module
—
Kernel Modules interface
Kernel modules allow the system administrator to dynamically add and remove
functionality from a running system. This also helps software developers add
new parts of the kernel without constantly rebooting to test their changes.
The kernel may automatically load software modules as needed to
perform requested operations. For example, an “xyzfs” module
can be loaded automatically when an attempt is made to mount an
“xyzfs” file system. Modules can also depend on other modules,
and dependent modules are automatically loaded. When a module is no longer
needed, it can be automatically unloaded.
An in-kernel linker resolves symbol references between the module
and the rest of the kernel.
The module
interface is accessed with the
modctl(2) system call. All
common operations involving kernel modules are handled by the
modload(8),
modunload(8), and
modstat(8) programs. Users
should never have to interact with
modctl(2) directly.
Virtual file systems may be added via the module
interface.
Many device drivers can be loaded as a kernel module. One potential problem
specific to block and character device drivers is that the device nodes must
exist for the devices to be accessed. These need to be created manually, after
the driver module has been successfully loaded. Most device driver modules do
not need any manual intervention to function properly.
Execution Interpreters can be loaded to provide support for executing binaries
not normally supported by the kernel. This also allows loading support for
executing foreign system binaries. Execution Interpreters may require that an
appropriate emulation module also be loaded.
Miscellaneous modules are modules for which there are not currently well-defined
or well-used interfaces for extension. They are provided for extension, and
the user-provided module initialization routine is expected to install the
necessary "hooks" into the rest of the operating system. An example
of a "miscellaneous module" might be a loader for card-specific VGA
drivers or alternate terminal emulations in an appropriately layered console
driver.
Alternate system security models also may be loaded using
module
.
The common build tool of NetBSD,
“build.sh”, automatically compiles and installs most modules
during a full system build and install. (The exceptions are some modules from
external sources which, due to licensing concerns, can be built only as
separately-loaded modules.) However, sometimes it is useful to update only
modules. The following example demonstrates one way to do this. It is assumed
that the source code is under /usr/src, while the
object and toolchain directories are under /usr/obj
and /usr/tools, respectively.
cd /usr/src/sys/modules
export OBJDIR=/usr/obj
export TOOLDIR=/usr/tools
make clean
make
make install
Alternatively, the “build.sh” tool can be used to
build only the modules.
cd /usr/src
./build.sh -O /usr/obj -T /usr/tools modules
The module
facility was designed to be similar in
functionality to the loadable kernel modules facility provided by SunOS 4.1.3.
The old LKM
interface was replaced by
module
in NetBSD 5.0.
The module
subsystem was implemented by
Andrew Doran ⟨ad@netbsd.org⟩.
The module
framework is still under active development.
At least two potential caveats can be mentioned.
- Kernel modules are built to operate only with a specific version of the
NetBSD kernel. When the kernel is updated to a new
version, the contents of the
/stand/${ARCH}/${VERSION}/modules/ directory
should be updated as well. (This location has been the subject of much
discussion, and may change in future versions of
NetBSD.)
- If an attempt is made to boot the operating system from a file system for
which the module is not built into the kernel, the boot may fail with the
message “Cannot mount root, error 79”. On certain
architectures (currently, i386 and amd64), one may be able to recover from
this error by using the “load xxxfs” command before trying
to boot. This command is only available on newer bootloaders.
The absence of required modules or the inability of the bootloader
to load the modules are common reasons for failures to boot a
MODULAR
kernel. It may be a good practice to
maintain a non-MODULAR kernel in the root file system for recovery
purposes.
A module becomes part of the kernel once loaded. Unlike in userland programs,
fatal errors in kernel modules may crash the operating system. There is no
memory protection between modules and the rest of the kernel. Hence, a
potential attacker with access to the
modctl(2) system call can
acquire total control over the system.
To avoid such security risks, new modules can only be loaded when
securelevel is less than or equal to zero, or if the
kernel was built with options INSECURE
. Refer to
secmodel_securelevel(9)
for additional details on the securelevel. Only use
modules from trusted sources.