are the ones with [.#] after them (eg. build[.#] or
   375 build-all[.#], and so on...).
   376 
   377 Note on // jobs |
   378 ----------------+
   379 
   380 The crosstool-NG script 'ct-ng' is a Makefile-script. It does *not* execute
   381 in parallel (there is not much to gain). When speaking of // jobs, we are
   382 refering to the number of // jobs when making the *components*. That is, we
   383 speak of the number of // jobs used to build gcc, glibc, and so on...
   384 
   385 
   386 _______________________
   387                       /
   388 Using the toolchain  /
   389 ____________________/
   390 
   391 Using the toolchain is as simple as adding the toolchain's bin directory in
   392 your PATH, such as:
   393   export PATH="${PATH}:/your/toolchain/path/bin"
   394 
   395 and then using the target tuple to tell the build systems to use your
   396 toolchain:
   397   ./configure --target=your-target-tuple
   398 or
   399   make CC=your-target-tuple-gcc
   400 or
   401   make CROSS_COMPILE=your-target-tuple-
   402 and so on...
   403 
   404 It is strongly advised not to use the toolchain sys-root directory as an
   405 install directory for your programs/packages. If you do so, you will not be
   406 able to use your toolchain for another project. It is even strongly advised
   407 that your toolchain is chmod-ed to read-only once successfully build, so that
   408 you don't go polluting your toolchain with your programs/packages' files.
   409 
   410 Thus, when you build a program/package, install it in a separate directory,
   411 eg. /your/root. This directory is the /image/ of what would be in the root file
   412 system of your target, and will contain all that your programs/packages have
   413 installed.
   414 
   415 The 'populate' script |
   416 ----------------------+
   417 
   418 When your root directory is ready, it is still missing some important bits: the
   419 toolchain's libraries. To populate your root directory with those libs, just
   420 run:
   421   your-target-tuple-populate -s /your/root -d /your/root-populated
   422 
   423 This will copy /your/root into /your/root-populated, and put the needed and only
   424 the needed libraries there. Thus you don't polute /your/root with any cruft that
   425 would no longer be needed should you have to remove stuff. /your/root always
   426 contains only those things you install in it.
   427 
   428 You can then use /your/root-populated to build up your file system image, a
   429 tarball, or to NFS-mount it from your target, or whatever you need.
   430 
   431 The populate script accepts the following options:
   432 
   433  -s src_dir
   434     Use 'src_dir' as the un-populated root directory.
   435 
   436  -d dst_dir
   437     Put the populated root directory in 'dst_dir'.
   438 
   439  -l lib1 [...]
   440     Always add specified libraries.
   441 
   442  -L file
   443     Always add libraries listed in 'file'.
   444 
   445  -f
   446     Remove 'dst_dir' if it previously existed; continue even if any library
   447     specified with -l or -L is missing.
   448 
   449  -v
   450     Be verbose, and tell what's going on (you can see exactly where libs are
   451     coming from).
   452 
   453  -h
   454     Print the help.
   455 
   456 See 'your-target-tuple-populate -h' for more information on the options.
   457 
   458 Here is how populate works:
   459 
   460   1) performs some sanity checks:
   461      - src_dir and dst_dir are specified
   462      - src_dir exists
   463      - unless forced, dst_dir does not exist
   464      - src_dir != dst_dir
   465 
   466   2) copy src_dir to dst_dir
   467 
   468   3) add forced libraries to dst_dir
   469      - build the list from -l and -L options
   470      - get forced libraries from the sysroot (see below for heuristics)
   471        - abort on the first missing library, unless -f is specified
   472 
   473   4) add all missing libraries to dst_dir
   474      - scan dst_dir for every ELF files that are 'executable' or
   475        'shared object'
   476      - list the "NEEDED Shared library" fields
   477        - check if the library is already in dst_dir/lib or dst_dir/usr/lib
   478        - if not, get the library from the sysroot
   479          - if it's in sysroot/lib, copy it to dst_dir/lib
   480          - if it's in sysroot/usr/lib, copy it to dst_dir/usr/lib
   481          - in both cases, use the SONAME of the library to create the file
   482            in dst_dir
   483          - if it was not found in the sysroot, this is an error.
   484 
   485 ___________________
   486                   /
   487 Toolchain types  /
   488 ________________/
   489 
   490 There are four kinds of toolchains you could encounter.
   491 
   492 First off, you must understand the following: when it comes to compilers there
   493 are up to four machines involved:
   494   1) the machine configuring the toolchain components: the config machine
   495   2) the machine building the toolchain components:    the build machine
   496   3) the machine running the toolchain:                the host machine
   497   4) the machine the toolchain is generating code for: the target machine
   498 
   499 We can most of the time assume that the config machine and the build machine
   500 are the same. Most of the time, this will be true. The only time it isn't
   501 is if you're using distributed compilation (such as distcc). Let's forget
   502 this for the sake of simplicity.
   503 
   504 So we're left with three machines:
   505  - build
   506  - host
   507  - target
   508 
   509 Any toolchain will involve those three machines. You can be as pretty sure of
   510 this as "2 and 2 are 4". Here is how they come into play:
   511 
   512 1) build == host == target
   513     This is a plain native toolchain, targetting the exact same machine as the
   514     one it is built on, and running again on this exact same machine. You have
   515     to build such a toolchain when you want to use an updated component, such
   516     as a newer gcc for example.
   517     crosstool-NG calls it "native".
   518 
   519 2) build == host != target
   520     This is a classic cross-toolchain, which is expected to be run on the same
   521     machine it is compiled on, and generate code to run on a second machine,
   522     the target.
   523     crosstool-NG calls it "cross".
   524 
   525 3) build != host == target
   526     Such a toolchain is also a native toolchain, as it targets the same machine
   527     as it runs on. But it is build on another machine. You want such a
   528     toolchain when porting to a new architecture, or if the build machine is
   529     much faster than the host machine.
   530     crosstool-NG calls it "cross-native".
   531 
   532 4) build != host != target
   533     This one is called a canadian-toolchain (*), and is tricky. The three
   534     machines in play are different. You might want such a toolchain if you
   535     have a fast build machine, but the users will use it on another machine,
   536     and will produce code to run on a third machine.
   537     crosstool-NG calls it "canadian".
   538 
   539 crosstool-NG can build all these kinds of toolchains (or is aiming at it,
   540 anyway!)
   541 
   542 (*) The term Canadian Cross came about because at the time that these issues
   543     were all being hashed out, Canada had three national political parties.
   544     http://en.wikipedia.org/wiki/Cross_compiler
   545 
   546 _____________
   547             /
   548 Internals  /
   549 __________/
   550 
   551 Internally, crosstool-NG is script-based. To ease usage, the frontend is
   552 Makefile-based.
   553 
   554 Makefile front-end |
   555 -------------------+
   556 
   557 The entry point to crosstool-NG is the Makefile script "ct-ng". Calling this
   558 script with an action will act exactly as if the Makefile was in the current
   559 working directory and make was called with the action as rule. Thus:
   560   ct-ng menuconfig
   561 
   562 is equivalent to having the Makefile in CWD, and calling:
   563   make menuconfig
   564 
   565 Having ct-ng as it is avoids copying the Makefile everywhere, and acts as a
   566 traditional command.
   567 
   568 ct-ng loads sub- Makefiles from the library directory $(CT_LIB_DIR), as set up
   569 at configuration time with ./configure.
   570 
   571 ct-ng also searches for config files, sub-tools, samples, scripts and patches in
   572 that library directory.
   573 
   574 Because of a stupid make behavior/bug I was unable to track down, implicit make
   575 rules are disabled: installing with --local would triger those rules, and mconf
   576 was unbuildable.
   577 
   578 Kconfig parser |
   579 ---------------+
   580 
   581 The kconfig language is a hacked version, vampirised from the Linux kernel
   582 (http://www.kernel.org/), and (heavily) adapted to my needs.
   583 
   584 The list of the most notable changes (at least the ones I remember) follows:
   585 - the CONFIG_ prefix has been replaced with CT_
   586 - a leading | in prompts is skipped, and subsequent leading spaces are not
   587   trimmed
   588 - otherwise leading spaces are silently trimmed
   589 
   590 The kconfig parsers (conf and mconf) are not installed pre-built, but as
   591 source files. Thus you can have the directory where crosstool-NG is installed,
   592 exported (via NFS or whatever) and have clients with different architectures
   593 use the same crosstool-NG installation, and most notably, the same set of
   594 patches.
   595 
   596 Architecture-specific |
   597 ----------------------+
   598 
   599 Note: this chapter is not really well written, and might thus be a little bit
   600 complex to understand. To get a better grasp of what an architecture is, the
   601 reader is kindly encouraged to look at the "arch/" sub-directory, and to the
   602 existing architectures to see how things are laid out.
   603 
   604 An architecture is defined by:
   605 
   606  - a human-readable name, in lower case letters, with numbers as appropriate.
   607    The underscore is allowed; space and special characters are not.
   608      Eg.: arm, x86_64
   609  - a file in "config/arch/", named after the architecture's name, and suffixed
   610    with ".in".
   611      Eg.: config/arch/arm.in
   612  - a file in "scripts/build/arch/", named after the architecture's name, and
   613    suffixed with ".sh".
   614      Eg.: scripts/build/arch/arm.sh
   615 
   616 The architecture's ".in" file API:
   617  > the config option "ARCH_%arch%" (where %arch% is to be replaced with the
   618    actual architecture name).
   619    That config option must have *neither* a type, *nor* a prompt! Also, it can
   620    *not* depend on any other config option (EXPERIMENTAL is managed as above).
   621      Eg.:
   622        config ARCH_arm
   623    + mandatory:
   624        defines a (terse) help entry for this architecture:
   625        Eg.:
   626          config ARCH_arm
   627            help
   628              The ARM architecture.
   629    + optional:
   630        selects adequate associated config options.
   631        Note: 64-bit architectures *shall* select ARCH_64
   632        Eg.:
   633          config ARCH_arm
   634            select ARCH_SUPPORTS_BOTH_ENDIAN
   635            select ARCH_DEFAULT_LE
   636            help
   637              The ARM architecture.
   638        Eg.:
   639          config ARCH_x86_64
   640             select ARCH_64
   641             help
   642               The x86_64 architecture.
   643 
   644  > other target-specific options, at your discretion. Note however that to
   645    avoid name-clashing, such options shall be prefixed with "ARCH_%arch%",
   646    where %arch% is again replaced by the actual architecture name.
   647    (Note: due to historical reasons, and lack of time to clean up the code,
   648     I may have left some config options that do not completely conform to
   649     this, as the architecture name was written all upper case. However, the
   650     prefix is unique among architectures, and does not cause harm).
   651 
   652 The architecture's ".sh" file API:
   653  > the function "CT_DoArchTupleValues"
   654    + parameters: none
   655    + environment:
   656      - all variables from the ".config" file,
   657      - the two variables "target_endian_eb" and "target_endian_el" which are
   658        the endianness suffixes
   659    + return value: 0 upon success, !0 upon failure
   660    + provides:
   661      - mandatory
   662      - the environment variable CT_TARGET_ARCH
   663      - contains:
   664        the architecture part of the target tuple.
   665        Eg.: "armeb" for big endian ARM
   666             "i386" for an i386
   667    + provides:
   668      - optional
   669      - the environment variable CT_TARGET_SYS
   670      - contains:
   671        the sytem part of the target tuple.
   672        Eg.: "gnu" for glibc on most architectures
   673             "gnueabi" for glibc on an ARM EABI
   674      - defaults to:
   675        - for glibc-based toolchain: "gnu"
   676        - for uClibc-based toolchain: "uclibc"
   677    + provides:
   678      - optional
   679      - the environment variable CT_KERNEL_ARCH
   680      - contains:
   681        the architecture name as understandable by the Linux kernel build
   682        system.
   683        Eg.: "arm" for an ARM
   684             "powerpc" for a PowerPC
   685             "i386" for an x86
   686      - defaults to:
   687        ${CT_ARCH}
   688    + provides:
   689      - optional
   690      - the environment variables to configure the cross-gcc (defaults)
   691        - CT_ARCH_WITH_ARCH    : the gcc ./configure switch to select architecture level         ( "--with-arch=${CT_ARCH_ARCH}"   )
   692        - CT_ARCH_WITH_ABI     : the gcc ./configure switch to select ABI level                  ( "--with-abi=${CT_ARCH_ABI}"     )
   693        - CT_ARCH_WITH_CPU     : the gcc ./configure switch to select CPU instruction set        ( "--with-cpu=${CT_ARCH_CPU}"     )
   694        - CT_ARCH_WITH_TUNE    : the gcc ./configure switch to select scheduling                 ( "--with-tune=${CT_ARCH_TUNE}"   )
   695        - CT_ARCH_WITH_FPU     : the gcc ./configure switch to select FPU type                   ( "--with-fpu=${CT_ARCH_FPU}"     )
   696        - CT_ARCH_WITH_FLOAT   : the gcc ./configure switch to select floating point arithmetics ( "--with-float=soft" or /empty/  )
   697    + provides:
   698      - optional
   699      - the environment variables to pass to the cross-gcc to build target binaries (defaults)
   700        - CT_ARCH_ARCH_CFLAG   : the gcc switch to select architecture level                     ( "-march=${CT_ARCH_ARCH}"            )
   701        - CT_ARCH_ABI_CFLAG    : the gcc switch to select ABI level                              ( "-mabi=${CT_ARCH_ABI}"              )
   702        - CT_ARCH_CPU_CFLAG    : the gcc switch to select CPU instruction set                    ( "-mcpu=${CT_ARCH_CPU}"              )
   703        - CT_ARCH_TUNE_CFLAG   : the gcc switch to select scheduling                             ( "-mtune=${CT_ARCH_TUNE}"            )
   704        - CT_ARCH_FPU_CFLAG    : the gcc switch to select FPU type                               ( "-mfpu=${CT_ARCH_FPU}"              )
   705        - CT_ARCH_FLOAT_CFLAG  : the gcc switch to choose floating point arithmetics             ( "-msoft-float" or /empty/           )
   706        - CT_ARCH_ENDIAN_CFLAG : the gcc switch to choose big or little endian                   ( "-mbig-endian" or "-mlittle-endian" )
   707      - default to:
   708        see above.
   709    + provides:
   710      - optional
   711      - the environement variables to configure the core and final compiler, specific to this architecture:
   712        - CT_ARCH_CC_CORE_EXTRA_CONFIG   : additional, architecture specific core gcc ./configure flags
   713        - CT_ARCH_CC_EXTRA_CONFIG        : additional, architecture specific final gcc ./configure flags
   714      - default to:
   715        - all empty
   716    + provides:
   717      - optional
   718      - the architecture-specific CFLAGS and LDFLAGS:
   719        - CT_ARCH_TARGET_CLFAGS
   720        - CT_ARCH_TARGET_LDFLAGS
   721      - default to:
   722        - all empty
   723 
   724 You can have a look at "config/arch/arm.in" and "scripts/build/arch/arm.sh" for
   725 a quite complete example of what an actual architecture description looks like.
   726 
   727 Kernel specific |
   728 ----------------+
   729 
   730 A kernel is defined by:
   731 
   732  - a human-readable name, in lower case letters, with numbers as appropriate.
   733    The underscore is allowed; space and special characters are not (although
   734    they are internally replaced with underscores.
   735      Eg.: linux, bare-metal
   736  - a file in "config/kernel/", named after the kernel name, and suffixed with
   737    ".in".
   738      Eg.: config/kernel/linux.in, config/kernel/bare-metal.in
   739  - a file in "scripts/build/kernel/", named after the kernel name, and suffixed
   740    with ".sh".
   741      Eg.: scripts/build/kernel/linux.sh, scripts/build/kernel/bare-metal.sh
   742 
   743 The kernel's ".in" file must contain:
   744  > an optional lines containing exactly "# EXPERIMENTAL", starting on the
   745    first column, and without any following space or other character.
   746    If this line is present, then this kernel is considered EXPERIMENTAL,
   747    and correct dependency on EXPERIMENTAL will be set.
   748 
   749  > the config option "KERNEL_%kernel_name%" (where %kernel_name% is to be
   750    replaced with the actual kernel name, with all special characters and
   751    spaces replaced by underscores).
   752    That config option must have *neither* a type, *nor* a prompt! Also, it can
   753    *not* depends on EXPERIMENTAL.
   754      Eg.: KERNEL_linux, KERNEL_bare_metal
   755    + mandatory:
   756        defines a (terse) help entry for this kernel.
   757        Eg.:
   758          config KERNEL_bare_metal
   759            help
   760              Build a compiler for use without any kernel.
   761    + optional:
   762        selects adequate associated config options.
   763        Eg.:
   764          config KERNEL_bare_metal
   765            select BARE_METAL
   766            help
   767              Build a compiler for use without any kernel.
   768 
   769  > other kernel specific options, at your discretion. Note however that, to
   770    avoid name-clashing, such options should be prefixed with
   771    "KERNEL_%kernel_name%", where %kernel_name% is again tp be replaced with
   772    the actual kernel name.
   773    (Note: due to historical reasons, and lack of time to clean up the code,
   774     I may have left some config options that do not completely conform to
   775     this, as the kernel name was written all upper case. However, the prefix
   776     is unique among kernels, and does not cause harm).
   777 
   778 The kernel's ".sh" file API:
   779  > is a bash script fragment
   780 
   781  > defines the function CT_DoKernelTupleValues
   782    + see the architecture's CT_DoArchTupleValues, except for:
   783    + set the environment variable CT_TARGET_KERNEL, the kernel part of the
   784      target tuple
   785    + return value: ignored
   786 
   787  > defines the function "do_kernel_get":
   788    + parameters: none
   789    + environment:
   790       - all variables from the ".config" file.
   791    + return value: 0 for success, !0 for failure.
   792    + behavior: download the kernel's sources, and store the tarball into
   793      "${CT_TARBALLS_DIR}". To this end, a functions is available, that
   794      abstracts downloading tarballs:
   795      - CT_DoGet <tarball_base_name> <URL1 [URL...]>
   796        Eg.: CT_DoGet linux-2.6.26.5 ftp://ftp.kernel.org/pub/linux/kernel/v2.6
   797      Note: retrieving sources from svn, cvs, git and the likes is not supported
   798      by CT_DoGet. You'll have to do this by hand, as it is done for eglibc in
   799      "scripts/build/libc/eglibc.sh"
   800 
   801  > defines the function "do_kernel_extract":
   802    + parameters: none
   803    + environment:
   804       - all variables from the ".config" file,
   805    + return value: 0 for success, !0 for failure.
   806    + behavior: extract the kernel's tarball into "${CT_SRC_DIR}", and apply
   807      required patches. To this end, a function is available, that abstracts
   808      extracting tarballs:
   809      - CT_ExtractAndPatch <tarball_base_name>
   810        Eg.: CT_ExtractAndPatch linux-2.6.26.5
   811 
   812  > defines the function "do_kernel_headers":
   813    + parameters: none
   814    + environment:
   815       - all variables from the ".config" file,
   816    + return value: 0 for success, !0 for failure.
   817    + behavior: install the kernel headers (if any) in "${CT_SYSROOT_DIR}/usr/include"
   818 
   819  > defines any kernel-specific helper functions
   820    These functions, if any, must be prefixed with "do_kernel_%CT_KERNEL%_",
   821    where '%CT_KERNEL%' is to be replaced with the actual kernel name, to avoid
   822    any name-clashing.
   823 
   824 You can have a look at "config/kernel/linux.in" and "scripts/build/kernel/linux.sh"
   825 as an example of what a complex kernel description looks like.
   826 
   827 Adding a new version of a component |
   828 ------------------------------------+
   829 
   830 When a new component, such as the Linux kernel, gcc or any other is released,
   831 adding the new version to crosstool-NG is quite easy. There is a script that
   832 will do all that for you:
   833   scripts/addToolVersion.sh
   834 
   835 Run it with no option to get some help.
   836 
   837 Build scripts |
   838 --------------+
   839 
   840 To Be Written later...