Linux kernel module makefile

Kernel modules need to be compiled a bit differently from regular userspace apps. Former kernel versions required us to care much about these settings, which are usually stored in Makefiles. Although hierarchically organized, many redundant settings accumulated in sublevel Makefiles and made them large and rather difficult to maintain. Fortunately, there is a new way of doing these things, called kbuild, and the build process for external loadable modules is now fully integrated into the standard kernel build mechanism. To learn more on how to compile modules which are not part of the official kernel (such as all the examples you’ll find in this guide), see file linux/Documentation/kbuild/modules.txt .

So, let’s look at a simple Makefile for compiling a module named hello-1.c :

Example 2-2. Makefile for a basic kernel module

obj-m += hello-1.o all: make -C /lib/modules/$(shell uname -r)/build M=$(PWD) modules clean: make -C /lib/modules/$(shell uname -r)/build M=$(PWD) clean

From a technical point of view just the first line is really necessary, the «all» and «clean» targets were added for pure convenience.

Now you can compile the module by issuing the command make . You should obtain an output which resembles the following:

/lkmpg-examples/02-HelloWorld# make make -C /lib/modules/2.6.11/build M=/root/lkmpg-examples/02-HelloWorld modules make[1]: Entering directory `/usr/src/linux-2.6.11′ CC [M] /root/lkmpg-examples/02-HelloWorld/hello-1.o Building modules, stage 2. MODPOST CC /root/lkmpg-examples/02-HelloWorld/hello-1.mod.o LD [M] /root/lkmpg-examples/02-HelloWorld/hello-1.ko make[1]: Leaving directory `/usr/src/linux-2.6.11′ hostname:

Note that kernel 2.6 introduces a new file naming convention: kernel modules now have a .ko extension (in place of the old .o extension) which easily distinguishes them from conventional object files. The reason for this is that they contain an additional .modinfo section that where additional information about the module is kept. We’ll soon see what this information is good for.

Use modinfo hello-*.ko to see what kind of information it is.

/lkmpg-examples/02-HelloWorld# modinfo hello-1.ko filename: hello-1.ko vermagic: 2.6.11 preempt PENTIUMII 4KSTACKS gcc-3.3 depends:

Nothing spectacular, so far. That changes once we’re using modinfo on one of our the later examples, hello-5.ko .

/lkmpg-examples/02-HelloWorld# modinfo hello-5.ko filename: hello-5.ko license: GPL author: Peter Jay Salzman vermagic: 2.6.11 preempt PENTIUMII 4KSTACKS gcc-3.3 depends: parm: myintArray:An array of integers (array of int) parm: mystring:A character string (charp) parm: mylong:A long integer (long) parm: myint:An integer (int) parm: myshort:A short integer (short) hostname:

Lot’s of useful information to see here. An author string for bugreports, license information, even a short description of the parameters it accepts.

Additional details about Makefiles for kernel modules are available in linux/Documentation/kbuild/makefiles.txt . Be sure to read this and the related files before starting to hack Makefiles. It’ll probably save you lots of work.

Now it is time to insert your freshly-compiled module it into the kernel with insmod ./hello-1.ko (ignore anything you see about tainted kernels; we’ll cover that shortly).

All modules loaded into the kernel are listed in /proc/modules . Go ahead and cat that file to see that your module is really a part of the kernel. Congratulations, you are now the author of Linux kernel code! When the novelty wears off, remove your module from the kernel by using rmmod hello-1 . Take a look at /var/log/messages just to see that it got logged to your system logfile.

Here’s another exercise for the reader. See that comment above the return statement in init_module() ? Change the return value to something negative, recompile and load the module again. What happens?

Источник

Пишем простой модуль ядра Linux

Захват Золотого Кольца-0

Linux предоставляет мощный и обширный API для приложений, но иногда его недостаточно. Для взаимодействия с оборудованием или осуществления операций с доступом к привилегированной информации в системе нужен драйвер ядра.

Модуль ядра Linux — это скомпилированный двоичный код, который вставляется непосредственно в ядро Linux, работая в кольце 0, внутреннем и наименее защищённом кольце выполнения команд в процессоре x86–64. Здесь код исполняется совершенно без всяких проверок, но зато на невероятной скорости и с доступом к любым ресурсам системы.

Не для простых смертных

Написание модуля ядра Linux — занятие не для слабонервных. Изменяя ядро, вы рискуете потерять данные. В коде ядра нет стандартной защиты, как в обычных приложениях Linux. Если сделать ошибку, то повесите всю систему.

Ситуация ухудшается тем, что проблема необязательно проявляется сразу. Если модуль вешает систему сразу после загрузки, то это наилучший сценарий сбоя. Чем больше там кода, тем выше риск бесконечных циклов и утечек памяти. Если вы неосторожны, то проблемы станут постепенно нарастать по мере работы машины. В конце концов важные структуры данных и даже буфера могут быть перезаписаны.

Можно в основном забыть традиционные парадигмы разработки приложений. Кроме загрузки и выгрузки модуля, вы будете писать код, который реагирует на системные события, а не работает по последовательному шаблону. При работе с ядром вы пишете API, а не сами приложения.

У вас также нет доступа к стандартной библиотеке. Хотя ядро предоставляет некоторые функции вроде printk (которая служит заменой printf ) и kmalloc (работает похоже на malloc ), в основном вы остаётесь наедине с железом. Вдобавок, после выгрузки модуля следует полностью почистить за собой. Здесь нет сборки мусора.

Необходимые компоненты

Прежде чем начать, следует убедиться в наличии всех необходимых инструментов для работы. Самое главное, нужна машина под Linux. Знаю, это неожиданно! Хотя подойдёт любой дистрибутив Linux, в этом примере я использую Ubuntu 16.04 LTS, так что в случае использования других дистрибутивов может понадобиться слегка изменить команды установки.

Во-вторых, нужна или отдельная физическая машина, или виртуальная машина. Лично я предпочитаю работать на виртуальной машине, но выбирайте сами. Не советую использовать свою основную машину из-за потери данных, когда сделаете ошибку. Я говорю «когда», а не «если», потому что вы обязательно подвесите машину хотя бы несколько раз в процессе. Ваши последние изменения в коде могут ещё находиться в буфере записи в момент паники ядра, так что могут повредиться и ваши исходники. Тестирование в виртуальной машине устраняет эти риски.

И наконец, нужно хотя бы немного знать C. Рабочая среда C++ слишком велика для ядра, так что необходимо писать на чистом голом C. Для взаимодействия с оборудованием не помешает и некоторое знание ассемблера.

Установка среды разработки

На Ubuntu нужно запустить:

Устанавливаем самые важные инструменты разработки и заголовки ядра, необходимые для данного примера.

Примеры ниже предполагают, что вы работаете из-под обычного пользователя, а не рута, но что у вас есть привилегии sudo. Sudo необходима для загрузки модулей ядра, но мы хотим работать по возможности за пределами рута.

Начинаем

Приступим к написанию кода. Подготовим нашу среду:

Запустите любимый редактор (в моём случае это vim) и создайте файл lkm_example.c следующего содержания:

Мы сконструировали самый простой возможный модуль, рассмотрим подробнее самые важные его части:

В include перечислены файлы заголовков, необходимые для разработки ядра Linux.
В MODULE_LICENSE можно установить разные значения, в зависимости от лицензии модуля. Для просмотра полного списка запустите:

Мы устанавливаем init (загрузка) и exit (выгрузка) как статические функции, которые возвращают целые числа.

Обратите внимание на использование printk вместо printf . Также параметры printk отличаются от printf . Например, флаг KERN_INFO для объявления приоритета журналирования для конкретной строки указывается без запятой. Ядро разбирается с этими вещами внутри функции printk для экономии памяти стека.

В конце файла можно вызвать module_init и module_exit и указать функции загрузки и выгрузки. Это даёт возможность произвольного именования функций.

Впрочем, пока мы не можем скомпилировать этот файл. Нужен Makefile. Такого базового примера пока достаточно. Обратите внимание, что make очень привередлив к пробелам и табам, так что убедитесь, что используете табы вместо пробелов где положено.

Если мы запускаем make , он должен успешно скомпилировать наш модуль. Результатом станет файл lkm_example.ko . Если выскакивают какие-то ошибки, проверьте, что кавычки в исходном коде установлены корректно, а не случайно в кодировке UTF-8.

Теперь можно внедрить модуль и проверить его. Для этого запускаем:

Если всё нормально, то вы ничего не увидите. Функция printk обеспечивает выдачу не в консоль, а в журнал ядра. Для просмотра нужно запустить:

Вы должны увидеть строку “Hello, World!” с меткой времени в начале. Это значит, что наш модуль ядра загрузился и успешно сделал запись в журнал ядра. Мы можем также проверить, что модуль ещё в памяти:

Для удаления модуля запускаем:

Если вы снова запустите dmesg, то увидите в журнале запись “Goodbye, World!”. Можно снова запустить lsmod и убедиться, что модуль выгрузился.

Как видите, эта процедура тестирования слегка утомительна, но её можно автоматизировать, добавив:

в конце Makefile, а потом запустив:

для тестирования модуля и проверки выдачи в журнал ядра без необходимости запускать отдельные команды.

Теперь у нас есть полностью функциональный, хотя и абсолютно тривиальный модуль ядра!

Немного интереснее

Копнём чуть глубже. Хотя модули ядра способны выполнять все виды задач, взаимодействие с приложениями — один из самых распространённых вариантов использования.

Поскольку приложениям запрещено просматривать память в пространстве ядра, для взаимодействия с ними приходится использовать API. Хотя технически есть несколько способов такого взаимодействия, наиболее привычный — создание файла устройства.

Вероятно, раньше вы уже имели дело с файлами устройств. Команды с упоминанием /dev/zero , /dev/null и тому подобного взаимодействуют с устройствами “zero” и “null”, которые возвращают ожидаемые значения.

В нашем примере мы возвращаем “Hello, World”. Хотя это не особенно полезная функция для приложений, она всё равно демонстрирует процесс взаимодействия с приложением через файл устройства.

Вот полный листинг:

Тестирование улучшенного примера

Теперь наш пример делает нечто большее, чем просто вывод сообщения при загрузке и выгрузке, так что понадобится менее строгая процедура тестирования. Изменим Makefile только для загрузки модуля, без его выгрузки.

Теперь после запуска make test вы увидите выдачу старшего номера устройства. В нашем примере его автоматически присваивает ядро. Однако этот номер нужен для создания нового устройства.

Возьмите номер, полученный в результате выполнения make test , и используйте его для создания файла устройства, чтобы можно было установить коммуникацию с нашим модулем ядра из пространства пользователя.

(в этом примере замените MAJOR значением, полученным в результате выполнения make test или dmesg )

Параметр c в команде mknod говорит mknod, что нам нужно создать файл символьного устройства.

Теперь мы можем получить содержимое с устройства:

или даже через команду dd :

Вы также можете получить доступ к этому файлу из приложений. Это необязательно должны быть скомпилированные приложения — даже у скриптов Python, Ruby и PHP есть доступ к этим данным.

Когда мы закончили с устройством, удаляем его и выгружаем модуль:

Заключение

Надеюсь, вам понравились наши шалости в пространстве ядра. Хотя показанные примеры примитивны, эти структуры можно использовать для создания собственных модулей, выполняющих очень сложные задачи.

Просто помните, что в пространстве ядра всё под вашу ответственность. Там для вашего кода нет поддержки или второго шанса. Если делаете проект для клиента, заранее запланируйте двойное, если не тройное время на отладку. Код ядра должен быть идеален, насколько это возможно, чтобы гарантировать цельность и надёжность систем, на которых он запускается.

Источник

Linux Kernel MakefilesВ¶

This document describes the Linux kernel Makefiles.

1 OverviewВ¶

The Makefiles have five parts:

The top Makefile reads the .config file, which comes from the kernel configuration process.

The top Makefile is responsible for building two major products: vmlinux (the resident kernel image) and modules (any module files). It builds these goals by recursively descending into the subdirectories of the kernel source tree. The list of subdirectories which are visited depends upon the kernel configuration. The top Makefile textually includes an arch Makefile with the name arch/$(ARCH)/Makefile. The arch Makefile supplies architecture-specific information to the top Makefile.

Each subdirectory has a kbuild Makefile which carries out the commands passed down from above. The kbuild Makefile uses information from the .config file to construct various file lists used by kbuild to build any built-in or modular targets.

scripts/Makefile.* contains all the definitions/rules etc. that are used to build the kernel based on the kbuild makefiles.

2 Who does whatВ¶

People have four different relationships with the kernel Makefiles.

Users are people who build kernels. These people type commands such as «make menuconfig» or «make». They usually do not read or edit any kernel Makefiles (or any other source files).

Normal developers are people who work on features such as device drivers, file systems, and network protocols. These people need to maintain the kbuild Makefiles for the subsystem they are working on. In order to do this effectively, they need some overall knowledge about the kernel Makefiles, plus detailed knowledge about the public interface for kbuild.

Arch developers are people who work on an entire architecture, such as sparc or ia64. Arch developers need to know about the arch Makefile as well as kbuild Makefiles.

Kbuild developers are people who work on the kernel build system itself. These people need to know about all aspects of the kernel Makefiles.

This document is aimed towards normal developers and arch developers.

3 The kbuild filesВ¶

Most Makefiles within the kernel are kbuild Makefiles that use the kbuild infrastructure. This chapter introduces the syntax used in the kbuild makefiles. The preferred name for the kbuild files are ‘Makefile’ but ‘Kbuild’ can be used and if both a ‘Makefile’ and a ‘Kbuild’ file exists, then the ‘Kbuild’ file will be used.

Section 3.1 «Goal definitions» is a quick intro, further chapters provide more details, with real examples.

3.1 Goal definitionsВ¶

Goal definitions are the main part (heart) of the kbuild Makefile. These lines define the files to be built, any special compilation options, and any subdirectories to be entered recursively.

The most simple kbuild makefile contains one line:

This tells kbuild that there is one object in that directory, named foo.o. foo.o will be built from foo.c or foo.S.

If foo.o shall be built as a module, the variable obj-m is used. Therefore the following pattern is often used:

$(CONFIG_FOO) evaluates to either y (for built-in) or m (for module). If CONFIG_FOO is neither y nor m, then the file will not be compiled nor linked.

3.2 Built-in object goals — obj-yВ¶

The kbuild Makefile specifies object files for vmlinux in the $(obj-y) lists. These lists depend on the kernel configuration.

Kbuild compiles all the $(obj-y) files. It then calls «$(AR) rcSTP» to merge these files into one built-in.a file. This is a thin archive without a symbol table. It will be later linked into vmlinux by scripts/link-vmlinux.sh

The order of files in $(obj-y) is significant. Duplicates in the lists are allowed: the first instance will be linked into built-in.a and succeeding instances will be ignored.

Link order is significant, because certain functions ( module_init() / __initcall) will be called during boot in the order they appear. So keep in mind that changing the link order may e.g. change the order in which your SCSI controllers are detected, and thus your disks are renumbered.

3.3 Loadable module goals — obj-mВ¶

$(obj-m) specifies object files which are built as loadable kernel modules.

A module may be built from one source file or several source files. In the case of one source file, the kbuild makefile simply adds the file to $(obj-m).

Note: In this example $(CONFIG_ISDN_PPP_BSDCOMP) evaluates to ‘m’

If a kernel module is built from several source files, you specify that you want to build a module in the same way as above; however, kbuild needs to know which object files you want to build your module from, so you have to tell it by setting a $( -y) variable.

In this example, the module name will be isdn.o. Kbuild will compile the objects listed in $(isdn-y) and then run «$(LD) -r» on the list of these files to generate isdn.o.

Due to kbuild recognizing $( -y) for composite objects, you can use the value of a CONFIG_ symbol to optionally include an object file as part of a composite object.

In this example, xattr.o, xattr_user.o and xattr_trusted.o are only part of the composite object ext2.o if $(CONFIG_EXT2_FS_XATTR) evaluates to ‘y’.

Note: Of course, when you are building objects into the kernel, the syntax above will also work. So, if you have CONFIG_EXT2_FS=y, kbuild will build an ext2.o file for you out of the individual parts and then link this into built-in.a, as you would expect.

3.4 Objects which export symbolsВ¶

3.5 Library file goals — lib-yВ¶

Objects listed with obj-* are used for modules, or combined in a built-in.a for that specific directory. There is also the possibility to list objects that will be included in a library, lib.a. All objects listed with lib-y are combined in a single library for that directory. Objects that are listed in obj-y and additionally listed in lib-y will not be included in the library, since they will be accessible anyway. For consistency, objects listed in lib-m will be included in lib.a.

Note that the same kbuild makefile may list files to be built-in and to be part of a library. Therefore the same directory may contain both a built-in.a and a lib.a file.

This will create a library lib.a based on delay.o. For kbuild to actually recognize that there is a lib.a being built, the directory shall be listed in libs-y.

See also «6.4 List directories to visit when descending».

Use of lib-y is normally restricted to lib/ and arch/*/lib .

3.6 Descending down in directoriesВ¶

A Makefile is only responsible for building objects in its own directory. Files in subdirectories should be taken care of by Makefiles in these subdirs. The build system will automatically invoke make recursively in subdirectories, provided you let it know of them.

To do so, obj-y and obj-m are used. ext2 lives in a separate directory, and the Makefile present in fs/ tells kbuild to descend down using the following assignment.

If CONFIG_EXT2_FS is set to either ‘y’ (built-in) or ‘m’ (modular) the corresponding obj- variable will be set, and kbuild will descend down in the ext2 directory. Kbuild only uses this information to decide that it needs to visit the directory, it is the Makefile in the subdirectory that specifies what is modular and what is built-in.

It is good practice to use a CONFIG_ variable when assigning directory names. This allows kbuild to totally skip the directory if the corresponding CONFIG_ option is neither ‘y’ nor ‘m’.

3.7 Compilation flagsВ¶

These three flags apply only to the kbuild makefile in which they are assigned. They are used for all the normal cc, as and ld invocations happening during a recursive build. Note: Flags with the same behaviour were previously named: EXTRA_CFLAGS, EXTRA_AFLAGS and EXTRA_LDFLAGS. They are still supported but their usage is deprecated.

ccflags-y specifies options for compiling with $(CC).

This variable is necessary because the top Makefile owns the variable $(KBUILD_CFLAGS) and uses it for compilation flags for the entire tree.

asflags-y specifies assembler options.

ldflags-y specifies options for linking with $(LD).

The two flags listed above are similar to ccflags-y and asflags-y. The difference is that the subdir- variants have effect for the kbuild file where they are present and all subdirectories. Options specified using subdir-* are added to the commandline before the options specified using the non-subdir variants.

CFLAGS_$@ and AFLAGS_$@ only apply to commands in current kbuild makefile.

$(CFLAGS_$@) specifies per-file options for $(CC). The $@ part has a literal value which specifies the file that it is for.

These two lines specify compilation flags for aha152x.o and gdth.o.

$(AFLAGS_$@) is a similar feature for source files in assembly languages.

3.9 Dependency trackingВ¶

Kbuild tracks dependencies on the following:

All prerequisite files (both *.c and *.h )
CONFIG_ options used in all prerequisite files
Command-line used to compile target

Thus, if you change an option to $(CC) all affected files will be re-compiled.

3.10 Special RulesВ¶

Special rules are used when the kbuild infrastructure does not provide the required support. A typical example is header files generated during the build process. Another example are the architecture-specific Makefiles which need special rules to prepare boot images etc.

Special rules are written as normal Make rules. Kbuild is not executing in the directory where the Makefile is located, so all special rules shall provide a relative path to prerequisite files and target files.

Two variables are used when defining special rules:

$(src) $(src) is a relative path which points to the directory where the Makefile is located. Always use $(src) when referring to files located in the src tree. $(obj)

$(obj) is a relative path which points to the directory where the target is saved. Always use $(obj) when referring to generated files.

This is a special rule, following the normal syntax required by make.

The target file depends on two prerequisite files. References to the target file are prefixed with $(obj), references to prerequisites are referenced with $(src) (because they are not generated files).

$(kecho) echoing information to user in a rule is often a good practice but when execution «make -s» one does not expect to see any output except for warnings/errors. To support this kbuild defines $(kecho) which will echo out the text following $(kecho) to stdout except if «make -s» is used.

3.11 $(CC) support functionsВ¶

as-option is used to check if $(CC) — when used to compile assembler ( *.S ) files — supports the given option. An optional second option may be specified if the first option is not supported.

In the above example, cflags-y will be assigned the option -Wa$(comma)-isa=$(isa-y) if it is supported by $(CC). The second argument is optional, and if supplied will be used if first argument is not supported.

cc-ldoption is used to check if $(CC) when used to link object files supports the given option. An optional second option may be specified if first option are not supported.

In the above example, vsyscall-flags will be assigned the option -Wl$(comma)—hash-style=sysv if it is supported by $(CC). The second argument is optional, and if supplied will be used if first argument is not supported.

as-instr as-instr checks if the assembler reports a specific instruction and then outputs either option1 or option2 C escapes are supported in the test instruction Note: as-instr-option uses KBUILD_AFLAGS for assembler options cc-option

cc-option is used to check if $(CC) supports a given option, and if not supported to use an optional second option.

In the above example, cflags-y will be assigned the option -march=pentium-mmx if supported by $(CC), otherwise -march=i586. The second argument to cc-option is optional, and if omitted, cflags-y will be assigned no value if first option is not supported. Note: cc-option uses KBUILD_CFLAGS for $(CC) options

cc-option-yn is used to check if gcc supports a given option and return ‘y’ if supported, otherwise ‘n’.

In the above example, $(biarch) is set to y if $(CC) supports the -m32 option. When $(biarch) equals ‘y’, the expanded variables $(aflags-y) and $(cflags-y) will be assigned the values -a32 and -m32, respectively. Note: cc-option-yn uses KBUILD_CFLAGS for $(CC) options

cc-disable-warning checks if gcc supports a given warning and returns the commandline switch to disable it. This special function is needed, because gcc 4.4 and later accept any unknown -Wno-* option and only warn about it if there is another warning in the source file.

In the above example, -Wno-unused-but-set-variable will be added to KBUILD_CFLAGS only if gcc really accepts it.

cc-ifversion tests the version of $(CC) and equals the fourth parameter if version expression is true, or the fifth (if given) if the version expression is false.

In this example, ccflags-y will be assigned the value -O1 if the $(CC) version is less than 4.2. cc-ifversion takes all the shell operators: -eq, -ne, -lt, -le, -gt, and -ge The third parameter may be a text as in this example, but it may also be an expanded variable or a macro.

cc-cross-prefix is used to check if there exists a $(CC) in path with one of the listed prefixes. The first prefix where there exist a prefix$(CC) in the PATH is returned — and if no prefix$(CC) is found then nothing is returned. Additional prefixes are separated by a single space in the call of cc-cross-prefix. This functionality is useful for architecture Makefiles that try to set CROSS_COMPILE to well-known values but may have several values to select between. It is recommended only to try to set CROSS_COMPILE if it is a cross build (host arch is different from target arch). And if CROSS_COMPILE is already set then leave it with the old value.

3.12 $(LD) support functionsВ¶

ld-option is used to check if $(LD) supports the supplied option. ld-option takes two options as arguments. The second argument is an optional option that can be used if the first option is not supported by $(LD).

4 Host Program supportВ¶

Kbuild supports building executables on the host for use during the compilation stage. Two steps are required in order to use a host executable.

The first step is to tell kbuild that a host program exists. This is done utilising the variable hostprogs-y.

The second step is to add an explicit dependency to the executable. This can be done in two ways. Either add the dependency in a rule, or utilise the variable $(always). Both possibilities are described in the following.

4.1 Simple Host ProgramВ¶

In some cases there is a need to compile and run a program on the computer where the build is running. The following line tells kbuild that the program bin2hex shall be built on the build host.

Kbuild assumes in the above example that bin2hex is made from a single c-source file named bin2hex.c located in the same directory as the Makefile.

4.2 Composite Host ProgramsВ¶

Host programs can be made up based on composite objects. The syntax used to define composite objects for host programs is similar to the syntax used for kernel objects. $( -objs) lists all objects used to link the final executable.

Objects with extension .o are compiled from the corresponding .c files. In the above example, checklist.c is compiled to checklist.o and lxdialog.c is compiled to lxdialog.o.

Finally, the two .o files are linked to the executable, lxdialog. Note: The syntax -y is not permitted for host-programs.

4.3 Using C++ for host programsВ¶

kbuild offers support for host programs written in C++. This was introduced solely to support kconfig, and is not recommended for general use.

In the example above the executable is composed of the C++ file qconf.cc — identified by $(qconf-cxxobjs).

If qconf is composed of a mixture of .c and .cc files, then an additional line can be used to identify this.

4.4 Controlling compiler options for host programsВ¶

When compiling host programs, it is possible to set specific flags. The programs will always be compiled utilising $(HOSTCC) passed the options specified in $(KBUILD_HOSTCFLAGS). To set flags that will take effect for all host programs created in that Makefile, use the variable HOST_EXTRACFLAGS.

To set specific flags for a single file the following construction is used:

It is also possible to specify additional options to the linker.

When linking qconf, it will be passed the extra option «-L$(QTDIR)/lib».

4.5 When host programs are actually builtВ¶

Kbuild will only build host-programs when they are referenced as a prerequisite. This is possible in two ways:

List the prerequisite explicitly in a special rule.

The target $(obj)/devlist.h will not be built before $(obj)/gen-devlist is updated. Note that references to the host programs in special rules must be prefixed with $(obj).

When there is no suitable special rule, and the host program shall be built when a makefile is entered, the $(always) variable shall be used.

This will tell kbuild to build lxdialog even if not referenced in any rule.

4.6 Using hostprogs-$(CONFIG_FOO)В¶

A typical pattern in a Kbuild file looks like this:

Kbuild knows about both ‘y’ for built-in and ‘m’ for module. So if a config symbol evaluates to ‘m’, kbuild will still build the binary. In other words, Kbuild handles hostprogs-m exactly like hostprogs-y. But only hostprogs-y is recommended to be used when no CONFIG symbols are involved.

5 Kbuild clean infrastructureВ¶

«make clean» deletes most generated files in the obj tree where the kernel is compiled. This includes generated files such as host programs. Kbuild knows targets listed in $(hostprogs-y), $(hostprogs-m), $(always), $(extra-y) and $(targets). They are all deleted during «make clean». Files matching the patterns «.[oas]», «.ko», plus some additional files generated by kbuild are deleted all over the kernel src tree when «make clean» is executed.

Additional files can be specified in kbuild makefiles by use of $(clean-files).

When executing «make clean», the file «crc32table.h» will be deleted. Kbuild will assume files to be in the same relative directory as the Makefile, except if prefixed with $(objtree).

To delete a directory hierarchy use:

This will delete the directory debian in the toplevel directory, including all subdirectories.

To exclude certain files from make clean, use the $(no-clean-files) variable. This is only a special case used in the top level Kbuild file:

Usually kbuild descends down in subdirectories due to «obj-* := dir/», but in the architecture makefiles where the kbuild infrastructure is not sufficient this sometimes needs to be explicit.

The above assignment instructs kbuild to descend down in the directory compressed/ when «make clean» is executed.

To support the clean infrastructure in the Makefiles that build the final bootimage there is an optional target named archclean:

When «make clean» is executed, make will descend down in arch/x86/boot, and clean as usual. The Makefile located in arch/x86/boot/ may use the subdir- trick to descend further down.

Note 1: arch/$(ARCH)/Makefile cannot use «subdir-«, because that file is included in the top level makefile, and the kbuild infrastructure is not operational at that point.

Note 2: All directories listed in core-y, libs-y, drivers-y and net-y will be visited during «make clean».

6 Architecture MakefilesВ¶

The top level Makefile sets up the environment and does the preparation, before starting to descend down in the individual directories. The top level makefile contains the generic part, whereas arch/$(ARCH)/Makefile contains what is required to set up kbuild for said architecture. To do so, arch/$(ARCH)/Makefile sets up a number of variables and defines a few targets.

When kbuild executes, the following steps are followed (roughly):

Configuration of the kernel => produce .config
Store kernel version in include/linux/version.h
Updating all other prerequisites to the target prepare: — Additional prerequisites are specified in arch/$(ARCH)/Makefile
Recursively descend down in all directories listed in init-* core* drivers-* net-* libs-* and build all targets. — The values of the above variables are expanded in arch/$(ARCH)/Makefile.
All object files are then linked and the resulting file vmlinux is located at the root of the obj tree. The very first objects linked are listed in head-y, assigned by arch/$(ARCH)/Makefile.
Finally, the architecture-specific part does any required post processing and builds the final bootimage. — This includes building boot records — Preparing initrd images and the like

6.1 Set variables to tweak the build to the architectureВ¶

Generic $(LD) options

Flags used for all invocations of the linker. Often specifying the emulation is sufficient.

Note: ldflags-y can be used to further customise the flags used. See chapter 3.7.

Options for $(LD) when linking vmlinux

LDFLAGS_vmlinux is used to specify additional flags to pass to the linker when linking the final vmlinux image. LDFLAGS_vmlinux uses the LDFLAGS_$@ support.

When $(call if_changed,objcopy) is used to translate a .o file, the flags specified in OBJCOPYFLAGS will be used. $(call if_changed,objcopy) is often used to generate raw binaries on vmlinux.

In this example, the binary $(obj)/image is a binary version of vmlinux. The usage of $(call if_changed,xxx) will be described later.

Default value — see top level Makefile Append or modify as required per architecture.

$(CC) compiler flags

Default value — see top level Makefile Append or modify as required per architecture.

Often, the KBUILD_CFLAGS variable depends on the configuration.

Many arch Makefiles dynamically run the target C compiler to probe supported options:

The first example utilises the trick that a config option expands to ‘y’ when selected.

Assembler options specific for built-in

$(KBUILD_AFLAGS_KERNEL) contains extra C compiler flags used to compile resident kernel code.

Assembler options specific for modules

$(KBUILD_AFLAGS_MODULE) is used to add arch-specific options that are used for assembler.

From commandline AFLAGS_MODULE shall be used (see kbuild.txt).

$(CC) options specific for built-in

$(KBUILD_CFLAGS_KERNEL) contains extra C compiler flags used to compile resident kernel code.

Options for $(CC) when building modules

$(KBUILD_CFLAGS_MODULE) is used to add arch-specific options that are used for $(CC). From commandline CFLAGS_MODULE shall be used (see kbuild.txt).

Options for $(LD) when linking modules

$(KBUILD_LDFLAGS_MODULE) is used to add arch-specific options used when linking modules. This is often a linker script.

From commandline LDFLAGS_MODULE shall be used (see kbuild.txt).

KBUILD_ARFLAGS Options for $(AR) when creating archives

ARCH_CPPFLAGS, ARCH_AFLAGS, ARCH_CFLAGS Overrides the kbuild defaults

6.2 Add prerequisites to archheadersВ¶

The archheaders: rule is used to generate header files that may be installed into user space by «make header_install».

It is run before «make archprepare» when run on the architecture itself.

6.3 Add prerequisites to archprepareВ¶

The archprepare: rule is used to list prerequisites that need to be built before starting to descend down in the subdirectories. This is usually used for header files containing assembler constants.

In this example, the file target maketools will be processed before descending down in the subdirectories. See also chapter XXX-TODO that describe how kbuild supports generating offset header files.

6.4 List directories to visit when descendingВ¶

An arch Makefile cooperates with the top Makefile to define variables which specify how to build the vmlinux file. Note that there is no corresponding arch-specific section for modules; the module-building machinery is all architecture-independent.

head-y, init-y, core-y, libs-y, drivers-y, net-y

$(head-y) lists objects to be linked first in vmlinux.

$(libs-y) lists directories where a lib.a archive can be located.

The rest list directories where a built-in.a object file can be located.

$(init-y) objects will be located after $(head-y).

Then the rest follows in this order:

The top level Makefile defines values for all generic directories, and arch/$(ARCH)/Makefile only adds architecture-specific directories.

6.5 Architecture-specific boot imagesВ¶

An arch Makefile specifies goals that take the vmlinux file, compress it, wrap it in bootstrapping code, and copy the resulting files somewhere. This includes various kinds of installation commands. The actual goals are not standardized across architectures.

It is common to locate any additional processing in a boot/ directory below arch/$(ARCH)/.

Kbuild does not provide any smart way to support building a target specified in boot/. Therefore arch/$(ARCH)/Makefile shall call make manually to build a target in boot/.

The recommended approach is to include shortcuts in arch/$(ARCH)/Makefile, and use the full path when calling down into the arch/$(ARCH)/boot/Makefile.

«$(Q)$(MAKE) $(build)= » is the recommended way to invoke make in a subdirectory.

There are no rules for naming architecture-specific targets, but executing «make help» will list all relevant targets. To support this, $(archhelp) must be defined.

When make is executed without arguments, the first goal encountered will be built. In the top level Makefile the first goal present is all:. An architecture shall always, per default, build a bootable image. In «make help», the default goal is highlighted with a ‘*’. Add a new prerequisite to all: to select a default goal different from vmlinux.

When «make» is executed without arguments, bzImage will be built.

6.6 Building non-kbuild targetsВ¶

extra-y specifies additional targets created in the current directory, in addition to any targets specified by obj-* .

Listing all targets in extra-y is required for two purposes:

Enable kbuild to check changes in command lines
- When $(call if_changed,xxx) is used
kbuild knows what files to delete during «make clean»

In this example, extra-y is used to list object files that shall be built, but shall not be linked as part of built-in.a.

This works as a weaker version of header-test-y, and accepts wildcard patterns. The typical usage is:

This specifies all the files that matches to ‘ * .h’ in the current directory, but the files in ‘header-test-‘ are excluded.

6.7 Commands useful for building a boot imageВ¶

Kbuild provides a few macros that are useful when building a boot image.

if_changed is the infrastructure used for the following commands.

When the rule is evaluated, it is checked to see if any files need an update, or the command line has changed since the last invocation. The latter will force a rebuild if any options to the executable have changed. Any target that utilises if_changed must be listed in $(targets), otherwise the command line check will fail, and the target will always be built. Assignments to $(targets) are without $(obj)/ prefix. if_changed may be used in conjunction with custom commands as defined in 6.8 «Custom kbuild commands».

Note: It is a typical mistake to forget the FORCE prerequisite. Another common pitfall is that whitespace is sometimes significant; for instance, the below will fail (note the extra space after the comma):

WRONG! $(call if_changed, ld/objcopy/gzip/. )

Note: if_changed should not be used more than once per target. It stores the executed command in a corresponding .cmd

file and multiple calls would result in overwrites and unwanted results when the target is up to date and only the tests on changed commands trigger execution of commands.

Link target. Often, LDFLAGS_$@ is used to set specific options to ld.

In this example, there are two possible targets, requiring different options to the linker. The linker options are specified using the LDFLAGS_$@ syntax — one for each potential target. $(targets) are assigned all potential targets, by which kbuild knows the targets and will:

check for commandline changes
delete target during make clean

The «: %: %.o» part of the prerequisite is a shorthand that frees us from listing the setup.o and bootsect.o files.

Note: It is a common mistake to forget the «targets :=» assignment, resulting in the target file being recompiled for no obvious reason. objcopy Copy binary. Uses OBJCOPYFLAGS usually specified in arch/$(ARCH)/Makefile. OBJCOPYFLAGS_$@ may be used to set additional options. gzip

Compress target. Use maximum compression to compress target.

Create flattened device tree blob object suitable for linking into vmlinux. Device tree blobs linked into vmlinux are placed in an init section in the image. Platform code must copy the blob to non-init memory prior to calling unflatten_device_tree().

To use this command, simply add *.dtb into obj-y or targets, or make some other target depend on %.dtb

A central rule exists to create $(obj)/%.dtb from $(src)/%.dts ; architecture Makefiles do no need to explicitly write out that rule.

6.8 Custom kbuild commandsВ¶

When kbuild is executing with KBUILD_VERBOSE=0, then only a shorthand of a command is normally displayed. To enable this behaviour for custom commands kbuild requires two variables to be set:

When updating the $(obj)/bzImage target, the line:

will be displayed with «make KBUILD_VERBOSE=0».

— 6.9 Preprocessing linker scripts

When the vmlinux image is built, the linker script arch/$(ARCH)/kernel/vmlinux.lds is used. The script is a preprocessed variant of the file vmlinux.lds.S located in the same directory. kbuild knows .lds files and includes a rule *lds.S -> *lds .

The assignment to $(always) is used to tell kbuild to build the target vmlinux.lds. The assignment to $(CPPFLAGS_vmlinux.lds) tells kbuild to use the specified options when building the target vmlinux.lds.

When building the *.lds target, kbuild uses the variables:

The kbuild infrastructure for *lds files is used in several architecture-specific files.

6.10 Generic header filesВ¶

6.11 Post-link passВ¶

If the file arch/xxx/Makefile.postlink exists, this makefile will be invoked for post-link objects (vmlinux and modules.ko) for architectures to run post-link passes on. Must also handle the clean target.

This pass runs after kallsyms generation. If the architecture needs to modify symbol locations, rather than manipulate the kallsyms, it may be easier to add another postlink target for .tmp_vmlinux? targets to be called from link-vmlinux.sh.

For example, powerpc uses this to check relocation sanity of the linked vmlinux file.

7 Kbuild syntax for exported headersВ¶

The kernel includes a set of headers that is exported to userspace. Many headers can be exported as-is but other headers require a minimal pre-processing before they are ready for user-space. The pre-processing does:

drop kernel-specific annotations
drop include of compiler.h
drop all sections that are kernel internal (guarded by ifdef __KERNEL__ )

7.1 no-export-headersВ¶

7.2 generic-yВ¶

If an architecture uses a verbatim copy of a header from include/asm-generic then this is listed in the file arch/$(ARCH)/include/asm/Kbuild like this:

During the prepare phase of the build a wrapper include file is generated in the directory:

When a header is exported where the architecture uses the generic header a similar wrapper is generated as part of the set of exported headers in the directory:

The generated wrapper will in both cases look like the following:

Источник