source: chapter10/kernel.xml@ c4108e5

multilib
Last change on this file since c4108e5 was c4108e5, checked in by Thomas Trepl <thomas@…>, 7 months ago

Automatic merge of trunk into multilib

  • Property mode set to 100644
File size: 22.1 KB
Line 
1<?xml version="1.0" encoding="ISO-8859-1"?>
2<!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
3 "http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [
4 <!ENTITY % general-entities SYSTEM "../general.ent">
5 %general-entities;
6]>
7
8<sect1 id="ch-bootable-kernel" role="wrap">
9 <?dbhtml filename="kernel.html"?>
10
11 <sect1info condition="script">
12 <productname>kernel</productname>
13 <productnumber>&linux-version;</productnumber>
14 <address>&linux-url;</address>
15 </sect1info>
16
17 <title>Linux-&linux-version;</title>
18
19 <indexterm zone="ch-bootable-kernel">
20 <primary sortas="a-Linux">Linux</primary>
21 </indexterm>
22
23 <sect2 role="package">
24 <title/>
25
26 <para>The Linux package contains the Linux kernel.</para>
27
28 <segmentedlist>
29 <segtitle>&buildtime;</segtitle>
30 <segtitle>&diskspace;</segtitle>
31
32 <seglistitem>
33 <seg>&linux-knl-sbu;</seg>
34 <seg>&linux-knl-du;</seg>
35 </seglistitem>
36 </segmentedlist>
37
38 </sect2>
39
40 <sect2 role="installation">
41 <title>Installation of the kernel</title>
42
43 <para>Building the kernel involves a few steps&mdash;configuration,
44 compilation, and installation. Read the <filename>README</filename> file
45 in the kernel source tree for alternative methods to the way this book
46 configures the kernel.</para>
47
48 <important>
49 <para>
50 Building the linux kernel for the first time is one of the most
51 challenging tasks in LFS. Getting it right depends on the specific
52 hardware for the target system and your specific needs. There are
53 almost 12,000 configuration items that are available for the kernel
54 although only about a third of them are needed for most computers. The
55 LFS editors recommend that users not familiar with this process follow
56 the procedures below fairly closely. The objective is to get an
57 initial system to a point where you can log in at the command line when
58 you reboot later in <xref linkend="ch-finish-reboot"/>. At this point
59 optimization and customization is not a goal.
60 </para>
61
62
63 <para>
64 For general information on kernel configuration see <ulink
65 url="&hints-root;kernel-configuration.txt"/>. Additional information
66 about configuring and building the kernel can be found at <ulink
67 url="&anduin-sources;/kernel-nutshell/"/>.
68 These references are a bit
69 dated, but still give a reasonable overview of the process.
70 </para>
71
72 <para>
73 If all else fails, you can ask for help on the <ulink
74 url="https://www.linuxfromscratch.org/mail.html">lfs-support</ulink>
75 mailing list. Note that subscribing is required in order for the list
76 to avoid spam.
77 </para>
78 </important>
79
80 <para>Prepare for compilation by running the following command:</para>
81
82<screen><userinput remap="pre">make mrproper</userinput></screen>
83
84 <para>This ensures that the kernel tree is absolutely clean. The
85 kernel team recommends that this command be issued prior to each
86 kernel compilation. Do not rely on the source tree being clean after
87 un-tarring.</para>
88
89 <para>There are several ways to configure the kernel options. Usually,
90 This is done through a menu-driven interface, for example:</para>
91
92<screen role="nodump"><userinput>make menuconfig</userinput></screen>
93
94 <variablelist>
95 <title>The meaning of optional make environment variables:</title>
96
97 <varlistentry>
98 <term><parameter>LANG=&lt;host_LANG_value&gt; LC_ALL=</parameter></term>
99 <listitem>
100 <para>This establishes the locale setting to the one used on the
101 host. This may be needed for a proper menuconfig ncurses interface
102 line drawing on a UTF-8 linux text console.</para>
103
104 <para>If used, be sure to replace
105 <replaceable>&lt;host_LANG_value&gt;</replaceable> by the value of
106 the <envar>$LANG</envar> variable from your host. You can
107 alternatively use instead the host's value of <envar>$LC_ALL</envar>
108 or <envar>$LC_CTYPE</envar>.</para>
109 </listitem>
110 </varlistentry>
111
112 <varlistentry>
113 <term><command>make menuconfig</command></term>
114 <listitem>
115 <para>This launches an ncurses menu-driven interface. For other
116 (graphical) interfaces, type <command>make help</command>.</para>
117 </listitem>
118 </varlistentry>
119 </variablelist>
120
121 <!-- Support for compiling a keymap into the kernel is deliberately removed -->
122
123
124 <note>
125 <?dbfo keep-together="auto"?>
126 <para>A good starting place for setting up the kernel configuration is to
127 run <command>make defconfig</command>. This will set the base
128 configuration to a good state that takes your current system architecture
129 into account.</para>
130
131 <para>Be sure to enable/disable/set the following features or the system might
132 not work correctly or boot at all:</para>
133
134 <!-- To editors: for updating kernel configuration, edit
135 kernel/*.toml and regenerate kernel/*.xml with
136 "make -C kernel KERNEL_TREE=</usr/src/linux-&linux-version> -->
137
138 <xi:include xmlns:xi="http://www.w3.org/2001/XInclude"
139 href="kernel/sysv.xml"/>
140
141 <xi:include xmlns:xi="http://www.w3.org/2001/XInclude"
142 href="kernel/systemd.xml"/>
143
144 <para>Enable some additional features if you are building a 64-bit
145 system. If you are using menuconfig, enable them in the order of
146 <parameter>CONFIG_PCI_MSI</parameter> first, then
147 <parameter>CONFIG_IRQ_REMAP</parameter>, at last
148 <parameter>CONFIG_X86_X2APIC</parameter> because an option only
149 shows up after its dependencies are selected.</para>
150
151 <xi:include xmlns:xi="http://www.w3.org/2001/XInclude"
152 href="kernel/x2apic.xml"/>
153
154 <para>If you are building a 32-bit system running on a hardware
155 with RAM more than 4GB, adjust the configuration so the kernel will
156 be able to use up to 64GB physical RAM:</para>
157
158 <xi:include xmlns:xi="http://www.w3.org/2001/XInclude"
159 href="kernel/highmem.xml"/>
160
161 <para>If the partition for the LFS system is in a NVME SSD (i. e. the
162 device node for the partition is <filename>/dev/nvme*</filename>
163 instead of <filename>/dev/sd*</filename>), enable NVME support or
164 the LFS system won't boot:</para>
165
166 <xi:include xmlns:xi="http://www.w3.org/2001/XInclude"
167 href="kernel/nvme.xml"/>
168 </note>
169
170 <note revision="systemd">
171 <para>While "The IPv6 Protocol" is not strictly
172 required, it is highly recommended by the systemd developers.</para>
173 </note>
174
175 <para>There are several other options that may be desired
176 depending on the requirements for the system. For a list of options needed
177 for BLFS packages, see the <ulink
178 url="&lfs-root;blfs/view/&short-version;/longindex.html#kernel-config-index">BLFS
179 Index of Kernel Settings</ulink>.</para>
180
181 <note>
182 <para>If your host hardware is using UEFI and you wish to boot the
183 LFS system with it, you should adjust some kernel configuration
184 following <ulink url="&blfs-book;postlfs/grub-setup.html#uefi-kernel">
185 the BLFS page</ulink> <emphasis role='bold'>even if you'll use the
186 UEFI bootloader from the host distro</emphasis>.</para>
187 </note>
188
189 <note arch="ml_32,ml_x32,ml_all">
190 <para>
191 The kernel on a multilib system needs to be able to
192 identify and start binaries compiled for different architectures
193 than the default.
194 </para>
195
196 <para arch="ml_32,ml_all">
197 If support for any 32bit ABI was built, make sure that the option
198 "IA32 Emulation" is selected. The option 'IA32 a.out support' is
199 optional.
200 </para>
201
202 <para arch="ml_x32,ml_all">
203 If support for the x32bit ABI was built, make sure that the option
204 "x32 ABI for 64-bit mode" is selected.
205 </para>
206
207<screen arch="ml_32">Binary Emulations ---&gt;
208 [*] IA32 Emulation [CONFIG_IA32_EMULATION]
209 &lt;M&gt; IA32 a.out support [CONFIG_IA32_AOUT]
210</screen>
211<screen arch="ml_x32">Binary Emulations ---&gt;
212 [*] x32 ABI for 64-bit mode [CONFIG_X86_X32]
213</screen>
214<screen arch="ml_all">Binary Emulations ---&gt;
215 [*] IA32 Emulation [CONFIG_IA32_EMULATION]
216 &lt;M&gt; IA32 a.out support [CONFIG_IA32_AOUT]
217 [*] x32 ABI for 64-bit mode [CONFIG_X86_X32]
218</screen>
219 </note>
220
221 <variablelist>
222 <title>The rationale for the above configuration items:</title>
223
224 <varlistentry>
225 <term><parameter>Randomize the address of the kernel image (KASLR)</parameter></term>
226 <listitem>
227 <para>Enable ASLR for kernel image, to mitigate some attacks based
228 on fixed addresses of sensitive data or code in the kernel.</para>
229 </listitem>
230 </varlistentry>
231
232 <varlistentry>
233 <term>
234 <parameter>
235 Compile the kernel with warnings as errors
236 </parameter>
237 </term>
238 <listitem>
239 <para>This may cause building failure if the compiler and/or
240 configuration are different from those of the kernel
241 developers.</para>
242 </listitem>
243 </varlistentry>
244
245 <varlistentry>
246 <term>
247 <parameter>
248 Enable kernel headers through /sys/kernel/kheaders.tar.xz
249 </parameter>
250 </term>
251 <listitem>
252 <para>This will require <command>cpio</command> building the kernel.
253 <command>cpio</command> is not installed by LFS.</para>
254 </listitem>
255 </varlistentry>
256
257 <varlistentry>
258 <term>
259 <parameter>
260 Configure standard kernel features (expert users)
261 </parameter>
262 </term>
263 <listitem>
264 <para>This will make some options show up in the configuration
265 interface but changing those options may be dangerous. Do not use
266 this unless you know what you are doing.</para>
267 </listitem>
268 </varlistentry>
269
270 <varlistentry>
271 <term><parameter>Strong Stack Protector</parameter></term>
272 <listitem>
273 <para>Enable SSP for the kernel. We've enabled it for the entire
274 userspace with <parameter>--enable-default-ssp</parameter>
275 configuring GCC, but the kernel does not use GCC default setting
276 for SSP. We enable it explicitly here.</para>
277 </listitem>
278 </varlistentry>
279
280 <varlistentry>
281 <term><parameter>Support for uevent helper</parameter></term>
282 <listitem>
283 <para>Having this option set may interfere with device
284 management when using Udev.</para>
285 </listitem>
286 </varlistentry>
287
288 <varlistentry>
289 <term><parameter>Maintain a devtmpfs</parameter></term>
290 <listitem>
291 <para>This will create automated device nodes which are populated by the
292 kernel, even without Udev running. Udev then runs on top of this,
293 managing permissions and adding symlinks. This configuration
294 item is required for all users of Udev.</para>
295 </listitem>
296 </varlistentry>
297
298 <varlistentry>
299 <term><parameter>Automount devtmpfs at /dev</parameter></term>
300 <listitem>
301 <para>This will mount the kernel view of the devices on /dev
302 upon switching to root filesystem just before starting
303 init.</para>
304 </listitem>
305 </varlistentry>
306
307 <varlistentry>
308 <term>
309 <parameter>
310 Enable legacy fbdev support for your modesetting driver
311 </parameter> and
312 <parameter>Framebuffer Console support</parameter>
313 </term>
314 <listitem>
315 <para>These are needed to display the Linux console on a
316 GPU driven by a DRI (Direct Rendering Infrastructure) driver.
317 If <option>CONFIG_DRM</option> (Direct Rendering Manager) is
318 enabled, you should enable these two options as well or you'll see
319 a blank screen once the DRI driver is loaded.</para>
320 </listitem>
321 </varlistentry>
322
323 <varlistentry>
324 <term><parameter>Support x2apic</parameter></term>
325 <listitem>
326 <para>Support running the interrupt controller of 64-bit x86
327 processors in x2APIC mode. x2APIC may be enabled by firmware on
328 64-bit x86 systems, and a kernel without this option enabled will
329 panic on boot if x2APIC is enabled by firmware. This option
330 has no effect, but also does no harm if x2APIC is disabled by the
331 firmware.</para>
332 </listitem>
333 </varlistentry>
334
335 </variablelist>
336
337 <para>Alternatively, <command>make oldconfig</command> may be more
338 appropriate in some situations. See the <filename>README</filename>
339 file for more information.</para>
340
341 <para>If desired, skip kernel configuration by copying the kernel
342 config file, <filename>.config</filename>, from the host system
343 (assuming it is available) to the unpacked <filename
344 class="directory">linux-&linux-version;</filename> directory. However,
345 we do not recommend this option. It is often better to explore all the
346 configuration menus and create the kernel configuration from
347 scratch.</para>
348
349 <para>Compile the kernel image and modules:</para>
350
351<screen><userinput remap="make">make</userinput></screen>
352
353 <para>If using kernel modules, module configuration in <filename
354 class="directory">/etc/modprobe.d</filename> may be required.
355 Information pertaining to modules and kernel configuration is
356 located in <xref linkend="ch-config-udev"/> and in the kernel
357 documentation in the <filename
358 class="directory">linux-&linux-version;/Documentation</filename> directory.
359 Also, <filename>modprobe.d(5)</filename> may be of interest.</para>
360
361 <para>Unless module support has been disabled in the kernel configuration,
362 install the modules with:</para>
363
364<screen><userinput remap="install">make modules_install</userinput></screen>
365
366 <para>After kernel compilation is complete, additional steps are
367 required to complete the installation. Some files need to be copied to
368 the <filename class="directory">/boot</filename> directory.</para>
369
370 <caution>
371 <para>If you've decided to use a separate &boot-dir; partition for the
372 LFS system (maybe sharing a &boot-dir; partition with the host
373 distro) , the files copied below should go there. The easiest way to
374 do that is to create the entry for &boot-dir; in &fstab; first (read
375 the previous section for details), then issue the following command
376 as the &root; user in the
377 <emphasis>chroot environment</emphasis>:</para>
378
379<screen role="nodump"><userinput>mount /boot</userinput></screen>
380
381 <para>The path to the device node is omitted in the command because
382 <command>mount</command> can read it from &fstab;.</para>
383 </caution>
384
385 <para>The path to the kernel image may vary depending on the platform being
386 used. The filename below can be changed to suit your taste, but the stem of
387 the filename should be <emphasis>vmlinuz</emphasis> to be compatible with
388 the automatic setup of the boot process described in the next section. The
389 following command assumes an x86 architecture:</para>
390
391<screen><userinput remap="install">cp -iv arch/x86/boot/bzImage /boot/vmlinuz-&linux-version;-lfs-&version;</userinput></screen>
392
393 <para><filename>System.map</filename> is a symbol file for the kernel.
394 It maps the function entry points of every function in the kernel API,
395 as well as the addresses of the kernel data structures for the running
396 kernel. It is used as a resource when investigating kernel problems.
397 Issue the following command to install the map file:</para>
398
399<screen><userinput remap="install">cp -iv System.map /boot/System.map-&linux-version;</userinput></screen>
400
401 <para>The kernel configuration file <filename>.config</filename>
402 produced by the <command>make menuconfig</command> step
403 above contains all the configuration selections for the kernel
404 that was just compiled. It is a good idea to keep this file for future
405 reference:</para>
406
407<screen><userinput remap="install">cp -iv .config /boot/config-&linux-version;</userinput></screen>
408
409 <para>Install the documentation for the Linux kernel:</para>
410
411<screen><userinput remap="install">cp -r Documentation -T /usr/share/doc/linux-&linux-version;</userinput></screen>
412
413 <para>It is important to note that the files in the kernel source
414 directory are not owned by <emphasis>root</emphasis>. Whenever a
415 package is unpacked as user <emphasis>root</emphasis> (like we did
416 inside chroot), the files have the user and group IDs of whatever
417 they were on the packager's computer. This is usually not a problem
418 for any other package to be installed because the source tree is
419 removed after the installation. However, the Linux source tree is
420 often retained for a long time. Because of this, there is a chance
421 that whatever user ID the packager used will be assigned to somebody
422 on the machine. That person would then have write access to the kernel
423 source.</para>
424
425 <note>
426 <para>In many cases, the configuration of the kernel will need to be
427 updated for packages that will be installed later in BLFS. Unlike
428 other packages, it is not necessary to remove the kernel source tree
429 after the newly built kernel is installed.</para>
430
431 <para>If the kernel source tree is going to be retained, run
432 <command>chown -R 0:0</command> on the <filename
433 class="directory">linux-&linux-version;</filename> directory to ensure
434 all files are owned by user <emphasis>root</emphasis>.</para>
435 </note>
436
437 <warning>
438 <para>Some kernel documentation recommends creating a symlink from
439 <filename class="symlink">/usr/src/linux</filename> pointing to the kernel
440 source directory. This is specific to kernels prior to the 2.6 series and
441 <emphasis>must not</emphasis> be created on an LFS system as it can cause
442 problems for packages you may wish to build once your base LFS system is
443 complete.</para>
444 </warning>
445
446 <warning>
447 <para>The headers in the system's <filename
448 class="directory">include</filename> directory (<filename
449 class="directory">/usr/include</filename>) should
450 <emphasis>always</emphasis> be the ones against which Glibc was compiled,
451 that is, the sanitised headers installed in <xref
452 linkend="ch-tools-linux-headers"/>. Therefore, they should
453 <emphasis>never</emphasis> be replaced by either the raw kernel headers
454 or any other kernel sanitized headers.</para>
455 </warning>
456
457 </sect2>
458
459 <sect2 id="conf-modprobe" role="configuration">
460 <title>Configuring Linux Module Load Order</title>
461
462 <indexterm zone="conf-modprobe">
463 <primary sortas="e-/etc/modprobe.d/usb.conf">/etc/modprobe.d/usb.conf</primary>
464 </indexterm>
465
466 <para>Most of the time Linux modules are loaded automatically, but
467 sometimes it needs some specific direction. The program that loads
468 modules, <command>modprobe</command> or <command>insmod</command>, uses
469 <filename>/etc/modprobe.d/usb.conf</filename> for this purpose. This file
470 needs to be created so that if the USB drivers (ehci_hcd, ohci_hcd and
471 uhci_hcd) have been built as modules, they will be loaded in the correct
472 order; ehci_hcd needs to be loaded prior to ohci_hcd and uhci_hcd in order
473 to avoid a warning being output at boot time.</para>
474
475 <para>Create a new file <filename>/etc/modprobe.d/usb.conf</filename> by running
476 the following:</para>
477
478<screen><userinput>install -v -m755 -d /etc/modprobe.d
479cat &gt; /etc/modprobe.d/usb.conf &lt;&lt; "EOF"
480<literal># Begin /etc/modprobe.d/usb.conf
481
482install ohci_hcd /sbin/modprobe ehci_hcd ; /sbin/modprobe -i ohci_hcd ; true
483install uhci_hcd /sbin/modprobe ehci_hcd ; /sbin/modprobe -i uhci_hcd ; true
484
485# End /etc/modprobe.d/usb.conf</literal>
486EOF</userinput></screen>
487
488 </sect2>
489
490 <sect2 id="contents-kernel" role="content">
491 <title>Contents of Linux</title>
492
493 <segmentedlist>
494 <segtitle>Installed files</segtitle>
495 <segtitle>Installed directories</segtitle>
496
497 <seglistitem>
498 <seg>config-&linux-version;,
499 vmlinuz-&linux-version;-lfs-&version;,
500 and System.map-&linux-version;</seg>
501 <seg>/lib/modules, /usr/share/doc/linux-&linux-version;</seg>
502 </seglistitem>
503 </segmentedlist>
504
505 <variablelist>
506 <bridgehead renderas="sect3">Short Descriptions</bridgehead>
507 <?dbfo list-presentation="list"?>
508 <?dbhtml list-presentation="table"?>
509
510 <varlistentry id="config">
511 <term><filename>config-&linux-version;</filename></term>
512 <listitem>
513 <para>Contains all the configuration selections for the kernel</para>
514 <indexterm zone="ch-bootable-kernel config">
515 <primary sortas="e-/boot/config">/boot/config-&linux-version;</primary>
516 </indexterm>
517 </listitem>
518 </varlistentry>
519
520 <varlistentry id="lfskernel">
521 <term><filename>vmlinuz-&linux-version;-lfs-&version;</filename></term>
522 <listitem>
523 <para>The engine of the Linux system. When turning on the computer,
524 the kernel is the first part of the operating system that gets loaded.
525 It detects and initializes all components of the computer's hardware,
526 then makes these components available as a tree of files to the
527 software and turns a single CPU into a multitasking machine capable
528 of running scores of programs seemingly at the same time</para>
529 <indexterm zone="ch-bootable-kernel lfskernel">
530 <primary sortas="b-lfskernel">lfskernel-&linux-version;</primary>
531 </indexterm>
532 </listitem>
533 </varlistentry>
534
535 <varlistentry id="System.map">
536 <term><filename>System.map-&linux-version;</filename></term>
537 <listitem>
538 <para>A list of addresses and symbols; it maps the entry points and
539 addresses of all the functions and data structures in the
540 kernel</para>
541 <indexterm zone="ch-bootable-kernel System.map">
542 <primary sortas="e-/boot/System.map">/boot/System.map-&linux-version;</primary>
543 </indexterm>
544 </listitem>
545 </varlistentry>
546
547 </variablelist>
548
549 </sect2>
550
551</sect1>
Note: See TracBrowser for help on using the repository browser.