2015/05/19 19:31
阅读数 17




Linux kernel release 2.6.xx

These are the release notes for Linux version 2.6.  Read them carefully,
as they tell you what this is all about, explain how to install the
kernel, and what to do if something goes wrong. 


  Linux is a Unix clone written from scratch by Linus Torvalds with
  assistance from a loosely-knit team of hackers across the Net.
  It aims towards POSIX compliance. 
Linux是Liuns Torvalds在一些来自网络的电脑爱好者团队的帮助下从头开始写的unix克隆版,它致力于POSIX标准。

  It has all the features you would expect in a modern fully-fledged
  Unix, including true multitasking, virtual memory, shared libraries,
  demand loading, shared copy-on-write executables, proper memory
  management and TCP/IP networking. 

  It is distributed under the GNU General Public License - see the
  accompanying COPYING file for more details. 


  Linux was first developed for 386/486-based PCs.  These days it also
  runs on ARMs, DEC Alphas, SUN Sparcs, M68000 machines (like Atari and
  Amiga), MIPS and PowerPC, and others.

Linux起初是为386/486架构的pc机开发的,但是现在他也可以运行在 ARMs, DEC Alphas, SUN Sparcs, M68000 machines (像 Atari and
  Amiga), MIPS , PowerPC和其他架构的机器上。


 - There is a lot of documentation available both in electronic form on
   the Internet and in books, both Linux-specific and pertaining to
   general UNIX questions.  I'd recommend looking into the documentation
   subdirectories on any Linux FTP site for the LDP (Linux Documentation
   Project) books.  This README is not meant to be documentation on the
   system: there are much better sources available.

 - There are various README files in the Documentation/ subdirectory:
   these typically contain kernel-specific installation notes for some 
   drivers for example. See ./Documentation/00-INDEX for a list of what
   is contained in each file.  Please read the Changes file, as it
   contains information about the problems, which may result by upgrading
   your kernel.

在主文件的子目录下有很多README文件,这些通常是包含着一些内核专有的安装信息,例如一些驱动程序。查看目录./Documentation/00-INDEX下的文件列表,这些文件介绍每一个文件都包含什么信息。请月的Changes file,因为它包含了一些关于问题的信息,这些问题将会是你更新你的内核后的结果。

 - The Documentation/DocBook/ subdirectory contains several guides for
   kernel developers and users.  These guides can be rendered in a
   number of formats:  PostScript (.ps), PDF, and HTML, among others.
   After installation, "make psdocs", "make pdfdocs", or "make htmldocs"
   will render the documentation in the requested format.
子目录 Documentation/DocBook/ 为内核开发者和使用这提供了一些指导文件。这些指导文件可以被转换成很多格式的文件:PostScript (.ps), PDF, and HTML,还有其他很多。在安装完成之后,使用"make psdocs", "make pdfdocs", 或者是 "make htmldocs"命令可以将文档转换成相应的格式。

INSTALLING the kernel:

 - If you install the full sources, put the kernel tarball in a
   directory where you have permissions (eg. your home directory) and
   unpack it:

gzip -cd linux-2.6.XX.tar.gz | tar xvf -

   Replace "XX" with the version number of the latest kernel.
gzip -cd linux-2.6.XX.tar.gz | tar xvf -

   Do NOT use the /usr/src/linux area! This area has a (usually
   incomplete) set of kernel headers that are used by the library header
   files.  They should match the library, and not get messed up by
   whatever the kernel-du-jour happens to be.
不要使用 /usr/src/linux区域,这块区域有一些内核头文件集(通常不完整),这些头文件被用作库的头文件。他们应该与库相符合,不要将他们弄乱。

 - You can also upgrade between 2.6.xx releases by patching.  Patches are
   distributed in the traditional gzip and the new bzip2 format.  To
   install by patching, get all the newer patch files, enter the
   top level directory of the kernel source (linux-2.6.xx) and execute:

gzip -cd ../patch-2.6.xx.gz | patch -p1

bzip2 -dc ../patch-2.6.xx.bz2 | patch -p1

   (repeat xx for all versions bigger than the version of your current
   source tree, _in_order_) and you should be ok.  You may want to remove
   the backup files (xxx~ or xxx.orig), and make sure that there are no
   failed patches (xxx# or xxx.rej). If there are, either you or me has
   made a mistake.
gzip -cd ../patch-2.6.xx.gz | patch -p1
bzip2 -dc ../patch-2.6.xx.bz2 | patch -p1
(重复的XX是比你当前内核版本号要高的版本号,按顺序进行排列),你应该可以。应可能想要去除备份文件(xxx~ 或者是 xxx.reg),一定要确认你没有失败的补丁(xxx# 或者是 xxx.reg).如果有的话,你和我都已经犯了错误。

   Alternatively, the script patch-kernel can be used to automate this
   process.  It determines the current kernel version and applies any
   patches found.

linux/scripts/patch-kernel linux

   The first argument in the command above is the location of the
   kernel source.  Patches are applied from the current directory, but
   an alternative directory can be specified as the second argument.
linux/scripts/patch-kernel linux

 - Make sure you have no stale .o files and dependencies lying around:

cd linux
make mrproper

   You should now have the sources correctly installed.
cd linux
make mrproper


   Compiling and running the 2.6.xx kernels requires up-to-date
   versions of various software packages.  Consult
   ./Documentation/Changes for the minimum version numbers required
   and how to get updates for these packages.  Beware that using
   excessively old versions of these packages can cause indirect
   errors that are very difficult to track down, so don't assume that
   you can just update packages when obvious problems arise during
   build or operation.

BUILD directory for the kernel:

   When compiling the kernel all output files will per default be
   stored together with the kernel source code.
   Using the option "make O=output/dir" allow you to specify an alternate
   place for the output files (including .config).
     kernel source code: /usr/src/linux-2.6.N
     build directory: /home/name/build/kernel

   To configure and build the kernel use:
   cd /usr/src/linux-2.6.N
   make O=/home/name/build/kernel menuconfig
   make O=/home/name/build/kernel
   sudo make O=/home/name/build/kernel install_modules install

   Please note: If the 'O=output/dir' option is used then it must be
   used for all invocations of make.
当编译内核的时候所有的输出文件都默认与内核源码保存在一起。使用选项"make O=output/dir"
Kernel source code: /usr/src/linux-2.6.N
build directory:    /home/name/build/kernel
cd /usr/src/linux-2.6.N
make O=/home/name/build/kernel menuconfig
make O=/home/name/build/kernel
sudo make O=/home/name/build/kernel install_modules install

CONFIGURING the kernel:

   Do not skip this step even if you are only upgrading one minor
   version.  New configuration options are added in each release, and
   odd problems will turn up if the configuration files are not set up
   as expected.  If you want to carry your existing configuration to a
   new version with minimal work, use "make oldconfig", which will
   only ask you for the answers to new questions.
如果你想要使用现在的配置应用到最新版本来使自己有最小工作量,使用命令"make config"

 - Alternate configuration commands are:
"make menuconfig"  Text based color menus, radiolists & dialogs.
"make xconfig"     X windows (Qt) based configuration tool.
"make gconfig"     X windows (Gtk) based configuration tool.
"make oldconfig"   Default all questions based on the contents of
  your existing ./.config file.
"make menuconfig"  带有颜色的菜单,选项列表和对话框的文本
"make xconfig"     带有配置工具的X窗体(Qt)
"make gconfig"     带有配置工具的X窗体(Gtk)
NOTES on "make config":
"make config"的注意事项:
- having unnecessary drivers will make the kernel bigger, and can
 under some circumstances lead to problems: probing for a
 nonexistent controller card may confuse your other controllers
- 配置不必要的驱动将会是内核更大,并且在一些环境下可能会导致一些问题:
- compiling the kernel with "Processor type" set higher than 386
 will result in a kernel that does NOT work on a 386.  The
 kernel will detect this on bootup, and give up.
- 编译内核的时候,将"professor type"设置的比386要高的话,将会导致内核在
- A kernel with math-emulation compiled in will still use the
 coprocessor if one is present: the math emulation will just
 never get used in that case.  The kernel will be slightly larger,
 but will work on different machines regardless of whether they
 have a math coprocessor or not. 
- 将数学仿真编译进内核将会使用存在的协处理器,如果协处理器不存在,那么
- the "kernel hacking" configuration details usually result in a
 bigger or slower kernel (or both), and can even make the kernel
 less stable by configuring some routines to actively try to
 break bad code to find kernel problems (kmalloc()).  Thus you
 should probably answer 'n' to the questions for
          "development", "experimental", or "debugging" features.
- "kernel hacking"的配置细节通常会导致内核更大或者是更慢(或者是两者
 - Check the top Makefile for further site-dependent configuration
   (default SVGA mode etc). 
 - 检查顶层的Makefile文件来获取更多的依赖配置(默认的SVGA模式等)。
COMPILING the kernel:

 - Make sure you have gcc 2.95.3 available.
   gcc 2.91.66 (egcs-1.1.2), and gcc are known to miscompile
   some parts of the kernel, and are *no longer supported*.
   Also remember to upgrade your binutils package (for as/ld/nm and company)
   if necessary. For more information, refer to ./Documentation/Changes.
 - 确定你有gcc 2.95.3编译器。
   gcc 2.91.66(还有-1.1.10)和gcc编译器可能会错误的编译内核的一部分
   程序,所以以不再被支持。如果必要的话也要记着取更新你的二进制包(例如 as/ld/

   Please note that you can still run a.out user programs with this kernel.

 - Do a "make bzImage" to create a compressed kernel image.  If you want
   to make a boot disk (without root filesystem or LILO), insert a floppy
   in your A: drive, and do a "make bzdisk".  It is also possible to do
   "make install" if you have lilo installed to suit the kernel makefiles,
   but you may want to check your particular lilo setup first. 
 - 使用命令"make bzImage"来创建一个压缩的内核镜像文件。如果你想制作一个启动盘
   的A驱下插入一个软盘,并且使用命令"make bzdisk"进行制作。如果你有安装好的lilo
   与内核文件相适应的话,命令"make install"也是可以的,但是首先你应该先检查你的

   To do the actual install you have to be root, but none of the normal
   build should require that. Don't take the name of root in vain.

 - In the unlikely event that your system cannot boot bzImage kernels you
   can still compile your kernel as zImage. However, since zImage support
   will be removed at some point in the future in favor of bzImage we
   encourage people having problems with booting bzImage kernels to report
   these, with detailed hardware configuration information, to the
   linux-kernel mailing list and to H. Peter Anvin <hpa+linux@zytor.com>.
 - 存在一些不太可能的事件,例如你的系统不能启动bzImage内核但是依然能够编译内核为bzImage文件。 然而尽管在未来zImage支持可能会被移除,但是为了支持bzImage.我们鼓励人们遇到启动bzImage问题的时候能够上传这些问题,同时上传具体的硬件配置信息。将这些信息以邮件的形式发送给Peter Anvin <hpa+linux@zytor.com>。

 - If you configured any of the parts of the kernel as `modules', you
   will have to do "make modules" followed by "make modules_install".
 - 如果你将内核的一些部分配置成'modules',你必须使用"make modules"和"make

 - Keep a backup kernel handy in case something goes wrong.  This is 
   especially true for the development releases, since each new release
   contains new code which has not been debugged.  Make sure you keep a
   backup of the modules corresponding to that kernel, as well.  If you
   are installing a new kernel with the same version number as your
   working kernel, make a backup of your modules directory before you
   do a "make modules_install".
 - 为了防止出现错误,一定要对内核进行备份。对一些开发发行版这种做法是尤其正确
   的。因为一些新的发行版包含一些新的代码,但这些代码可能还没有被调试。还要备份一些与内核相对应的模块。如果你安装一个具有相同版本号的新内核作为你的运行内核,在你使用命令"make modules_install"之前备份你的模块目录。

 - In order to boot your new kernel, you'll need to copy the kernel
   image (found in .../linux/arch/i386/boot/bzImage after compilation)
   to the place where your regular bootable kernel is found.
 - 为了启动你的新内核,你需要将内核镜像(编译完成后,在目录.../linux/arch/i386

   For some, this is on a floppy disk, in which case you can copy the
   kernel bzImage file to /dev/fd0 to make a bootable floppy.

   If you boot Linux from the hard drive, chances are you use LILO which
   uses the kernel image as specified in the file /etc/lilo.conf.  The
   kernel image file is usually /vmlinuz, /boot/vmlinuz, /bzImage or
   /boot/bzImage.  To use the new kernel, save a copy of the old image
   and copy the new image over the old one.  Then, you MUST RERUN LILO
   to update the loading map!! If you don't, you won't be able to boot
   the new kernel image.

   Reinstalling LILO is usually a matter of running /sbin/lilo. 
   You may wish to edit /etc/lilo.conf to specify an entry for your
   old kernel image (say, /vmlinux.old) in case the new one does not
   work.  See the LILO docs for more information. 
   重新安装LILO通常的做法是运行 /sbin/lilo命令。你可能希望便器 /etc/lilo.conf

   After reinstalling LILO, you should be all set.  Shutdown the system,
   reboot, and enjoy!

   If you ever need to change the default root device, video mode,
   ramdisk size, etc.  in the kernel image, use the 'rdev' program (or
   alternatively the LILO boot options when appropriate).  No need to
   recompile the kernel to change these parameters. 

 - Reboot with the new kernel and enjoy. 


 - If you have problems that seem to be due to kernel bugs, please check
   the file MAINTAINERS to see if there is a particular person associated
   with the part of the kernel that you are having trouble with. If there
   isn't anyone listed there, then the second best thing is to mail
   them to me (torvalds@osdl.org), and possibly to any other relevant
   mailing-list or to the newsgroup.
 - 如果你遇到了一些问题,并且这些问题看上去像是内核bug,请查找文件MAINTAINGS来查看是否有与你所遇到问题的这部分内核相关的人,如果人物列表中没有,那么最好的方式就是给我发邮件(torvalds@osdl.org),或者是给其他相关人员发送邮件。

 - In all bug-reports, *please* tell what kernel you are talking about,
   how to duplicate the problem, and what your setup is (use your common
   sense).  If the problem is new, tell me so, and if the problem is
   old, please try to tell me when you first noticed it.
 - 在所有的上传的bugs中,请告诉我们你正在讨论的是哪一个版本的内核,如何重复

 - If the bug results in a message like
 - 如果一个bug导致了像以下类似的消息

unable to handle kernel paging request at address C0000010
Oops: 0002
eax: xxxxxxxx   ebx: xxxxxxxx   ecx: xxxxxxxx   edx: xxxxxxxx
esi: xxxxxxxx   edi: xxxxxxxx   ebp: xxxxxxxx
ds: xxxx  es: xxxx  fs: xxxx  gs: xxxx
Pid: xx, process nr: xx
xx xx xx xx xx xx xx xx xx xx
    不能够在地址 C0000010处处理内核页请求
    Oops: 0002
eax: xxxxxxxx   ebx: xxxxxxxx   ecx: xxxxxxxx   edx: xxxxxxxx
esi: xxxxxxxx   edi: xxxxxxxx   ebp: xxxxxxxx
ds: xxxx  es: xxxx  fs: xxxx  gs: xxxx
Pid: xx, process nr: xx
xx xx xx xx xx xx xx xx xx xx

   or similar kernel debugging information on your screen or in your
   system log, please duplicate it *exactly*.  The dump may look
   incomprehensible to you, but it does contain information that may
   help debugging the problem.  The text above the dump is also
   important: it tells something about why the kernel dumped code (in
   the above example it's due to a bad kernel pointer). More information
   on making sense of the dump is in Documentation/oops-tracing.txt
   或者是在你的屏幕上显示类似的内核调试信息或者是在你的系统日志中,请确切的重复这些调试信息给我们。转储对你来说可能不是很理解,但是她确实包含着一些能够帮助调试问题的信息。上述转储文本也确实很重要:他告诉我们为什么内核放弃代码(在上述例子中,他是由于错误的内核指针导致的)。要想获取更多的能够理解转储的信息,请查看文档 /oops-tracing.txt.
 - You can use the "ksymoops" program to make sense of the dump.  This
   utility can be downloaded from
   Alternately you can do the dump lookup by hand:
 - 你可以使用"ksymoops"程序来理解转储。这个工具可以从

 - In debugging dumps like the above, it helps enormously if you can
   look up what the EIP value means.  The hex value as such doesn't help
   me or anybody else very much: it will depend on your particular
   kernel setup.  What you should do is take the hex value from the EIP
   line (ignore the "0010:"), and look it up in the kernel namelist to
   see which kernel function contains the offending address.
 - 像上述方式调试内核的时候,如果你能够查看EIP值代表什么,那么他的帮助是巨大

   To find out the kernel function name, you'll need to find the system
   binary associated with the kernel that exhibited the symptom.  This is
   the file 'linux/vmlinux'.  To extract the namelist and match it against
   the EIP from the kernel crash, do:

nm vmlinux | sort | less
     nm vmlinuz | sort | less

   This will give you a list of kernel addresses sorted in ascending
   order, from which it is simple to find the function that contains the
   offending address.  Note that the address given by the kernel
   debugging messages will not necessarily match exactly with the
   function addresses (in fact, that is very unlikely), so you can't
   just 'grep' the list: the list will, however, give you the starting
   point of each kernel function, so by looking for the function that
   has a starting address lower than the one you are searching for but
   is followed by a function with a higher address you will find the one
   you want.  In fact, it may be a good idea to include a bit of
   "context" in your problem report, giving a few lines around the
   interesting one. 

   If you for some reason cannot do the above (you have a pre-compiled
   kernel image or similar), telling me as much about your setup as
   possible will help. 

 - Alternately, you can use gdb on a running kernel. (read-only; i.e. you
   cannot change values or set break points.) To do this, first compile the
   kernel with -g; edit arch/i386/Makefile appropriately, then do a "make
   clean". You'll also need to enable CONFIG_PROC_FS (via "make config").
 - 除此之后,在运行着的内核中你也可以使用gdb调试程序。(只读的;你不能改变值或者是设置断点。)为了能够实现调试,初次编译内核的时候使用参数 -g;适当的编辑 arch/i386/Makefile文件,然后使用命令"make clean".你也需要是CONFIG_PROC_FS可用(通过"make config")。

   After you've rebooted with the new kernel, do "gdb vmlinux /proc/kcore".
   You can now use all the usual gdb commands. The command to look up the
   point where your system crashed is "l *0xXXXXXXXX". (Replace the XXXes
   with the EIP value.)
   在你重新启动新内核之后,使用命令"gdb vmlimuz /proc/kcore".现在你可以使用所有的常用的gdb命令。在你系统崩溃的地方使用命令"l *0xXXXXXXXX"来查看指针。(使用EIP值替换掉XXXes)

   gdb'ing a non-running kernel currently fails because gdb (wrongly)
   disregards the starting offset for which the kernel is compiled.

0 收藏
0 评论
0 收藏