lastdragon/linux-kernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Resguardo inicial

Browse Source
This commit is contained in:
lastdragon
2026-09-06 12:53:21 -06:00
parent 662efe3c2a
commit ad2088d178
94853 changed files with 43200404 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
linux-7.2.3/Documentation/kbuild/Kconfig.recursion-issue-01
+57
View File
@@ -0,0 +1,57 @@
# Simple Kconfig recursive issue
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# Test with:
#
# make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-01 allnoconfig
#
# This Kconfig file has a simple recursive dependency issue. In order to
# understand why this recursive dependency issue occurs lets consider what
# Kconfig needs to address. We iterate over what Kconfig needs to address
# by stepping through the questions it needs to address sequentially.
#
# * What values are possible for CORE?
#
# CORE_BELL_A_ADVANCED selects CORE, which means that it influences the values
# that are possible for CORE. So for example if CORE_BELL_A_ADVANCED is 'y',
# CORE must be 'y' too.
#
# * What influences CORE_BELL_A_ADVANCED?
#
# As the name implies CORE_BELL_A_ADVANCED is an advanced feature of
# CORE_BELL_A so naturally it depends on CORE_BELL_A. So if CORE_BELL_A is 'y'
# we know CORE_BELL_A_ADVANCED can be 'y' too.
#
# * What influences CORE_BELL_A?
#
# CORE_BELL_A depends on CORE, so CORE influences CORE_BELL_A.
#
# But that is a problem, because this means that in order to determine
# what values are possible for CORE we ended up needing to address questions
# regarding possible values of CORE itself again. Answering the original
# question of what are the possible values of CORE would make the kconfig
# tools run in a loop. When this happens Kconfig exits and complains about
# the "recursive dependency detected" error.
#
# Reading the Documentation/kbuild/Kconfig.recursion-issue-01 file it may be
# obvious that an easy solution to this problem should just be the removal
# of the "select CORE" from CORE_BELL_A_ADVANCED as that is implicit already
# since CORE_BELL_A depends on CORE. Recursive dependency issues are not always
# so trivial to resolve, we provide another example below of practical
# implications of this recursive issue where the solution is perhaps not so
# easy to understand. Note that matching semantics on the dependency on
# CORE also consist of a solution to this recursive problem.
mainmenu "Simple example to demo kconfig recursive dependency issue"
config CORE
tristate
config CORE_BELL_A
tristate
depends on CORE
config CORE_BELL_A_ADVANCED
tristate
depends on CORE_BELL_A
select CORE
linux-7.2.3/Documentation/kbuild/Kconfig.recursion-issue-02
+63
View File
@@ -0,0 +1,63 @@
# Cumulative Kconfig recursive issue
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# Test with:
#
# make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-02 allnoconfig
#
# The recursive limitations with Kconfig has some non intuitive implications on
# kconfig semantics which are documented here. One known practical implication
# of the recursive limitation is that drivers cannot negate features from other
# drivers if they share a common core requirement and use disjoint semantics to
# annotate those requirements, ie, some drivers use "depends on" while others
# use "select". For instance it means if a driver A and driver B share the same
# core requirement, and one uses "select" while the other uses "depends on" to
# annotate this, all features that driver A selects cannot now be negated by
# driver B.
#
# A perhaps not so obvious implication of this is that, if semantics on these
# core requirements are not carefully synced, as drivers evolve features
# they select or depend on end up becoming shared requirements which cannot be
# negated by other drivers.
#
# The example provided in Documentation/kbuild/Kconfig.recursion-issue-02
# describes a simple driver core layout of example features a kernel might
# have. Let's assume we have some CORE functionality, then the kernel has a
# series of bells and whistles it desires to implement, its not so advanced so
# it only supports bells at this time: CORE_BELL_A and CORE_BELL_B. If
# CORE_BELL_A has some advanced feature CORE_BELL_A_ADVANCED which selects
# CORE_BELL_A then CORE_BELL_A ends up becoming a common BELL feature which
# other bells in the system cannot negate. The reason for this issue is
# due to the disjoint use of semantics on expressing each bell's relationship
# with CORE, one uses "depends on" while the other uses "select". Another
# more important reason is that kconfig does not check for dependencies listed
# under 'select' for a symbol, when such symbols are selected kconfig them
# as mandatory required symbols. For more details on the heavy handed nature
# of select refer to Documentation/kbuild/Kconfig.select-break
#
# To fix this the "depends on CORE" must be changed to "select CORE", or the
# "select CORE" must be changed to "depends on CORE".
#
# For an example real world scenario issue refer to the attempt to remove
# "select FW_LOADER" [0], in the end the simple alternative solution to this
# problem consisted on matching semantics with newly introduced features.
#
# [0] https://lore.kernel.org/r/1432241149-8762-1-git-send-email-mcgrof@do-not-panic.com
mainmenu "Simple example to demo cumulative kconfig recursive dependency implication"
config CORE
tristate
config CORE_BELL_A
tristate
depends on CORE
config CORE_BELL_A_ADVANCED
tristate
select CORE_BELL_A
config CORE_BELL_B
tristate
depends on !CORE_BELL_A
select CORE
linux-7.2.3/Documentation/kbuild/Kconfig.select-break
+33
View File
@@ -0,0 +1,33 @@
# Select broken dependency issue
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#
# Test with:
#
# make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.select-break menuconfig
#
# kconfig will not complain and enable this layout for configuration. This is
# currently a feature of kconfig, given select was designed to be heavy handed.
# Kconfig currently does not check the list of symbols listed on a symbol's
# "select" list, this is done on purpose to help load a set of known required
# symbols. Because of this use of select should be used with caution. An
# example of this issue is below.
#
# The option B and C are clearly contradicting with respect to A.
# However, when A is set, C can be set as well because Kconfig does not
# visit the dependencies of the select target (in this case B). And since
# Kconfig does not visit the dependencies, it breaks the dependencies of B
# (!A).
mainmenu "Simple example to demo kconfig select broken dependency issue"
config A
bool "CONFIG A"
config B
bool "CONFIG B"
depends on !A
config C
bool "CONFIG C"
depends on A
select B
linux-7.2.3/Documentation/kbuild/bash-completion.rst
+65
View File
@@ -0,0 +1,65 @@
.. SPDX-License-Identifier: GPL-2.0-only
==========================
Bash completion for Kbuild
==========================
The kernel build system is written using Makefiles, and Bash completion
for the `make` command is available through the `bash-completion`_ project.
However, the Makefiles for the kernel build are complex. The generic completion
rules for the `make` command do not provide meaningful suggestions for the
kernel build system, except for the options of the `make` command itself.
To enhance completion for various variables and targets, the kernel source
includes its own completion script at `scripts/bash-completion/make`.
This script provides additional completions when working within the kernel tree.
Outside the kernel tree, it defaults to the generic completion rules for the
`make` command.
Prerequisites
=============
The script relies on helper functions provided by `bash-completion`_ project.
Please ensure it is installed on your system. On most distributions, you can
install the `bash-completion` package through the standard package manager.
How to use
==========
You can source the script directly::
$ source scripts/bash-completion/make
Or, you can copy it into the search path for Bash completion scripts.
For example::
$ mkdir -p ~/.local/share/bash-completion/completions
$ cp scripts/bash-completion/make ~/.local/share/bash-completion/completions/
Details
=======
The additional completion for Kbuild is enabled in the following cases:
- You are in the root directory of the kernel source.
- You are in the top-level build directory created by the O= option
(checked via the `source` symlink pointing to the kernel source).
- The -C make option specifies the kernel source or build directory.
- The -f make option specifies a file in the kernel source or build directory.
If none of the above are met, it falls back to the generic completion rules.
The completion supports:
- Commonly used targets, such as `all`, `menuconfig`, `dtbs`, etc.
- Make (or environment) variables, such as `ARCH`, `LLVM`, etc.
- Single-target builds (`foo/bar/baz.o`)
- Configuration files (`*_defconfig` and `*.config`)
Some variables offer intelligent behavior. For instance, `CROSS_COMPILE=`
followed by a TAB displays installed toolchains. The list of defconfig files
shown depends on the value of the `ARCH=` variable.
.. _bash-completion: https://github.com/scop/bash-completion/
linux-7.2.3/Documentation/kbuild/gcc-plugins.rst
+133
View File
@@ -0,0 +1,133 @@
=========================
GCC plugin infrastructure
=========================
Introduction
============
GCC plugins are loadable modules that provide extra features to the
compiler [1]_. They are useful for runtime instrumentation and static analysis.
We can analyse, change and add further code during compilation via
callbacks [2]_, GIMPLE [3]_, IPA [4]_ and RTL passes [5]_.
The GCC plugin infrastructure of the kernel supports building out-of-tree
modules, cross-compilation and building in a separate directory.
Plugin source files have to be compilable by a C++ compiler.
Currently the GCC plugin infrastructure supports only some architectures.
Grep "select HAVE_GCC_PLUGINS" to find out which architectures support
GCC plugins.
This infrastructure was ported from grsecurity [6]_ and PaX [7]_.
--
.. [1] https://gcc.gnu.org/onlinedocs/gccint/Plugins.html
.. [2] https://gcc.gnu.org/onlinedocs/gccint/Plugin-API.html#Plugin-API
.. [3] https://gcc.gnu.org/onlinedocs/gccint/GIMPLE.html
.. [4] https://gcc.gnu.org/onlinedocs/gccint/IPA.html
.. [5] https://gcc.gnu.org/onlinedocs/gccint/RTL.html
.. [6] https://grsecurity.net/
.. [7] https://pax.grsecurity.net/
Purpose
=======
GCC plugins are designed to provide a place to experiment with potential
compiler features that are neither in GCC nor Clang upstream. Once
their utility is proven, the goal is to upstream the feature into GCC
(and Clang), and then to finally remove them from the kernel once the
feature is available in all supported versions of GCC.
Specifically, new plugins should implement only features that have no
upstream compiler support (in either GCC or Clang).
When a feature exists in Clang but not GCC, effort should be made to
bring the feature to upstream GCC (rather than just as a kernel-specific
GCC plugin), so the entire ecosystem can benefit from it.
Similarly, even if a feature provided by a GCC plugin does *not* exist
in Clang, but the feature is proven to be useful, effort should be spent
to upstream the feature to GCC (and Clang).
After a feature is available in upstream GCC, the plugin will be made
unbuildable for the corresponding GCC version (and later). Once all
kernel-supported versions of GCC provide the feature, the plugin will
be removed from the kernel.
Files
=====
**$(src)/scripts/gcc-plugins**
This is the directory of the GCC plugins.
**$(src)/scripts/gcc-plugins/gcc-common.h**
This is a compatibility header for GCC plugins.
It should be always included instead of individual gcc headers.
**$(src)/scripts/gcc-plugins/gcc-generate-gimple-pass.h,
$(src)/scripts/gcc-plugins/gcc-generate-ipa-pass.h,
$(src)/scripts/gcc-plugins/gcc-generate-simple_ipa-pass.h,
$(src)/scripts/gcc-plugins/gcc-generate-rtl-pass.h**
These headers automatically generate the registration structures for
GIMPLE, SIMPLE_IPA, IPA and RTL passes.
They should be preferred to creating the structures by hand.
Usage
=====
You must install the gcc plugin headers for your gcc version,
e.g., on Ubuntu for gcc-10::
apt-get install gcc-10-plugin-dev
Or on Fedora::
dnf install gcc-plugin-devel libmpc-devel
Or on Fedora when using cross-compilers that include plugins::
dnf install libmpc-devel
Enable the GCC plugin infrastructure and some plugin(s) you want to use
in the kernel config::
CONFIG_GCC_PLUGINS=y
CONFIG_GCC_PLUGIN_LATENT_ENTROPY=y
...
Run gcc (native or cross-compiler) to ensure plugin headers are detected::
gcc -print-file-name=plugin
CROSS_COMPILE=arm-linux-gnu- ${CROSS_COMPILE}gcc -print-file-name=plugin
The word "plugin" means they are not detected::
plugin
A full path means they are detected::
/usr/lib/gcc/x86_64-redhat-linux/12/plugin
To compile the minimum tool set including the plugin(s)::
make scripts
or just run the kernel make and compile the whole kernel with
the cyclomatic complexity GCC plugin.
4. How to add a new GCC plugin
==============================
The GCC plugins are in scripts/gcc-plugins/. You need to put plugin source files
right under scripts/gcc-plugins/. Creating subdirectories is not supported.
It must be added to scripts/gcc-plugins/Makefile, scripts/Makefile.gcc-plugins
and a relevant Kconfig file.
linux-7.2.3/Documentation/kbuild/gendwarfksyms.rst
+418
View File
@@ -0,0 +1,418 @@
=======================
DWARF module versioning
=======================
Introduction
============
When CONFIG_MODVERSIONS is enabled, symbol versions for modules
are typically calculated from preprocessed source code using the
**genksyms** tool. However, this is incompatible with languages such
as Rust, where the source code has insufficient information about
the resulting ABI. With CONFIG_GENDWARFKSYMS (and CONFIG_DEBUG_INFO)
selected, **gendwarfksyms** is used instead to calculate symbol versions
from the DWARF debugging information, which contains the necessary
details about the final module ABI.
Dependencies
------------
gendwarfksyms depends on the libelf, libdw, and zlib libraries.
Here are a few examples of how to install these dependencies:
* Arch Linux and derivatives::
sudo pacman --needed -S libelf zlib
* Debian, Ubuntu, and derivatives::
sudo apt install libelf-dev libdw-dev zlib1g-dev
* Fedora and derivatives::
sudo dnf install elfutils-libelf-devel elfutils-devel zlib-devel
* openSUSE and derivatives::
sudo zypper install libelf-devel libdw-devel zlib-devel
Usage
-----
gendwarfksyms accepts a list of object files on the command line, and a
list of symbol names (one per line) in standard input::
Usage: gendwarfksyms [options] elf-object-file ... < symbol-list
Options:
-d, --debug Print debugging information
--dump-dies Dump DWARF DIE contents
--dump-die-map Print debugging information about die_map changes
--dump-types Dump type strings
--dump-versions Dump expanded type strings used for symbol versions
-s, --stable Support kABI stability features
-T, --symtypes file Write a symtypes file
-h, --help Print this message
Type information availability
=============================
While symbols are typically exported in the same translation unit (TU)
where they're defined, it's also perfectly fine for a TU to export
external symbols. For example, this is done when calculating symbol
versions for exports in stand-alone assembly code.
To ensure the compiler emits the necessary DWARF type information in the
TU where symbols are actually exported, gendwarfksyms adds a pointer
to exported symbols in the `EXPORT_SYMBOL()` macro using the following
macro::
#define __GENDWARFKSYMS_EXPORT(sym) \
static typeof(sym) *__gendwarfksyms_ptr_##sym __used \
__section(".discard.gendwarfksyms") = &sym;
When a symbol pointer is found in DWARF, gendwarfksyms can use its
type for calculating symbol versions even if the symbol is defined
elsewhere. The name of the symbol pointer is expected to start with
`__gendwarfksyms_ptr_`, followed by the name of the exported symbol.
Symtypes output format
======================
Similarly to genksyms, gendwarfksyms supports writing a symtypes
file for each processed object that contain types for exported
symbols and each referenced type that was used in calculating symbol
versions. These files can be useful when trying to determine what
exactly caused symbol versions to change between builds. To generate
symtypes files during a kernel build, set `KBUILD_SYMTYPES=1`.
Matching the existing format, the first column of each line contains
either a type reference or a symbol name. Type references have a
one-letter prefix followed by "#" and the name of the type. Four
reference types are supported::
e#<type> = enum
s#<type> = struct
t#<type> = typedef
u#<type> = union
Type names with spaces in them are wrapped in single quotes, e.g.::
s#'core::result::Result<u8, core::num::error::ParseIntError>'
The rest of the line contains a type string. Unlike with genksyms that
produces C-style type strings, gendwarfksyms uses the same simple parsed
DWARF format produced by **--dump-dies**, but with type references
instead of fully expanded strings.
Maintaining a stable kABI
=========================
Distribution maintainers often need the ability to make ABI compatible
changes to kernel data structures due to LTS updates or backports. Using
the traditional `#ifndef __GENKSYMS__` to hide these changes from symbol
versioning won't work when processing object files. To support this
use case, gendwarfksyms provides kABI stability features designed to
hide changes that won't affect the ABI when calculating versions. These
features are all gated behind the **--stable** command line flag and are
not used in the mainline kernel. To use stable features during a kernel
build, set `KBUILD_GENDWARFKSYMS_STABLE=1`.
Examples for using these features are provided in the
**scripts/gendwarfksyms/examples** directory, including helper macros
for source code annotation. Note that as these features are only used to
transform the inputs for symbol versioning, the user is responsible for
ensuring that their changes actually won't break the ABI.
kABI rules
----------
kABI rules allow distributions to fine-tune certain parts
of gendwarfksyms output and thus control how symbol
versions are calculated. These rules are defined in the
`.discard.gendwarfksyms.kabi_rules` section of the object file and
consist of simple null-terminated strings with the following structure::
version\0type\0target\0value\0
This string sequence is repeated as many times as needed to express all
the rules. The fields are as follows:
- `version`: Ensures backward compatibility for future changes to the
structure. Currently expected to be "1".
- `type`: Indicates the type of rule being applied.
- `target`: Specifies the target of the rule, typically the fully
qualified name of the DWARF Debugging Information Entry (DIE).
- `value`: Provides rule-specific data.
The following helper macros, for example, can be used to specify rules
in the source code::
#define ___KABI_RULE(hint, target, value) \
static const char __PASTE(__gendwarfksyms_rule_, \
__COUNTER__)[] __used __aligned(1) \
__section(".discard.gendwarfksyms.kabi_rules") = \
"1\0" #hint "\0" target "\0" value
#define __KABI_RULE(hint, target, value) \
___KABI_RULE(hint, #target, #value)
Currently, only the rules discussed in this section are supported, but
the format is extensible enough to allow further rules to be added as
need arises.
Managing definition visibility
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A declaration can change into a full definition when additional includes
are pulled into the translation unit. This changes the versions of any
symbol that references the type even if the ABI remains unchanged. As
it may not be possible to drop includes without breaking the build, the
`declonly` rule can be used to specify a type as declaration-only, even
if the debugging information contains the full definition.
The rule fields are expected to be as follows:
- `type`: "declonly"
- `target`: The fully qualified name of the target data structure
(as shown in **--dump-dies** output).
- `value`: This field is ignored.
Using the `__KABI_RULE` macro, this rule can be defined as::
#define KABI_DECLONLY(fqn) __KABI_RULE(declonly, fqn, )
Example usage::
struct s {
/* definition */
};
KABI_DECLONLY(s);
Adding enumerators
~~~~~~~~~~~~~~~~~~
For enums, all enumerators and their values are included in calculating
symbol versions, which becomes a problem if we later need to add more
enumerators without changing symbol versions. The `enumerator_ignore`
rule allows us to hide named enumerators from the input.
The rule fields are expected to be as follows:
- `type`: "enumerator_ignore"
- `target`: The fully qualified name of the target enum
(as shown in **--dump-dies** output) and the name of the
enumerator field separated by a space.
- `value`: This field is ignored.
Using the `__KABI_RULE` macro, this rule can be defined as::
#define KABI_ENUMERATOR_IGNORE(fqn, field) \
__KABI_RULE(enumerator_ignore, fqn field, )
Example usage::
enum e {
A, B, C, D,
};
KABI_ENUMERATOR_IGNORE(e, B);
KABI_ENUMERATOR_IGNORE(e, C);
If the enum additionally includes an end marker and new values must
be added in the middle, we may need to use the old value for the last
enumerator when calculating versions. The `enumerator_value` rule allows
us to override the value of an enumerator for version calculation:
- `type`: "enumerator_value"
- `target`: The fully qualified name of the target enum
(as shown in **--dump-dies** output) and the name of the
enumerator field separated by a space.
- `value`: Integer value used for the field.
Using the `__KABI_RULE` macro, this rule can be defined as::
#define KABI_ENUMERATOR_VALUE(fqn, field, value) \
__KABI_RULE(enumerator_value, fqn field, value)
Example usage::
enum e {
A, B, C, LAST,
};
KABI_ENUMERATOR_IGNORE(e, C);
KABI_ENUMERATOR_VALUE(e, LAST, 2);
Managing structure size changes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A data structure can be partially opaque to modules if its allocation is
handled by the core kernel, and modules only need to access some of its
members. In this situation, it's possible to append new members to the
structure without breaking the ABI, as long as the layout for the original
members remains unchanged.
To append new members, we can hide them from symbol versioning as
described in section :ref:`Hiding members <hiding_members>`, but we can't
hide the increase in structure size. The `byte_size` rule allows us to
override the structure size used for symbol versioning.
The rule fields are expected to be as follows:
- `type`: "byte_size"
- `target`: The fully qualified name of the target data structure
(as shown in **--dump-dies** output).
- `value`: A positive decimal number indicating the structure size
in bytes.
Using the `__KABI_RULE` macro, this rule can be defined as::
#define KABI_BYTE_SIZE(fqn, value) \
__KABI_RULE(byte_size, fqn, value)
Example usage::
struct s {
/* Unchanged original members */
unsigned long a;
void *p;
/* Appended new members */
KABI_IGNORE(0, unsigned long n);
};
KABI_BYTE_SIZE(s, 16);
Overriding type strings
~~~~~~~~~~~~~~~~~~~~~~~
In rare situations where distributions must make significant changes to
otherwise opaque data structures that have inadvertently been included
in the published ABI, keeping symbol versions stable using the more
targeted kABI rules can become tedious. The `type_string` rule allows us
to override the full type string for a type or a symbol, and even add
types for versioning that no longer exist in the kernel.
The rule fields are expected to be as follows:
- `type`: "type_string"
- `target`: The fully qualified name of the target data structure
(as shown in **--dump-dies** output) or symbol.
- `value`: A valid type string (as shown in **--symtypes**) output)
to use instead of the real type.
Using the `__KABI_RULE` macro, this rule can be defined as::
#define KABI_TYPE_STRING(type, str) \
___KABI_RULE("type_string", type, str)
Example usage::
/* Override type for a structure */
KABI_TYPE_STRING("s#s",
"structure_type s { "
"member base_type int byte_size(4) "
"encoding(5) n "
"data_member_location(0) "
"} byte_size(8)");
/* Override type for a symbol */
KABI_TYPE_STRING("my_symbol", "variable s#s");
The `type_string` rule should be used only as a last resort if maintaining
a stable symbol versions cannot be reasonably achieved using other
means. Overriding a type string increases the risk of actual ABI breakages
going unnoticed as it hides all changes to the type.
Adding structure members
------------------------
Perhaps the most common ABI compatible change is adding a member to a
kernel data structure. When changes to a structure are anticipated,
distribution maintainers can pre-emptively reserve space in the
structure and take it into use later without breaking the ABI. If
changes are needed to data structures without reserved space, existing
alignment holes can potentially be used instead. While kABI rules could
be added for these type of changes, using unions is typically a more
natural method. This section describes gendwarfksyms support for using
reserved space in data structures and hiding members that don't change
the ABI when calculating symbol versions.
Reserving space and replacing members
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Space is typically reserved for later use by appending integer types, or
arrays, to the end of the data structure, but any type can be used. Each
reserved member needs a unique name, but as the actual purpose is usually
not known at the time the space is reserved, for convenience, names that
start with `__kabi_` are left out when calculating symbol versions::
struct s {
long a;
long __kabi_reserved_0; /* reserved for future use */
};
The reserved space can be taken into use by wrapping the member in a
union, which includes the original type and the replacement member::
struct s {
long a;
union {
long __kabi_reserved_0; /* original type */
struct b b; /* replaced field */
};
};
If the `__kabi_` naming scheme was used when reserving space, the name
of the first member of the union must start with `__kabi_reserved`. This
ensures the original type is used when calculating versions, but the name
is again left out. The rest of the union is ignored.
If we're replacing a member that doesn't follow this naming convention,
we also need to preserve the original name to avoid changing versions,
which we can do by changing the first union member's name to start with
`__kabi_renamed` followed by the original name.
The examples include `KABI_(RESERVE|USE|REPLACE)*` macros that help
simplify the process and also ensure the replacement member is correctly
aligned and its size won't exceed the reserved space.
.. _hiding_members:
Hiding members
~~~~~~~~~~~~~~
Predicting which structures will require changes during the support
timeframe isn't always possible, in which case one might have to resort
to placing new members into existing alignment holes::
struct s {
int a;
/* a 4-byte alignment hole */
unsigned long b;
};
While this won't change the size of the data structure, one needs to
be able to hide the added members from symbol versioning. Similarly
to reserved fields, this can be accomplished by wrapping the added
member to a union where one of the fields has a name starting with
`__kabi_ignored`::
struct s {
int a;
union {
char __kabi_ignored_0;
int n;
};
unsigned long b;
};
With **--stable**, both versions produce the same symbol version. The
examples include a `KABI_IGNORE` macro to simplify the code.
linux-7.2.3/Documentation/kbuild/headers_install.rst
+44
View File
@@ -0,0 +1,44 @@
=============================================
Exporting kernel headers for use by userspace
=============================================
The "make headers_install" command exports the kernel's header files in a
form suitable for use by userspace programs.
The linux kernel's exported header files describe the API for user space
programs attempting to use kernel services. These kernel header files are
used by the system's C library (such as glibc or uClibc) to define available
system calls, as well as constants and structures to be used with these
system calls. The C library's header files include the kernel header files
from the "linux" subdirectory. The system's libc headers are usually
installed at the default location /usr/include and the kernel headers in
subdirectories under that (most notably /usr/include/linux and
/usr/include/asm).
Kernel headers are backwards compatible, but not forwards compatible. This
means that a program built against a C library using older kernel headers
should run on a newer kernel (although it may not have access to new
features), but a program built against newer kernel headers may not work on an
older kernel.
The "make headers_install" command can be run in the top level directory of the
kernel source code (or using a standard out-of-tree build). It takes two
optional arguments::
make headers_install ARCH=i386 INSTALL_HDR_PATH=/usr
ARCH indicates which architecture to produce headers for, and defaults to the
current architecture. The linux/asm directory of the exported kernel headers
is platform-specific, to see a complete list of supported architectures use
the command::
ls -d include/asm-* | sed 's/.*-//'
INSTALL_HDR_PATH indicates where to install the headers. It defaults to
"./usr".
An 'include' directory is automatically created inside INSTALL_HDR_PATH and
headers are installed in 'INSTALL_HDR_PATH/include'.
The kernel header export infrastructure is maintained by David Woodhouse
<dwmw2@infradead.org>.
linux-7.2.3/Documentation/kbuild/index.rst
+26
View File
@@ -0,0 +1,26 @@
.. SPDX-License-Identifier: GPL-2.0
===================
Kernel Build System
===================
.. toctree::
:maxdepth: 1
kconfig-language
kconfig-macro-language
kbuild
kconfig
makefiles
modules
headers_install
issues
reproducible-builds
gcc-plugins
llvm
gendwarfksyms
bash-completion
linux-7.2.3/Documentation/kbuild/issues.rst
+15
View File
@@ -0,0 +1,15 @@
================
Recursion issues
================
issue #1
--------
.. literalinclude:: Kconfig.recursion-issue-01
:language: kconfig
issue #2
--------
.. literalinclude:: Kconfig.recursion-issue-02
:language: kconfig
linux-7.2.3/Documentation/kbuild/kbuild.rst
+349
View File
@@ -0,0 +1,349 @@
======
Kbuild
======
Output files
============
modules.order
-------------
This file records the order in which modules appear in Makefiles. This
is used by modprobe to deterministically resolve aliases that match
multiple modules.
modules.builtin
---------------
This file lists all modules that are built into the kernel. This is used
by modprobe to not fail when trying to load something builtin.
modules.builtin.modinfo
-----------------------
This file contains modinfo from all modules that are built into the kernel.
Unlike modinfo of a separate module, all fields are prefixed with module name.
modules.builtin.ranges
----------------------
This file contains address offset ranges (per ELF section) for all modules
that are built into the kernel. Together with System.map, it can be used
to associate module names with symbols.
Environment variables
=====================
KCPPFLAGS
---------
Additional options to pass when preprocessing. The preprocessing options
will be used in all cases where kbuild does preprocessing including
building C files and assembler files.
KAFLAGS
-------
Additional options to the assembler (for built-in and modules).
AFLAGS_MODULE
-------------
Additional assembler options for modules.
AFLAGS_KERNEL
-------------
Additional assembler options for built-in.
KCFLAGS
-------
Additional options to the C compiler (for built-in and modules).
KRUSTFLAGS
----------
Additional options to the Rust compiler (for built-in and modules).
CFLAGS_KERNEL
-------------
Additional options for $(CC) when used to compile
code that is compiled as built-in.
CFLAGS_MODULE
-------------
Additional module specific options to use for $(CC).
RUSTFLAGS_KERNEL
----------------
Additional options for $(RUSTC) when used to compile
code that is compiled as built-in.
RUSTFLAGS_MODULE
----------------
Additional module specific options to use for $(RUSTC).
LDFLAGS_MODULE
--------------
Additional options used for $(LD) when linking modules.
HOSTCFLAGS
----------
Additional flags to be passed to $(HOSTCC) when building host programs.
HOSTCXXFLAGS
------------
Additional flags to be passed to $(HOSTCXX) when building host programs.
HOSTRUSTFLAGS
-------------
Additional flags to be passed to $(HOSTRUSTC) when building host programs.
PROCMACROLDFLAGS
----------------
Flags to be passed when linking Rust proc macros. Since proc macros are loaded
by rustc at build time, they must be linked in a way that is compatible with
the rustc toolchain being used.
For instance, it can be useful when rustc uses a different C library than
the one the user wants to use for host programs.
If unset, it defaults to the flags passed when linking host programs.
HOSTLDFLAGS
-----------
Additional flags to be passed when linking host programs.
HOSTLDLIBS
----------
Additional libraries to link against when building host programs.
.. _userkbuildflags:
USERCFLAGS
----------
Additional options used for $(CC) when compiling userprogs.
USERLDFLAGS
-----------
Additional options used for $(LD) when linking userprogs. userprogs are linked
with CC, so $(USERLDFLAGS) should include "-Wl," prefix as applicable.
KBUILD_KCONFIG
--------------
Set the top-level Kconfig file to the value of this environment
variable. The default name is "Kconfig".
KBUILD_VERBOSE
--------------
Set the kbuild verbosity. Can be assigned same values as "V=...".
See make help for the full list.
Setting "V=..." takes precedence over KBUILD_VERBOSE.
KBUILD_EXTMOD
-------------
Set the directory to look for the kernel source when building external
modules.
Setting "M=..." takes precedence over KBUILD_EXTMOD.
KBUILD_OUTPUT
-------------
Specify the output directory when building the kernel.
This variable can also be used to point to the kernel output directory when
building external modules against a pre-built kernel in a separate build
directory. Please note that this does NOT specify the output directory for the
external modules themselves. (Use KBUILD_EXTMOD_OUTPUT for that purpose.)
The output directory can also be specified using "O=...".
Setting "O=..." takes precedence over KBUILD_OUTPUT.
KBUILD_EXTMOD_OUTPUT
--------------------
Specify the output directory for external modules.
Setting "MO=..." takes precedence over KBUILD_EXTMOD_OUTPUT.
KBUILD_EXTRA_WARN
-----------------
Specify the extra build checks. The same value can be assigned by passing
W=... from the command line.
See `make help` for the list of the supported values.
Setting "W=..." takes precedence over KBUILD_EXTRA_WARN.
KBUILD_DEBARCH
--------------
For the deb-pkg target, allows overriding the normal heuristics deployed by
deb-pkg. Normally deb-pkg attempts to guess the right architecture based on
the UTS_MACHINE variable, and on some architectures also the kernel config.
The value of KBUILD_DEBARCH is assumed (not checked) to be a valid Debian
architecture.
KDOCFLAGS
---------
Specify extra (warning/error) flags for kernel-doc checks during the build,
see tools/docs/kernel-doc for which flags are supported. Note that this doesn't
(currently) apply to documentation builds.
ARCH
----
Set ARCH to the architecture to be built.
In most cases the name of the architecture is the same as the
directory name found in the arch/ directory.
But some architectures such as x86 and sparc have aliases.
- x86: i386 for 32 bit, x86_64 for 64 bit
- parisc: parisc64 for 64 bit
- sparc: sparc32 for 32 bit, sparc64 for 64 bit
CROSS_COMPILE
-------------
Specify an optional fixed part of the binutils filename.
CROSS_COMPILE can be a part of the filename or the full path.
CROSS_COMPILE is also used for ccache in some setups.
CF
--
Additional options for sparse.
CF is often used on the command-line like this::
make CF=-Wbitwise C=2
INSTALL_PATH
------------
INSTALL_PATH specifies where to place the updated kernel and system map
images. Default is /boot, but you can set it to other values.
INSTALLKERNEL
-------------
Install script called when using "make install".
The default name is "installkernel".
The script will be called with the following arguments:
- $1 - kernel version
- $2 - kernel image file
- $3 - kernel map file
- $4 - default install path (use root directory if blank)
The implementation of "make install" is architecture specific
and it may differ from the above.
INSTALLKERNEL is provided to enable the possibility to
specify a custom installer when cross compiling a kernel.
MODLIB
------
Specify where to install modules.
The default value is::
$(INSTALL_MOD_PATH)/lib/modules/$(KERNELRELEASE)
The value can be overridden in which case the default value is ignored.
INSTALL_MOD_PATH
----------------
INSTALL_MOD_PATH specifies a prefix to MODLIB for module directory
relocations required by build roots. This is not defined in the
makefile but the argument can be passed to make if needed.
INSTALL_MOD_STRIP
-----------------
INSTALL_MOD_STRIP, if defined, will cause modules to be
stripped after they are installed. If INSTALL_MOD_STRIP is '1', then
the default option --strip-debug will be used. Otherwise,
INSTALL_MOD_STRIP value will be used as the options to the strip command.
INSTALL_HDR_PATH
----------------
INSTALL_HDR_PATH specifies where to install user space headers when
executing "make headers_*".
The default value is::
$(objtree)/usr
$(objtree) is the directory where output files are saved.
The output directory is often set using "O=..." on the commandline.
The value can be overridden in which case the default value is ignored.
INSTALL_DTBS_PATH
-----------------
INSTALL_DTBS_PATH specifies where to install device tree blobs for
relocations required by build roots. This is not defined in the
makefile but the argument can be passed to make if needed.
KBUILD_ABS_SRCTREE
--------------------------------------------------
Kbuild uses a relative path to point to the tree when possible. For instance,
when building in the source tree, the source tree path is '.'
Setting this flag requests Kbuild to use absolute path to the source tree.
There are some useful cases to do so, like when generating tag files with
absolute path entries etc.
KBUILD_SIGN_PIN
---------------
This variable allows a passphrase or PIN to be passed to the sign-file
utility when signing kernel modules, if the private key requires such.
KBUILD_MODPOST_WARN
-------------------
KBUILD_MODPOST_WARN can be set to avoid errors in case of undefined
symbols in the final module linking stage. It changes such errors
into warnings.
KBUILD_MODPOST_NOFINAL
----------------------
KBUILD_MODPOST_NOFINAL can be set to skip the final link of modules.
This is solely useful to speed up test compiles.
KBUILD_EXTRA_SYMBOLS
--------------------
For modules that use symbols from other modules.
See more details in modules.rst.
ALLSOURCE_ARCHS
---------------
For tags/TAGS/cscope targets, you can specify more than one arch
to be included in the databases, separated by blank space. E.g.::
$ make ALLSOURCE_ARCHS="x86 mips arm" tags
To get all available archs you can also specify all. E.g.::
$ make ALLSOURCE_ARCHS=all tags
IGNORE_DIRS
-----------
For tags/TAGS/cscope targets, you can choose which directories won't
be included in the databases, separated by blank space. E.g.::
$ make IGNORE_DIRS="drivers/gpu/drm/radeon tools" cscope
KBUILD_BUILD_TIMESTAMP
----------------------
Setting this to a date string overrides the timestamp used in the
UTS_VERSION definition (uname -v in the running kernel). The value has to
be a string that can be passed to date -d. E.g.::
$ KBUILD_BUILD_TIMESTAMP="Mon Oct 13 00:00:00 UTC 2025" make
The default value is the output of the date command at one point during
build. If provided, this timestamp will also be used for mtime fields
within any initramfs archive. Initramfs mtimes are 32-bit, so dates before
the 1970 Unix epoch, or after 2106-02-07 06:28:15 UTC will fail.
KBUILD_BUILD_USER, KBUILD_BUILD_HOST
------------------------------------
These two variables allow to override the user@host string displayed during
boot and in /proc/version. The default value is the output of the commands
whoami and host, respectively.
LLVM
----
If this variable is set to 1, Kbuild will use Clang and LLVM utilities instead
of GCC and GNU binutils to build the kernel.
linux-7.2.3/Documentation/kbuild/kconfig-language.rst
+827
View File
@@ -0,0 +1,827 @@
================
Kconfig Language
================
Introduction
------------
The configuration database is a collection of configuration options
organized in a tree structure::
+- Code maturity level options
| +- Prompt for development and/or incomplete code/drivers
+- General setup
| +- Networking support
| +- System V IPC
| +- BSD Process Accounting
| +- Sysctl support
+- Loadable module support
| +- Enable loadable module support
| +- Set version information on all module symbols
| +- Kernel module loader
+- ...
Every entry has its own dependencies. These dependencies are used
to determine the visibility of an entry. Any child entry is only
visible if its parent entry is also visible.
Menu entries
------------
Most entries define a config option; all other entries help to organize
them. A single configuration option is defined like this::
config MODVERSIONS
bool "Set version information on all module symbols"
depends on MODULES
help
Usually, modules have to be recompiled whenever you switch to a new
kernel. ...
Every line starts with a key word and can be followed by multiple
arguments. "config" starts a new config entry. The following lines
define attributes for this config option. Attributes can be the type of
the config option, input prompt, dependencies, help text and default
values. A config option can be defined multiple times with the same
name, but every definition can have only a single input prompt and the
type must not conflict.
Menu attributes
---------------
A menu entry can have a number of attributes. Not all of them are
applicable everywhere (see syntax).
- type definition: "bool"/"tristate"/"string"/"hex"/"int"
Every config option must have a type. There are only two basic types:
tristate and string; the other types are based on these two. The type
definition optionally accepts an input prompt, so these two examples
are equivalent::
bool "Networking support"
and::
bool
prompt "Networking support"
- input prompt: "prompt" <prompt> ["if" <expr>]
Every menu entry can have at most one prompt, which is used to display
to the user. Optionally dependencies only for this prompt can be added
with "if". If a prompt is not present, the config option is a non-visible
symbol, meaning its value cannot be directly changed by the user (such as
altering the value in ``.config``) and the option will not appear in any
config menus. Its value can only be set via "default" and "select" (see
below).
- default value: "default" <expr> ["if" <expr>]
A config option can have any number of default values. If multiple
default values are visible, only the first defined one is active.
Default values are not limited to the menu entry where they are
defined. This means the default can be defined somewhere else or be
overridden by an earlier definition.
The default value is only assigned to the config symbol if no other
value was set by the user (via the input prompt above). If an input
prompt is visible the default value is presented to the user and can
be overridden by him.
Optionally, dependencies only for this default value can be added with
"if".
The default value deliberately defaults to 'n' in order to avoid bloating the
build. With few exceptions, new config options should not change this. The
intent is for "make oldconfig" to add as little as possible to the config from
release to release.
Note:
Things that merit "default y/m" include:
a) A new Kconfig option for something that used to always be built
should be "default y".
b) A new gatekeeping Kconfig option that hides/shows other Kconfig
options (but does not generate any code of its own), should be
"default y" so people will see those other options.
c) Sub-driver behavior or similar options for a driver that is
"default n". This allows you to provide sane defaults.
d) Hardware or infrastructure that everybody expects, such as CONFIG_NET
or CONFIG_BLOCK. These are rare exceptions.
- type definition + default value::
"def_bool"/"def_tristate" <expr> ["if" <expr>]
This is a shorthand notation for a type definition plus a value.
Optionally dependencies for this default value can be added with "if".
- dependencies: "depends on" <expr> ["if" <expr>]
This defines a dependency for this menu entry. If multiple
dependencies are defined, they are connected with '&&'. Dependencies
are applied to all other options within this menu entry (which also
accept an "if" expression), so these two examples are equivalent::
bool "foo" if BAR
default y if BAR
and::
depends on BAR
bool "foo"
default y
The dependency definition itself may be conditional by appending "if"
followed by an expression. For example::
config FOO
tristate
depends on BAR if BAZ
meaning that FOO is constrained by the value of BAR only if BAZ is
also set.
- reverse dependencies: "select" <symbol> ["if" <expr>]
While normal dependencies reduce the upper limit of a symbol (see
below), reverse dependencies can be used to force a lower limit of
another symbol. The value of the current menu symbol is used as the
minimal value <symbol> can be set to. If <symbol> is selected multiple
times, the limit is set to the largest selection.
Reverse dependencies can only be used with boolean or tristate
symbols.
Note:
select should be used with care. select will force
a symbol to a value without visiting the dependencies.
By abusing select you are able to select a symbol FOO even
if FOO depends on BAR that is not set.
In general use select only for non-visible symbols
(no prompts anywhere) and for symbols with no dependencies.
That will limit the usefulness but on the other hand avoid
the illegal configurations all over.
If "select" <symbol> is followed by "if" <expr>, <symbol> will be
selected by the logical AND of the value of the current menu symbol
and <expr>. This means, the lower limit can be downgraded due to the
presence of "if" <expr>. This behavior may seem weird, but we rely on
it. (The future of this behavior is undecided.)
- weak reverse dependencies: "imply" <symbol> ["if" <expr>]
This is similar to "select" as it enforces a lower limit on another
symbol except that the "implied" symbol's value may still be set to n
from a direct dependency or with a visible prompt.
Given the following example::
config FOO
tristate "foo"
imply BAZ
config BAZ
tristate "baz"
depends on BAR
The following values are possible:
=== === ============= ==============
FOO BAR BAZ's default choice for BAZ
=== === ============= ==============
n y n N/m/y
m y m M/y/n
y y y Y/m/n
n m n N/m
m m m M/n
y m m M/n
y n * N
=== === ============= ==============
This is useful e.g. with multiple drivers that want to indicate their
ability to hook into a secondary subsystem while allowing the user to
configure that subsystem out without also having to unset these drivers.
Note: If the feature provided by BAZ is highly desirable for FOO,
FOO should imply not only BAZ, but also its dependency BAR::
config FOO
tristate "foo"
imply BAR
imply BAZ
Note: If "imply" <symbol> is followed by "if" <expr>, the default of <symbol>
will be the logical AND of the value of the current menu symbol and <expr>.
(The future of this behavior is undecided.)
- limiting menu display: "visible if" <expr>
This attribute is only applicable to menu blocks, if the condition is
false, the menu block is not displayed to the user (the symbols
contained there can still be selected by other symbols, though). It is
similar to a conditional "prompt" attribute for individual menu
entries. Default value of "visible" is true.
- numerical ranges: "range" <symbol> <symbol> ["if" <expr>]
This allows limiting the range of possible input values for int
and hex symbols. The user can only input a value which is larger than
or equal to the first symbol and smaller than or equal to the second
symbol.
- help text: "help"
This defines a help text. The end of the help text is determined by
the indentation level, this means it ends at the first line which has
a smaller indentation than the first line of the help text.
- module attribute: "modules"
This declares the symbol to be used as the MODULES symbol, which
enables the third modular state for all config symbols.
At most one symbol may have the "modules" option set.
- transitional attribute: "transitional"
This declares the symbol as transitional, meaning it should be processed
during configuration but omitted from newly written .config files.
Transitional symbols are useful for backward compatibility during config
option migrations - they allow olddefconfig to process existing .config
files while ensuring the old option doesn't appear in new configurations.
A transitional symbol:
- Has no prompt (is not visible to users in menus)
- Is processed normally during configuration (values are read and used)
- Can be referenced in default expressions of other symbols
- Is not written to new .config files
- Cannot have any other properties (it is a pass-through option)
Example migration from OLD_NAME to NEW_NAME::
config NEW_NAME
bool "New option name"
default OLD_NAME
help
This replaces the old CONFIG_OLD_NAME option.
config OLD_NAME
bool
transitional
help
Transitional config for OLD_NAME to NEW_NAME migration.
With this setup, existing .config files with "CONFIG_OLD_NAME=y" will
result in "CONFIG_NEW_NAME=y" being set, while CONFIG_OLD_NAME will be
omitted from newly written .config files.
Menu dependencies
-----------------
Dependencies define the visibility of a menu entry and can also reduce
the input range of tristate symbols. The tristate logic used in the
expressions uses one more state than normal boolean logic to express the
module state. Dependency expressions have the following syntax::
<expr> ::= <symbol> (1)
<symbol> '=' <symbol> (2)
<symbol> '!=' <symbol> (3)
<symbol1> '<' <symbol2> (4)
<symbol1> '>' <symbol2> (4)
<symbol1> '<=' <symbol2> (4)
<symbol1> '>=' <symbol2> (4)
'(' <expr> ')' (5)
'!' <expr> (6)
<expr> '&&' <expr> (7)
<expr> '||' <expr> (8)
Expressions are listed in decreasing order of precedence.
(1) Convert the symbol into an expression. Boolean and tristate symbols
are simply converted into the respective expression values. All
other symbol types result in 'n'.
(2) If the values of both symbols are equal, it returns 'y',
otherwise 'n'.
(3) If the values of both symbols are equal, it returns 'n',
otherwise 'y'.
(4) If value of <symbol1> is respectively lower, greater, lower-or-equal,
or greater-or-equal than value of <symbol2>, it returns 'y',
otherwise 'n'.
(5) Returns the value of the expression. Used to override precedence.
(6) Returns the result of (2-/expr/).
(7) Returns the result of min(/expr/, /expr/).
(8) Returns the result of max(/expr/, /expr/).
An expression can have a value of 'n', 'm' or 'y' (or 0, 1, 2
respectively for calculations). A menu entry becomes visible when its
expression evaluates to 'm' or 'y'.
There are two types of symbols: constant and non-constant symbols.
Non-constant symbols are the most common ones and are defined with the
'config' statement. Non-constant symbols consist entirely of alphanumeric
characters or underscores.
Constant symbols are only part of expressions. Constant symbols are
always surrounded by single or double quotes. Within the quote, any
other character is allowed and the quotes can be escaped using '\'.
Menu structure
--------------
The position of a menu entry in the tree is determined in two ways. First
it can be specified explicitly::
menu "Network device support"
depends on NET
config NETDEVICES
...
endmenu
All entries within the "menu" ... "endmenu" block become a submenu of
"Network device support". All subentries inherit the dependencies from
the menu entry, e.g. this means the dependency "NET" is added to the
dependency list of the config option NETDEVICES.
The other way to generate the menu structure is done by analyzing the
dependencies. If a menu entry somehow depends on the previous entry, it
can be made a submenu of it. First, the previous (parent) symbol must
be part of the dependency list and then one of these two conditions
must be true:
- the child entry must become invisible, if the parent is set to 'n'
- the child entry must only be visible, if the parent is visible::
config MODULES
bool "Enable loadable module support"
config MODVERSIONS
bool "Set version information on all module symbols"
depends on MODULES
comment "module support disabled"
depends on !MODULES
MODVERSIONS directly depends on MODULES, this means it's only visible if
MODULES is different from 'n'. The comment on the other hand is only
visible when MODULES is set to 'n'.
Kconfig syntax
--------------
The configuration file describes a series of menu entries, where every
line starts with a keyword (except help texts). The following keywords
end a menu entry:
- config
- menuconfig
- choice/endchoice
- comment
- menu/endmenu
- if/endif
- source
The first five also start the definition of a menu entry.
config::
"config" <symbol>
<config options>
This defines a config symbol <symbol> and accepts any of above
attributes as options.
menuconfig::
"menuconfig" <symbol>
<config options>
This is similar to the simple config entry above, but it also gives a
hint to front ends, that all suboptions should be displayed as a
separate list of options. To make sure all the suboptions will really
show up under the menuconfig entry and not outside of it, every item
from the <config options> list must depend on the menuconfig symbol.
In practice, this is achieved by using one of the next two constructs::
(1):
menuconfig M
if M
config C1
config C2
endif
(2):
menuconfig M
config C1
depends on M
config C2
depends on M
In the following examples (3) and (4), C1 and C2 still have the M
dependency, but will not appear under menuconfig M anymore, because
of C0, which doesn't depend on M::
(3):
menuconfig M
config C0
if M
config C1
config C2
endif
(4):
menuconfig M
config C0
config C1
depends on M
config C2
depends on M
choices::
"choice"
<choice options>
<choice block>
"endchoice"
This defines a choice group and accepts "prompt", "default", "depends on", and
"help" attributes as options.
A choice only allows a single config entry to be selected.
comment::
"comment" <prompt>
<comment options>
This defines a comment which is displayed to the user during the
configuration process and is also echoed to the output files. The only
possible options are dependencies.
menu::
"menu" <prompt>
<menu options>
<menu block>
"endmenu"
This defines a menu block, see "Menu structure" above for more
information. The only possible options are dependencies and "visible"
attributes.
if::
"if" <expr>
<if block>
"endif"
This defines an if block. The dependency expression <expr> is appended
to all enclosed menu entries.
source::
"source" <prompt>
This reads the specified configuration file. This file is always parsed.
mainmenu::
"mainmenu" <prompt>
This sets the config program's title bar if the config program chooses
to use it. It should be placed at the top of the configuration, before any
other statement.
'#' Kconfig source file comment:
An unquoted '#' character anywhere in a source file line indicates
the beginning of a source file comment. The remainder of that line
is a comment.
Kconfig hints
-------------
This is a collection of Kconfig tips, most of which aren't obvious at
first glance and most of which have become idioms in several Kconfig
files.
Adding common features and make the usage configurable
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
It is a common idiom to implement a feature/functionality that are
relevant for some architectures but not all.
The recommended way to do so is to use a config variable named HAVE_*
that is defined in a common Kconfig file and selected by the relevant
architectures.
An example is the generic IOMAP functionality.
We would in lib/Kconfig see::
# Generic IOMAP is used to ...
config HAVE_GENERIC_IOMAP
config GENERIC_IOMAP
depends on HAVE_GENERIC_IOMAP && FOO
And in lib/Makefile we would see::
obj-$(CONFIG_GENERIC_IOMAP) += iomap.o
For each architecture using the generic IOMAP functionality we would see::
config X86
select ...
select HAVE_GENERIC_IOMAP
select ...
Note: we use the existing config option and avoid creating a new
config variable to select HAVE_GENERIC_IOMAP.
Note: the use of the internal config variable HAVE_GENERIC_IOMAP, it is
introduced to overcome the limitation of select which will force a
config option to 'y' no matter the dependencies.
The dependencies are moved to the symbol GENERIC_IOMAP and we avoid the
situation where select forces a symbol equals to 'y'.
Adding features that need compiler support
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
There are several features that need compiler support. The recommended way
to describe the dependency on the compiler feature is to use "depends on"
followed by a test macro::
config STACKPROTECTOR
bool "Stack Protector buffer overflow detection"
depends on $(cc-option,-fstack-protector)
...
If you need to expose a compiler capability to makefiles and/or C source files,
`CC_HAS_` is the recommended prefix for the config option::
config CC_HAS_FOO
def_bool $(success,$(srctree)/scripts/cc-check-foo.sh $(CC))
Build as module only
~~~~~~~~~~~~~~~~~~~~
To restrict a component build to module-only, qualify its config symbol
with "depends on m". E.g.::
config FOO
depends on BAR && m
limits FOO to module (=m) or disabled (=n).
Compile-testing
~~~~~~~~~~~~~~~
If a config symbol has a dependency, but the code controlled by the config
symbol can still be compiled if the dependency is not met, it is encouraged to
increase build coverage by adding an "|| COMPILE_TEST" clause to the
dependency. This is especially useful for drivers for more exotic hardware, as
it allows continuous-integration systems to compile-test the code on a more
common system, and detect bugs that way.
Note that compile-tested code should avoid crashing when run on a system where
the dependency is not met.
Architecture and platform dependencies
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Due to the presence of stubs, most drivers can now be compiled on most
architectures. However, this does not mean it makes sense to have all drivers
available everywhere, as the actual hardware may only exist on specific
architectures and platforms. This is especially true for on-SoC IP cores,
which may be limited to a specific vendor or SoC family.
To prevent asking the user about drivers that cannot be used on the system(s)
the user is compiling a kernel for, and if it makes sense, config symbols
controlling the compilation of a driver should contain proper dependencies,
limiting the visibility of the symbol to (a superset of) the platform(s) the
driver can be used on. The dependency can be an architecture (e.g. ARM) or
platform (e.g. ARCH_OMAP4) dependency. This makes life simpler not only for
distro config owners, but also for every single developer or user who
configures a kernel.
Such a dependency can be relaxed by combining it with the compile-testing rule
above, leading to:
config FOO
bool "Support for foo hardware"
depends on ARCH_FOO_VENDOR || COMPILE_TEST
Optional dependencies
~~~~~~~~~~~~~~~~~~~~~
Some drivers are able to optionally use a feature from another module
or build cleanly with that module disabled, but cause a link failure
when trying to use that loadable module from a built-in driver.
The recommended way to express this optional dependency in Kconfig logic
uses the conditional form::
config FOO
tristate "Support for foo hardware"
depends on BAR if BAR
This slightly counterintuitive style is also widely used::
config FOO
tristate "Support for foo hardware"
depends on BAR || !BAR
This means that there is either a dependency on BAR that disallows
the combination of FOO=y with BAR=m, or BAR is completely disabled. The BAR
module must provide all the stubs for !BAR case.
For a more formalized approach if there are multiple drivers that have
the same dependency, a helper symbol can be used, like::
config FOO
tristate "Support for foo hardware"
depends on BAR_OPTIONAL
config BAR_OPTIONAL
def_tristate BAR || !BAR
Much less favorable way to express optional dependency is IS_REACHABLE() within
the module code, useful for example when the module BAR does not provide
!BAR stubs::
foo_init()
{
if (IS_REACHABLE(CONFIG_BAR))
bar_register(&foo);
...
}
IS_REACHABLE() is generally discouraged, because the code will be silently
discarded, when CONFIG_BAR=m and this code is built-in. This is not what users
usually expect when enabling BAR as module.
Kconfig recursive dependency limitations
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If you've hit the Kconfig error: "recursive dependency detected" you've run
into a recursive dependency issue with Kconfig, a recursive dependency can be
summarized as a circular dependency. The kconfig tools need to ensure that
Kconfig files comply with specified configuration requirements. In order to do
that kconfig must determine the values that are possible for all Kconfig
symbols, this is currently not possible if there is a circular relation
between two or more Kconfig symbols. For more details refer to the "Simple
Kconfig recursive issue" subsection below. Kconfig does not do recursive
dependency resolution; this has a few implications for Kconfig file writers.
We'll first explain why this issues exists and then provide an example
technical limitation which this brings upon Kconfig developers. Eager
developers wishing to try to address this limitation should read the next
subsections.
Simple Kconfig recursive issue
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Read: Documentation/kbuild/Kconfig.recursion-issue-01
Test with::
make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-01 allnoconfig
Cumulative Kconfig recursive issue
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Read: Documentation/kbuild/Kconfig.recursion-issue-02
Test with::
make KBUILD_KCONFIG=Documentation/kbuild/Kconfig.recursion-issue-02 allnoconfig
Practical solutions to kconfig recursive issue
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Developers who run into the recursive Kconfig issue have two options
at their disposal. We document them below and also provide a list of
historical issues resolved through these different solutions.
a) Remove any superfluous "select FOO" or "depends on FOO"
b) Match dependency semantics:
b1) Swap all "select FOO" to "depends on FOO" or,
b2) Swap all "depends on FOO" to "select FOO"
The resolution to a) can be tested with the sample Kconfig file
Documentation/kbuild/Kconfig.recursion-issue-01 through the removal
of the "select CORE" from CORE_BELL_A_ADVANCED as that is implicit already
since CORE_BELL_A depends on CORE. At times it may not be possible to remove
some dependency criteria, for such cases you can work with solution b).
The two different resolutions for b) can be tested in the sample Kconfig file
Documentation/kbuild/Kconfig.recursion-issue-02.
Below is a list of examples of prior fixes for these types of recursive issues;
all errors appear to involve one or more "select" statements and one or more
"depends on".
============ ===================================
commit fix
============ ===================================
06b718c01208 select A -> depends on A
c22eacfe82f9 depends on A -> depends on B
6a91e854442c select A -> depends on A
118c565a8f2e select A -> select B
f004e5594705 select A -> depends on A
c7861f37b4c6 depends on A -> (null)
80c69915e5fb select A -> (null) (1)
c2218e26c0d0 select A -> depends on A (1)
d6ae99d04e1c select A -> depends on A
95ca19cf8cbf select A -> depends on A
8f057d7bca54 depends on A -> (null)
8f057d7bca54 depends on A -> select A
a0701f04846e select A -> depends on A
0c8b92f7f259 depends on A -> (null)
e4e9e0540928 select A -> depends on A (2)
7453ea886e87 depends on A > (null) (1)
7b1fff7e4fdf select A -> depends on A
86c747d2a4f0 select A -> depends on A
d9f9ab51e55e select A -> depends on A
0c51a4d8abd6 depends on A -> select A (3)
e98062ed6dc4 select A -> depends on A (3)
91e5d284a7f1 select A -> (null)
============ ===================================
(1) Partial (or no) quote of error.
(2) That seems to be the gist of that fix.
(3) Same error.
Future kconfig work
~~~~~~~~~~~~~~~~~~~
Work on kconfig is welcomed on both areas of clarifying semantics and on
evaluating the use of a full SAT solver for it. A full SAT solver can be
desirable to enable more complex dependency mappings and / or queries,
for instance one possible use case for a SAT solver could be that of handling
the current known recursive dependency issues. It is not known if this would
address such issues but such evaluation is desirable. If support for a full SAT
solver proves too complex or that it cannot address recursive dependency issues
Kconfig should have at least clear and well defined semantics which also
addresses and documents limitations or requirements such as the ones dealing
with recursive dependencies.
Further work on both of these areas is welcomed on Kconfig. We elaborate
on both of these in the next two subsections.
Semantics of Kconfig
~~~~~~~~~~~~~~~~~~~~
The use of Kconfig is broad, Linux is now only one of Kconfig's users:
one study has completed a broad analysis of Kconfig use in 12 projects [0]_.
Despite its widespread use, and although this document does a reasonable job
in documenting basic Kconfig syntax a more precise definition of Kconfig
semantics is welcomed. One project deduced Kconfig semantics through
the use of the xconfig configurator [1]_. Work should be done to confirm if
the deduced semantics matches our intended Kconfig design goals.
Another project formalized a denotational semantics of a core subset of
the Kconfig language [10]_.
Having well defined semantics can be useful for tools for practical
evaluation of dependencies, for instance one such case was work to
express in boolean abstraction of the inferred semantics of Kconfig to
translate Kconfig logic into boolean formulas and run a SAT solver on this to
find dead code / features (always inactive), 114 dead features were found in
Linux using this methodology [1]_ (Section 8: Threats to validity).
The kismet tool, based on the semantics in [10]_, finds abuses of reverse
dependencies and has led to dozens of committed fixes to Linux Kconfig files [11]_.
Confirming this could prove useful as Kconfig stands as one of the leading
industrial variability modeling languages [1]_ [2]_. Its study would help
evaluate practical uses of such languages, their use was only theoretical
and real world requirements were not well understood. As it stands though
only reverse engineering techniques have been used to deduce semantics from
variability modeling languages such as Kconfig [3]_.
.. [0] https://www.eng.uwaterloo.ca/~shshe/kconfig_semantics.pdf
.. [1] https://gsd.uwaterloo.ca/sites/default/files/vm-2013-berger.pdf
.. [2] https://gsd.uwaterloo.ca/sites/default/files/ase241-berger_0.pdf
.. [3] https://gsd.uwaterloo.ca/sites/default/files/icse2011.pdf
Full SAT solver for Kconfig
~~~~~~~~~~~~~~~~~~~~~~~~~~~
Although SAT solvers [4]_ haven't yet been used by Kconfig directly, as noted
in the previous subsection, work has been done however to express in boolean
abstraction the inferred semantics of Kconfig to translate Kconfig logic into
boolean formulas and run a SAT solver on it [5]_. Another known related project
is CADOS [6]_ (former VAMOS [7]_) and the tools, mainly undertaker [8]_, which
has been introduced first with [9]_. The basic concept of undertaker is to
extract variability models from Kconfig and put them together with a
propositional formula extracted from CPP #ifdefs and build-rules into a SAT
solver in order to find dead code, dead files, and dead symbols. If using a SAT
solver is desirable on Kconfig one approach would be to evaluate repurposing
such efforts somehow on Kconfig. There is enough interest from mentors of
existing projects to not only help advise how to integrate this work upstream
but also help maintain it long term. Interested developers should visit:
https://kernelnewbies.org/KernelProjects/kconfig-sat
.. [4] https://www.cs.cornell.edu/~sabhar/chapters/SATSolvers-KR-Handbook.pdf
.. [5] https://gsd.uwaterloo.ca/sites/default/files/vm-2013-berger.pdf
.. [6] https://cados.cs.fau.de
.. [7] https://vamos.cs.fau.de
.. [8] https://undertaker.cs.fau.de
.. [9] https://www4.cs.fau.de/Publications/2011/tartler_11_eurosys.pdf
.. [10] https://paulgazzillo.com/papers/esecfse21.pdf
.. [11] https://github.com/paulgazz/kmax
linux-7.2.3/Documentation/kbuild/kconfig-macro-language.rst
+247
View File
@@ -0,0 +1,247 @@
======================
Kconfig macro language
======================
Concept
-------
The basic idea was inspired by Make. When we look at Make, we notice sort of
two languages in one. One language describes dependency graphs consisting of
targets and prerequisites. The other is a macro language for performing textual
substitution.
There is clear distinction between the two language stages. For example, you
can write a makefile like follows::
APP := foo
SRC := foo.c
CC := gcc
$(APP): $(SRC)
$(CC) -o $(APP) $(SRC)
The macro language replaces the variable references with their expanded form,
and handles as if the source file were input like follows::
foo: foo.c
gcc -o foo foo.c
Then, Make analyzes the dependency graph and determines the targets to be
updated.
The idea is quite similar in Kconfig - it is possible to describe a Kconfig
file like this::
CC := gcc
config CC_HAS_FOO
def_bool $(shell, $(srctree)/scripts/gcc-check-foo.sh $(CC))
The macro language in Kconfig processes the source file into the following
intermediate::
config CC_HAS_FOO
def_bool y
Then, Kconfig moves onto the evaluation stage to resolve inter-symbol
dependency as explained in kconfig-language.rst.
Variables
---------
Like in Make, a variable in Kconfig works as a macro variable. A macro
variable is expanded "in place" to yield a text string that may then be
expanded further. To get the value of a variable, enclose the variable name in
$( ). The parentheses are required even for single-letter variable names; $X is
a syntax error. The curly brace form as in ${CC} is not supported either.
There are two types of variables: simply expanded variables and recursively
expanded variables.
A simply expanded variable is defined using the := assignment operator. Its
righthand side is expanded immediately upon reading the line from the Kconfig
file.
A recursively expanded variable is defined using the = assignment operator.
Its righthand side is simply stored as the value of the variable without
expanding it in any way. Instead, the expansion is performed when the variable
is used.
There is another type of assignment operator; += is used to append text to a
variable. The righthand side of += is expanded immediately if the lefthand
side was originally defined as a simple variable. Otherwise, its evaluation is
deferred.
The variable reference can take parameters, in the following form::
$(name,arg1,arg2,arg3)
You can consider the parameterized reference as a function. (more precisely,
"user-defined function" in contrast to "built-in function" listed below).
Useful functions must be expanded when they are used since the same function is
expanded differently if different parameters are passed. Hence, a user-defined
function is defined using the = assignment operator. The parameters are
referenced within the body definition with $(1), $(2), etc.
In fact, recursively expanded variables and user-defined functions are the same
internally. (In other words, "variable" is "function with zero argument".)
When we say "variable" in a broad sense, it includes "user-defined function".
Built-in functions
------------------
Like Make, Kconfig provides several built-in functions. Every function takes a
particular number of arguments.
In Make, every built-in function takes at least one argument. Kconfig allows
zero argument for built-in functions, such as $(filename), $(lineno). You could
consider those as "built-in variable", but it is just a matter of how we call
it after all. Let's say "built-in function" here to refer to natively supported
functionality.
Kconfig currently supports the following built-in functions.
- $(shell,command)
The "shell" function accepts a single argument that is expanded and passed
to a subshell for execution. The standard output of the command is then read
and returned as the value of the function. Every newline in the output is
replaced with a space. Any trailing newlines are deleted. The standard error
is not returned, nor is any program exit status.
- $(info,text)
The "info" function takes a single argument and prints it to stdout.
It evaluates to an empty string.
- $(warning-if,condition,text)
The "warning-if" function takes two arguments. If the condition part is "y",
the text part is sent to stderr. The text is prefixed with the name of the
current Kconfig file and the current line number.
- $(error-if,condition,text)
The "error-if" function is similar to "warning-if", but it terminates the
parsing immediately if the condition part is "y".
- $(filename)
The 'filename' takes no argument, and $(filename) is expanded to the file
name being parsed.
- $(lineno)
The 'lineno' takes no argument, and $(lineno) is expanded to the line number
being parsed.
Make vs Kconfig
---------------
Kconfig adopts Make-like macro language, but the function call syntax is
slightly different.
A function call in Make looks like this::
$(func-name arg1,arg2,arg3)
The function name and the first argument are separated by at least one
whitespace. Then, leading whitespaces are trimmed from the first argument,
while whitespaces in the other arguments are kept. You need to use a kind of
trick to start the first parameter with spaces. For example, if you want
to make "info" function print " hello", you can write like follows::
empty :=
space := $(empty) $(empty)
$(info $(space)$(space)hello)
Kconfig uses only commas for delimiters, and keeps all whitespaces in the
function call. Some people prefer putting a space after each comma delimiter::
$(func-name, arg1, arg2, arg3)
In this case, "func-name" will receive " arg1", " arg2", " arg3". The presence
of leading spaces may matter depending on the function. The same applies to
Make - for example, $(subst .c, .o, $(sources)) is a typical mistake; it
replaces ".c" with " .o".
In Make, a user-defined function is referenced by using a built-in function,
'call', like this::
$(call my-func,arg1,arg2,arg3)
Kconfig invokes user-defined functions and built-in functions in the same way.
The omission of 'call' makes the syntax shorter.
In Make, some functions treat commas verbatim instead of argument separators.
For example, $(shell echo hello, world) runs the command "echo hello, world".
Likewise, $(info hello, world) prints "hello, world" to stdout. You could say
this is _useful_ inconsistency.
In Kconfig, for simpler implementation and grammatical consistency, commas that
appear in the $( ) context are always delimiters. It means::
$(shell, echo hello, world)
is an error because it is passing two parameters where the 'shell' function
accepts only one. To pass commas in arguments, you can use the following trick::
comma := ,
$(shell, echo hello$(comma) world)
Caveats
-------
A variable (or function) cannot be expanded across tokens. So, you cannot use
a variable as a shorthand for an expression that consists of multiple tokens.
The following works::
RANGE_MIN := 1
RANGE_MAX := 3
config FOO
int "foo"
range $(RANGE_MIN) $(RANGE_MAX)
But, the following does not work::
RANGES := 1 3
config FOO
int "foo"
range $(RANGES)
A variable cannot be expanded to any keyword in Kconfig. The following does
not work::
MY_TYPE := tristate
config FOO
$(MY_TYPE) "foo"
default y
Obviously from the design, $(shell command) is expanded in the textual
substitution phase. You cannot pass symbols to the 'shell' function.
The following does not work as expected::
config ENDIAN_FLAG
string
default "-mbig-endian" if CPU_BIG_ENDIAN
default "-mlittle-endian" if CPU_LITTLE_ENDIAN
config CC_HAS_ENDIAN_FLAG
def_bool $(shell $(srctree)/scripts/gcc-check-flag ENDIAN_FLAG)
Instead, you can do like follows so that any function call is statically
expanded::
config CC_HAS_ENDIAN_FLAG
bool
default $(shell $(srctree)/scripts/gcc-check-flag -mbig-endian) if CPU_BIG_ENDIAN
default $(shell $(srctree)/scripts/gcc-check-flag -mlittle-endian) if CPU_LITTLE_ENDIAN
linux-7.2.3/Documentation/kbuild/kconfig.rst
+300
View File
@@ -0,0 +1,300 @@
=================================
Configuration targets and editors
=================================
This file contains some assistance for using ``make *config``.
Use ``make help`` to list all of the possible configuration targets.
The xconfig ('qconf'), menuconfig ('mconf'), and nconfig ('nconf')
programs also have embedded help text. Be sure to check that for
navigation, search, and other general help text.
The gconfig ('gconf') program has limited help text.
General
=======
New kernel releases often introduce new config symbols. Often more
important, new kernel releases may rename config symbols. When
this happens, using a previously working .config file and running
"make oldconfig" won't necessarily produce a working new kernel
for you, so you may find that you need to see what NEW kernel
symbols have been introduced.
To see a list of new config symbols, use::
cp user/some/old.config .config
make listnewconfig
and the config program will list any new symbols, one per line.
Alternatively, you can use the brute force method::
make oldconfig
scripts/diffconfig .config.old .config | less
Environment variables
=====================
Environment variables for ``*config``:
``KCONFIG_CONFIG``
This environment variable can be used to specify a default kernel config
file name to override the default name of ".config".
``KCONFIG_DEFCONFIG_LIST``
This environment variable specifies a list of config files which can be
used as a base configuration in case the .config does not exist yet.
Entries in the list are separated with whitespaces to each other, and
the first one that exists is used.
``KCONFIG_OVERWRITECONFIG``
If you set KCONFIG_OVERWRITECONFIG in the environment, Kconfig will not
break symlinks when .config is a symlink to somewhere else.
``KCONFIG_WARN_UNKNOWN_SYMBOLS``
This environment variable makes Kconfig warn about all unrecognized
symbols in the config input.
``KCONFIG_WARN_CHANGED_INPUT``
If set to a non-blank value, Kconfig prints optional warnings for
user-provided values that change after Kconfig resolves dependencies
or applies other constraints such as ranges.
``KCONFIG_WERROR``
If set, Kconfig treats warnings as errors.
``CONFIG_``
If you set ``CONFIG_`` in the environment, Kconfig will prefix all symbols
with its value when saving the configuration, instead of using the
default, ``CONFIG_``.
Environment variables for ``{allyes/allmod/allno/alldef/rand}config``:
``KCONFIG_ALLCONFIG``
The allyesconfig/allmodconfig/alldefconfig/allnoconfig/randconfig variants
can also use the environment variable KCONFIG_ALLCONFIG as a flag or a
filename that contains config symbols that the user requires to be set to a
specific value. If KCONFIG_ALLCONFIG is used without a filename where
KCONFIG_ALLCONFIG == "" or KCONFIG_ALLCONFIG == "1", ``make *config``
checks for a file named "all{yes/mod/no/def/random}.config"
(corresponding to the ``*config`` command that was used) for symbol values
that are to be forced. If this file is not found, it checks for a
file named "all.config" to contain forced values.
This enables you to create "miniature" config (miniconfig) or custom
config files containing just the config symbols that you are interested
in. Then the kernel config system generates the full .config file,
including symbols of your miniconfig file.
This ``KCONFIG_ALLCONFIG`` file is a config file which contains
(usually a subset of all) preset config symbols. These variable
settings are still subject to normal dependency checks.
Examples::
KCONFIG_ALLCONFIG=custom-notebook.config make allnoconfig
or::
KCONFIG_ALLCONFIG=mini.config make allnoconfig
or::
make KCONFIG_ALLCONFIG=mini.config allnoconfig
These examples will disable most options (allnoconfig) but enable or
disable the options that are explicitly listed in the specified
mini-config files.
Environment variables for ``randconfig``:
``KCONFIG_SEED``
You can set this to the integer value used to seed the RNG, if you want
to somehow debug the behaviour of the kconfig parser/frontends.
If not set, the current time will be used.
``KCONFIG_PROBABILITY``
This variable can be used to skew the probabilities. This variable can
be unset or empty, or set to three different formats:
======================= ================== =====================
KCONFIG_PROBABILITY y:n split y:m:n split
======================= ================== =====================
unset or empty 50 : 50 33 : 33 : 34
N N : 100-N N/2 : N/2 : 100-N
[1] N:M N+M : 100-(N+M) N : M : 100-(N+M)
[2] N:M:L N : 100-N M : L : 100-(M+L)
======================= ================== =====================
where N, M and L are integers (in base 10) in the range [0,100], and so
that:
[1] N+M is in the range [0,100]
[2] M+L is in the range [0,100]
Examples::
KCONFIG_PROBABILITY=10
10% of booleans will be set to 'y', 90% to 'n'
5% of tristates will be set to 'y', 5% to 'm', 90% to 'n'
KCONFIG_PROBABILITY=15:25
40% of booleans will be set to 'y', 60% to 'n'
15% of tristates will be set to 'y', 25% to 'm', 60% to 'n'
KCONFIG_PROBABILITY=10:15:15
10% of booleans will be set to 'y', 90% to 'n'
15% of tristates will be set to 'y', 15% to 'm', 70% to 'n'
Environment variables for ``syncconfig``:
``KCONFIG_NOSILENTUPDATE``
If this variable has a non-blank value, it prevents silent kernel
config updates (requires explicit updates).
``KCONFIG_AUTOCONFIG``
This environment variable can be set to specify the path & name of the
"auto.conf" file. Its default value is "include/config/auto.conf".
``KCONFIG_AUTOHEADER``
This environment variable can be set to specify the path & name of the
"autoconf.h" (header) file.
Its default value is "include/generated/autoconf.h".
menuconfig
==========
Searching in menuconfig:
The Search function searches for kernel configuration symbol
names, so you have to know something close to what you are
looking for.
Example::
/hotplug
This lists all config symbols that contain "hotplug",
e.g., HOTPLUG_CPU, MEMORY_HOTPLUG.
For search help, enter / followed by TAB-TAB (to highlight
<Help>) and Enter. This will tell you that you can also use
regular expressions (regexes) in the search string, so if you
are not interested in MEMORY_HOTPLUG, you could try::
/^hotplug
When searching, symbols are sorted thus:
- first, exact matches, sorted alphabetically (an exact match
is when the search matches the complete symbol name);
- then, other matches, sorted alphabetically.
For example, ^ATH.K matches:
ATH5K ATH9K ATH5K_AHB ATH5K_DEBUG [...] ATH6KL ATH6KL_DEBUG
[...] ATH9K_AHB ATH9K_BTCOEX_SUPPORT ATH9K_COMMON [...]
of which only ATH5K and ATH9K match exactly and so are sorted
first (and in alphabetical order), then come all other symbols,
sorted in alphabetical order.
In this menu, pressing the key in the (#) prefix will jump
directly to that location. You will be returned to the current
search results after exiting this new menu.
User interface options for 'menuconfig':
``MENUCONFIG_COLOR``
It is possible to select different color themes using the variable
MENUCONFIG_COLOR. To select a theme use::
make MENUCONFIG_COLOR=<theme> menuconfig
Available themes are::
- mono => selects colors suitable for monochrome displays
- blackbg => selects a color scheme with black background
- classic => theme with blue background. The classic look
- bluetitle => a LCD friendly version of classic. (default)
``MENUCONFIG_MODE``
This mode shows all sub-menus in one large tree.
Example::
make MENUCONFIG_MODE=single_menu menuconfig
nconfig
=======
nconfig is an alternate text-based configurator. It lists function
keys across the bottom of the terminal (window) that execute commands.
You can also just use the corresponding numeric key to execute the
commands unless you are in a data entry window. E.g., instead of F6
for Save, you can just press 6.
Use F1 for Global help or F3 for the Short help menu.
Searching in nconfig:
You can search either in the menu entry "prompt" strings
or in the configuration symbols.
Use / to begin a search through the menu entries. This does
not support regular expressions. Use <Down> or <Up> for
Next hit and Previous hit, respectively. Use <Esc> to
terminate the search mode.
F8 (SymSearch) searches the configuration symbols for the
given string or regular expression (regex).
In the SymSearch, pressing the key in the (#) prefix will
jump directly to that location. You will be returned to the
current search results after exiting this new menu.
Environment variables:
``NCONFIG_MODE``
This mode shows all sub-menus in one large tree.
Example::
make NCONFIG_MODE=single_menu nconfig
xconfig
=======
Searching in xconfig:
The Search function searches for kernel configuration symbol
names, so you have to know something close to what you are
looking for.
Example::
Ctrl-F hotplug
or::
Menu: File, Search, hotplug
lists all config symbol entries that contain "hotplug" in
the symbol name. In this Search dialog, you may change the
config setting for any of the entries that are not grayed out.
You can also enter a different search string without having
to return to the main menu.
gconfig
=======
Searching in gconfig:
There is no search command in gconfig. However, gconfig does
have several different viewing choices, modes, and options.
linux-7.2.3/Documentation/kbuild/llvm.rst
+221
View File
@@ -0,0 +1,221 @@
.. _kbuild_llvm:
==============================
Building Linux with Clang/LLVM
==============================
This document covers how to build the Linux kernel with Clang and LLVM
utilities.
About
-----
The Linux kernel has always traditionally been compiled with GNU toolchains
such as GCC and binutils. Ongoing work has allowed for `Clang
<https://clang.llvm.org/>`_ and `LLVM <https://llvm.org/>`_ utilities to be
used as viable substitutes. Distributions such as `Android
<https://www.android.com/>`_, `ChromeOS
<https://www.chromium.org/chromium-os>`_, `OpenMandriva
<https://www.openmandriva.org/>`_, and `Chimera Linux
<https://chimera-linux.org/>`_ use Clang built kernels. Google's and Meta's
datacenter fleets also run kernels built with Clang.
`LLVM is a collection of toolchain components implemented in terms of C++
objects <https://www.aosabook.org/en/llvm.html>`_. Clang is a front-end to LLVM
that supports C and the GNU C extensions required by the kernel, and is
pronounced "klang," not "see-lang."
Building with LLVM
------------------
Invoke ``make`` via::
make LLVM=1
to compile for the host target. For cross compiling::
make LLVM=1 ARCH=arm64
The LLVM= argument
------------------
LLVM has substitutes for GNU binutils utilities. They can be enabled
individually. The full list of supported make variables::
make CC=clang LD=ld.lld AR=llvm-ar NM=llvm-nm STRIP=llvm-strip \
OBJCOPY=llvm-objcopy OBJDUMP=llvm-objdump READELF=llvm-readelf \
HOSTCC=clang HOSTCXX=clang++ HOSTAR=llvm-ar HOSTLD=ld.lld
``LLVM=1`` expands to the above.
If your LLVM tools are not available in your PATH, you can supply their
location using the LLVM variable with a trailing slash::
make LLVM=/path/to/llvm/
which will use ``/path/to/llvm/clang``, ``/path/to/llvm/ld.lld``, etc. The
following may also be used::
PATH=/path/to/llvm:$PATH make LLVM=1
If your LLVM tools have a version suffix and you want to test with that
explicit version rather than the unsuffixed executables like ``LLVM=1``, you
can pass the suffix using the ``LLVM`` variable::
make LLVM=-14
which will use ``clang-14``, ``ld.lld-14``, etc.
To support combinations of out of tree paths with version suffixes, we
recommend::
PATH=/path/to/llvm/:$PATH make LLVM=-14
The same value used for ``LLVM=`` should be set for each invocation of ``make``
if configuring and building via distinct commands. ``LLVM=`` should also be set
as an environment variable when running scripts that will eventually run
``make``.
Cross Compiling
---------------
A single Clang compiler binary (and corresponding LLVM utilities) will
typically contain all supported back ends, which can help simplify cross
compiling especially when ``LLVM=1`` is used. If you use only LLVM tools,
``CROSS_COMPILE`` or target-triple-prefixes become unnecessary. Example::
make LLVM=1 ARCH=arm64
As an example of mixing LLVM and GNU utilities, for a target like ``ARCH=s390``
which does not yet have ``ld.lld`` or ``llvm-objcopy`` support, you could
invoke ``make`` via::
make LLVM=1 ARCH=s390 LD=s390x-linux-gnu-ld.bfd \
OBJCOPY=s390x-linux-gnu-objcopy
This example will invoke ``s390x-linux-gnu-ld.bfd`` as the linker and
``s390x-linux-gnu-objcopy``, so ensure those are reachable in your ``$PATH``.
``CROSS_COMPILE`` is not used to prefix the Clang compiler binary (or
corresponding LLVM utilities) as is the case for GNU utilities when ``LLVM=1``
is not set.
The LLVM_IAS= argument
----------------------
Clang can assemble assembler code. You can pass ``LLVM_IAS=0`` to disable this
behavior and have Clang invoke the corresponding non-integrated assembler
instead. Example::
make LLVM=1 LLVM_IAS=0
``CROSS_COMPILE`` is necessary when cross compiling and ``LLVM_IAS=0``
is used in order to set ``--prefix=`` for the compiler to find the
corresponding non-integrated assembler (typically, you don't want to use the
system assembler when targeting another architecture). Example::
make LLVM=1 ARCH=arm LLVM_IAS=0 CROSS_COMPILE=arm-linux-gnueabi-
Ccache
------
``ccache`` can be used with ``clang`` to improve subsequent builds, (though
KBUILD_BUILD_TIMESTAMP_ should be set to a deterministic value between builds
in order to avoid 100% cache misses, see Reproducible_builds_ for more info)::
KBUILD_BUILD_TIMESTAMP='' make LLVM=1 CC="ccache clang"
.. _KBUILD_BUILD_TIMESTAMP: kbuild.html#kbuild-build-timestamp
.. _Reproducible_builds: reproducible-builds.html#timestamps
Supported Architectures
-----------------------
LLVM does not target all of the architectures that Linux supports and
just because a target is supported in LLVM does not mean that the kernel
will build or work without any issues. Below is a general summary of
architectures that currently work with ``CC=clang`` or ``LLVM=1``. Level
of support corresponds to "S" values in the MAINTAINERS files. If an
architecture is not present, it either means that LLVM does not target
it or there are known issues. Using the latest stable version of LLVM or
even the development tree will generally yield the best results.
An architecture's ``defconfig`` is generally expected to work well,
certain configurations may have problems that have not been uncovered
yet. Bug reports are always welcome at the issue tracker below!
.. list-table::
:widths: 10 10 10
:header-rows: 1
* - Architecture
- Level of support
- ``make`` command
* - arm
- Supported
- ``LLVM=1``
* - arm64
- Supported
- ``LLVM=1``
* - hexagon
- Maintained
- ``LLVM=1``
* - loongarch
- Maintained
- ``LLVM=1``
* - mips
- Maintained
- ``LLVM=1``
* - powerpc
- Maintained
- ``LLVM=1``
* - riscv
- Supported
- ``LLVM=1``
* - s390
- Maintained
- ``LLVM=1`` (LLVM >= 18.1.0), ``CC=clang`` (LLVM < 18.1.0)
* - sparc (sparc64 only)
- Maintained
- ``CC=clang LLVM_IAS=0`` (LLVM >= 20)
* - um (User Mode)
- Maintained
- ``LLVM=1``
* - x86
- Supported
- ``LLVM=1``
Getting Help
------------
- `Website <https://clangbuiltlinux.github.io/>`_
- `Mailing List <https://lore.kernel.org/llvm/>`_: <llvm@lists.linux.dev>
- `Old Mailing List Archives <https://groups.google.com/g/clang-built-linux>`_
- `Issue Tracker <https://github.com/ClangBuiltLinux/linux/issues>`_
- IRC: #clangbuiltlinux on irc.libera.chat
- `Telegram <https://t.me/ClangBuiltLinux>`_: @ClangBuiltLinux
- `Wiki <https://github.com/ClangBuiltLinux/linux/wiki>`_
- `Beginner Bugs <https://github.com/ClangBuiltLinux/linux/issues?q=is%3Aopen+is%3Aissue+label%3A%22good+first+issue%22>`_
.. _getting_llvm:
Getting LLVM
-------------
We provide prebuilt stable versions of LLVM on `kernel.org
<https://kernel.org/pub/tools/llvm/>`_. These have been optimized with profile
data for building Linux kernels, which should improve kernel build times
relative to other distributions of LLVM.
Below are links that may be useful for building LLVM from source or procuring
it through a distribution's package manager.
- https://releases.llvm.org/download.html
- https://github.com/llvm/llvm-project
- https://llvm.org/docs/GettingStarted.html
- https://llvm.org/docs/CMake.html
- https://apt.llvm.org/
- https://www.archlinux.org/packages/extra/x86_64/llvm/
- https://github.com/ClangBuiltLinux/tc-build
- https://github.com/ClangBuiltLinux/linux/wiki/Building-Clang-from-source
- https://android.googlesource.com/platform/prebuilts/clang/host/linux-x86/
linux-7.2.3/Documentation/kbuild/makefiles.rst
+1723
View File
File diff suppressed because it is too large Load Diff
linux-7.2.3/Documentation/kbuild/modules.rst
+513
View File
@@ -0,0 +1,513 @@
=========================
Building External Modules
=========================
This document describes how to build an out-of-tree kernel module.
Introduction
============
"kbuild" is the build system used by the Linux kernel. Modules must use
kbuild to stay compatible with changes in the build infrastructure and
to pick up the right flags to the compiler. Functionality for building modules
both in-tree and out-of-tree is provided. The method for building
either is similar, and all modules are initially developed and built
out-of-tree.
Covered in this document is information aimed at developers interested
in building out-of-tree (or "external") modules. The author of an
external module should supply a makefile that hides most of the
complexity, so one only has to type "make" to build the module. This is
easily accomplished, and a complete example will be presented in
section `Creating a Kbuild File for an External Module`_.
How to Build External Modules
=============================
To build external modules, you must have a prebuilt kernel available
that contains the configuration and header files used in the build.
Also, the kernel must have been built with modules enabled. If you are
using a distribution kernel, there will be a package for the kernel you
are running provided by your distribution.
An alternative is to use the "make" target "modules_prepare." This will
make sure the kernel contains the information required. The target
exists solely as a simple way to prepare a kernel source tree for
building external modules.
NOTE: "modules_prepare" will not build Module.symvers even if
CONFIG_MODVERSIONS is set; therefore, a full kernel build needs to be
executed to make module versioning work.
Command Syntax
--------------
The command to build an external module is::
$ make -C <path_to_kernel_dir> M=$PWD
The kbuild system knows that an external module is being built
due to the "M=<dir>" option given in the command.
To build against the running kernel use::
$ make -C /lib/modules/`uname -r`/build M=$PWD
Then to install the module(s) just built, add the target
"modules_install" to the command::
$ make -C /lib/modules/`uname -r`/build M=$PWD modules_install
Starting from Linux 6.13, you can use the -f option instead of -C. This
will avoid unnecessary change of the working directory. The external
module will be output to the directory where you invoke make.
$ make -f /lib/modules/`uname -r`/build/Makefile M=$PWD
Options
-------
($KDIR refers to the path of the kernel source directory, or the path
of the kernel output directory if the kernel was built in a separate
build directory.)
You can optionally pass MO= option if you want to build the modules in
a separate directory.
make -C $KDIR M=$PWD [MO=$BUILD_DIR]
-C $KDIR
The directory that contains the kernel and relevant build
artifacts used for building an external module.
"make" will actually change to the specified directory
when executing and will change back when finished.
M=$PWD
Informs kbuild that an external module is being built.
The value given to "M" is the absolute path of the
directory where the external module (kbuild file) is
located.
MO=$BUILD_DIR
Specifies a separate output directory for the external module.
Targets
-------
When building an external module, only a subset of the "make"
targets are available.
make -C $KDIR M=$PWD [target]
The default will build the module(s) located in the current
directory, so a target does not need to be specified. All
output files will also be generated in this directory. No
attempts are made to update the kernel source, and it is a
precondition that a successful "make" has been executed for the
kernel.
modules
The default target for external modules. It has the
same functionality as if no target was specified. See
description above.
modules_install
Install the external module(s). The default location is
/lib/modules/<kernel_release>/updates/, but a prefix may
be added with INSTALL_MOD_PATH (discussed in section
`Module Installation`_).
clean
Remove all generated files in the module directory only.
help
List the available targets for external modules.
Building Separate Files
-----------------------
It is possible to build single files that are part of a module.
This works equally well for the kernel, a module, and even for
external modules.
Example (The module foo.ko, consist of bar.o and baz.o)::
make -C $KDIR M=$PWD bar.lst
make -C $KDIR M=$PWD baz.o
make -C $KDIR M=$PWD foo.ko
make -C $KDIR M=$PWD ./
Creating a Kbuild File for an External Module
=============================================
In the last section we saw the command to build a module for the
running kernel. The module is not actually built, however, because a
build file is required. Contained in this file will be the name of
the module(s) being built, along with the list of requisite source
files. The file may be as simple as a single line::
obj-m := <module_name>.o
The kbuild system will build <module_name>.o from <module_name>.c,
and, after linking, will result in the kernel module <module_name>.ko.
The above line can be put in either a "Kbuild" file or a "Makefile."
When the module is built from multiple sources, an additional line is
needed listing the files::
<module_name>-y := <src1>.o <src2>.o ...
NOTE: Further documentation describing the syntax used by kbuild is
located in Documentation/kbuild/makefiles.rst.
The examples below demonstrate how to create a build file for the
module 8123.ko, which is built from the following files::
8123_if.c
8123_if.h
8123_pci.c
Shared Makefile
---------------
An external module always includes a wrapper makefile that
supports building the module using "make" with no arguments.
This target is not used by kbuild; it is only for convenience.
Additional functionality, such as test targets, can be included
but should be filtered out from kbuild due to possible name
clashes.
Example 1::
--> filename: Makefile
ifneq ($(KERNELRELEASE),)
# kbuild part of makefile
obj-m := 8123.o
8123-y := 8123_if.o 8123_pci.o
else
# normal makefile
KDIR ?= /lib/modules/`uname -r`/build
default:
$(MAKE) -C $(KDIR) M=$$PWD
endif
The check for KERNELRELEASE is used to separate the two parts
of the makefile. In the example, kbuild will only see the two
assignments, whereas "make" will see everything except these
two assignments. This is due to two passes made on the file:
the first pass is by the "make" instance run on the command
line; the second pass is by the kbuild system, which is
initiated by the parameterized "make" in the default target.
Separate Kbuild File and Makefile
---------------------------------
Kbuild will first look for a file named "Kbuild", and if it is not
found, it will then look for "Makefile". Utilizing a "Kbuild" file
allows us to split up the "Makefile" from example 1 into two files:
Example 2::
--> filename: Kbuild
obj-m := 8123.o
8123-y := 8123_if.o 8123_pci.o
--> filename: Makefile
KDIR ?= /lib/modules/`uname -r`/build
default:
$(MAKE) -C $(KDIR) M=$$PWD
The split in example 2 is questionable due to the simplicity of
each file; however, some external modules use makefiles
consisting of several hundred lines, and here it really pays
off to separate the kbuild part from the rest.
Linux 6.13 and later support another way. The external module Makefile
can include the kernel Makefile directly, rather than invoking sub Make.
Example 3::
--> filename: Kbuild
obj-m := 8123.o
8123-y := 8123_if.o 8123_pci.o
--> filename: Makefile
KDIR ?= /lib/modules/$(shell uname -r)/build
export KBUILD_EXTMOD := $(realpath $(dir $(lastword $(MAKEFILE_LIST))))
include $(KDIR)/Makefile
Building Multiple Modules
-------------------------
kbuild supports building multiple modules with a single build
file. For example, if you wanted to build two modules, foo.ko
and bar.ko, the kbuild lines would be::
obj-m := foo.o bar.o
foo-y := <foo_srcs>
bar-y := <bar_srcs>
It is that simple!
Include Files
=============
Within the kernel, header files are kept in standard locations
according to the following rule:
* If the header file only describes the internal interface of a
module, then the file is placed in the same directory as the
source files.
* If the header file describes an interface used by other parts
of the kernel that are located in different directories, then
the file is placed in include/linux/.
NOTE:
There are two notable exceptions to this rule: larger
subsystems have their own directory under include/, such as
include/scsi; and architecture specific headers are located
under arch/$(SRCARCH)/include/.
Kernel Includes
---------------
To include a header file located under include/linux/, simply
use::
#include <linux/module.h>
kbuild will add options to the compiler so the relevant directories
are searched.
Single Subdirectory
-------------------
External modules tend to place header files in a separate
include/ directory where their source is located, although this
is not the usual kernel style. To inform kbuild of the
directory, use either ccflags-y or CFLAGS_<filename>.o.
Using the example from section 3, if we moved 8123_if.h to a
subdirectory named include, the resulting kbuild file would
look like::
--> filename: Kbuild
obj-m := 8123.o
ccflags-y := -I $(src)/include
8123-y := 8123_if.o 8123_pci.o
Several Subdirectories
----------------------
kbuild can handle files that are spread over several directories.
Consider the following example::
.
|__ src
| |__ complex_main.c
| |__ hal
| |__ hardwareif.c
| |__ include
| |__ hardwareif.h
|__ include
|__ complex.h
To build the module complex.ko, we then need the following
kbuild file::
--> filename: Kbuild
obj-m := complex.o
complex-y := src/complex_main.o
complex-y += src/hal/hardwareif.o
ccflags-y := -I$(src)/include
ccflags-y += -I$(src)/src/hal/include
As you can see, kbuild knows how to handle object files located
in other directories. The trick is to specify the directory
relative to the kbuild file's location. That being said, this
is NOT recommended practice.
For the header files, kbuild must be explicitly told where to
look. When kbuild executes, the current directory is always the
root of the kernel tree (the argument to "-C") and therefore an
absolute path is needed. $(src) provides the absolute path by
pointing to the directory where the currently executing kbuild
file is located.
Module Installation
===================
Modules which are included in the kernel are installed in the
directory:
/lib/modules/$(KERNELRELEASE)/kernel/
And external modules are installed in:
/lib/modules/$(KERNELRELEASE)/updates/
INSTALL_MOD_PATH
----------------
Above are the default directories but as always some level of
customization is possible. A prefix can be added to the
installation path using the variable INSTALL_MOD_PATH::
$ make INSTALL_MOD_PATH=/frodo modules_install
=> Install dir: /frodo/lib/modules/$(KERNELRELEASE)/kernel/
INSTALL_MOD_PATH may be set as an ordinary shell variable or,
as shown above, can be specified on the command line when
calling "make." This has effect when installing both in-tree
and out-of-tree modules.
INSTALL_MOD_DIR
---------------
External modules are by default installed to a directory under
/lib/modules/$(KERNELRELEASE)/updates/, but you may wish to
locate modules for a specific functionality in a separate
directory. For this purpose, use INSTALL_MOD_DIR to specify an
alternative name to "updates."::
$ make INSTALL_MOD_DIR=gandalf -C $KDIR \
M=$PWD modules_install
=> Install dir: /lib/modules/$(KERNELRELEASE)/gandalf/
Module Versioning
=================
Module versioning is enabled by the CONFIG_MODVERSIONS tag, and is used
as a simple ABI consistency check. A CRC value of the full prototype
for an exported symbol is created. When a module is loaded/used, the
CRC values contained in the kernel are compared with similar values in
the module; if they are not equal, the kernel refuses to load the
module.
Module.symvers contains a list of all exported symbols from a kernel
build.
Symbols From the Kernel (vmlinux + modules)
-------------------------------------------
During a kernel build, a file named Module.symvers will be
generated. Module.symvers contains all exported symbols from
the kernel and compiled modules. For each symbol, the
corresponding CRC value is also stored.
The syntax of the Module.symvers file is::
<CRC> <Symbol> <Module> <Export Type> <Namespace>
0xe1cc2a05 usb_stor_suspend drivers/usb/storage/usb-storage EXPORT_SYMBOL_GPL USB_STORAGE
The fields are separated by tabs and values may be empty (e.g.
if no namespace is defined for an exported symbol).
For a kernel build without CONFIG_MODVERSIONS enabled, the CRC
would read 0x00000000.
Module.symvers serves two purposes:
1) It lists all exported symbols from vmlinux and all modules.
2) It lists the CRC if CONFIG_MODVERSIONS is enabled.
Version Information Formats
---------------------------
Exported symbols have information stored in the __ksymtab and
__kflagstab sections. Symbol names and namespaces are stored in
__ksymtab_strings section, using a format similar to the string
table used for ELF. If CONFIG_MODVERSIONS is enabled, the CRCs
corresponding to exported symbols will be added to the
__kcrctab section.
If CONFIG_BASIC_MODVERSIONS is enabled (default with
CONFIG_MODVERSIONS), imported symbols will have their symbol name and
CRC stored in the __versions section of the importing module. This
mode only supports symbols of length up to 64 bytes.
If CONFIG_EXTENDED_MODVERSIONS is enabled (required to enable both
CONFIG_MODVERSIONS and CONFIG_RUST at the same time), imported symbols
will have their symbol name recorded in the __version_ext_names
section as a series of concatenated, null-terminated strings. CRCs for
these symbols will be recorded in the __version_ext_crcs section.
Symbols and External Modules
----------------------------
When building an external module, the build system needs access
to the symbols from the kernel to check if all external symbols
are defined. This is done in the MODPOST step. modpost obtains
the symbols by reading Module.symvers from the kernel source
tree. During the MODPOST step, a new Module.symvers file will be
written containing all exported symbols from that external module.
Symbols From Another External Module
------------------------------------
Sometimes, an external module uses exported symbols from
another external module. Kbuild needs to have full knowledge of
all symbols to avoid spitting out warnings about undefined
symbols. Two solutions exist for this situation.
NOTE: The method with a top-level kbuild file is recommended
but may be impractical in certain situations.
Use a top-level kbuild file
If you have two modules, foo.ko and bar.ko, where
foo.ko needs symbols from bar.ko, you can use a
common top-level kbuild file so both modules are
compiled in the same build. Consider the following
directory layout::
./foo/ <= contains foo.ko
./bar/ <= contains bar.ko
The top-level kbuild file would then look like::
#./Kbuild (or ./Makefile):
obj-m := foo/ bar/
And executing::
$ make -C $KDIR M=$PWD
will then do the expected and compile both modules with
full knowledge of symbols from either module.
Use "make" variable KBUILD_EXTRA_SYMBOLS
If it is impractical to add a top-level kbuild file,
you can assign a space separated list
of files to KBUILD_EXTRA_SYMBOLS in your build file.
These files will be loaded by modpost during the
initialization of its symbol tables.
Tips & Tricks
=============
Testing for CONFIG_FOO_BAR
--------------------------
Modules often need to check for certain `CONFIG_` options to
decide if a specific feature is included in the module. In
kbuild this is done by referencing the `CONFIG_` variable
directly::
#fs/ext2/Makefile
obj-$(CONFIG_EXT2_FS) += ext2.o
ext2-y := balloc.o bitmap.o dir.o
ext2-$(CONFIG_EXT2_FS_XATTR) += xattr.o
linux-7.2.3/Documentation/kbuild/reproducible-builds.rst
+143
View File
@@ -0,0 +1,143 @@
===================
Reproducible builds
===================
It is generally desirable that building the same source code with
the same set of tools is reproducible, i.e. the output is always
exactly the same. This makes it possible to verify that the build
infrastructure for a binary distribution or embedded system has not
been subverted. This can also make it easier to verify that a source
or tool change does not make any difference to the resulting binaries.
The `Reproducible Builds project`_ has more information about this
general topic. This document covers the various reasons why building
the kernel may be unreproducible, and how to avoid them.
Timestamps
----------
The kernel embeds timestamps in three places:
* The version string exposed by ``uname()`` and included in
``/proc/version``
* File timestamps in the embedded initramfs
* If enabled via ``CONFIG_IKHEADERS``, file timestamps of kernel
headers embedded in the kernel or respective module,
exposed via ``/sys/kernel/kheaders.tar.xz``
By default the timestamp is the current time and in the case of
``kheaders`` the various files' modification times. This must
be overridden using the `KBUILD_BUILD_TIMESTAMP`_ variable.
If you are building from a git commit, you could use its commit date.
The kernel does *not* use the ``__DATE__`` and ``__TIME__`` macros,
and enables warnings if they are used. If you incorporate external
code that does use these, you must override the timestamp they
correspond to by setting the `SOURCE_DATE_EPOCH`_ environment
variable.
User, host
----------
The kernel embeds the building user and host names in
``/proc/version``. These must be overridden using the
`KBUILD_BUILD_USER and KBUILD_BUILD_HOST`_ variables. If you are
building from a git commit, you could use its committer address.
Absolute filenames
------------------
When the kernel is built out-of-tree, debug information may include
absolute filenames for the source files and build directory. These must
be overridden by including a ``-fdebug-prefix-map`` option for each in
the `KCFLAGS`_ and `KAFLAGS`_ variables to cover both ``.c`` and ``.S``
files.
Depending on the compiler used, the ``__FILE__`` macro may also expand
to an absolute filename in an out-of-tree build. Kbuild automatically
uses the ``-fmacro-prefix-map`` option to prevent this, if it is
supported.
The Reproducible Builds web site has more information about these
`prefix-map options`_.
Some CONFIG options such as `CONFIG_DEBUG_EFI` embed absolute paths in
object files. Such options should be disabled.
Generated files in source packages
----------------------------------
The build processes for some programs under the ``tools/``
subdirectory do not completely support out-of-tree builds. This may
cause a later source package build using e.g. ``make rpm-pkg`` to
include generated files. You should ensure the source tree is
pristine by running ``make mrproper`` or ``git clean -d -f -x`` before
building a source package.
Module signing
--------------
If you enable ``CONFIG_MODULE_SIG_ALL``, the default behaviour is to
generate a different temporary key for each build, resulting in the
modules being unreproducible. However, including a signing key with
your source would presumably defeat the purpose of signing modules.
One approach to this is to divide up the build process so that the
unreproducible parts can be treated as sources:
1. Generate a persistent signing key. Add the certificate for the key
to the kernel source.
2. Set the ``CONFIG_SYSTEM_TRUSTED_KEYS`` symbol to include the
signing key's certificate, set ``CONFIG_MODULE_SIG_KEY`` to an
empty string, and disable ``CONFIG_MODULE_SIG_ALL``.
Build the kernel and modules.
3. Create detached signatures for the modules, and publish them as
sources.
4. Perform a second build that attaches the module signatures. It
can either rebuild the modules or use the output of step 2.
Structure randomisation
-----------------------
If you enable ``CONFIG_RANDSTRUCT``, you will need to pre-generate
the random seed in ``scripts/basic/randstruct.seed`` so the same
value is used by each build. See ``scripts/gen-randstruct-seed.sh``
for details.
Debug info conflicts
--------------------
This is not a problem of unreproducibility, but of generated files
being *too* reproducible.
Once you set all the necessary variables for a reproducible build, a
vDSO's debug information may be identical even for different kernel
versions. This can result in file conflicts between debug information
packages for the different kernel versions.
To avoid this, you can make the vDSO different for different
kernel versions by including an arbitrary string of "salt" in it.
This is specified by the Kconfig symbol ``CONFIG_BUILD_SALT``.
Git
---
Uncommitted changes or different commit ids in git can also lead
to different compilation results. For example, after executing
``git reset HEAD^``, even if the code is the same, the
``include/config/kernel.release`` generated during compilation
will be different, which will eventually lead to binary differences.
See ``scripts/setlocalversion`` for details.
.. _KBUILD_BUILD_TIMESTAMP: kbuild.html#kbuild-build-timestamp
.. _KBUILD_BUILD_USER and KBUILD_BUILD_HOST: kbuild.html#kbuild-build-user-kbuild-build-host
.. _KCFLAGS: kbuild.html#kcflags
.. _KAFLAGS: kbuild.html#kaflags
.. _prefix-map options: https://reproducible-builds.org/docs/build-path/
.. _Reproducible Builds project: https://reproducible-builds.org/
.. _SOURCE_DATE_EPOCH: https://reproducible-builds.org/docs/source-date-epoch/