Shulou(Shulou.com)06/01 Report--

This article focuses on "how to solve the problem of Linux kernel compilation failure", interested friends may wish to take a look. The method introduced in this paper is simple, fast and practical. Let's let the editor take you to learn how to solve the problem of Linux kernel compilation failure.

Kernel common sense

We need to know one thing first. What does a kernel have?

For example, we:

The code is as follows:

Rpm-ql kernel

You can see the following information

The code is as follows:

/ boot/System.map-2.6.24.4

/ boot/config-2.6.24.4

/ boot/vmlinuz-2.6.24.4

/ lib/modules/lib/modules/2.6.24.4

/ lib/modules/2.6.24.4/build

/ lib/modules/2.6.24.4/kernel

See, in fact, there is nothing, mainly the above files under boot, and the module files under / lib/modules.

The System.map kernel symbol mapping table, as its name implies, is a list of symbols (that is, functions in the kernel) that can be associated with its address.

Config kernel compiled configuration file, make oldconfig needs to use this file.

Vmlinuz this is the real bootable, compressed kernel.

Well, as we know from above, this has nothing to do with anything else in the system, so to delete a kernel is to delete these files.

You can refer to [Linux] I wrote a few days ago to compile a custom kernel rpm package for RHEL.

Understanding and checking of Initrd

What I'm going to talk about today is the thing that most affects startup, initrd. You will think, what is this, there is nothing in the kernel, and the other one is generated according to each different hardware after the kernel package is installed. The hardware is different. This is different. The following is an official introduction:

Initrd is an abbreviation for "initial ramdisk". Initrd is generally used to temporarily boot the hardware to a state where the actual kernel vmlinuz can take over and continue to boot. Initrd- 2.4.7-10.img in the figure is mainly used to load drivers for file systems such as ext3 and scsi devices.

For example, if you use a scsi hard disk and the kernel vmlinuz does not have a driver for this scsi hardware, the kernel cannot load the root file system until the scsi module is mounted, but the scsi module is stored under / lib/modules of the root file system. To solve this problem, you can boot an initrd kernel that can read the actual kernel and correct the scsi boot question with initrd. Initrd-2.4.7-10.img is a file compressed by gzip, and initrd implements the functions of loading some modules and installing the file system.

So now we know that if it's a strange piece of hardware, but it doesn't work properly, initrd doesn't load by default, or the kernel doesn't add it.

Let's classify and deal with the causes of the failure.

After the new kernel fails to boot: use modprobe and lsmod in the old system to confirm the modules to be loaded.

We know that initrd affects startup, so how can we make sure that the hardware is fully loaded. Initrd is largely done by lsmod modules and / etc/modprobe instructions. Of course, you can also use lspci and the like to confirm.

As follows

The code is as follows:

Cat / etc/modprobe.conf

Alias scsi_hostadapter mptbase

Alias scsi_hostadapter1 mptscsi

Alias scsi_hostadapter2 mptspi

Alias scsi_hostadapter3 mptsas

Alias scsi_hostadapter4 mptscsih

Alias scsi_hostadapter5 ata_piix

# lsmod

Module Size Used by

Iptable_filter 7300 1

Ip_tables 18132 1 iptable_filter

Xt_tcpudp 7680 1

X_tables 19204 2 ip_tables,xt_tcpudp

Ipv6 245092 32

We found that this machine has some hardware that is not available on other machines, which is related to mpt. If the kernel is recompiled above and these modules are not loaded, it will certainly fail to start.

Use the initrd that can be started normally on the old system to confirm the modules that the new kernel needs to load.

Okay。 We can still see it above. What if we can't see it, but the startup fails?

We can unlock the initrd.img to see which modules are loaded originally.

The code is as follows:

Cd / boot/

Mkdir tmp

Cp initrd*.img tmp

Cd / tmp

Zcat intrd**img | cpio-id

Cat init

All right, let's focus on the following.

The code is as follows:

Echo "Loading scsi_mod.ko module"

Insmod / lib/scsi_mod.ko

Echo "Loading sd_mod.ko module"

Insmod / lib/sd_mod.ko

Echo "Loading mptbase.ko module"

Insmod / lib/mptbase.ko

Echo "Loading mptscsi.ko module"

Insmod / lib/mptscsi.ko

Echo "Loading mptspi.ko module"

Insmod / lib/mptspi.ko

Echo "Loading mptsas.ko module"

Insmod / lib/mptsas.ko

Echo "Loading mptscsih.ko module"

Insmod / lib/mptscsih.ko

Echo "Loading jbd.ko module"

Insmod / lib/jbd.ko

Echo "Loading ext3.ko module"

Insmod / lib/ext3.ko

Based on the above, and your original, see if you don't have the modules mentioned in this information. No, just add it quickly. After dealing with these methods, 99% can be started, and I can't help you if you can't start it.

Modify the kernel of initrd again

The code is as follows:

Gzip-cd initrd-2.6.34.2.img | cpio-imd-- quiet find. | | cpio-co | gzip-9 > / tmp/initrd-2.6.34.2.img |

Compile the kernel what you can know.

Rename the new core

Before compiling the kernel, we can modify the version information in Makefile and open / usr/src/Linux/Makefile. At the beginning, there is a variable EXTRAVERSION that can be defined by itself. After changing this variable, such as "EXTRAVERSION=-ChinaCache", the compiled core version number will be displayed as 2.6.24-ChinaCache.

But in fact, from version 2.6.8 onwards, personalized strings can be added after the kernel version number. So there is no need to modify Makefile: () Local version-append to kernel release

If you not only modify the EXTRAVERSION in Makefile, but also define local version. Then the string defined by local version will be at the end, immediately after the value of EXTRAVERSION.

Depmod

Function: analyze the dependencies of loadable modules and generate modules.dep files and mapping files.

Usage:

The code is as follows:

Depmod [- b basedir] [- e] [- F System.map] [- n] [- v] [version] [- A]

Depmod [- e] [- F System.map] [- n] [- v] [version] [filename...]

Description:

The Linux kernel module can provide services to other modules (using EXPORT_SYMBOL in the code), which is called "symbols". If the second module uses this symbol, the module obviously depends on the first module. These dependencies are very complicated.

Depmod reads all the modules in the / lib/modules/version directory, checks the symbol exported by each module and the required symbol, and then creates a list of dependencies. By default, the list is written to the modules.dep file in the / lib/moudules / version directory. If the filename in the command is specified, only those specified modules are checked (not very useful).

If the version parameter is provided in the command, the dependency is generated using the directory specified by version instead of the current kernel version (returned by uname-r).

Use of the mkinitrd command

Function: create an image file to load ramdisk.

Usage: mkinitrd [- fv] [--omit-scsi-modules] [--version] [--preload=] [--with=& lt; module name >] [image file] [Kernel version]

Description: mkinitrd can create an image file that can be loaded into ramdisk when Linux boots.

Parameters:

-f if the specified image asks for a name that duplicates the existing file, the existing file is overwritten.

-v displays detailed information when executed.

-omit-scsi-modules do not load the SCSI module.

-preload= specifies the module to load.

-with= specifies the module to load.

-version displays version information

At this point, I believe you have a deeper understanding of "how to solve the problem of Linux kernel compilation failure". You might as well do it in practice. Here is the website, more related content can enter the relevant channels to inquire, follow us, continue to learn!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.