[2024-feb-29] Sad news: Eric Layton aka Nocturnal Slacker aka vtel57 passed away on Feb 26th, shortly after hospitalization. He was one of our Wiki's most prominent admins. He will be missed.

Welcome to the Slackware Documentation Project

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision
Previous revision
slackbook:linux_kernel [2012/09/17 03:33 (UTC)] – [The Linux Kernel] Aded original text and formatting mfillpotslackbook:linux_kernel [2014/11/05 02:25 (UTC)] (current) – [Compiling A Kernel and Why to do So] nocturnal.slacker
Line 1: Line 1:
 <!-- Add your text below. We strongly advise to start with a Headline (see button bar above). --> <!-- Add your text below. We strongly advise to start with a Headline (see button bar above). -->
 ====== The Linux Kernel ====== ====== The Linux Kernel ======
 +
 +
 +===== What Does the Kernel Do? =====
  
 You've probably heard people talking about compiling the kernel or You've probably heard people talking about compiling the kernel or
Line 15: Line 18:
 the kernel. This makes the kernel both the single most important piece the kernel. This makes the kernel both the single most important piece
 of software on your computer and the most complex. of software on your computer and the most complex.
-===== What Does the Kernel Do? ===== 
  
 ===== Working with Modules ===== ===== Working with Modules =====
  
 +The complexity of a modern linux kernel is staggering. The source code
 +for the kernel weighs in at nearly 400MB uncompressed. There are
 +thousands of developers, hundreds of options, and if everything were
 +built together, the kernel would soon pass 100MB in size itself. In
 +order to keep the size of the kernel down (as well as the amount of RAM
 +needed for the kernel), most of the kernel options are built as
 +modules. You can think of these modules as device drivers which can be
 +inserted or removed from a running kernel at will. In truth, many of
 +them aren't device drivers at all, but contain support for things such
 +as network protocols, security measures, and even filesystems. In
 +short, nearly any piece of the linux kernel can be built as a loadable
 +module.
 +
 +It's important to realize that Slackware will automatically handle
 +loading most modules for you. When your system boots,
 +**//udevd//**(8) is started and begins to probe your
 +system's hardware. For each device it finds, it loads the proper module
 +and created a device node in ''/dev''. This usually
 +means that you will not need to load any modules in order to use your
 +computer, but occasionally this is necessary.
 +
 +So what modules are currently loaded on your computer and how do we
 +load and unload them? Fortunately we have a full suite of tools for
 +handling this. As you might have guessed, the tool for listing modules
 +is **//lsmod//**(8).
 +
 +
 +<code>
 +darkstar:~# lsmod
 +Module                  Size  Used by
 +nls_utf8                1952  1 
 +cifs                  240600 
 +i915                  168584 
 +drm                   168128  3 i915
 +i2c_algo_bit            6468  1 i915
 +tun                    12740  1 
 +... many more lines ommitted ...
 +</code>
 +
 +
 +In addition to showing you what modules are loaded, it displays the
 +size of each module and tells you what other modules are using it.
 +
 +There are two applications for loading modules:
 +**//insmod//**(8) and
 +**//modprobe//**(8). Both will load modules and
 +report any errors (such as loading a module for a device that isn't
 +present in your system), but **//modprobe//** is
 +preferred because it can load any module dependencies. Using either is
 +straight-forward.
 +
 +
 +<code>
 +darkstar:~# insmod ext3
 +darkstar:~# modprobe ext4
 +darkstar:~# lsmod | grep ext
 +ext4                  239928 
 +jbd2                   59088  1 ext4
 +crc16                   1984  1 ext4
 +ext3                  139408 
 +jbd                    48520  1 ext3
 +mbcache                 8068  2 ext4,ext3
 +</code>
 +
 +Removing modules can be a tricky process, and once again we have two
 +programs for removing them: **//rmmod//**(8) and
 +**//modprobe//**.  In order to remove a module with
 +modprobe, you'll need to use the //-r// argument.
 +
 +
 +<code>
 +darkstar:~# rmmod ext3
 +darkstar:~# modprobe -r ext4
 +darkstar:~# lsmod | grep ext
 +</code>
 ===== Compiling A Kernel and Why to do So ===== ===== Compiling A Kernel and Why to do So =====
  
 +Most Slackware users will never need to compile a kernel. The huge and
 +generic kernels contain virtually all the support you will need.
 +
 +
 +However, some users may need to compile a kernel. If your computer
 +contains bleeding edge hardware, a newer kernel may offer improved
 +support. Sometimes a kernel patch may be available that corrects a
 +problem you are experiencing. In these cases a kernel compile is
 +probably warranted. Users who simply want the latest and greatest
 +version or who believe using a custom compiled kernel will give them
 +greater performance can certainly upgrade, but are unlikely to
 +actually notice any major changes. 
 +
 +
 +If you still think compiling your own kernel is something you want or
 +need to do, this section should walk you through the many steps.
 +Compiling and installing a kernel is not that difficult, but there are
 +a number of mistakes that can be made along the way, many of which can
 +prevent your computer from booting and cause major frustration.
 +
 +
 +The first step is ensuring you have the kernel source code installed
 +on your system.  The kernel source package is included in the
 +//"k"// disk set in the Slackware installer, or you can download
 +another version from [[
 +http://www.kernel.org/|http://www.kernel.org/]].
 +Traditionally, the kernel source is located in
 +''/usr/src/linux'', a symbolic link that
 +points to the specific kernel release used, but this is by no means
 +set in stone. You can place the kernel source code virtually anywhere
 +without encountering any problems.
 +
 +
 +<code>
 +darkstar:~# ls -l /usr/src
 +lrwxrwxrwx  1 root root   14 2009-07-22 19:59 linux -> linux-2.6.29.6/
 +drwxr-xr-x 23 root root 4096 2010-03-17 19:00 linux-2.6.29.6/
 +</code>
 +
 +
 +The most difficult part of any kernel compile is the kernel
 +configuration. There are hundreds of options, many of which can
 +optionally be compiled into modules. This means there are thousands of
 +ways to configure a kernel. Fortunately, there are a few handy tricks
 +that can keep you from running into too much trouble.  The kernel
 +configuration file is ''.config''. If you are very
 +brave, you can manually edit this file with a text editor, but I highly
 +recommend you use the kernel's built-in tools for manipulating
 +''.config''.
 +
 +
 +Unless you are very familiar with configuring kernels, you should
 +always start with a solid base configuration and modify it. This
 +prevents you from skipping an important option that might force you to
 +compile the kernel again and again until you get it right.  The best
 +kernel ''.config'' files to start with are those used
 +by Slackware's default kernels. You can find them on your Slackware
 +install disks or at your favorite mirror in the
 +''kernels/'' directory.
 +
 +
 +<code>
 +darkstar:~# mount /mnt/cdrom
 +darkstar:~# cd /mnt/cdrom/kernels
 +darkstar:/mnt/cdrom/kernels# ls
 +VERSIONS.TXT  huge.s/  generic.s/  speakup.s/
 +darkstar:/mnt/cdrom/kernels# ls genric.s
 +System.map.gz  bzImage  config
 +</code>
 +
 +
 +You can replace the default ''.config'' file easily by
 +copying or downloading the ''config'' file for the
 +kernel you wish to use as a base. Here I am using Slackware's
 +recommended generic.s kernel for a base, but you may wish to use the
 +huge.s config file. The generic kernel builds more things as modules
 +and thus creates a smaller kernel image, but it usually requires the
 +use of an initrd.
 +
 +
 +<code>
 +darkstar:/mnt/cdrom/kernels# cp generic.s/config /usr/src/linux/.config
 +</code>
 +
 +<note important>
 + 
 +The Slackware kernel file lacks the //"dot"// while the kernel file
 +includes it. If you forget, or simply copy the
 +''config'' to
 +''/usr/src'' whatever
 +''.config'' file was already present will be used
 +instead.
 +</note>
 +
 +If you want to use the configuration for the currently running kernel
 +as your base, you may be able to locate it at
 +''/proc/config.gz''. This is a special kernel-related
 +file that includes the entire kernel configuration in a compressed
 +format and requires that your kernel was built to support it.
 +
 +
 +<code>
 +darkstar:~# zcat /proc/config.gz > /usr/src/linux/.config
 +</code>
 +
 +
 +Now that we've created a solid base configuration, it's time to make
 +any configuration changes we want. The entire kernel build process from
 +configuration to compilation is performed with the
 +**//make//**(1) command and special arguments to it.
 +Each argument performs a different function.
 +
 +
 +If you are upgrading to a newer kernel release, you will definitely
 +want to use the //oldconfig// argument. This will step through
 +your base ''.config'' and look for missing elements
 +that usually indicates that the new kernel release contains additional
 +options. Since options are added at virtually every kernel release,
 +this is generally a good thing to do.
 +
 +
 +<code>
 +darkstar:/usr/src/linux# make oldconfig
 +scripts/kconfig/conf -o arch/x86/Kconfig
 +*
 +* Restart config...
 +*
 +*
 +* File systems
 +*
 +Second extended fs support (EXT2_FS) [M/n/y/?] m
 +  Ext2 extended attributes (EXT2_FS_XATTR) [N/y/?] n
 +  Ext2 execute in place support (EXT2_FS_XIP) [N/y/?] n
 +Ext3 journalling file system support (EXT3_FS) [M/n/y/?] m
 +  Ext3 extended attributes (EXT3_FS_XATTR) [Y/n/?] y
 +    Ext3 POSIX Access Control Lists (EXT3_FS_POSIX_ACL) [Y/n/?] y
 +    Ext3 Security Labels (EXT3_FS_SECURITY) [Y/n/?] y
 +The Extended 4 (ext4) filesystem (EXT4_FS) [N/m/y/?] (NEW) m
 +</code>
 +
 +
 +Here you can see that the new kernel I am compiling has added support
 +for a new filesystem: ext4. //oldconfig// has gone through my
 +original configuration, kept all the old options exactly as they were
 +set, and prompted me on what to do with new options. Typically it is
 +safe to choose the default option, but you may wish change this.
 +//oldconfig// is a very handy tool for presenting you with only
 +new configuration options, making it ideal for users who simply have to
 +try out the latest kernel release.
 +
 +
 +For more serious configuration tasks, there are a multitude of options.
 +The linux kernel can be configured in three primary ways. The first is
 +//config//, which will step through each and every option one by
 +one and ask what you would like to do. This is so tedious that hardly
 +anyone ever uses it anymore.
 +
 +
 +<code>
 +darkstar:/usr/src/linux# make config
 +scripts/kconfig/conf arch/x86/Kconfig
 +*
 +* Linux Kernel Configuration
 +*
 +*
 +* General setup
 +*
 +Prompt for development and/or incomplete code/drivers (EXPERIMENTAL) [Y/n/?] Y
 +Local version - append to kernel release (LOCALVERSION) [] -test
 +Automatically append version information to the version string (LOCALVERSION_AUTO) [N/y/?] n
 +Support for paging of anonymous memory (swap) (SWAP) [Y/n/?]
 +</code>
 +
 +Fortunately, there are two much easier ways to configure your kernel,
 +//menuconfig// and //xconfig//. Both of these create a
 +menu-driven program that lets you select and de-select options without
 +having to step through each one. //menuconfig// is the most
 +commonly used method, and the one I recommend. //xconfig// is
 +only useful if you are attempting to compile the kernel from a
 +graphical user interface within **//X//**. Both are
 +so similar however, that we are only going to document
 +//menuconfig//.
 +
 +
 +Running make menuconfig from a terminal will
 +present you with the friendly curses-driven interface you see below.
 +Each kernel section is given its own submenu, and you can navigate with
 +the arrow keys.
 +
 +
 +{{ :slackbook:make-menuconfig-w.png |make-menuconfig}}
 +
 +<note warning>
 +  
 +If you are compiling a kernel that is the same release as the
 +stock Slackware kernel, you must set the //"Local version"//
 +option. This is found on the //"General setup"// submenu. Failure to
 +set this will result in your kernel compile over-writing all the
 +modules used by the stock kernels. This can quickly render your
 +system unbootable.
 +</note>
 +
 +Once you've finished configuring the kernel, it's time to begin
 +compiling it. There are many different methods for this, but the most
 +reliable is to use //bzImage//. When you pass this argument to
 +**//make//**, the kernel compilation will begin and
 +you will see lots of data scroll through the terminal until either the
 +compile process is complete or a fatal error is encountered.
 +
 +
 +<code>
 +darkstar:/usr/src/linux# make bzImage
 +scripts/kconfig/conf -s arch/x86/Kconfig
 +  CHK     include/linux/version.h
 +  CHK     include/linux/utsrelease.h
 +  SYMLINK include/asm -> include/asm-x86
 +  CALL    scripts/checksyscalls.sh
 +  CC      scripts/mod/empty.o
 +  HOSTCC  scripts/mod/mk_elfconfig
 +  MKELF   scripts/mod/elfconfig.h
 +  HOSTCC  scripts/mod/file2alias.o
 +... many hundreds of lines ommitted ...
 +</code>
 +
 +
 +If the process ends in an error, you should check your kernel
 +configuration first. Compile errors are usually caused by a fault
 +''.config'' file. Assuming everything went alright,
 +we're still not entirely finished, as we need to build the modules.
 +
 +
 +<code>
 +
 +darkstar:/usr/src/linux# make modules
 +  CHK     include/linux/version.h
 +  CHK     include/linux/utsrelease.h
 +  SYMLINK include/asm -> include/asm-x86
 +  CALL    scripts/checksyscalls.sh
 +  HOSTCC  scripts/mod/file2alias.o
 +... many thousands of lines omitted ...
 +</code>
 +
 +
 +If both the kernel and the modules compiles finished sucessfully, we're
 +ready to install them. The kernel image needs to be copied into a safe
 +location, typically the ''/boot'' directory, and you
 +should give it a unique name to avoid overwriting any other kernel
 +images located there. Traditionaly kernel images are named
 +''vmlinuz'' with the kernel release and local version
 +appended.
 +
 +
 +<code>
 +
 +  darkstar:/usr/src/linux# cat arch/x86/boot/bzImage > /boot/vmlinuz-release_number-local_version
 +  darkstar:/usr/src/linux# make modules_install
 +</code>
 +
 +
 +Once these steps have been completed, you will have a new kernel image
 +located under ''/boot'' and a new kernel modules
 +directory under ''/lib/modules''. In order to use
 +this new kernel, you will need to edit ''lilo.conf'',
 +create an initrd for it (only if you need to load one or more of this
 +kernel's modules to boot), and run **//lilo//** to
 +update the boot loader. When you reboot, if all went according to plan,
 +you should have an option to boot with your newly compiled kernel. If
 +something went wrong, you may be spending some time fixing the problem.
 ====== Chapter Navigation ====== ====== Chapter Navigation ======
  
 slackbook:linux_kernel ()