yann@1: File.........: overview.txt yann@197: Content......: Overview of how crosstool-NG works. yann@92: Copyrigth....: (C) 2007 Yann E. MORIN yann@192: License......: Creative Commons Attribution Share Alike (CC-by-sa), v2.5 yann@92: yann@1: ________________ yann@1: / yann@1: Introduction / yann@1: _____________/ yann@1: yann@1: crosstool-NG aims at building toolchains. Toolchains are an essential component yann@1: in a software development project. It will compile, assemble and link the code yann@1: that is being developped. Some pieces of the toolchain will eventually end up yann@1: in the resulting binary/ies: static libraries are but an example. yann@1: yann@1: So, a toolchain is a very sensitive piece of software, as any bug in one of the yann@1: components, or a poorly configured component, can lead to execution problems, yann@1: ranging from poor performance, to applications ending unexpectedly, to yann@1: mis-behaving software (which more than often is hard to detect), to hardware yann@1: damage, or even to human risks (which is more than regretable). yann@1: yann@1: Toolchains are made of different piece of software, each being quite complex yann@1: and requiring specially crafted options to build and work seamlessly. This yann@1: is usually not that easy, even in the not-so-trivial case of native toolchains. yann@1: The work reaches a higher degree of complexity when it comes to cross- yann@40: compilation, where it can become quite a nightmare... yann@1: yann@40: Some cross-toolchains exist on the internet, and can be used for general yann@1: development, but they have a number of limitations: yann@1: - they can be general purpose, in that they are configured for the majority: yann@1: no optimisation for your specific target, yann@1: - they can be prepared for a specific target and thus are not easy to use, yann@1: nor optimised for, or even supporting your target, yann@1: - they often are using ageing components (compiler, C library, etc...) not yann@1: supporting special features of your shiny new processor; yann@1: On the other side, these toolchain offer some advantages: yann@1: - they are ready to use and quite easy to install and setup, yann@1: - they are proven if used by a wide community. yann@1: yann@1: But once you want to get all the juice out of your specific hardware, you will yann@197: want to build your own toolchain. This is where crosstool-NG comes into play. yann@1: yann@1: There are also a number of tools that builds toolchains for specific needs, yann@1: which is not really scalable. Examples are: yann@1: - buildroot (buildroot.uclibc.org) whose main puprpose is to build root file yann@1: systems, hence the name. But once you have your toolchain with buildroot, yann@1: part of it is installed in the root-to-be, so if you want to build a whole yann@1: new root, you either have to save the existing one as a template and yann@1: restore it later, or restart again from scratch. This is not convenient, yann@1: - ptxdist (www.pengutronix.de/software/ptxdist), whose purpose is very yann@1: similar to buildroot, yann@1: - other projects (openembeded.org for example), which is again used to yann@1: build root file systems. yann@1: yann@1: crosstool-NG is really targetted at building toolchains, and only toolchains. yann@1: It is then up to you to use it the way you want. yann@1: yann@1: ___________ yann@1: / yann@1: History / yann@1: ________/ yann@1: yann@1: crosstool was first 'conceived' by Dan Kegel, which offered it to the community, yann@1: as a set of scripts, a repository of patches, and some pre-configured, general yann@1: purpose setup files to be used to configure crosstool. This is available at yann@203: http://www.kegel.com/crosstool, and the subversion repository is hosted on yann@203: google at http://code.google.com/p/crosstool/. yann@1: yann@1: At the time of writing, crosstool only supports building with one C library, yann@1: namely glibc, and one C compiler, gcc; it is cripled with historical support yann@92: for legacy components, and is some kind of a mess to upgrade. Also, submited yann@203: patches take a loooong time before they are integrated mainline. yann@1: yann@1: I once managed to add support for uClibc-based toolchains, but it did not make yann@1: into mainline, mostly because I don't have time to port the patch forward to yann@1: the new versions, due in part to the big effort it was taking. yann@1: yann@1: So I decided to clean up crosstool in the state it was, re-order the things yann@1: in place, and add appropriate support for what I needed, that is uClibc yann@203: support. That was a disaster, as inclusion into mainline is slow as hell, yann@203: and the changes were so numerous. yann@1: yann@1: The only option left to me was rewrite crosstool from scratch. I decided to go yann@197: this way, and name the new implementation crosstool-NG, standing for crosstool yann@197: Next Generation, as many other comunity projects do, and as a wink at the TV yann@197: series "Star Trek: The Next Generation". ;-) yann@1: yann@168: ____________________________ yann@168: / yann@168: Configuring crosstool-NG / yann@168: _________________________/ yann@168: yann@168: crosstool-NG is configured the same way you configure your Linux kernel: by yann@168: using a curses-based menu. It is assumed you now how to handle this. yann@168: yann@168: To enter the menu, type: yann@192: ct-ng menuconfig yann@168: yann@203: Almost every config item has a help entry. Read them carefully. yann@168: yann@168: String and number options can refer to environment variables. In such a case, yann@192: you must use the shell syntax: ${VAR}. You shall neither single- nor double- yann@192: quote the string options. yann@168: yann@192: There are three environment variables that are computed by crosstool-NG, and yann@168: that you can use: yann@168: yann@168: CT_TARGET: yann@168: It represents the target triplet you are building for. You can use it for yann@168: example in the installation/prefix directory, such as: yann@168: /opt/x-tools/${CT_TARGET} yann@168: yann@168: CT_TOP_DIR: yann@182: The top directory where crosstool-NG is running. You shouldn't need it in yann@182: most cases. There is one case where you may need it: if you have local yann@182: patches and you store them in your running directory, you can refer to them yann@168: by using CT_TOP_DIR, such as: yann@168: ${CT_TOP_DIR}/patches.myproject yann@168: yann@168: CT_VERSION: yann@192: The version of crosstool-NG you are using. Not much use for you, but it's yann@168: there if you need it. yann@168: yann@168: yann@168: Interesting config options | yann@168: ---------------------------* yann@168: yann@168: CT_LOCAL_TARBALLS_DIR: yann@168: If you already have sone tarballs in a direcotry, enter it here. That will yann@197: speed up the retrieving phase, where crosstool-NG would otherwise download yann@168: those tarballs. yann@168: yann@168: CT_PREFIX_DIR: yann@168: This is where the toolchain will be installed in (and for now, where it yann@168: will run from). yann@168: yann@168: CT_TARGET_VENDOR: yann@168: An identifier for your toolchain, will take place in the vendor part of the yann@168: target triplet. It shall *not* contain spaces or dashes. Usually, keep it yann@168: to a one-word string, or use underscores to separate words if you need. yann@168: Avoid dots, commas, and special characters. yann@168: yann@168: CT_TARGET_ALIAS: yann@168: An alias for the toolchian. It will be used as a prefix to the toolchain yann@168: tools. For example, you will have ${CT_TARGET_ALIAS}-gcc yann@168: yann@246: Also, if you think you don't see enough versions, you can try to enable one of yann@246: those: yann@246: yann@246: CT_OBSOLETE: yann@246: Show obsolete versions or tools. Most of the time, you don't want to base yann@246: your toolchain on too old a version (of gcc, for example). But at times, it yann@246: can come handy to use such an old version for regression tests. Those old yann@246: versions are hidden behind CT_BSOLETE. yann@246: yann@246: CT_EXPERIMENTAL: yann@246: Show experimental versions or tools. Again, you might not want to base your yann@246: toolchain on too recent tools (eg. gcc) for production. But if you need a yann@246: feature present only in a recent version, or a new tool, you can find them yann@246: hidden behind CT_EXPERIMENTAL. yann@246: yann@246: CT_BROKEN: yann@246: Show broken versions or tools. Some usefull tools are currently broken: they yann@246: won't compile, run, or worse, cause defects when running. But if you are yann@246: brave enough, you can try and debug them. They are hidden behind CT_BROKEN, yann@246: which itself is hiddent behind EXPERIMENTAL. yann@246: yann@168: ________________________ yann@168: / yann@168: Running crosstool-NG / yann@168: _____________________/ yann@1: yann@197: crosstool-NG is configured by a configurator presenting a menu-stuctured set of yann@1: options. These options let you specify the way you want your toolchain built, yann@1: where you want it installed, what architecture and specific processor it yann@1: will support, the version of the components you want to use, etc... The yann@1: value for those options are then stored in a configuration file. yann@1: yann@168: To build the toolchain, simply type: yann@203: ct-ng build yann@135: yann@135: This will use the above configuration to retrieve, extract and patch the yann@135: components, build, install and eventually test your newly built toolchain. yann@1: yann@1: You are then free to add the toolchain /bin directory in your PATH to use yann@1: it at will. yann@1: yann@135: In any case, you can get some terse help. Just type: yann@192: ct-ng help yann@203: or: yann@203: man 1 ct-ng yann@135: yann@135: yann@135: Stoping and restarting a build | yann@135: -------------------------------* yann@135: yann@135: If you want to stop the build after a step you are debugging, you can pass the yann@135: variable STOP to make: yann@192: ct-ng STOP=some_step yann@135: yann@135: Conversely, if you want to restart a build at a specific step you are yann@135: debugging, you can pass the RESTART variable to make: yann@192: ct-ng RESTART=some_step yann@135: yann@136: Alternatively, you can call make with the name of a step to just do that step: yann@192: ct-ng libc_headers yann@136: is equivalent to: yann@192: ct-ng RESTART=libs_headers STOP=libc_headers yann@136: yann@136: The shortcuts -step_name and step_name- allow to respectively stop or restart yann@136: at that step. Thus: yann@192: ct-ng -libc_headers and: ct-ng libc_headers- yann@136: are equivalent to: yann@192: ct-ng STOP=libc_headers and: ct-ng RESTART=libc_headers yann@136: yann@181: To obtain the list of acceptable steps, please call: yann@192: ct-ng liststeps yann@181: yann@168: Note that in order to restart a build, you'll have to say 'Y' to the config yann@168: option CT_DEBUG_CT_SAVE_STEPS, and that the previous build effectively went yann@168: that far. yann@92: yann@92: yann@168: Testing all toolchains at once | yann@168: -------------------------------* yann@92: yann@168: You can test-build all samples; simply call: yann@192: ct-ng regtest yann@40: yann@227: _______________________ yann@227: / yann@227: Using the toolchain / yann@227: ____________________/ yann@227: yann@227: Using the toolchain is as simple as adding the toolchain's bin directory in yann@227: your PATH, such as: yann@227: export PATH="${PATH}:/your/toolchain/path/bin" yann@227: yann@227: and then using the target triplet to tell the build systems to use your yann@227: toolchain: yann@227: ./configure --target=your-target-triplet yann@227: make CC=your-target-triplet-gcc yann@227: make CROSS_COMPILE=your-target-triplet- yann@227: and so on... yann@227: yann@227: When your root directory is ready, it is still missing some important bits: the yann@227: toolchain's libraries. To populate your root directory with those libs, just yann@227: run: yann@227: your-target-triplet-populate -s /your/root -d /your/root-populated yann@227: yann@227: This will copy /your/root into /your/root-populated, and put the needed and only yann@227: the needed libraries there. Thus you don't polute /your/root with any cruft that yann@227: would no longer be needed should you have to remove stuff. /your/root always yann@227: contains only those things you install in it. yann@227: yann@227: You can then use /your/root-populated to build up your file system image, a yann@227: tarball, or to NFS-mount it from your target, or whatever you need. yann@227: yann@40: ___________________ yann@40: / yann@40: Toolchain types / yann@40: ________________/ yann@40: yann@40: There are four kinds of toolchains you could encounter. yann@40: yann@40: First off, you must understand the following: when it comes to compilers there yann@40: are up to four machines involved: yann@40: 1) the machine configuring the toolchain components: the config machine yann@40: 2) the machine building the toolchain components: the build machine yann@40: 3) the machine running the toolchain: the host machine yann@203: 4) the machine the toolchain is generating code for: the target machine yann@40: yann@40: We can most of the time assume that the config machine and the build machine yann@40: are the same. Most of the time, this will be true. The only time it isn't yann@40: is if you're using distributed compilation (such as distcc). Let's forget yann@40: this for the sake of simplicity. yann@40: yann@40: So we're left with three machines: yann@40: - build yann@40: - host yann@40: - target yann@40: yann@40: Any toolchain will involve those three machines. You can be as pretty sure of yann@40: this as "2 and 2 are 4". Here is how they come into play: yann@40: yann@40: 1) build == host == target yann@40: This is a plain native toolchain, targetting the exact same machine as the yann@40: one it is built on, and running again on this exact same machine. You have yann@40: to build such a toolchain when you want to use an updated component, such yann@40: as a newer gcc for example. yann@197: crosstool-NG calls it "native". yann@40: yann@40: 2) build == host != target yann@40: This is a classic cross-toolchain, which is expected to be run on the same yann@40: machine it is compiled on, and generate code to run on a second machine, yann@40: the target. yann@197: crosstool-NG calls it "cross". yann@40: yann@40: 3) build != host == target yann@40: Such a toolchain is also a native toolchain, as it targets the same machine yann@40: as it runs on. But it is build on another machine. You want such a yann@40: toolchain when porting to a new architecture, or if the build machine is yann@40: much faster than the host machine. yann@197: crosstool-NG calls it "cross-native". yann@40: yann@40: 4) build != host != target yann@92: This one is called a canadian-toolchain (*), and is tricky. The three yann@40: machines in play are different. You might want such a toolchain if you yann@40: have a fast build machine, but the users will use it on another machine, yann@40: and will produce code to run on a third machine. yann@197: crosstool-NG calls it "canadian". yann@40: yann@197: crosstool-NG can build all these kinds of toolchains (or is aiming at it, yann@197: anyway!) yann@40: yann@40: (*) The term Canadian Cross came about because at the time that these issues yann@40: were all being hashed out, Canada had three national political parties. yann@40: http://en.wikipedia.org/wiki/Cross_compiler yann@40: yann@1: _____________ yann@1: / yann@1: Internals / yann@1: __________/ yann@1: yann@92: Internally, crosstool-NG is script-based. To ease usage, the frontend is yann@92: Makefile-based. yann@92: yann@92: Makefile front-end | yann@92: -------------------* yann@92: yann@203: The entry point to crosstool-NG is the Makefile script "ct-ng". Calling this yann@203: script with an action will act exactly as if the Makefile was in the current yann@203: working directory and make was called with the action as rule. Thus: yann@203: ct-ng menuconfig yann@203: is equivalent to having the Makefile in CWD, and calling: yann@203: make menuconfig yann@203: yann@203: Having ct-ng as it is avoids copying the Makefile everywhere, and acts as a yann@203: traditional command. yann@203: yann@203: ct-ng loads sub- Makefiles from the library directory $(CT_LIB_DIR), as set up yann@203: at configuration time with ./configure. yann@203: yann@203: ct-ng also search for config files, sub-tools, samples, scripts and patches in yann@203: that library directory. yann@92: yann@182: Kconfig parser | yann@182: ---------------* yann@92: yann@92: The kconfig language is a hacked version, vampirised from the toybox project yann@182: by Rob LANDLEY (http://www.landley.net/code/toybox/), itself coming from the yann@182: Linux kernel (http://www.linux.org/ http://www.kernel.org/), and (heavily) yann@182: adapted to my needs. yann@92: yann@203: The kconfig parsers (conf and mconf) are not installed pre-built, but as yann@203: source files. Thus you can have the directory where crosstool-NG is installed, yann@203: exported (via NFS or whatever) and have clients with different architectures yann@203: use the same crosstool-NG installation, and most notably, the same set of yann@203: patches. yann@203: yann@203: Build scripts | yann@203: --------------* yann@203: yann@203: To Be Written later...