xref: /linux/README (revision 90890687859ea658759e653c4e70ed7e9e1a6217)
1	Linux kernel release 2.6.xx
2
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.
6
7WHAT IS LINUX?
8
9  Linux is a Unix clone written from scratch by Linus Torvalds with
10  assistance from a loosely-knit team of hackers across the Net.
11  It aims towards POSIX compliance.
12
13  It has all the features you would expect in a modern fully-fledged
14  Unix, including true multitasking, virtual memory, shared libraries,
15  demand loading, shared copy-on-write executables, proper memory
16  management and TCP/IP networking.
17
18  It is distributed under the GNU General Public License - see the
19  accompanying COPYING file for more details.
20
21ON WHAT HARDWARE DOES IT RUN?
22
23  Linux was first developed for 386/486-based PCs.  These days it also
24  runs on ARMs, DEC Alphas, SUN Sparcs, M68000 machines (like Atari and
25  Amiga), MIPS and PowerPC, and others.
26
27DOCUMENTATION:
28
29 - There is a lot of documentation available both in electronic form on
30   the Internet and in books, both Linux-specific and pertaining to
31   general UNIX questions.  I'd recommend looking into the documentation
32   subdirectories on any Linux FTP site for the LDP (Linux Documentation
33   Project) books.  This README is not meant to be documentation on the
34   system: there are much better sources available.
35
36 - There are various README files in the Documentation/ subdirectory:
37   these typically contain kernel-specific installation notes for some
38   drivers for example. See Documentation/00-INDEX for a list of what
39   is contained in each file.  Please read the Changes file, as it
40   contains information about the problems, which may result by upgrading
41   your kernel.
42
43 - The Documentation/DocBook/ subdirectory contains several guides for
44   kernel developers and users.  These guides can be rendered in a
45   number of formats:  PostScript (.ps), PDF, and HTML, among others.
46   After installation, "make psdocs", "make pdfdocs", or "make htmldocs"
47   will render the documentation in the requested format.
48
49INSTALLING the kernel:
50
51 - If you install the full sources, put the kernel tarball in a
52   directory where you have permissions (eg. your home directory) and
53   unpack it:
54
55		gzip -cd linux-2.6.XX.tar.gz | tar xvf -
56
57   Replace "XX" with the version number of the latest kernel.
58
59   Do NOT use the /usr/src/linux area! This area has a (usually
60   incomplete) set of kernel headers that are used by the library header
61   files.  They should match the library, and not get messed up by
62   whatever the kernel-du-jour happens to be.
63
64 - You can also upgrade between 2.6.xx releases by patching.  Patches are
65   distributed in the traditional gzip and the new bzip2 format.  To
66   install by patching, get all the newer patch files, enter the
67   top level directory of the kernel source (linux-2.6.xx) and execute:
68
69		gzip -cd ../patch-2.6.xx.gz | patch -p1
70
71   or
72		bzip2 -dc ../patch-2.6.xx.bz2 | patch -p1
73
74   (repeat xx for all versions bigger than the version of your current
75   source tree, _in_order_) and you should be ok.  You may want to remove
76   the backup files (xxx~ or xxx.orig), and make sure that there are no
77   failed patches (xxx# or xxx.rej). If there are, either you or me has
78   made a mistake.
79
80   Alternatively, the script patch-kernel can be used to automate this
81   process.  It determines the current kernel version and applies any
82   patches found.
83
84		linux/scripts/patch-kernel linux
85
86   The first argument in the command above is the location of the
87   kernel source.  Patches are applied from the current directory, but
88   an alternative directory can be specified as the second argument.
89
90 - If you are upgrading between releases using the stable series patches
91   (for example, patch-2.6.xx.y), note that these "dot-releases" are
92   not incremental and must be applied to the 2.6.xx base tree. For
93   example, if your base kernel is 2.6.12 and you want to apply the
94   2.6.12.3 patch, you do not and indeed must not first apply the
95   2.6.12.1 and 2.6.12.2 patches. Similarly, if you are running kernel
96   version 2.6.12.2 and want to jump to 2.6.12.3, you must first
97   reverse the 2.6.12.2 patch (that is, patch -R) _before_ applying
98   the 2.6.12.3 patch.
99
100 - Make sure you have no stale .o files and dependencies lying around:
101
102		cd linux
103		make mrproper
104
105   You should now have the sources correctly installed.
106
107SOFTWARE REQUIREMENTS
108
109   Compiling and running the 2.6.xx kernels requires up-to-date
110   versions of various software packages.  Consult
111   Documentation/Changes for the minimum version numbers required
112   and how to get updates for these packages.  Beware that using
113   excessively old versions of these packages can cause indirect
114   errors that are very difficult to track down, so don't assume that
115   you can just update packages when obvious problems arise during
116   build or operation.
117
118BUILD directory for the kernel:
119
120   When compiling the kernel all output files will per default be
121   stored together with the kernel source code.
122   Using the option "make O=output/dir" allow you to specify an alternate
123   place for the output files (including .config).
124   Example:
125     kernel source code:	/usr/src/linux-2.6.N
126     build directory:		/home/name/build/kernel
127
128   To configure and build the kernel use:
129   cd /usr/src/linux-2.6.N
130   make O=/home/name/build/kernel menuconfig
131   make O=/home/name/build/kernel
132   sudo make O=/home/name/build/kernel modules_install install
133
134   Please note: If the 'O=output/dir' option is used then it must be
135   used for all invocations of make.
136
137CONFIGURING the kernel:
138
139   Do not skip this step even if you are only upgrading one minor
140   version.  New configuration options are added in each release, and
141   odd problems will turn up if the configuration files are not set up
142   as expected.  If you want to carry your existing configuration to a
143   new version with minimal work, use "make oldconfig", which will
144   only ask you for the answers to new questions.
145
146 - Alternate configuration commands are:
147	"make menuconfig"  Text based color menus, radiolists & dialogs.
148	"make xconfig"     X windows (Qt) based configuration tool.
149	"make gconfig"     X windows (Gtk) based configuration tool.
150	"make oldconfig"   Default all questions based on the contents of
151			   your existing ./.config file.
152	"make silentoldconfig"
153			   Like above, but avoids cluttering the screen
154			   with questions already answered.
155
156	NOTES on "make config":
157	- having unnecessary drivers will make the kernel bigger, and can
158	  under some circumstances lead to problems: probing for a
159	  nonexistent controller card may confuse your other controllers
160	- compiling the kernel with "Processor type" set higher than 386
161	  will result in a kernel that does NOT work on a 386.  The
162	  kernel will detect this on bootup, and give up.
163	- A kernel with math-emulation compiled in will still use the
164	  coprocessor if one is present: the math emulation will just
165	  never get used in that case.  The kernel will be slightly larger,
166	  but will work on different machines regardless of whether they
167	  have a math coprocessor or not.
168	- the "kernel hacking" configuration details usually result in a
169	  bigger or slower kernel (or both), and can even make the kernel
170	  less stable by configuring some routines to actively try to
171	  break bad code to find kernel problems (kmalloc()).  Thus you
172	  should probably answer 'n' to the questions for
173          "development", "experimental", or "debugging" features.
174
175COMPILING the kernel:
176
177 - Make sure you have gcc 2.95.3 available.
178   gcc 2.91.66 (egcs-1.1.2), and gcc 2.7.2.3 are known to miscompile
179   some parts of the kernel, and are *no longer supported*.
180   Also remember to upgrade your binutils package (for as/ld/nm and company)
181   if necessary. For more information, refer to Documentation/Changes.
182
183   Please note that you can still run a.out user programs with this kernel.
184
185 - Do a "make" to create a compressed kernel image. It is also
186   possible to do "make install" if you have lilo installed to suit the
187   kernel makefiles, but you may want to check your particular lilo setup first.
188
189   To do the actual install you have to be root, but none of the normal
190   build should require that. Don't take the name of root in vain.
191
192 - If you configured any of the parts of the kernel as `modules', you
193   will also have to do "make modules_install".
194
195 - Keep a backup kernel handy in case something goes wrong.  This is
196   especially true for the development releases, since each new release
197   contains new code which has not been debugged.  Make sure you keep a
198   backup of the modules corresponding to that kernel, as well.  If you
199   are installing a new kernel with the same version number as your
200   working kernel, make a backup of your modules directory before you
201   do a "make modules_install".
202   Alternatively, before compiling, use the kernel config option
203   "LOCALVERSION" to append a unique suffix to the regular kernel version.
204   LOCALVERSION can be set in the "General Setup" menu.
205
206 - In order to boot your new kernel, you'll need to copy the kernel
207   image (e.g. .../linux/arch/i386/boot/bzImage after compilation)
208   to the place where your regular bootable kernel is found.
209
210 - Booting a kernel directly from a floppy without the assistance of a
211   bootloader such as LILO, is no longer supported.
212
213   If you boot Linux from the hard drive, chances are you use LILO which
214   uses the kernel image as specified in the file /etc/lilo.conf.  The
215   kernel image file is usually /vmlinuz, /boot/vmlinuz, /bzImage or
216   /boot/bzImage.  To use the new kernel, save a copy of the old image
217   and copy the new image over the old one.  Then, you MUST RERUN LILO
218   to update the loading map!! If you don't, you won't be able to boot
219   the new kernel image.
220
221   Reinstalling LILO is usually a matter of running /sbin/lilo.
222   You may wish to edit /etc/lilo.conf to specify an entry for your
223   old kernel image (say, /vmlinux.old) in case the new one does not
224   work.  See the LILO docs for more information.
225
226   After reinstalling LILO, you should be all set.  Shutdown the system,
227   reboot, and enjoy!
228
229   If you ever need to change the default root device, video mode,
230   ramdisk size, etc.  in the kernel image, use the 'rdev' program (or
231   alternatively the LILO boot options when appropriate).  No need to
232   recompile the kernel to change these parameters.
233
234 - Reboot with the new kernel and enjoy.
235
236IF SOMETHING GOES WRONG:
237
238 - If you have problems that seem to be due to kernel bugs, please check
239   the file MAINTAINERS to see if there is a particular person associated
240   with the part of the kernel that you are having trouble with. If there
241   isn't anyone listed there, then the second best thing is to mail
242   them to me (torvalds@osdl.org), and possibly to any other relevant
243   mailing-list or to the newsgroup.
244
245 - In all bug-reports, *please* tell what kernel you are talking about,
246   how to duplicate the problem, and what your setup is (use your common
247   sense).  If the problem is new, tell me so, and if the problem is
248   old, please try to tell me when you first noticed it.
249
250 - If the bug results in a message like
251
252	unable to handle kernel paging request at address C0000010
253	Oops: 0002
254	EIP:   0010:XXXXXXXX
255	eax: xxxxxxxx   ebx: xxxxxxxx   ecx: xxxxxxxx   edx: xxxxxxxx
256	esi: xxxxxxxx   edi: xxxxxxxx   ebp: xxxxxxxx
257	ds: xxxx  es: xxxx  fs: xxxx  gs: xxxx
258	Pid: xx, process nr: xx
259	xx xx xx xx xx xx xx xx xx xx
260
261   or similar kernel debugging information on your screen or in your
262   system log, please duplicate it *exactly*.  The dump may look
263   incomprehensible to you, but it does contain information that may
264   help debugging the problem.  The text above the dump is also
265   important: it tells something about why the kernel dumped code (in
266   the above example it's due to a bad kernel pointer). More information
267   on making sense of the dump is in Documentation/oops-tracing.txt
268
269 - If you compiled the kernel with CONFIG_KALLSYMS you can send the dump
270   as is, otherwise you will have to use the "ksymoops" program to make
271   sense of the dump.  This utility can be downloaded from
272   ftp://ftp.<country>.kernel.org/pub/linux/utils/kernel/ksymoops.
273   Alternately you can do the dump lookup by hand:
274
275 - In debugging dumps like the above, it helps enormously if you can
276   look up what the EIP value means.  The hex value as such doesn't help
277   me or anybody else very much: it will depend on your particular
278   kernel setup.  What you should do is take the hex value from the EIP
279   line (ignore the "0010:"), and look it up in the kernel namelist to
280   see which kernel function contains the offending address.
281
282   To find out the kernel function name, you'll need to find the system
283   binary associated with the kernel that exhibited the symptom.  This is
284   the file 'linux/vmlinux'.  To extract the namelist and match it against
285   the EIP from the kernel crash, do:
286
287		nm vmlinux | sort | less
288
289   This will give you a list of kernel addresses sorted in ascending
290   order, from which it is simple to find the function that contains the
291   offending address.  Note that the address given by the kernel
292   debugging messages will not necessarily match exactly with the
293   function addresses (in fact, that is very unlikely), so you can't
294   just 'grep' the list: the list will, however, give you the starting
295   point of each kernel function, so by looking for the function that
296   has a starting address lower than the one you are searching for but
297   is followed by a function with a higher address you will find the one
298   you want.  In fact, it may be a good idea to include a bit of
299   "context" in your problem report, giving a few lines around the
300   interesting one.
301
302   If you for some reason cannot do the above (you have a pre-compiled
303   kernel image or similar), telling me as much about your setup as
304   possible will help.
305
306 - Alternately, you can use gdb on a running kernel. (read-only; i.e. you
307   cannot change values or set break points.) To do this, first compile the
308   kernel with -g; edit arch/i386/Makefile appropriately, then do a "make
309   clean". You'll also need to enable CONFIG_PROC_FS (via "make config").
310
311   After you've rebooted with the new kernel, do "gdb vmlinux /proc/kcore".
312   You can now use all the usual gdb commands. The command to look up the
313   point where your system crashed is "l *0xXXXXXXXX". (Replace the XXXes
314   with the EIP value.)
315
316   gdb'ing a non-running kernel currently fails because gdb (wrongly)
317   disregards the starting offset for which the kernel is compiled.
318
319