1    Linux kernel release 2.6.xx <>
3These are the release notes for Linux version 2.6. Read them carefully,
4as they tell you what this is all about, explain how to install the
5kernel, and what to do if something goes wrong.
9  Linux is a clone of the operating system Unix, written from scratch by
10  Linus Torvalds with assistance from a loosely-knit team of hackers across
11  the Net. It aims towards POSIX and Single UNIX Specification compliance.
13  It has all the features you would expect in a modern fully-fledged Unix,
14  including true multitasking, virtual memory, shared libraries, demand
15  loading, shared copy-on-write executables, proper memory management,
16  and multistack networking including IPv4 and IPv6.
18  It is distributed under the GNU General Public License - see the
19  accompanying COPYING file for more details.
23  Although originally developed first for 32-bit x86-based PCs (386 or higher),
24  today Linux also runs on (at least) the Compaq Alpha AXP, Sun SPARC and
25  UltraSPARC, Motorola 68000, PowerPC, PowerPC64, ARM, Hitachi SuperH, Cell,
26  IBM S/390, MIPS, HP PA-RISC, Intel IA-64, DEC VAX, AMD x86-64, AXIS CRIS,
27  Xtensa, AVR32 and Renesas M32R architectures.
29  Linux is easily portable to most general-purpose 32- or 64-bit architectures
30  as long as they have a paged memory management unit (PMMU) and a port of the
31  GNU C compiler (gcc) (part of The GNU Compiler Collection, GCC). Linux has
32  also been ported to a number of architectures without a PMMU, although
33  functionality is then obviously somewhat limited.
34  Linux has also been ported to itself. You can now run the kernel as a
35  userspace application - this is called UserMode Linux (UML).
39 - There is a lot of documentation available both in electronic form on
40   the Internet and in books, both Linux-specific and pertaining to
41   general UNIX questions. I'd recommend looking into the documentation
42   subdirectories on any Linux FTP site for the LDP (Linux Documentation
43   Project) books. This README is not meant to be documentation on the
44   system: there are much better sources available.
46 - There are various README files in the Documentation/ subdirectory:
47   these typically contain kernel-specific installation notes for some
48   drivers for example. See Documentation/00-INDEX for a list of what
49   is contained in each file. Please read the Changes file, as it
50   contains information about the problems, which may result by upgrading
51   your kernel.
53 - The Documentation/DocBook/ subdirectory contains several guides for
54   kernel developers and users. These guides can be rendered in a
55   number of formats: PostScript (.ps), PDF, HTML, & man-pages, among others.
56   After installation, "make psdocs", "make pdfdocs", "make htmldocs",
57   or "make mandocs" will render the documentation in the requested format.
59INSTALLING the kernel source:
61 - If you install the full sources, put the kernel tarball in a
62   directory where you have permissions (eg. your home directory) and
63   unpack it:
65        gzip -cd linux-2.6.XX.tar.gz | tar xvf -
67   or
68        bzip2 -dc linux-2.6.XX.tar.bz2 | tar xvf -
71   Replace "XX" with the version number of the latest kernel.
73   Do NOT use the /usr/src/linux area! This area has a (usually
74   incomplete) set of kernel headers that are used by the library header
75   files. They should match the library, and not get messed up by
76   whatever the kernel-du-jour happens to be.
78 - You can also upgrade between 2.6.xx releases by patching. Patches are
79   distributed in the traditional gzip and the newer bzip2 format. To
80   install by patching, get all the newer patch files, enter the
81   top level directory of the kernel source (linux-2.6.xx) and execute:
83        gzip -cd ../patch-2.6.xx.gz | patch -p1
85   or
86        bzip2 -dc ../patch-2.6.xx.bz2 | patch -p1
88   (repeat xx for all versions bigger than the version of your current
89   source tree, _in_order_) and you should be ok. You may want to remove
90   the backup files (xxx~ or xxx.orig), and make sure that there are no
91   failed patches (xxx# or xxx.rej). If there are, either you or me has
92   made a mistake.
94   Unlike patches for the 2.6.x kernels, patches for the 2.6.x.y kernels
95   (also known as the -stable kernels) are not incremental but instead apply
96   directly to the base 2.6.x kernel. Please read
97   Documentation/applying-patches.txt for more information.
99   Alternatively, the script patch-kernel can be used to automate this
100   process. It determines the current kernel version and applies any
101   patches found.
103        linux/scripts/patch-kernel linux
105   The first argument in the command above is the location of the
106   kernel source. Patches are applied from the current directory, but
107   an alternative directory can be specified as the second argument.
109 - If you are upgrading between releases using the stable series patches
110   (for example, patch-2.6.xx.y), note that these "dot-releases" are
111   not incremental and must be applied to the 2.6.xx base tree. For
112   example, if your base kernel is 2.6.12 and you want to apply the
113 patch, you do not and indeed must not first apply the
114 and patches. Similarly, if you are running kernel
115   version and want to jump to, you must first
116   reverse the patch (that is, patch -R) _before_ applying
117   the patch.
118   You can read more on this in Documentation/applying-patches.txt
120 - Make sure you have no stale .o files and dependencies lying around:
122        cd linux
123        make mrproper
125   You should now have the sources correctly installed.
129   Compiling and running the 2.6.xx kernels requires up-to-date
130   versions of various software packages. Consult
131   Documentation/Changes for the minimum version numbers required
132   and how to get updates for these packages. Beware that using
133   excessively old versions of these packages can cause indirect
134   errors that are very difficult to track down, so don't assume that
135   you can just update packages when obvious problems arise during
136   build or operation.
138BUILD directory for the kernel:
140   When compiling the kernel all output files will per default be
141   stored together with the kernel source code.
142   Using the option "make O=output/dir" allow you to specify an alternate
143   place for the output files (including .config).
144   Example:
145     kernel source code: /usr/src/linux-2.6.N
146     build directory: /home/name/build/kernel
148   To configure and build the kernel use:
149   cd /usr/src/linux-2.6.N
150   make O=/home/name/build/kernel menuconfig
151   make O=/home/name/build/kernel
152   sudo make O=/home/name/build/kernel modules_install install
154   Please note: If the 'O=output/dir' option is used then it must be
155   used for all invocations of make.
157CONFIGURING the kernel:
159   Do not skip this step even if you are only upgrading one minor
160   version. New configuration options are added in each release, and
161   odd problems will turn up if the configuration files are not set up
162   as expected. If you want to carry your existing configuration to a
163   new version with minimal work, use "make oldconfig", which will
164   only ask you for the answers to new questions.
166 - Alternate configuration commands are:
167    "make config" Plain text interface.
168    "make menuconfig" Text based color menus, radiolists & dialogs.
169    "make xconfig" X windows (Qt) based configuration tool.
170    "make gconfig" X windows (Gtk) based configuration tool.
171    "make oldconfig" Default all questions based on the contents of
172               your existing ./.config file and asking about
173               new config symbols.
174    "make silentoldconfig"
175               Like above, but avoids cluttering the screen
176               with questions already answered.
177               Additionally updates the dependencies.
178    "make defconfig" Create a ./.config file by using the default
179               symbol values from either arch/$ARCH/defconfig
180               or arch/$ARCH/configs/${PLATFORM}_defconfig,
181               depending on the architecture.
182    "make ${PLATFORM}_defconfig"
183              Create a ./.config file by using the default
184              symbol values from
185              arch/$ARCH/configs/${PLATFORM}_defconfig.
186              Use "make help" to get a list of all available
187              platforms of your architecture.
188    "make allyesconfig"
189               Create a ./.config file by setting symbol
190               values to 'y' as much as possible.
191    "make allmodconfig"
192               Create a ./.config file by setting symbol
193               values to 'm' as much as possible.
194    "make allnoconfig" Create a ./.config file by setting symbol
195               values to 'n' as much as possible.
196    "make randconfig" Create a ./.config file by setting symbol
197               values to random values.
199   You can find more information on using the Linux kernel config tools
200   in Documentation/kbuild/kconfig.txt.
202    NOTES on "make config":
203    - having unnecessary drivers will make the kernel bigger, and can
204      under some circumstances lead to problems: probing for a
205      nonexistent controller card may confuse your other controllers
206    - compiling the kernel with "Processor type" set higher than 386
207      will result in a kernel that does NOT work on a 386. The
208      kernel will detect this on bootup, and give up.
209    - A kernel with math-emulation compiled in will still use the
210      coprocessor if one is present: the math emulation will just
211      never get used in that case. The kernel will be slightly larger,
212      but will work on different machines regardless of whether they
213      have a math coprocessor or not.
214    - the "kernel hacking" configuration details usually result in a
215      bigger or slower kernel (or both), and can even make the kernel
216      less stable by configuring some routines to actively try to
217      break bad code to find kernel problems (kmalloc()). Thus you
218      should probably answer 'n' to the questions for
219          "development", "experimental", or "debugging" features.
221COMPILING the kernel:
223 - Make sure you have at least gcc 3.2 available.
224   For more information, refer to Documentation/Changes.
226   Please note that you can still run a.out user programs with this kernel.
228 - Do a "make" to create a compressed kernel image. It is also
229   possible to do "make install" if you have lilo installed to suit the
230   kernel makefiles, but you may want to check your particular lilo setup first.
232   To do the actual install you have to be root, but none of the normal
233   build should require that. Don't take the name of root in vain.
235 - If you configured any of the parts of the kernel as `modules', you
236   will also have to do "make modules_install".
238 - Verbose kernel compile/build output:
240   Normally the kernel build system runs in a fairly quiet mode (but not
241   totally silent). However, sometimes you or other kernel developers need
242   to see compile, link, or other commands exactly as they are executed.
243   For this, use "verbose" build mode. This is done by inserting
244   "V=1" in the "make" command. E.g.:
246    make V=1 all
248   To have the build system also tell the reason for the rebuild of each
249   target, use "V=2". The default is "V=0".
251 - Keep a backup kernel handy in case something goes wrong. This is
252   especially true for the development releases, since each new release
253   contains new code which has not been debugged. Make sure you keep a
254   backup of the modules corresponding to that kernel, as well. If you
255   are installing a new kernel with the same version number as your
256   working kernel, make a backup of your modules directory before you
257   do a "make modules_install".
258   Alternatively, before compiling, use the kernel config option
259   "LOCALVERSION" to append a unique suffix to the regular kernel version.
260   LOCALVERSION can be set in the "General Setup" menu.
262 - In order to boot your new kernel, you'll need to copy the kernel
263   image (e.g. .../linux/arch/i386/boot/bzImage after compilation)
264   to the place where your regular bootable kernel is found.
266 - Booting a kernel directly from a floppy without the assistance of a
267   bootloader such as LILO, is no longer supported.
269   If you boot Linux from the hard drive, chances are you use LILO which
270   uses the kernel image as specified in the file /etc/lilo.conf. The
271   kernel image file is usually /vmlinuz, /boot/vmlinuz, /bzImage or
272   /boot/bzImage. To use the new kernel, save a copy of the old image
273   and copy the new image over the old one. Then, you MUST RERUN LILO
274   to update the loading map!! If you don't, you won't be able to boot
275   the new kernel image.
277   Reinstalling LILO is usually a matter of running /sbin/lilo.
278   You may wish to edit /etc/lilo.conf to specify an entry for your
279   old kernel image (say, /vmlinux.old) in case the new one does not
280   work. See the LILO docs for more information.
282   After reinstalling LILO, you should be all set. Shutdown the system,
283   reboot, and enjoy!
285   If you ever need to change the default root device, video mode,
286   ramdisk size, etc. in the kernel image, use the 'rdev' program (or
287   alternatively the LILO boot options when appropriate). No need to
288   recompile the kernel to change these parameters.
290 - Reboot with the new kernel and enjoy.
294 - If you have problems that seem to be due to kernel bugs, please check
295   the file MAINTAINERS to see if there is a particular person associated
296   with the part of the kernel that you are having trouble with. If there
297   isn't anyone listed there, then the second best thing is to mail
298   them to me (, and possibly to any other
299   relevant mailing-list or to the newsgroup.
301 - In all bug-reports, *please* tell what kernel you are talking about,
302   how to duplicate the problem, and what your setup is (use your common
303   sense). If the problem is new, tell me so, and if the problem is
304   old, please try to tell me when you first noticed it.
306 - If the bug results in a message like
308    unable to handle kernel paging request at address C0000010
309    Oops: 0002
310    EIP: 0010:XXXXXXXX
311    eax: xxxxxxxx ebx: xxxxxxxx ecx: xxxxxxxx edx: xxxxxxxx
312    esi: xxxxxxxx edi: xxxxxxxx ebp: xxxxxxxx
313    ds: xxxx es: xxxx fs: xxxx gs: xxxx
314    Pid: xx, process nr: xx
315    xx xx xx xx xx xx xx xx xx xx
317   or similar kernel debugging information on your screen or in your
318   system log, please duplicate it *exactly*. The dump may look
319   incomprehensible to you, but it does contain information that may
320   help debugging the problem. The text above the dump is also
321   important: it tells something about why the kernel dumped code (in
322   the above example it's due to a bad kernel pointer). More information
323   on making sense of the dump is in Documentation/oops-tracing.txt
325 - If you compiled the kernel with CONFIG_KALLSYMS you can send the dump
326   as is, otherwise you will have to use the "ksymoops" program to make
327   sense of the dump (but compiling with CONFIG_KALLSYMS is usually preferred).
328   This utility can be downloaded from
329   ftp://ftp.<country> .
330   Alternately you can do the dump lookup by hand:
332 - In debugging dumps like the above, it helps enormously if you can
333   look up what the EIP value means. The hex value as such doesn't help
334   me or anybody else very much: it will depend on your particular
335   kernel setup. What you should do is take the hex value from the EIP
336   line (ignore the "0010:"), and look it up in the kernel namelist to
337   see which kernel function contains the offending address.
339   To find out the kernel function name, you'll need to find the system
340   binary associated with the kernel that exhibited the symptom. This is
341   the file 'linux/vmlinux'. To extract the namelist and match it against
342   the EIP from the kernel crash, do:
344        nm vmlinux | sort | less
346   This will give you a list of kernel addresses sorted in ascending
347   order, from which it is simple to find the function that contains the
348   offending address. Note that the address given by the kernel
349   debugging messages will not necessarily match exactly with the
350   function addresses (in fact, that is very unlikely), so you can't
351   just 'grep' the list: the list will, however, give you the starting
352   point of each kernel function, so by looking for the function that
353   has a starting address lower than the one you are searching for but
354   is followed by a function with a higher address you will find the one
355   you want. In fact, it may be a good idea to include a bit of
356   "context" in your problem report, giving a few lines around the
357   interesting one.
359   If you for some reason cannot do the above (you have a pre-compiled
360   kernel image or similar), telling me as much about your setup as
361   possible will help. Please read the REPORTING-BUGS document for details.
363 - Alternately, you can use gdb on a running kernel. (read-only; i.e. you
364   cannot change values or set break points.) To do this, first compile the
365   kernel with -g; edit arch/i386/Makefile appropriately, then do a "make
366   clean". You'll also need to enable CONFIG_PROC_FS (via "make config").
368   After you've rebooted with the new kernel, do "gdb vmlinux /proc/kcore".
369   You can now use all the usual gdb commands. The command to look up the
370   point where your system crashed is "l *0xXXXXXXXX". (Replace the XXXes
371   with the EIP value.)
373   gdb'ing a non-running kernel currently fails because gdb (wrongly)
374   disregards the starting offset for which the kernel is compiled.

Archive Download this file