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
14 Installing crosstool-NG
17 Preparing for packaging
20 Configuring crosstool-NG
21 Interesting config options
22 Re-building an existing toolchain
24 Stopping and restarting a build
25 Testing all toolchains at once
26 Overriding the number of // jobs
31 Seemingly-native toolchains
36 Adding a new version of a component
44 crosstool-NG aims at building toolchains. Toolchains are an essential component
45 in a software development project. It will compile, assemble and link the code
46 that is being developed. Some pieces of the toolchain will eventually end up
47 in the resulting binary/ies: static libraries are but an example.
49 So, a toolchain is a very sensitive piece of software, as any bug in one of the
50 components, or a poorly configured component, can lead to execution problems,
51 ranging from poor performance, to applications ending unexpectedly, to
52 mis-behaving software (which more than often is hard to detect), to hardware
53 damage, or even to human risks (which is more than regrettable).
55 Toolchains are made of different piece of software, each being quite complex
56 and requiring specially crafted options to build and work seamlessly. This
57 is usually not that easy, even in the not-so-trivial case of native toolchains.
58 The work reaches a higher degree of complexity when it comes to cross-
59 compilation, where it can become quite a nightmare...
61 Some cross-toolchains exist on the internet, and can be used for general
62 development, but they have a number of limitations:
63 - they can be general purpose, in that they are configured for the majority:
64 no optimisation for your specific target,
65 - they can be prepared for a specific target and thus are not easy to use,
66 nor optimised for, or even supporting your target,
67 - they often are using aging components (compiler, C library, etc...) not
68 supporting special features of your shiny new processor;
69 On the other side, these toolchain offer some advantages:
70 - they are ready to use and quite easy to install and setup,
71 - they are proven if used by a wide community.
73 But once you want to get all the juice out of your specific hardware, you will
74 want to build your own toolchain. This is where crosstool-NG comes into play.
76 There are also a number of tools that build toolchains for specific needs,
77 which are not really scalable. Examples are:
78 - buildroot (buildroot.uclibc.org) whose main purpose is to build root file
79 systems, hence the name. But once you have your toolchain with buildroot,
80 part of it is installed in the root-to-be, so if you want to build a whole
81 new root, you either have to save the existing one as a template and
82 restore it later, or restart again from scratch. This is not convenient,
83 - ptxdist (www.pengutronix.de/software/ptxdist), whose purpose is very
85 - other projects (openembedded.org for example), which are again used to
86 build root file systems.
88 crosstool-NG is really targeted at building toolchains, and only toolchains.
89 It is then up to you to use it the way you want.
96 crosstool was first 'conceived' by Dan Kegel, who offered it to the community
97 as a set of scripts, a repository of patches, and some pre-configured, general
98 purpose setup files to be used to configure crosstool. This is available at
99 http://www.kegel.com/crosstool, and the subversion repository is hosted on
100 google at http://code.google.com/p/crosstool/.
102 I once managed to add support for uClibc-based toolchains, but it did not make
103 into mainline, mostly because I didn't have time to port the patch forward to
104 the new versions, due in part to the big effort it was taking.
106 So I decided to clean up crosstool in the state it was, re-order the things
107 in place, add appropriate support for what I needed, that is uClibc support
108 and a menu-driven configuration, named the new implementation crosstool-NG,
109 (standing for crosstool Next Generation, as many other comunity projects do,
110 and as a wink at the TV series "Star Trek: The Next Generation" ;-) ) and
111 made it available to the community, in case it was of interest to any one.
113 ___________________________
115 Installing crosstool-NG /
116 ________________________/
118 There are two ways you can use crosstool-NG:
119 - build and install it, then get rid of the sources like you'd do for most
121 - or only build it and run from the source directory.
123 The former should be used if you got crosstool-NG from a packaged tarball, see
124 "Install method", below, while the latter is most useful for developpers that
125 checked the code out from SVN, and want to submit patches, see "The Hacker's
131 If you go for the install, then you just follow the classical, but yet easy
133 ./configure --prefix=/some/place
136 export PATH="${PATH}:/some/place/bin"
138 You can then get rid of crosstool-NG source. Next create a directory to serve
139 as a working place, cd in there and run:
142 See below for complete usage.
147 If you go the hacker's way, then the usage is a bit different, although very
152 Now, *do not* remove crosstool-NG sources. They are needed to run crosstool-NG!
153 Stay in the directory holding the sources, and run:
156 See below for complete usage.
158 Now, provided you checked-out the code, you can send me your interesting changes
162 and mailing me the result! :-P
164 Preparing for packaging |
165 ------------------------+
167 If you plan on packaging crosstool-NG, you surely don't want to install it
168 in your root file system. The install procedure of crosstool-NG honors the
171 ./configure --prefix=/usr
173 make DESTDIR=/packaging/place install
178 crosstool-NG comes with a shell script fragment that defines bash-compatible
179 completion. That shell fragment is currently not installed automatically, but
182 To install the shell script fragment, you have two options:
183 - install system-wide, most probably by copying ct-ng.comp into
184 /etc/bash_completion.d/
185 - install for a single user, by copying ct-ng.comp into ${HOME}/ and
186 sourcing this file from your ${HOME}/.bashrc
191 Some people contibuted code that couldn't get merged for various reasons. This
192 code is available as patches in the contrib/ sub-directory. These patches are
193 to be applied to the source of crosstool-NG, prior to installing.
195 An easy way to use contributed code is to pass the --with-contrib= option to
196 ./configure. The possible values depend upon which contributions are packaged
197 with your version, but you can get with it with passing one of those two
200 will list all available contributions
203 will select all avalaible contributions
205 There is no guarantee that a particuliar contribution applies to the current
206 version of crosstool-ng, or that it will work at all. Use contributions at
209 ____________________________
211 Configuring crosstool-NG /
212 _________________________/
214 crosstool-NG is configured with a configurator presenting a menu-stuctured set
215 of options. These options let you specify the way you want your toolchain
216 built, where you want it installed, what architecture and specific processor it
217 will support, the version of the components you want to use, etc... The
218 value for those options are then stored in a configuration file.
220 The configurator works the same way you configure your Linux kernel. It is
221 assumed you now how to handle this.
223 To enter the menu, type:
226 Almost every config item has a help entry. Read them carefully.
228 String and number options can refer to environment variables. In such a case,
229 you must use the shell syntax: ${VAR}. You shall neither single- nor double-
230 quote the string/number options.
232 There are three environment variables that are computed by crosstool-NG, and
236 It represents the target tuple you are building for. You can use it for
237 example in the installation/prefix directory, such as:
238 /opt/x-tools/${CT_TARGET}
241 The top directory where crosstool-NG is running. You shouldn't need it in
242 most cases. There is one case where you may need it: if you have local
243 patches and you store them in your running directory, you can refer to them
244 by using CT_TOP_DIR, such as:
245 ${CT_TOP_DIR}/patches.myproject
248 The version of crosstool-NG you are using. Not much use for you, but it's
249 there if you need it.
251 Interesting config options |
252 ---------------------------+
254 CT_LOCAL_TARBALLS_DIR:
255 If you already have some tarballs in a direcotry, enter it here. That will
256 speed up the retrieving phase, where crosstool-NG would otherwise download
260 This is where the toolchain will be installed in (and for now, where it
261 will run from). Common use is to add the target tuple in the directory
262 path, such as (see above):
263 /opt/x-tools/${CT_TARGET}
266 An identifier for your toolchain, will take place in the vendor part of the
267 target tuple. It shall *not* contain spaces or dashes. Usually, keep it
268 to a one-word string, or use underscores to separate words if you need.
269 Avoid dots, commas, and special characters.
272 An alias for the toolchian. It will be used as a prefix to the toolchain
273 tools. For example, you will have ${CT_TARGET_ALIAS}-gcc
275 Also, if you think you don't see enough versions, you can try to enable one of
279 Show obsolete versions or tools. Most of the time, you don't want to base
280 your toolchain on too old a version (of gcc, for example). But at times, it
281 can come handy to use such an old version for regression tests. Those old
282 versions are hidden behind CT_OBSOLETE. Those versions (or features) are so
283 marked because maintaining support for those in crosstool-NG would be too
284 costly, time-wise, and time is dear.
287 Show experimental versions or tools. Again, you might not want to base your
288 toolchain on too recent tools (eg. gcc) for production. But if you need a
289 feature present only in a recent version, or a new tool, you can find them
290 hidden behind CT_EXPERIMENTAL. Those versions (or features) did not (yet)
291 receive thorough testing in crosstool-NG, and/or are not mature enough to
294 Re-building an existing toolchain |
295 ----------------------------------+
297 If you have an existing toolchain, you can re-use the options used to build it
298 to create a new toolchain. That needs a very little bit of effort on your side
299 but is quite easy. The options to build a toolchain are saved with the
300 toolchain, and you can retrieve this configuration by running:
303 This will dump the configuration to stdout, so to rebuild a toolchain with this
304 configuration, the following is all you need to do:
305 ${CT_TARGET}-config >.config
308 Then, you can review and change the configuration by running:
311 ________________________
313 Running crosstool-NG /
314 _____________________/
316 To build the toolchain, simply type:
319 This will use the above configuration to retrieve, extract and patch the
320 components, build, install and eventually test your newly built toolchain.
322 You are then free to add the toolchain /bin directory in your PATH to use
325 In any case, you can get some terse help. Just type:
330 Stopping and restarting a build |
331 --------------------------------+
333 If you want to stop the build after a step you are debugging, you can pass the
334 variable STOP to make:
337 Conversely, if you want to restart a build at a specific step you are
338 debugging, you can pass the RESTART variable to make:
339 ct-ng RESTART=some_step
341 Alternatively, you can call make with the name of a step to just do that step:
344 ct-ng RESTART=libc_headers STOP=libc_headers
346 The shortcuts +step_name and step_name+ allow to respectively stop or restart
348 ct-ng +libc_headers and: ct-ng libc_headers+
350 ct-ng STOP=libc_headers and: ct-ng RESTART=libc_headers
352 To obtain the list of acceptable steps, please call:
355 Note that in order to restart a build, you'll have to say 'Y' to the config
356 option CT_DEBUG_CT_SAVE_STEPS, and that the previous build effectively went
359 Building all toolchains at once |
360 --------------------------------+
362 You can build all samples; simply call:
365 Overriding the number of // jobs |
366 ---------------------------------+
368 If you want to override the number of jobs to run in // (the -j option to
369 make), you can either re-enter the menuconfig, or simply add it on the command
373 which tells crosstool-NG to override the number of // jobs to 4.
375 You can see the actions that support overriding the number of // jobs in
376 the help menu. Those are the ones with [.#] after them (eg. build[.#] or
377 build-all[.#], and so on...).
382 The crosstool-NG script 'ct-ng' is a Makefile-script. It does *not* execute
383 in parallel (there is not much to gain). When speaking of // jobs, we are
384 refering to the number of // jobs when making the *components*. That is, we
385 speak of the number of // jobs used to build gcc, glibc, and so on...
390 Starting with gcc-4.3 come two new dependencies: GMP and MPFR. With gcc-4.4,
391 come three new ones: GMP, PPL and CLooG/ppl. These are libraries that enable
392 advanced features to gcc. Additionally, some of the libraries can be used by
393 binutils and gdb. Unfortunately, not all systems on which crosstool-NG runs
394 have all of those libraries. And for those that do, the versions of those
395 libraries may be older than the version required by gcc.
397 This is why crosstool-NG builds its own set of libraries as part of the
400 The libraries are built as shared libraries, because building them as static
401 libraries has some short-comings. This poses no problem at build time, as
402 crosstool-NG correctly points gcc (and binutils and gdb) to the correct
403 place where our own version of the libraries are installed. But it poses
404 a problem when gcc et al. are run: the place where the libraries are is most
405 probably not known to the host dynamic linker. Still worse, if the host system
406 has its own versions, then ld.so would load the wrong library!
408 So we have to force the dynamic linker to load the correct version. We do this
409 by using the LD_LIBRARY_PATH variable, that informs the dynamic linker where
410 to look for shared libraries prior to searching its standard places. But we
411 can't impose that burden on all the system (because it'd be a nightmare to
412 configure, and because two toolchains on the same system may use different
413 versions of the libraries); so we have to do it on a per-toolchain basis.
415 So we rename all binaries of the toolchain (by adding a dot '.' as their first
416 character), and add a small program, the so-called "tools wrapper", that
417 correctly sets LD_LIBRARY_PATH prior to running the real tool.
419 First, the wrapper was written as a POSIX-compliant shell script. That shell
420 script is very simple, if not trivial, and works great. The only drawback is
421 that it does not work on host systems that lack a shell, for example the
422 MingW32 environment. To solve the issue, the wrapper has been re-written in C,
423 and compiled at build time. This C wrapper is much more complex than the shell
424 script, and although it sems to be working, it's been only lightly tested.
425 Some of the expected short-comings with this C wrapper are;
426 - multi-byte file names may not be handled correctly
427 - it's really big for what it does
429 So, the default wrapper installed with your toolchain is the shell script.
430 If you know that your system is missing a shell, then you shall use the C
431 wrapper (and report back whether it works, or does not work, for you).
434 _______________________
436 Using the toolchain /
437 ____________________/
439 Using the toolchain is as simple as adding the toolchain's bin directory in
441 export PATH="${PATH}:/your/toolchain/path/bin"
443 and then using the target tuple to tell the build systems to use your
445 ./configure --target=your-target-tuple
447 make CC=your-target-tuple-gcc
449 make CROSS_COMPILE=your-target-tuple-
452 It is strongly advised not to use the toolchain sys-root directory as an
453 install directory for your programs/packages. If you do so, you will not be
454 able to use your toolchain for another project. It is even strongly advised
455 that your toolchain is chmod-ed to read-only once successfully build, so that
456 you don't go polluting your toolchain with your programs/packages' files.
458 Thus, when you build a program/package, install it in a separate directory,
459 eg. /your/root. This directory is the /image/ of what would be in the root file
460 system of your target, and will contain all that your programs/packages have
463 The 'populate' script |
464 ----------------------+
466 When your root directory is ready, it is still missing some important bits: the
467 toolchain's libraries. To populate your root directory with those libs, just
469 your-target-tuple-populate -s /your/root -d /your/root-populated
471 This will copy /your/root into /your/root-populated, and put the needed and only
472 the needed libraries there. Thus you don't polute /your/root with any cruft that
473 would no longer be needed should you have to remove stuff. /your/root always
474 contains only those things you install in it.
476 You can then use /your/root-populated to build up your file system image, a
477 tarball, or to NFS-mount it from your target, or whatever you need.
479 The populate script accepts the following options:
482 Use 'src_dir' as the un-populated root directory.
485 Put the populated root directory in 'dst_dir'.
488 Always add specified libraries.
491 Always add libraries listed in 'file'.
494 Remove 'dst_dir' if it previously existed; continue even if any library
495 specified with -l or -L is missing.
498 Be verbose, and tell what's going on (you can see exactly where libs are
504 See 'your-target-tuple-populate -h' for more information on the options.
506 Here is how populate works:
508 1) performs some sanity checks:
509 - src_dir and dst_dir are specified
511 - unless forced, dst_dir does not exist
514 2) copy src_dir to dst_dir
516 3) add forced libraries to dst_dir
517 - build the list from -l and -L options
518 - get forced libraries from the sysroot (see below for heuristics)
519 - abort on the first missing library, unless -f is specified
521 4) add all missing libraries to dst_dir
522 - scan dst_dir for every ELF files that are 'executable' or
524 - list the "NEEDED Shared library" fields
525 - check if the library is already in dst_dir/lib or dst_dir/usr/lib
526 - if not, get the library from the sysroot
527 - if it's in sysroot/lib, copy it to dst_dir/lib
528 - if it's in sysroot/usr/lib, copy it to dst_dir/usr/lib
529 - in both cases, use the SONAME of the library to create the file
531 - if it was not found in the sysroot, this is an error.
538 There are four kinds of toolchains you could encounter.
540 First off, you must understand the following: when it comes to compilers there
541 are up to four machines involved:
542 1) the machine configuring the toolchain components: the config machine
543 2) the machine building the toolchain components: the build machine
544 3) the machine running the toolchain: the host machine
545 4) the machine the toolchain is generating code for: the target machine
547 We can most of the time assume that the config machine and the build machine
548 are the same. Most of the time, this will be true. The only time it isn't
549 is if you're using distributed compilation (such as distcc). Let's forget
550 this for the sake of simplicity.
552 So we're left with three machines:
557 Any toolchain will involve those three machines. You can be as pretty sure of
558 this as "2 and 2 are 4". Here is how they come into play:
560 1) build == host == target
561 This is a plain native toolchain, targetting the exact same machine as the
562 one it is built on, and running again on this exact same machine. You have
563 to build such a toolchain when you want to use an updated component, such
564 as a newer gcc for example.
565 crosstool-NG calls it "native".
567 2) build == host != target
568 This is a classic cross-toolchain, which is expected to be run on the same
569 machine it is compiled on, and generate code to run on a second machine,
571 crosstool-NG calls it "cross".
573 3) build != host == target
574 Such a toolchain is also a native toolchain, as it targets the same machine
575 as it runs on. But it is build on another machine. You want such a
576 toolchain when porting to a new architecture, or if the build machine is
577 much faster than the host machine.
578 crosstool-NG calls it "cross-native".
580 4) build != host != target
581 This one is called a canadian-toolchain (*), and is tricky. The three
582 machines in play are different. You might want such a toolchain if you
583 have a fast build machine, but the users will use it on another machine,
584 and will produce code to run on a third machine.
585 crosstool-NG calls it "canadian".
587 crosstool-NG can build all these kinds of toolchains (or is aiming at it,
590 (*) The term Canadian Cross came about because at the time that these issues
591 were all being hashed out, Canada had three national political parties.
592 http://en.wikipedia.org/wiki/Cross_compiler
594 Seemingly-native toolchains |
595 ----------------------------+
597 Seemingly-native toolchains are toolchains that target the same architecture
598 as the one it is built on, and on which it will run, but the machine tuple
599 may be different (eg i686 vs. i386, or x86_64-unknown-linux-gnu vs.
600 x86_64-pc-linux-gnu). This also applies if the target architecture is of the
601 same kind (eg. x86 vs. x86_64, or ppc vs. ppc64).
603 Such toolchain is tricky to build, as the configure scripts may incorrectly
604 assume that files (headers and libs) from the build (or host) machine can be
605 used by the cross-compiler it is going to build. The problem seems to arise
606 only with glibc (and eglibc?) starting with version 2.7.
614 Internally, crosstool-NG is script-based. To ease usage, the frontend is
620 The entry point to crosstool-NG is the Makefile script "ct-ng". Calling this
621 script with an action will act exactly as if the Makefile was in the current
622 working directory and make was called with the action as rule. Thus:
625 is equivalent to having the Makefile in CWD, and calling:
628 Having ct-ng as it is avoids copying the Makefile everywhere, and acts as a
631 ct-ng loads sub- Makefiles from the library directory $(CT_LIB_DIR), as set up
632 at configuration time with ./configure.
634 ct-ng also searches for config files, sub-tools, samples, scripts and patches in
635 that library directory.
637 Because of a stupid make behavior/bug I was unable to track down, implicit make
638 rules are disabled: installing with --local would triger those rules, and mconf
644 The kconfig language is a hacked version, vampirised from the Linux kernel
645 (http://www.kernel.org/), and (heavily) adapted to my needs.
647 The list of the most notable changes (at least the ones I remember) follows:
648 - the CONFIG_ prefix has been replaced with CT_
649 - a leading | in prompts is skipped, and subsequent leading spaces are not
651 - otherwise leading spaces are silently trimmed
653 The kconfig parsers (conf and mconf) are not installed pre-built, but as
654 source files. Thus you can have the directory where crosstool-NG is installed,
655 exported (via NFS or whatever) and have clients with different architectures
656 use the same crosstool-NG installation, and most notably, the same set of
659 Architecture-specific |
660 ----------------------+
662 Note: this chapter is not really well written, and might thus be a little bit
663 complex to understand. To get a better grasp of what an architecture is, the
664 reader is kindly encouraged to look at the "arch/" sub-directory, and to the
665 existing architectures to see how things are laid out.
667 An architecture is defined by:
669 - a human-readable name, in lower case letters, with numbers as appropriate.
670 The underscore is allowed; space and special characters are not.
672 - a file in "config/arch/", named after the architecture's name, and suffixed
674 Eg.: config/arch/arm.in
675 - a file in "scripts/build/arch/", named after the architecture's name, and
677 Eg.: scripts/build/arch/arm.sh
679 The architecture's ".in" file API:
680 > the config option "ARCH_%arch%" (where %arch% is to be replaced with the
681 actual architecture name).
682 That config option must have *neither* a type, *nor* a prompt! Also, it can
683 *not* depend on any other config option (EXPERIMENTAL is managed as above).
687 defines a (terse) help entry for this architecture:
691 The ARM architecture.
693 selects adequate associated config options.
694 Note: 64-bit architectures *shall* select ARCH_64
697 select ARCH_SUPPORTS_BOTH_ENDIAN
698 select ARCH_DEFAULT_LE
700 The ARM architecture.
705 The x86_64 architecture.
707 > other target-specific options, at your discretion. Note however that to
708 avoid name-clashing, such options shall be prefixed with "ARCH_%arch%",
709 where %arch% is again replaced by the actual architecture name.
710 (Note: due to historical reasons, and lack of time to clean up the code,
711 I may have left some config options that do not completely conform to
712 this, as the architecture name was written all upper case. However, the
713 prefix is unique among architectures, and does not cause harm).
715 The architecture's ".sh" file API:
716 > the function "CT_DoArchTupleValues"
719 - all variables from the ".config" file,
720 - the two variables "target_endian_eb" and "target_endian_el" which are
721 the endianness suffixes
722 + return value: 0 upon success, !0 upon failure
725 - the environment variable CT_TARGET_ARCH
727 the architecture part of the target tuple.
728 Eg.: "armeb" for big endian ARM
732 - the environment variable CT_TARGET_SYS
734 the sytem part of the target tuple.
735 Eg.: "gnu" for glibc on most architectures
736 "gnueabi" for glibc on an ARM EABI
738 - for glibc-based toolchain: "gnu"
739 - for uClibc-based toolchain: "uclibc"
742 - the environment variable CT_KERNEL_ARCH
744 the architecture name as understandable by the Linux kernel build
746 Eg.: "arm" for an ARM
747 "powerpc" for a PowerPC
753 - the environment variables to configure the cross-gcc (defaults)
754 - CT_ARCH_WITH_ARCH : the gcc ./configure switch to select architecture level ( "--with-arch=${CT_ARCH_ARCH}" )
755 - CT_ARCH_WITH_ABI : the gcc ./configure switch to select ABI level ( "--with-abi=${CT_ARCH_ABI}" )
756 - CT_ARCH_WITH_CPU : the gcc ./configure switch to select CPU instruction set ( "--with-cpu=${CT_ARCH_CPU}" )
757 - CT_ARCH_WITH_TUNE : the gcc ./configure switch to select scheduling ( "--with-tune=${CT_ARCH_TUNE}" )
758 - CT_ARCH_WITH_FPU : the gcc ./configure switch to select FPU type ( "--with-fpu=${CT_ARCH_FPU}" )
759 - CT_ARCH_WITH_FLOAT : the gcc ./configure switch to select floating point arithmetics ( "--with-float=soft" or /empty/ )
762 - the environment variables to pass to the cross-gcc to build target binaries (defaults)
763 - CT_ARCH_ARCH_CFLAG : the gcc switch to select architecture level ( "-march=${CT_ARCH_ARCH}" )
764 - CT_ARCH_ABI_CFLAG : the gcc switch to select ABI level ( "-mabi=${CT_ARCH_ABI}" )
765 - CT_ARCH_CPU_CFLAG : the gcc switch to select CPU instruction set ( "-mcpu=${CT_ARCH_CPU}" )
766 - CT_ARCH_TUNE_CFLAG : the gcc switch to select scheduling ( "-mtune=${CT_ARCH_TUNE}" )
767 - CT_ARCH_FPU_CFLAG : the gcc switch to select FPU type ( "-mfpu=${CT_ARCH_FPU}" )
768 - CT_ARCH_FLOAT_CFLAG : the gcc switch to choose floating point arithmetics ( "-msoft-float" or /empty/ )
769 - CT_ARCH_ENDIAN_CFLAG : the gcc switch to choose big or little endian ( "-mbig-endian" or "-mlittle-endian" )
774 - the environement variables to configure the core and final compiler, specific to this architecture:
775 - CT_ARCH_CC_CORE_EXTRA_CONFIG : additional, architecture specific core gcc ./configure flags
776 - CT_ARCH_CC_EXTRA_CONFIG : additional, architecture specific final gcc ./configure flags
781 - the architecture-specific CFLAGS and LDFLAGS:
782 - CT_ARCH_TARGET_CLFAGS
783 - CT_ARCH_TARGET_LDFLAGS
787 You can have a look at "config/arch/arm.in" and "scripts/build/arch/arm.sh" for
788 a quite complete example of what an actual architecture description looks like.
793 A kernel is defined by:
795 - a human-readable name, in lower case letters, with numbers as appropriate.
796 The underscore is allowed; space and special characters are not (although
797 they are internally replaced with underscores.
798 Eg.: linux, bare-metal
799 - a file in "config/kernel/", named after the kernel name, and suffixed with
801 Eg.: config/kernel/linux.in, config/kernel/bare-metal.in
802 - a file in "scripts/build/kernel/", named after the kernel name, and suffixed
804 Eg.: scripts/build/kernel/linux.sh, scripts/build/kernel/bare-metal.sh
806 The kernel's ".in" file must contain:
807 > an optional lines containing exactly "# EXPERIMENTAL", starting on the
808 first column, and without any following space or other character.
809 If this line is present, then this kernel is considered EXPERIMENTAL,
810 and correct dependency on EXPERIMENTAL will be set.
812 > the config option "KERNEL_%kernel_name%" (where %kernel_name% is to be
813 replaced with the actual kernel name, with all special characters and
814 spaces replaced by underscores).
815 That config option must have *neither* a type, *nor* a prompt! Also, it can
816 *not* depends on EXPERIMENTAL.
817 Eg.: KERNEL_linux, KERNEL_bare_metal
819 defines a (terse) help entry for this kernel.
821 config KERNEL_bare_metal
823 Build a compiler for use without any kernel.
825 selects adequate associated config options.
827 config KERNEL_bare_metal
830 Build a compiler for use without any kernel.
832 > other kernel specific options, at your discretion. Note however that, to
833 avoid name-clashing, such options should be prefixed with
834 "KERNEL_%kernel_name%", where %kernel_name% is again tp be replaced with
835 the actual kernel name.
836 (Note: due to historical reasons, and lack of time to clean up the code,
837 I may have left some config options that do not completely conform to
838 this, as the kernel name was written all upper case. However, the prefix
839 is unique among kernels, and does not cause harm).
841 The kernel's ".sh" file API:
842 > is a bash script fragment
844 > defines the function CT_DoKernelTupleValues
845 + see the architecture's CT_DoArchTupleValues, except for:
846 + set the environment variable CT_TARGET_KERNEL, the kernel part of the
848 + return value: ignored
850 > defines the function "do_kernel_get":
853 - all variables from the ".config" file.
854 + return value: 0 for success, !0 for failure.
855 + behavior: download the kernel's sources, and store the tarball into
856 "${CT_TARBALLS_DIR}". To this end, a functions is available, that
857 abstracts downloading tarballs:
858 - CT_DoGet <tarball_base_name> <URL1 [URL...]>
859 Eg.: CT_DoGet linux-2.6.26.5 ftp://ftp.kernel.org/pub/linux/kernel/v2.6
860 Note: retrieving sources from svn, cvs, git and the likes is not supported
861 by CT_DoGet. You'll have to do this by hand, as it is done for eglibc in
862 "scripts/build/libc/eglibc.sh"
864 > defines the function "do_kernel_extract":
867 - all variables from the ".config" file,
868 + return value: 0 for success, !0 for failure.
869 + behavior: extract the kernel's tarball into "${CT_SRC_DIR}", and apply
870 required patches. To this end, a function is available, that abstracts
872 - CT_ExtractAndPatch <tarball_base_name>
873 Eg.: CT_ExtractAndPatch linux-2.6.26.5
875 > defines the function "do_kernel_headers":
878 - all variables from the ".config" file,
879 + return value: 0 for success, !0 for failure.
880 + behavior: install the kernel headers (if any) in "${CT_SYSROOT_DIR}/usr/include"
882 > defines any kernel-specific helper functions
883 These functions, if any, must be prefixed with "do_kernel_%CT_KERNEL%_",
884 where '%CT_KERNEL%' is to be replaced with the actual kernel name, to avoid
887 You can have a look at "config/kernel/linux.in" and "scripts/build/kernel/linux.sh"
888 as an example of what a complex kernel description looks like.
890 Adding a new version of a component |
891 ------------------------------------+
893 When a new component, such as the Linux kernel, gcc or any other is released,
894 adding the new version to crosstool-NG is quite easy. There is a script that
895 will do all that for you:
896 scripts/addToolVersion.sh
898 Run it with no option to get some help.
903 To Be Written later...