source: postlfs/config/firmware.xml@ dc96204

<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE sect1 PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
   "http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd" [
  <!ENTITY % general-entities SYSTEM "../../general.ent">
  %general-entities;
]>

<sect1 id="postlfs-firmware" xreflabel="About Firmware">
  <?dbhtml filename="firmware.html"?>

  <sect1info>
    <othername>$LastChangedBy$</othername>
    <date>$Date$</date>
  </sect1info>

  <title>About Firmware</title>

  <indexterm zone="postlfs-firmware">
    <primary sortas="e-lib-firmware">/lib/firmware</primary>
  </indexterm>

  <para>
    On some recent PCs it can be necessary, or desirable, to load firmware
    to make them work at their best. There is a directory, <filename
    class="directory">/lib/firmware</filename>, where the kernel or kernel
    drivers look for firmware images.
  </para>

  <para>
    Currently, most firmware can be found at a <userinput>git</userinput>
    repository: <ulink url=
      "http://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/tree/"/>.
    For convenience, the LFS Project has created a mirror, updated daily, where
    these firmware files can be accessed via <userinput>wget</userinput> or a
    web browser at <ulink url="&sources-anduin-http;/linux-firmware/"/>.
  </para>

  <para>
    To get the firmware, either point a browser to one of the above
    repositories and then download the item(s) which you need, or install
    <xref linkend="git"/> and clone that repository.
  </para>

  <para>
    For some other firmware, particularly for Intel microcode and certain
    wifi devices, the needed firmware is not available in the above repository.
    Some of this will be addressed below, but a search of the Internet for
    needed firmware is sometimes necessary.
  </para>

  <para>
    Firmware files are conventionally referred to as blobs because you cannot
    determine what they will do. Note that firmware is distributed under
    various different licenses which do not permit disassembly or
    reverse-engineering.
  </para>

  <para>
    Firmware for PCs falls into four categories:
  </para>

  <itemizedlist spacing="compact">
    <listitem>
      <para>
        Updates to the CPU to work around errata, usually referred to as
        microcode.
      </para>
    </listitem>
    <listitem>
      <para>
        Firmware for video controllers. On x86 machines this seems to mostly
        apply to ATI devices (Radeon and AMDGPU chips) and Nvidia Maxwell
        and Pascal cards which all require firmware to be able to use KMS
        (kernel modesetting - the preferred option) as well as for Xorg. For
        earlier radeon chips (before the R600), the firmware is still in the
        kernel.
      </para>
    </listitem>
    <listitem>
      <para>
        Firmware updates for wired network ports. Mostly they work even
        without the updates, but probably they will work better with
        the updated firmware. For some modern laptops, firmware for both
        wired ethernet (e.g. rtl_nic) and also for bluetooth devices (e.g. qca)
        is <emphasis>required</emphasis> before the wired network can be used.
      </para>
    </listitem>
    <listitem>
      <para>
        Firmware for other devices, such as wifi. These devices are not
        required for the PC to boot, but need the firmware before these devices
        can be used.
      </para>
    </listitem>
  </itemizedlist>

  <note>
    <para>
      Although not needed to load a firmware blob, the following
      tools may be useful for determining, obtaining, or preparing the needed
      firmware in order to load it into the system:
      <xref linkend="cpio"/>,
      <xref linkend="git"/>,
      <xref linkend="pciutils"/>, and
      <xref linkend="wget"/>
    </para>
  </note>

  <para condition="html" role="usernotes">User Notes:
  <ulink url="&blfs-wiki;/aboutfirmware"/></para>

  <sect2 id="cpu-microcode">
    <title>Microcode updates for CPUs</title>

    <para>
      In general, microcode can be loaded by the BIOS or UEFI, and it might be
      updated by upgrading to a newer version of those. On linux, you can also
      load the microcode from the kernel if you are using an AMD family 10h or
      later processor (first introduced late 2007), or an Intel processor from
      1998 and later (Pentium4, Core, etc), if updated microcode has been
      released. These updates only last until the machine is powered off, so
      they need to be applied on every boot.
    </para>

    <para>
      Intel provide updates of their microcode for Haswell and later
      processors as new vulnerabilities come to light, and have in the past
      provided updates for processors from SandyBridge onwards, although those
      are no-longer supported for new fixes. New versions of AMD
      firmware are rare and usually only apply to a few models, although
      motherboard manufacturers get extra updates which maybe update microcode
      along with the changes to support newer CPUs and faster memory.
    </para>

    <para>
      There are two ways of loading the microcode, described as 'early' and
      'late'. Early loading happens before userspace has been started, late
      loading happens after userspace has started. Not surprisingly, early
      loading is preferred, (see e.g. an explanatory comment in a kernel
      commit noted at <ulink url="https://lwn.net/Articles/530346/">
        x86/microcode: Early load microcode</ulink> on LWN.)  Indeed, it
      is needed to work around one particular erratum in early Intel Haswell
      processors which had TSX enabled.  (See <ulink url=
      "http://www.anandtech.com/show/8376/intel-disables-tsx-instructions-erratum-found-in-haswell-haswelleep-broadwellyi/">
        Intel Disables TSX Instructions: Erratum Found in Haswell,
        Haswell-E/EP, Broadwell-Y
      </ulink>.) Without this update glibc can do the wrong thing in uncommon
      situations. 
    </para>

    <para>
      It is still possible to manually force late loading of microcode, either
      for testing or to prevent having to reboot. You will need to reconfigure
      your kernel for either method. The instructions here will create a
      kernel <filename>.config</filename> to suite early loading, before
      forcing late loading to see if there is any microcode. If there is,
      the instructions then show you how to create an initrd for early loading.
    </para>

    <para>
      To confirm what processor(s) you have (if more than one, they will be
      identical) look in /proc/cpuinfo.
    </para>

    <para>
      If you are creating an initrd to update firmware for different machines,
      as a distro would do, go down to 'Early loading of microcode' and cat all
      the Intel blobs to GenuineIntel.bin or cat all the AMD blobs to
      AuthenticAMD.bin. This creates a larger initrd - for all Intel machines in
      the 20200609 update the size is 3.0 MB compared to typically 24 KB for one
      machine.
    </para>

    <sect3 id="intel-microcode">
      <title>Intel Microcode for the CPU</title>

      <para>
        The first step is to get the most recent version of the Intel
        microcode.  This must be done by navigating to <ulink url=
         'https://github.com/intel/Intel-Linux-Processor-Microcode-Data-Files/releases/'/>
        and downloading the latest file there.  As of this writing the most
        recent version of the microcode is microcode-20200609.  Extract this
        file in the normal way, the microcode is in the <filename>intel-ucode
        </filename> directory, containing various blobs with names in the form
        XX-YY-ZZ. There are also various other files, and a releasenote.
      </para>

      <para>
        In the past, intel did not provide any details of which blobs had
        changed versions, but now the release note details this.
      </para>

      <para>
        The recent firmware for older processors is provided to deal with
        vulnerabilities which have now been made public, and for some of these
        such as Microarchitectural Data Sampling (MDS) you might wish to
        increase the protection by disabling hyperthreading, or alternatively
        to disable the kernel's default mitigation because of its impact on
        compile times. Please read the online documentation at <ulink url=
        'https://www.kernel.org/doc/html/latest/admin-guide/hw-vuln/index.html'/>.
      </para>

      <para>
        The documentation on the latest SRBDS (Special Register Buffer Data
        Sampling) vulnerabilities/fixes will be documented in kernels 5.4.46,
        5.6.18, 5.7.2, 5.8.0 and later.
      </para>

      <para>
        Now you need to determine your processor's identity to see if there
        is any microcode for it. Determine the decimal values of the cpu family,
        model and stepping by running the following command (it will also report
        the current microcode version):
      </para>

<screen><userinput>head -n7 /proc/cpuinfo</userinput></screen>

      <para>
        Convert the cpu family, model and stepping to pairs of hexadecimal
        digits. For a Skylake i3 6100 (described as Intel(R) Core(TM) i3-6100
        CPU) the relevant values are cpu family 6, model 94, stepping 3 so in
        this case the required identification is 06-5e-03. A look at the blobs
        will show that there is one for this CPU (although for older issues it
        might have already been applied by the BIOS). If there is a blob for
        your system then test if it will be applied by copying it (replace
        &lt;XX-YY-ZZ&gt; by the identifier for your CPU) to where the
        kernel can find it:
      </para>

<screen><userinput>mkdir -pv /lib/firmware/intel-ucode
cp -v intel-ucode/&lt;XX-YY-ZZ&gt; /lib/firmware/intel-ucode</userinput></screen>

      <para>
        Now that the Intel microcode has been prepared, use the following
        options when you configure the kernel to load Intel microcode:
      </para>

<screen><literal>General Setup ---&gt;
  [y] Initial RAM filesystem and RAM disk (initramfs/initrd) support [CONFIG_BLK_DEV_INITRD]
Processor type and features  ---&gt;
  [y] CPU microcode loading support  [CONFIG_MICROCODE]
  [y]      Intel microcode loading support [CONFIG_MICROCODE_INTEL]</literal></screen>

      <para>
        After you have successfully booted the new system, force late loading
        by using the command:
      </para>

<screen><userinput>echo 1 > /sys/devices/system/cpu/microcode/reload</userinput></screen>

      <para>
        Then use the following command to see if anything was loaded:
        (N.B. the dates when microcode was created may be months ahead of when
        it was released.)
      </para>

<screen><userinput>dmesg | grep -e 'microcode' -e 'Linux version' -e 'Command line'</userinput></screen>

      <para>
        This reformatted example for an old (20191115) verison of the microcode
        was created by temporarily booting without
        microcode, to show the current Firmware Bug message, then the late load
        shows it being updated to revision 0xd6.
      </para>

<screen><literal>[    0.000000] Linux version 5.4.2 (lfs@leshp) (gcc version 9.2.0 (GCC))
               #1 SMP PREEMPT Wed Dec 18 11:52:13 GMT 2019
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-5.4.2-sda11 root=/dev/sda11 ro
[    0.020218] [Firmware Bug]: TSC_DEADLINE disabled due to Errata; please update microcode
               to version: 0xb2 (or later)
[    0.153861] MDS: Vulnerable: Clear CPU buffers attempted, no microcode
[    0.550009] microcode: sig=0x506e3, pf=0x2, revision=0x74
[    0.550036] microcode: Microcode Update Driver: v2.2.
[  277.673064] microcode: updated to revision 0xd6, date = 2019-10-03
[  277.674231] x86/CPU: CPU features have changed after loading microcode, but might not take effect</literal></screen>

      <para>
        If the microcode was not updated, there is no new microcode for this
        system's processor. If it did get updated, you can now proceed to
        <xref linkend='early-microcode'/>.
      </para>

    </sect3>

    <sect3 id="and-microcode">
      <title>AMD Microcode for the CPU</title>

      <para>
        Begin by downloading a container of firmware for your CPU family
        from <ulink url=
          '&sources-anduin-http;/linux-firmware/amd-ucode/'/>.
        The family is always specified in hex. Families 10h to 14h (16 to 20)
        are in microcode_amd.bin.  Families 15h, 16h and 17h have their own
        containers.  Create the required directory and put the firmware you
        downloaded into it as the <systemitem
        class="username">root</systemitem> user:
      </para>

<screen><userinput>mkdir -pv /lib/firmware/amd-ucode
cp -v microcode_amd* /lib/firmware/amd-ucode</userinput></screen>

      <para>
        When you configure the kernel, use the following options
        to load AMD microcode:
      </para>

<screen><literal>General Setup ---&gt;
  [y] Initial RAM filesystem and RAM disk (initramfs/initrd) support [CONFIG_BLK_DEV_INITRD]
Processor type and features  ---&gt;
  [y] CPU microcode loading support  [CONFIG_MICROCODE]
  [y]      AMD microcode loading support [CONFIG_MICROCODE_AMD]</literal></screen>

      <para>
        After you have successfully booted the new system, force late loading
        by using the command:
      </para>

<screen><userinput>echo 1 > /sys/devices/system/cpu/microcode/reload</userinput></screen>

      <para>
        Then use the following command to see if anything was loaded:
      </para>

<screen><userinput>dmesg | grep -e 'microcode' -e 'Linux version' -e 'Command line'</userinput></screen>
      <para>
        This historic example from an old Athlon(tm) II X2 shows it has been
        updated. At that time, all CPUs were still reported in the microcode
        details on AMD machines (the current position for AMD machines where
        newer microcode is available is unknown) :
      </para>

<screen><literal>[    0.000000] Linux version 4.15.3 (ken@testserver) (gcc version 7.3.0 (GCC))
               #1 SMP Sun Feb 18 02:08:12 GMT 2018
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-4.15.3-sda5 root=/dev/sda5 ro
[    0.307619] microcode: CPU0: patch_level=0x010000b6
[    0.307671] microcode: CPU1: patch_level=0x010000b6
[    0.307743] microcode: Microcode Update Driver: v2.2.
[  187.928891] microcode: CPU0: new patch_level=0x010000c8
[  187.928899] microcode: CPU1: new patch_level=0x010000c8</literal></screen>

      <para>
        If the microcode was not updated, there is no new microcode for
        this system's processor. If it did get updated, you can now proceed to
        <xref linkend='early-microcode'/>.
      </para>

    </sect3>

    <sect3 id="early-microcode">
      <title>Early loading of microcode</title>

      <para>
        If you have established that updated microcode is available for
        your system, it is time to prepare it for early loading. This requires
        an additional package, <xref linkend='cpio'/> and the creation of an
        initrd which will need to be added to grub.cfg.
      </para>

      <para>
        It does not matter where you prepare the initrd, and once it is
        working you can apply the same initrd to later LFS systems or newer
        kernels on this same machine, at least until any newer microcode is
        released. Use the following commands:
      </para>

<screen><userinput>mkdir -p initrd/kernel/x86/microcode
cd initrd</userinput></screen>

      <para>
        For an AMD machine, use the following command (replace
        &lt;MYCONTAINER&gt; with the name of the container for your CPU's
        family):
      </para>

<screen><userinput>cp -v /lib/firmware/amd-ucode/&lt;MYCONTAINER&gt; kernel/x86/microcode/AuthenticAMD.bin</userinput></screen>

      <para>
        Or for an Intel machine copy the appropriate blob using this command:
      </para>

<screen><userinput>cp -v /lib/firmware/intel-ucode/&lt;XX-YY-ZZ&gt; kernel/x86/microcode/GenuineIntel.bin</userinput></screen>

      <caution>
        <para>
          On some Skylake machines with hex Model Number '4e' (78 decimal) the upgrade to
          microcode version '0xdc' is reported to cause the machine to hang in early boot,
          and the fix is to revert to version 0xd6 which was shipped in the 20191115
          microcode release.
        </para>

        <para>
          Doing that, while making the machine usable, will mean that the SRBDS
          mitigations are not present.
        </para>

        <para>
          The Skylake machines with hex Model Number '5e' (94 decimal) use similar
          firmware versions, but boot correctly with the mitigation applied.
        </para>
      </caution>

      <para>
        Now prepare the initrd:
      </para>

<screen><userinput>find . | cpio -o -H newc &gt; /boot/microcode.img</userinput></screen>

      <para>
        You now need to add a new entry to /boot/grub/grub.cfg and
        here you should add a new line after the linux line within the stanza.
        If /boot is a separate mountpoint:
       </para>

<screen><userinput>initrd /microcode.img</userinput></screen>

      <para>
        or this if it is not:
      </para>

<screen><userinput>initrd /boot/microcode.img</userinput></screen>

      <para>
        If you are already booting with an initrd (see <xref
        linkend="initramfs"/>), you should run <command>mkinitramfs</command> 
        again after putting the appropriate blob or container into <filename
        class="directory">/lib/firmware</filename> as explained above.
        Alternatively, you can have both initrd on the same line, such as
        <userinput>initrd /microcode.img /other-initrd.img</userinput> (adapt
        that as above if /boot is not a separate mountpoint).
      </para>

      <para>
        You can now reboot with the added initrd, and then use the same
        command to check that the early load worked:
      </para>

<screen><userinput>dmesg | grep -e 'microcode' -e 'Linux version' -e 'Command line'</userinput></screen>

      <para>
        If you updated to address vulnerabilities, you can look at <filename
        class="directory">/sys/devices/system/cpu/vulnerabilities/</filename>
        to see what is now reported.
      </para>

      <para>
        The places and times where early loading happens are very different
        in AMD and Intel machines. First, an Intel (Haswell) example with early loading:
      </para>

<screen><literal>[    0.000000] microcode: microcode updated early to revision 0x28, date = 2019-11-12
[    0.000000] Linux version 5.6.2 (ken@plexi) (gcc version 9.2.0 (GCC)) #2 SMP PREEMPT Tue Apr 7 21:34:32 BST 2020
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-5.6.2-sda10 root=/dev/sda10 ro resume=/dev/sdb1
[    0.371462] microcode: sig=0x306c3, pf=0x2, revision=0x28
[    0.371491] microcode: Microcode Update Driver: v2.2.</literal></screen>


      <para>
        A historic AMD example:
      </para>

<screen><literal>[    0.000000] Linux version 4.15.3 (ken@testserver) (gcc version 7.3.0 (GCC))
               #2 SMP Sun Feb 18 02:32:03 GMT 2018
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-4.15.3-sda5 root=/dev/sda5 ro
[    0.307619] microcode: microcode updated early to new patch_level=0x010000c8
[    0.307678] microcode: CPU0: patch_level=0x010000c8
[    0.307723] microcode: CPU1: patch_level=0x010000c8
[    0.307795] microcode: Microcode Update Driver: v2.2.</literal></screen>

    </sect3>

  </sect2>

  <sect2 id="video-firmware">
    <title>Firmware for Video Cards</title>

    <sect3 id="ati-video-firmware">
      <title>Firmware for ATI video chips (R600 and later)</title>

      <para>
        These instructions do NOT apply to old radeons before the R600
        family. For those, the firmware is in the kernel's <filename
        class='directory'>/lib/firmware/</filename> directory. Nor do they
        apply if you intend to avoid a graphical setup such as Xorg and are
        content to use the default 80x25 display rather than a framebuffer.
      </para>

      <para>
        Early radeon devices only needed a single 2K blob of firmware. Recent
        devices need several different blobs, and some of them are much bigger.
        The total size of the radeon firmware directory is over 500K &mdash;
        on a large modern system you can probably spare the space, but it is
        still redundant to install all the unused files each time you build
        a system.
      </para>

      <para>
        A better approach is to install <xref linkend='pciutils'/> and then
        use <userinput>lspci</userinput> to identify which VGA controller is
        installed.
      </para>

      <para>
        With that information, check the RadeonFeature page of the Xorg wiki
        for <ulink url="http://wiki.x.org/wiki/RadeonFeature/#index5h2">Decoder
        ring for engineering vs marketing names</ulink> to identify the family
        (you may need to know this for the Xorg driver in BLFS &mdash;
        Southern Islands and Sea Islands use the radeonsi driver) and the
        specific model.
      </para>

      <para>
        Now that you know which controller you are using, consult the
        <ulink url="https://wiki.gentoo.org/wiki/Radeon#Firmware">
           Radeon</ulink> page of the Gentoo wiki which has a table listing
        the required firmware blobs for the various chipsets. Note that
        Southern Islands and Sea Islands chips use different firmware for
        kernel 3.17 and later compared to earlier kernels. Identify and
        download the required blobs then install them:
      </para>

<screen><userinput>mkdir -pv /lib/firmware/radeon
cp -v &lt;YOUR_BLOBS&gt; /lib/firmware/radeon</userinput></screen>

      <para>
        There are actually two ways of installing this firmware. BLFS, in the
        'Kernel Configuration for additional firmware' section part of the
        <xref linkend="xorg-ati-driver"/> section gives an example of
        compiling the firmware into the kernel - that is slightly faster to
        load, but uses more kernel memory. Here we will use the alternative
        method of making the radeon driver a module. In your kernel config
        set the following:
      </para>

<screen><literal>Device Drivers ---&gt;
  Graphics support ---&gt;
      Direct Rendering Manager ---&gt;
        &lt;*&gt; Direct Rendering Manager (XFree86 ... support)  [CONFIG_DRM]
      &lt;m&gt; ATI Radeon                                        [CONFIG_DRM_RADEON]</literal></screen>

      <para>
        Loading several large blobs from /lib/firmware takes a noticeable
        time, during which the screen will be blank. If you do not enable the
        penguin framebuffer logo, or change the console size by using a bigger
        font, that probably does not matter. If desired, you can slightly
        reduce the time if you follow the alternate method of specifying 'y'
        for CONFIG_DRM_RADEON covered in BLFS at the link above &mdash; you
        must specify each needed radeon blob if you do that.
      </para>

    </sect3>

    <sect3 id="nvidia-video-firmware">
      <title>Firmware for Nvidia video chips</title>

      <para>
        Some Nvidia graphics chips need firmware updates to take advantage 
        of all the card's capability.  These are generally the GeForce 8, 9,
        9300, and 200-900 series chips.  For more exact information, see
        <ulink url=
          "https://nouveau.freedesktop.org/wiki/VideoAcceleration/#firmware"/>.
      </para>
    
      <para>
        First, the kernel Nvidia driver must be activated:
      </para>

<screen><literal>Device Drivers ---&gt;
  Graphics support ---&gt;
      Direct Rendering Manager ---&gt;
        &lt;*&gt; Direct Rendering Manager (XFree86 ... support)  [CONFIG_DRM]
      &lt;*/m&gt; Nouveau (NVIDIA) cards                          [CONFIG_DRM_NOUVEAU]</literal></screen>

      <para>
        The steps to install the Nvidia firmware are:
      </para>

<screen><userinput>wget https://raw.github.com/imirkin/re-vp2/master/extract_firmware.py
wget http://us.download.nvidia.com/XFree86/Linux-x86/325.15/NVIDIA-Linux-x86-325.15.run
sh NVIDIA-Linux-x86-325.15.run --extract-only
python extract_firmware.py 
mkdir -p /lib/firmware/nouveau
cp -d nv* vuc-* /lib/firmware/nouveau/</userinput></screen>

    </sect3>
  </sect2>

  <sect2 id="nic-firmware">
    <title>Firmware for Network Interfaces</title>

    <para>
      The kernel likes to load firmware for some network drivers, particularly
      those from Realtek (the /lib/linux-firmware/rtl_nic/) directory, but
      they generally appear to work without it. Therefore, you can boot the
      kernel, check dmesg for messages about this missing firmware, and if
      necessary download the firmware and put it in the specified directory in
      <filename class="directory">/lib/firmware</filename> so that it will
      be found on subsequent boots. Note that with current kernels this
      works whether or not the driver is compiled in or built as a module,
      there is no need to build this firmware into the kernel.
      Here is an example where the R8169 driver has been compiled in but the
      firmware was not made available. Once the firmware had been provided,
      there was no mention of it on later boots.
    </para>

<screen><literal>dmesg | grep firmware | grep r8169
[    7.018028] r8169 0000:01:00.0: Direct firmware load for rtl_nic/rtl8168g-2.fw failed with error -2
[    7.018036] r8169 0000:01:00.0 eth0: unable to load firmware patch rtl_nic/rtl8168g-2.fw (-2)</literal></screen>

  </sect2>

  <sect2 id="other-firmware">
    <title>Firmware for Other Devices</title>

    <para>
      Identifying the correct firmware will typically require you to install
      <xref linkend='pciutils'/>, and then use <userinput>lspci</userinput>
      to identify the device. You should then search online to check which
      module it uses, which firmware, and where to obtain the firmware &mdash;
      not all of it is in linux-firmware.
    </para>

    <para>
      If possible, you should begin by using a wired connection when you first
      boot your LFS system. To use a wireless connection you will need to
      use a network tools such as <xref linkend='wireless_tools'/> and <xref
      linkend='wpa_supplicant'/>.
    </para>

    <para>
      Firmware may also be needed for other devices such as some SCSI
      controllers, bluetooth adaptors, or TV recorders. The same principles
      apply.
    </para>

  </sect2>

</sect1>