path: root/documentation/kernel-dev/kernel-dev-advanced.xml

<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd"
[<!ENTITY % poky SYSTEM "../poky.ent"> %poky; ] >

<chapter id='kernel-dev-advanced'>
<title>Working with Advanced Metadata</title>

<section id='kernel-dev-advanced-overview'>
    <title>Overview</title>

    <para>
        In addition to supporting configuration fragments and patches, the
        Yocto Project kernel tools also support rich
        <ulink url='&YOCTO_DOCS_REF_URL;#metadata'>Metadata</ulink> that you can
        use to define complex policies and Board Support Package (BSP) support.
        The purpose of the Metadata and the tools that manage it, known as
        the kern-tools (<filename>kern-tools-native_git.bb</filename>), is
        to help you manage the complexity of the configuration and sources
        used to support multiple BSPs and Linux kernel types.
    </para>
</section>

<section id='using-kernel-metadata-in-a-recipe'>
    <title>Using Kernel Metadata in a Recipe</title>

    <para>
        The kernel sources in the Yocto Project contain kernel Metadata, which
        is located in the <filename>meta</filename> branches of the kernel
        source Git repositories.
        This Metadata defines Board Support Packages (BSPs) that
        correspond to definitions in linux-yocto recipes for the same BSPs.
        A BSP consists of an aggregation of kernel policy and enabled
        hardware-specific features.
        The BSP can be influenced from within the linux-yocto recipe.
        <note>
            Linux kernel source that contains kernel Metadata is said to be
            "linux-yocto style" kernel source.
            A Linux kernel recipe that inherits from the
            <filename>linux-yocto.inc</filename> include file is said to be a
            "linux-yocto style" recipe.
        </note>
    </para>

    <para>
        Every linux-yocto style recipe must define the
        <ulink url='&YOCTO_DOCS_REF_URL;#var-KMACHINE'><filename>KMACHINE</filename></ulink>
        variable.
        This variable is typically set to the same value as the
        <ulink url='&YOCTO_DOCS_REF_URL;#var-MACHINE'><filename>MACHINE</filename></ulink>
        variable, which is used by
        <ulink url='&YOCTO_DOCS_REF_URL;#bitbake-term'>BitBake</ulink>.
        However, in some cases, the variable might instead refer to the
        underlying platform of the <filename>MACHINE</filename>.
    </para>

    <para>
        Multiple BSPs can reuse the same <filename>KMACHINE</filename>
        name if they are built using the same BSP description.
        The "ep108-zynqmp" and "qemuzynqmp" BSP combination
        in the <filename>meta-xilinx</filename>
        layer is a good example of two BSPs using the same
        <filename>KMACHINE</filename> value (i.e. "zynqmp").
        See the <link linkend='bsp-descriptions'>BSP Descriptions</link> section
        for more information.
    </para>

    <para>
        Every linux-yocto style recipe must also indicate the Linux kernel
        source repository branch used to build the Linux kernel.
        The <ulink url='&YOCTO_DOCS_REF_URL;#var-KBRANCH'><filename>KBRANCH</filename></ulink>
        variable must be set to indicate the branch.
        <note>
            You can use the <filename>KBRANCH</filename> value to define an
            alternate branch typically with a machine override as shown here
            from the <filename>meta-emenlow</filename> layer:
            <literallayout class='monospaced'>
     KBRANCH_emenlow-noemgd = "standard/base"
            </literallayout>
        </note>
    </para>

    <para>
        The linux-yocto style recipes can optionally define the following
        variables:
        <literallayout class='monospaced'>
     KERNEL_FEATURES
     LINUX_KERNEL_TYPE
        </literallayout>
    </para>

    <para>
        <ulink url='&YOCTO_DOCS_REF_URL;#var-LINUX_KERNEL_TYPE'><filename>LINUX_KERNEL_TYPE</filename></ulink>
        defines the kernel type to be
        used in assembling the configuration.
        If you do not specify a <filename>LINUX_KERNEL_TYPE</filename>,
        it defaults to "standard".
        Together with
        <ulink url='&YOCTO_DOCS_REF_URL;#var-KMACHINE'><filename>KMACHINE</filename></ulink>,
        <filename>LINUX_KERNEL_TYPE</filename> defines the search
        arguments used by the kernel tools to find the
        appropriate description within the kernel Metadata with which to
        build out the sources and configuration.
        The linux-yocto recipes define "standard", "tiny", and "preempt-rt"
        kernel types.
        See the "<link linkend='kernel-types'>Kernel Types</link>" section
        for more information on kernel types.
    </para>

    <para>
        During the build, the kern-tools search for the BSP description
        file that most closely matches the <filename>KMACHINE</filename>
        and <filename>LINUX_KERNEL_TYPE</filename> variables passed in from the
        recipe.
        The tools use the first BSP description it finds that match
        both variables.
        If the tools cannot find a match, they issue a warning such as
        the following:
        <literallayout class='monospaced'>
     WARNING: Can't find any BSP hardware or required configuration fragments.
     WARNING: Looked at meta/cfg/broken/emenlow-broken/hdw_frags.txt and
              meta/cfg/broken/emenlow-broken/required_frags.txt in directory:
              meta/cfg/broken/emenlow-broken
        </literallayout>
        In this example, <filename>KMACHINE</filename> was set to "emenlow-broken"
        and <filename>LINUX_KERNEL_TYPE</filename> was set to "broken".
    </para>

    <para>
        The tools first search for the <filename>KMACHINE</filename> and
        then for the <filename>LINUX_KERNEL_TYPE</filename>.
        If the tools cannot find a partial match, they will use the
        sources from the <filename>KBRANCH</filename> and any configuration
        specified in the
        <ulink url='&YOCTO_DOCS_REF_URL;#var-SRC_URI'><filename>SRC_URI</filename></ulink>.
    </para>

    <para>
        You can use the
        <ulink url='&YOCTO_DOCS_REF_URL;#var-KERNEL_FEATURES'><filename>KERNEL_FEATURES</filename></ulink>
        variable
        to include features (configuration fragments, patches, or both) that
        are not already included by the <filename>KMACHINE</filename> and
        <filename>LINUX_KERNEL_TYPE</filename> variable combination.
        For example, to include a feature specified as
        "features/netfilter/netfilter.scc",
        specify:
        <literallayout class='monospaced'>
     KERNEL_FEATURES += "features/netfilter/netfilter.scc"
        </literallayout>
        To include a feature called "cfg/sound.scc" just for the
        <filename>qemux86</filename> machine, specify:
        <literallayout class='monospaced'>
     KERNEL_FEATURES_append_qemux86 = " cfg/sound.scc"
        </literallayout>
        The value of the entries in <filename>KERNEL_FEATURES</filename>
        are dependent on their location within the kernel Metadata itself.
        The examples here are taken from the <filename>meta</filename>
        branch of the <filename>linux-yocto-3.19</filename> repository.
        Within that branch, "features" and "cfg" are subdirectories of the
        <filename>meta/cfg/kernel-cache</filename> directory.
        For more information, see the
        "<link linkend='kernel-metadata-syntax'>Kernel Metadata Syntax</link>" section.
        <note>
            The processing of the these variables has evolved some between the
	        0.9 and 1.3 releases of the Yocto Project and associated
	        kern-tools sources.
            The descriptions in this section are accurate for 1.3 and later
	        releases of the Yocto Project.
        </note>
    </para>
</section>

<section id='kernel-metadata-syntax'>
    <title>Kernel Metadata Syntax</title>

    <para>
        The kernel Metadata consists of three primary types of files:
        <filename>scc</filename>
        <footnote>
            <para>
                <filename>scc</filename> stands for Series Configuration
                Control, but the naming has less significance in the
                current implementation of the tooling than it had in the
                past.
                Consider <filename>scc</filename> files to be description files.
            </para>
        </footnote>
        description files, configuration fragments, and patches.
        The <filename>scc</filename> files define variables and include or
        otherwise reference any of the three file types.
        The description files are used to aggregate all types of kernel
        Metadata into
        what ultimately describes the sources and the configuration required
        to build a Linux kernel tailored to a specific machine.
    </para>

    <para>
        The <filename>scc</filename> description files are used to define two
        fundamental types of kernel Metadata:
        <itemizedlist>
            <listitem><para>Features</para></listitem>
            <listitem><para>Board Support Packages (BSPs)</para></listitem>
        </itemizedlist>
    </para>

    <para>
        Features aggregate sources in the form of patches and configuration
        fragments into a modular reusable unit.
        You can use features to implement conceptually separate kernel
        Metadata descriptions such as pure configuration fragments,
        simple patches, complex features, and kernel types.
        <link linkend='kernel-types'>Kernel types</link> define general
        kernel features and policy to be reused in the BSPs.
    </para>

    <para>
        BSPs define hardware-specific features and aggregate them with kernel
        types to form the final description of what will be assembled and built.
    </para>

    <para>
        While the kernel Metadata syntax does not enforce any logical
        separation of configuration fragments, patches, features or kernel
        types, best practices dictate a logical separation of these types
        of Metadata.
        The following Metadata file hierarchy is recommended:
        <literallayout class='monospaced'>
     <replaceable>base</replaceable>/
        bsp/
        cfg/
        features/
        ktypes/
        patches/
        </literallayout>
    </para>

    <para>
        The <filename>bsp</filename> directory contains the
        <link linkend='bsp-descriptions'>BSP descriptions</link>.
        The remaining directories all contain "features".
        Separating <filename>bsp</filename> from the rest of the structure
        aids conceptualizing intended usage.
    </para>

    <para>
        Use these guidelines to help place your <filename>scc</filename>
        description files within the structure:
        <itemizedlist>
            <listitem><para>If your file contains
                only configuration fragments, place the file in the
                <filename>cfg</filename> directory.</para></listitem>
            <listitem><para>If your file contains
                only source-code fixes, place the file in the
                <filename>patches</filename> directory.</para></listitem>
            <listitem><para>If your file encapsulates
                a major feature, often combining sources and configurations,
                place the file in <filename>features</filename> directory.
                </para></listitem>
            <listitem><para>If your file aggregates
                non-hardware configuration and patches in order to define a
                base kernel policy or major kernel type to be reused across
                multiple BSPs, place the file in <filename>ktypes</filename>
                directory.
                </para></listitem>
        </itemizedlist>
    </para>

    <para>
        These distinctions can easily become blurred - especially as
        out-of-tree features slowly merge upstream over time.
        Also, remember that how the description files are placed is
        a purely logical organization and has no impact on the functionality
        of the kernel Metadata.
        There is no impact because all of <filename>cfg</filename>,
        <filename>features</filename>, <filename>patches</filename>, and
        <filename>ktypes</filename>, contain "features" as far as the kernel
        tools are concerned.
    </para>

    <para>
        Paths used in kernel Metadata files are relative to
        <filename>&lt;base&gt;</filename>, which is either
        <ulink url='&YOCTO_DOCS_REF_URL;#var-FILESEXTRAPATHS'><filename>FILESEXTRAPATHS</filename></ulink>
        if you are creating Metadata in
        <link linkend='recipe-space-metadata'>recipe-space</link>,
        or <filename>meta/cfg/kernel-cache/</filename> if you are creating
        <link linkend='metadata-outside-the-recipe-space'>Metadata outside of the recipe-space</link>.
    </para>

    <section id='configuration'>
        <title>Configuration</title>

        <para>
            The simplest unit of kernel Metadata is the configuration-only
            feature.
            This feature consists of one or more Linux kernel configuration
            parameters in a configuration fragment file
            (<filename>.cfg</filename>) and a <filename>.scc</filename> file
            that describes the fragment.
        </para>

        <para>
            The Symmetric Multi-Processing (SMP) fragment included in the
            <filename>linux-yocto-3.19</filename> Git repository
            consists of the following two files:
            <literallayout class='monospaced'>
     cfg/smp.scc:
        define KFEATURE_DESCRIPTION "Enable SMP"
        define KFEATURE_COMPATIBILITY all

        kconf hardware smp.cfg

     cfg/smp.cfg:
        CONFIG_SMP=y
        CONFIG_SCHED_SMT=y
        # Increase default NR_CPUS from 8 to 64 so that platform with
        # more than 8 processors can be all activated at boot time
        CONFIG_NR_CPUS=64
            </literallayout>
            You can find information on configuration fragment files in the
            "<link linkend='creating-config-fragments'>Creating Configuration Fragments</link>"
            section.
        </para>

        <para>
            <ulink url='&YOCTO_DOCS_REF_URL;#var-KFEATURE_DESCRIPTION'><filename>KFEATURE_DESCRIPTION</filename></ulink>
            provides a short description of the fragment.
            Higher level kernel tools use this description.
        </para>

        <para>
            The <filename>kconf</filename> command is used to include the
            actual configuration fragment in an <filename>.scc</filename>
            file, and the "hardware" keyword identifies the fragment as
            being hardware enabling, as opposed to general policy,
            which would use the "non-hardware" keyword.
            The distinction is made for the benefit of the configuration
            validation tools, which warn you if a hardware fragment
            overrides a policy set by a non-hardware fragment.
            <note>
                The description file can include multiple
                <filename>kconf</filename> statements, one per fragment.
            </note>
        </para>

        <para>
            As described in the
            "<link linkend='validating-configuration'>Validating Configuration</link>"
            section, you can use the following BitBake command to audit your
            configuration:
            <literallayout class='monospaced'>
     $ bitbake linux-yocto -c kernel_configcheck -f
            </literallayout>
        </para>
    </section>

    <section id='patches'>
        <title>Patches</title>

        <para>
            Patch descriptions are very similar to configuration fragment
            descriptions, which are described in the previous section.
            However, instead of a <filename>.cfg</filename> file, these
            descriptions work with source patches.
        </para>

        <para>
            A typical patch includes a description file and the patch itself:
            <literallayout class='monospaced'>
     patches/mypatch.scc:
        patch mypatch.patch

     patches/mypatch.patch:
        <replaceable>typical-patch</replaceable>
            </literallayout>
            You can create the typical <filename>.patch</filename>
            file using <filename>diff -Nurp</filename> or
            <filename>git format-patch</filename>.
        </para>

        <para>
            The description file can include multiple patch statements,
            one per patch.
        </para>
    </section>

    <section id='features'>
        <title>Features</title>

        <para>
            Features are complex kernel Metadata types that consist
            of configuration fragments (<filename>kconf</filename>), patches
            (<filename>patch</filename>), and possibly other feature
            description files (<filename>include</filename>).
        </para>

        <para>
            Here is an example that shows a feature description file:
            <literallayout class='monospaced'>
     features/myfeature.scc
        define KFEATURE_DESCRIPTION "Enable myfeature"

        patch 0001-myfeature-core.patch
        patch 0002-myfeature-interface.patch

        include cfg/myfeature_dependency.scc
        kconf non-hardware myfeature.cfg
            </literallayout>
            This example shows how the <filename>patch</filename> and
            <filename>kconf</filename> commands are used as well as
            how an additional feature description file is included.
        </para>

        <para>
            Typically, features are less granular than configuration
            fragments and are more likely than configuration fragments
            and patches to be the types of things you want to specify
            in the <filename>KERNEL_FEATURES</filename> variable of the
            Linux kernel recipe.
            See the "<link linkend='using-kernel-metadata-in-a-recipe'>Using Kernel Metadata in a Recipe</link>"
            section earlier in the manual.
        </para>
    </section>

    <section id='kernel-types'>
        <title>Kernel Types</title>

        <para>
            A kernel type defines a high-level kernel policy by
            aggregating non-hardware configuration fragments with
            patches you want to use when building a Linux kernels of a
            specific type.
            Syntactically, kernel types are no different than features
            as described in the "<link linkend='features'>Features</link>"
            section.
            The <filename>LINUX_KERNEL_TYPE</filename> variable in the kernel
            recipe selects the kernel type.
            See the "<link linkend='using-kernel-metadata-in-a-recipe'>Using Kernel Metadata in a Recipe</link>"
            section for more information.
        </para>

        <para>
            As an example, the <filename>linux-yocto-3.19</filename>
            tree defines three kernel types: "standard",
            "tiny", and "preempt-rt":
            <itemizedlist>
                <listitem><para>"standard":
                    Includes the generic Linux kernel policy of the Yocto
                    Project linux-yocto kernel recipes.
                    This policy includes, among other things, which file
                    systems, networking options, core kernel features, and
                    debugging and tracing options are supported.
                    </para></listitem>
                <listitem><para>"preempt-rt":
                    Applies the <filename>PREEMPT_RT</filename>
                    patches and the configuration options required to
                    build a real-time Linux kernel.
                    This kernel type inherits from the "standard" kernel type.
                    </para></listitem>
                <listitem><para>"tiny":
                    Defines a bare minimum configuration meant to serve as a
                    base for very small Linux kernels.
                    The "tiny" kernel type is independent from the "standard"
                    configuration.
                    Although the "tiny" kernel type does not currently include
                    any source changes, it might in the future.
                    </para></listitem>
            </itemizedlist>
        </para>

        <para>
            The "standard" kernel type is defined by
            <filename>standard.scc</filename>:
            <literallayout class='monospaced'>
     # Include this kernel type fragment to get the standard features and
     # configuration values.

     # Include all standard features
     include standard-nocfg.scc

     kconf non-hardware standard.cfg

     # individual cfg block section
     include cfg/fs/devtmpfs.scc
     include cfg/fs/debugfs.scc
     include cfg/fs/btrfs.scc
     include cfg/fs/ext2.scc
     include cfg/fs/ext3.scc
     include cfg/fs/ext4.scc

     include cfg/net/ipv6.scc
     include cfg/net/ip_nf.scc
     include cfg/net/ip6_nf.scc
     include cfg/net/bridge.scc
            </literallayout>
        </para>

        <para>
            As with any <filename>.scc</filename> file, a
            kernel type definition can aggregate other
            <filename>.scc</filename> files with
            <filename>include</filename> commands.
            These definitions can also directly pull in
            configuration fragments and patches with the
            <filename>kconf</filename> and <filename>patch</filename>
            commands, respectively.
        </para>

        <note>
            It is not strictly necessary to create a kernel type
            <filename>.scc</filename> file.
            The Board Support Package (BSP) file can implicitly define
            the kernel type using a <filename>define
            <ulink url='&YOCTO_DOCS_REF_URL;#var-KTYPE'>KTYPE</ulink> myktype</filename>
            line.
            See the "<link linkend='bsp-descriptions'>BSP Descriptions</link>"
            section for more information.
        </note>
    </section>

    <section id='bsp-descriptions'>
        <title>BSP Descriptions</title>

        <para>
            BSP descriptions (i.e. <filename>*.scc</filename> files)
            combine kernel types with hardware-specific features.
            The hardware-specific Metadata is typically defined
            independently in the BSP layer, and then aggregated with each
            supported kernel type.
            <note>
                For BSPs supported by the Yocto Project, the BSP description
                files are located in the <filename>bsp</filename> directory
                of the <filename>yocto-kernel-cache</filename> repository
                organized under the "Yocto Linux Kernel" heading in the
                <ulink url='http://git.yoctoproject.org/cgit/cgit.cgi'>Yocto Project Source Repositories</ulink>.
            </note>
        </para>

        <para>
            This section provides a BSP description structural overview along
            with aggregation concepts as well as a detailed example using
            a BSP supported by the Yocto Project (i.e. Minnow Board).
        </para>

        <section id='bsp-description-file-overview'>
            <title>Overview</title>

            <para>
                For simplicity, consider the following top-level BSP
                description file.
                Top-level BSP descriptions files employ both a structure
                and naming convention for consistency.
                The naming convention for the file is as follows:
                <literallayout class='monospaced'>
     <replaceable>bsp_name</replaceable>-<replaceable>kernel_type</replaceable>.scc
                </literallayout>
                Here are some example top-level BSP filenames for the
                Minnow Board BSP, which is supported by the Yocto Project:
                <literallayout class='monospaced'>
     minnow-standard.scc
     minnow-preempt-rt.scc
     minnow-tiny.scc
                </literallayout>
                Each file uses the BSP name followed by the kernel type.
            </para>

            <para>
                is simple BSP description file whose name has the
                form
                <replaceable>mybsp</replaceable><filename>-standard</filename>
                and supports the <replaceable>mybsp</replaceable> machine using
                a standard kernel:
                <literallayout class='monospaced'>
     define KMACHINE <replaceable>mybsp</replaceable>
     define KTYPE standard
     define KARCH i386

     include ktypes/standard

     include <replaceable>mybsp</replaceable>.scc

     kconf hardware <replaceable>mybsp</replaceable>-<replaceable>extra</replaceable>.cfg
                </literallayout>
                Every top-level BSP description file should define the
                <ulink url='&YOCTO_DOCS_REF_URL;#var-KMACHINE'><filename>KMACHINE</filename></ulink>,
                <ulink url='&YOCTO_DOCS_REF_URL;#var-KTYPE'><filename>KTYPE</filename></ulink>,
                and <ulink url='&YOCTO_DOCS_REF_URL;#var-KARCH'><filename>KARCH</filename></ulink>
                variables.
                These variables allow the OpenEmbedded build system to identify
                the description as meeting the criteria set by the recipe being
                built.
                This simple example supports the "mybsp" machine for the "standard"
                kernel and the "i386" architecture.
            </para>

            <para>
                Be aware that a hard link between the
                <filename>KTYPE</filename> variable and a kernel type description
                file does not exist.
                Thus, if you do not have kernel types defined in your kernel
                Metadata, you only need to ensure that the kernel recipe's
                <ulink url='&YOCTO_DOCS_REF_URL;#var-LINUX_KERNEL_TYPE'><filename>LINUX_KERNEL_TYPE</filename></ulink>
                variable and the <filename>KTYPE</filename> variable in the
                BSP description file match.
                <note>
                    Future versions of the tooling make the specification of
                    <filename>KTYPE</filename> in the BSP optional.
                </note>
            </para>

            <para>
                To separate your kernel policy from your hardware configuration,
                you include a kernel type (<filename>ktype</filename>), such as
                "standard".
                In the previous example, this is done using the following:
                <literallayout class='monospaced'>
     include ktypes/standard
                </literallayout>
                In the previous example, <filename>ktypes/standard.scc</filename>
                aggregates all the configuration fragments, patches, and
                features that make up your standard kernel policy.
                See the "<link linkend='kernel-types'>Kernel Types</link>" section
                for more information.
            </para>

            <para>
                To aggregate common configurations and features specific to the
                kernel for <replaceable>mybsp</replaceable>, use the following:
                <literallayout class='monospaced'>
     include <replaceable>mybsp</replaceable>.scc
                </literallayout>
                For information on how to break a complete
                <filename>.config</filename> file into the various
                configuration fragments, see the
                "<link linkend='creating-config-fragments'>Creating Configuration Fragments</link>"
                section.
            </para>

            <para>
                Finally, if you have any configurations specific to the
                hardware that are not in a <filename>*.scc</filename> file,
                you can include them as follows:
                <literallayout class='monospaced'>
     kconf hardware <replaceable>mybsp</replaceable>-<replaceable>extra</replaceable>.cfg
                </literallayout>
            </para>
        </section>

        <section id='bsp-description-file-example-minnow'>
            <title>Example</title>

            <para>
                Many real-world examples are more complex.
                Like any other <filename>.scc</filename> file, BSP
                descriptions can aggregate features.
                Consider the Minnow BSP definition from the
                <filename>linux-yocto-4.4</filename> in the
                Yocto Project
                <ulink url='&YOCTO_DOCS_REF_URL;#source-repositories'>Source Repositories</ulink>
                (i.e.
                <filename>yocto-kernel-cache/bsp/minnow</filename>):
                <literallayout class='monospaced'>
     minnow.scc:
         include cfg/x86.scc
         include features/eg20t/eg20t.scc
         include cfg/dmaengine.scc
         include features/power/intel.scc
         include cfg/efi.scc
         include features/usb/ehci-hcd.scc
         include features/usb/ohci-hcd.scc
         include features/usb/usb-gadgets.scc
         include features/usb/touchscreen-composite.scc
         include cfg/timer/hpet.scc
         include features/leds/leds.scc
         include features/spi/spidev.scc
         include features/i2c/i2cdev.scc
         include features/mei/mei-txe.scc

         # Earlyprintk and port debug requires 8250
         kconf hardware cfg/8250.cfg

         kconf hardware minnow.cfg
         kconf hardware minnow-dev.cfg
                </literallayout>
            </para>

            <para>
                The <filename>minnow.scc</filename> description file includes
                a hardware configuration fragment
                (<filename>minnow.cfg</filename>) specific to the Minnow
                BSP as well as several more general configuration
                fragments and features enabling hardware found on the
                machine.
                This <filename>minnow.scc</filename> description file is then
                included in each of the three
                "minnow" description files for the supported kernel types
                (i.e. "standard", "preempt-rt", and "tiny").
                Consider the "minnow" description for the "standard" kernel
                type:
                <literallayout class='monospaced'>
     minnow-standard.scc:
         define KMACHINE minnow
         define KTYPE standard
         define KARCH i386

         include ktypes/standard

         include minnow.scc

         # Extra minnow configs above the minimal defined in minnow.scc
         include cfg/efi-ext.scc
         include features/media/media-all.scc
         include features/sound/snd_hda_intel.scc

         # The following should really be in standard.scc
         # USB live-image support
         include cfg/usb-mass-storage.scc
         include cfg/boot-live.scc

         # Basic profiling
         include features/latencytop/latencytop.scc
         include features/profiling/profiling.scc

         # Requested drivers that don't have an existing scc
         kconf hardware minnow-drivers-extra.cfg
                </literallayout>
                The <filename>include</filename> command midway through the file
                includes the <filename>minnow.scc</filename> description that
                defines all enabled hardware for the BSP that is common to
                all kernel types.
                Using this command significantly reduces duplication.
            </para>

            <para>
                Now consider the "minnow" description for the "tiny" kernel
                type:
                <literallayout class='monospaced'>
     minnow-tiny.scc:
        define KMACHINE minnow
        define KTYPE tiny
        define KARCH i386

        include ktypes/tiny

        include minnow.scc
                </literallayout>
                As you might expect, the "tiny" description includes quite a
                bit less.
                In fact, it includes only the minimal policy defined by the
                "tiny" kernel type and the hardware-specific configuration
                required for booting the machine along with the most basic
                functionality of the system as defined in the base "minnow"
                description file.
            </para>

            <para>
                Notice again the three critical variables:
                <filename>KMACHINE</filename>, <filename>KTYPE</filename>,
                and <filename>KARCH</filename>.
                Of these variables, only the <filename>KTYPE</filename> has changed.
                It is now set to "tiny".
            </para>
        </section>
    </section>
</section>

<section id='kernel-metadata-location'>
    <title>Kernel Metadata Location</title>

    <para>
        Kernel Metadata always exists outside of the kernel tree either
        defined in a kernel recipe (recipe-space) or outside of the recipe.
        Where you choose to define the Metadata depends on what you want
        to do and how you intend to work.
        Regardless of where you define the kernel Metadata, the syntax used
        applies equally.
    </para>

    <para>
        If you are unfamiliar with the Linux kernel and only wish
        to apply a configuration and possibly a couple of patches provided to
        you by others, the recipe-space method is recommended.
        This method is also a good approach if you are working with Linux kernel
        sources you do not control or if you just do not want to maintain a
        Linux kernel Git repository on your own.
        For partial information on how you can define kernel Metadata in
        the recipe-space, see the
        "<link linkend='modifying-an-existing-recipe'>Modifying an Existing Recipe</link>"
        section.
    </para>

    <para>
        Conversely, if you are actively developing a kernel and are already
        maintaining a Linux kernel Git repository of your own, you might find
        it more convenient to work with kernel Metadata kept outside the
        recipe-space.
        Working with Metadata in this area can make iterative development of
        the Linux kernel more efficient outside of the BitBake environment.
    </para>

    <section id='recipe-space-metadata'>
        <title>Recipe-Space Metadata</title>

        <para>
            When stored in recipe-space, the kernel Metadata files reside in a
            directory hierarchy below
            <ulink url='&YOCTO_DOCS_REF_URL;#var-FILESEXTRAPATHS'><filename>FILESEXTRAPATHS</filename></ulink>.
            For a linux-yocto recipe or for a Linux kernel recipe derived
            by copying and modifying
            <filename>oe-core/meta-skeleton/recipes-kernel/linux/linux-yocto-custom.bb</filename>
            to a recipe in your layer, <filename>FILESEXTRAPATHS</filename>
            is typically set to
            <filename>${</filename><ulink url='&YOCTO_DOCS_REF_URL;#var-THISDIR'><filename>THISDIR</filename></ulink><filename>}/${</filename><ulink url='&YOCTO_DOCS_REF_URL;#var-PN'><filename>PN</filename></ulink><filename>}</filename>.
            See the "<link linkend='modifying-an-existing-recipe'>Modifying an Existing Recipe</link>"
            section for more information.
        </para>

        <para>
            Here is an example that shows a trivial tree of kernel Metadata
            stored in recipe-space within a BSP layer:
            <literallayout class='monospaced'>
     meta-<replaceable>my_bsp_layer</replaceable>/
     `-- recipes-kernel
         `-- linux
             `-- linux-yocto
                 |-- bsp-standard.scc
                 |-- bsp.cfg
                 `-- standard.cfg
            </literallayout>
        </para>

        <para>
            When the Metadata is stored in recipe-space, you must take
            steps to ensure BitBake has the necessary information to decide
            what files to fetch and when they need to be fetched again.
            It is only necessary to specify the <filename>.scc</filename>
            files on the
            <ulink url='&YOCTO_DOCS_REF_URL;#var-SRC_URI'><filename>SRC_URI</filename></ulink>.
            BitBake parses them and fetches any files referenced in the
            <filename>.scc</filename> files by the <filename>include</filename>,
            <filename>patch</filename>, or <filename>kconf</filename> commands.
            Because of this, it is necessary to bump the recipe
            <ulink url='&YOCTO_DOCS_REF_URL;#var-PR'><filename>PR</filename></ulink>
            value when changing the content of files not explicitly listed
            in the <filename>SRC_URI</filename>.
        </para>

        <para>
            If the BSP description is in recipe space, you cannot simply list
            the <filename>*.scc</filename> in the <filename>SRC_URI</filename>
            statement.
            You need to use the following form from your kernel append file:
            <literallayout class='monospaced'>
     SRC_URI_append_<replaceable>myplatform</replaceable> = " \
        file://<replaceable>myplatform</replaceable>;type=kmeta;destsuffix=<replaceable>myplatform</replaceable> \
        "
            </literallayout>
        </para>
    </section>

    <section id='metadata-outside-the-recipe-space'>
        <title>Metadata Outside the Recipe-Space</title>

        <para>
            When stored outside of the recipe-space, the kernel Metadata
            files reside in a separate repository.
            The OpenEmbedded build system adds the Metadata to the build as
            a "ktype=meta" repository through the
            <ulink url='&YOCTO_DOCS_REF_URL;#var-SRC_URI'><filename>SRC_URI</filename></ulink>
            variable.
            As an example, consider the following <filename>SRC_URI</filename>
            statement from the <filename>linux-yocto_4.4.bb</filename>
            kernel recipe:
            <literallayout class='monospaced'>
     SRC_URI = "git://git.yoctoproject.org/linux-yocto-4.4.git;name=machine;branch=${KBRANCH}; \
                git://git.yoctoproject.org/yocto-kernel-cache;type=kmeta;name=meta;branch=yocto-4.4;destsuffix=${KMETA}"
            </literallayout>
            <filename>${KMETA}</filename>, in this context, is simply used to
            name the directory into which the Git fetcher places the Metadata.
            This behavior is no different than any multi-repository
            <filename>SRC_URI</filename> statement used in a recipe (e.g.
            see the previous section).
        </para>

        <para>
            You can keep kernel Metadata in a "kernel-cache", which is a
            directory containing configuration fragments.
            As with any Metadata kept outside the recipe-space, you simply
            need to use the <filename>SRC_URI</filename> statement with the
            "type=kmeta" attribute.
            Doing so makes the kernel Metadata available during the
            configuration phase.
        </para>

<!--


        <para>
            Following is an example that shows how a trivial tree of Metadata
            is stored in a custom Linux kernel Git repository:
            <literallayout class='monospaced'>
     meta/
     `&dash;&dash; cfg
         `&dash;&dash; kernel-cache
             |&dash;&dash; bsp-standard.scc
             |&dash;&dash; bsp.cfg
             `&dash;&dash; standard.cfg
            </literallayout>
        </para>

        <para>
            To use a branch different from where the sources reside,
            specify the branch in the <filename>KMETA</filename> variable
            in your Linux kernel recipe.
            Here is an example:
            <literallayout class='monospaced'>
     KMETA = "meta"
            </literallayout>
            To use the same branch as the sources, set
            <filename>KMETA</filename> to an empty string:
            <literallayout class='monospaced'>
     KMETA = ""
            </literallayout>
            If you are working with your own sources and want to create an
            orphan <filename>meta</filename> branch, use these commands
            from within your Linux kernel Git repository:
            <literallayout class='monospaced'>
     $ git checkout &dash;&dash;orphan meta
     $ git rm -rf .
     $ git commit &dash;&dash;allow-empty -m "Create orphan meta branch"
            </literallayout>
        </para>
-->

        <para>
            If you modify the Metadata, you must not forget to update the
            <ulink url='&YOCTO_DOCS_REF_URL;#var-SRCREV'><filename>SRCREV</filename></ulink>
            statements in the kernel's recipe.
            In particular, you need to update the
            <filename>SRCREV_meta</filename> variable to match the commit in
            the <filename>KMETA</filename> branch you wish to use.
            Changing the data in these branches and not updating the
            <filename>SRCREV</filename> statements to match will cause the
            build to fetch an older commit.
        </para>
    </section>
</section>

<section id='organizing-your-source'>
    <title>Organizing Your Source</title>

    <para>
        Many recipes based on the <filename>linux-yocto-custom.bb</filename>
        recipe use Linux kernel sources that have only a single
        branch - "master".
        This type of repository structure is fine for linear development
        supporting a single machine and architecture.
        However, if you work with multiple boards and architectures,
        a kernel source repository with multiple branches is more
        efficient.
        For example, suppose you need a series of patches for one board to boot.
        Sometimes, these patches are works-in-progress or fundamentally wrong,
        yet they are still necessary for specific boards.
        In these situations, you most likely do not want to include these
        patches in every kernel you build (i.e. have the patches as part of
        the lone "master" branch).
        It is situations like these that give rise to multiple branches used
        within a Linux kernel sources Git repository.
    </para>

    <para>
        Repository organization strategies exist that maximize source reuse,
        remove redundancy, and logically order your changes.
        This section presents strategies for the following cases:
        <itemizedlist>
            <listitem><para>Encapsulating patches in a feature description
                and only including the patches in the BSP descriptions of
                the applicable boards.</para></listitem>
            <listitem><para>Creating a machine branch in your
                kernel source repository and applying the patches on that
                branch only.</para></listitem>
            <listitem><para>Creating a feature branch in your
                kernel source repository and merging that branch into your
                BSP when needed.</para></listitem>
        </itemizedlist>
    </para>

    <para>
        The approach you take is entirely up to you
        and depends on what works best for your development model.
    </para>

    <section id='encapsulating-patches'>
        <title>Encapsulating Patches</title>

        <para>
            if you are reusing patches from an external tree and are not
            working on the patches, you might find the encapsulated feature
            to be appropriate.
            Given this scenario, you do not need to create any branches in the
            source repository.
            Rather, you just take the static patches you need and encapsulate
            them within a feature description.
            Once you have the feature description, you simply include that into
            the BSP description as described in the
            "<link linkend='bsp-descriptions'>BSP Descriptions</link>"
            section.
        </para>

        <para>
            You can find information on how to create patches and BSP
            descriptions in the "<link linkend='patches'>Patches</link>" and
            "<link linkend='bsp-descriptions'>BSP Descriptions</link>"
            sections.
        </para>
    </section>

    <section id='machine-branches'>
        <title>Machine Branches</title>

        <para>
            When you have multiple machines and architectures to support,
            or you are actively working on board support, it is more
            efficient to create branches in the repository based on
            individual machines.
            Having machine branches allows common source to remain in the
            "master" branch with any features specific to a machine stored
            in the appropriate machine branch.
            This organization method frees you from continually reintegrating
            your patches into a feature.
        </para>

        <para>
            Once you have a new branch, you can set up your kernel Metadata
            to use the branch a couple different ways.
            In the recipe, you can specify the new branch as the
            <filename>KBRANCH</filename> to use for the board as
            follows:
            <literallayout class='monospaced'>
     KBRANCH = "mynewbranch"
            </literallayout>
            Another method is to use the <filename>branch</filename> command
            in the BSP description:
            <literallayout class='monospaced'>
     mybsp.scc:
        define KMACHINE mybsp
        define KTYPE standard
        define KARCH i386
        include standard.scc

        branch mynewbranch

        include mybsp-hw.scc
            </literallayout>
        </para>

        <para>
            If you find
            yourself with numerous branches, you might consider using a
            hierarchical branching system similar to what the linux-yocto Linux
            kernel repositories use:
            <literallayout class='monospaced'>
     <replaceable>common</replaceable>/<replaceable>kernel_type</replaceable>/<replaceable>machine</replaceable>
            </literallayout>
        </para>

        <para>
            If you had two kernel types, "standard" and "small" for
            instance, three machines, and <replaceable>common</replaceable>
            as <filename>mydir</filename>, the branches in your
            Git repository might look like this:
            <literallayout class='monospaced'>
     mydir/base
     mydir/standard/base
     mydir/standard/machine_a
     mydir/standard/machine_b
     mydir/standard/machine_c
     mydir/small/base
     mydir/small/machine_a
            </literallayout>
        </para>

        <para>
            This organization can help clarify the branch relationships.
            In this case, <filename>mydir/standard/machine_a</filename>
            includes everything in <filename>mydir/base</filename> and
            <filename>mydir/standard/base</filename>.
            The "standard" and "small" branches add sources specific to those
            kernel types that for whatever reason are not appropriate for the
            other branches.
            <note>The "base" branches are an artifact of the way Git manages
                its data internally on the filesystem: Git will not allow you
                to use <filename>mydir/standard</filename> and
                <filename>mydir/standard/machine_a</filename> because it
                would have to create a file and a directory named "standard".
            </note>
        </para>
    </section>

    <section id='feature-branches'>
        <title>Feature Branches</title>

        <para>
            When you are actively developing new features, it can be more
            efficient to work with that feature as a branch, rather than
            as a set of patches that have to be regularly updated.
            The Yocto Project Linux kernel tools provide for this with
            the <filename>git merge</filename> command.
        </para>

        <para>
            To merge a feature branch into a BSP, insert the
            <filename>git merge</filename> command after any
            <filename>branch</filename> commands:
            <literallayout class='monospaced'>
     mybsp.scc:
        define KMACHINE mybsp
        define KTYPE standard
        define KARCH i386
        include standard.scc

        branch mynewbranch
        git merge myfeature

        include mybsp-hw.scc
            </literallayout>
        </para>
    </section>
</section>

<section id='scc-reference'>
    <title>SCC Description File Reference</title>

    <para>
        This section provides a brief reference for the commands you can use
        within an SCC description file (<filename>.scc</filename>):
        <itemizedlist>
            <listitem><para><filename>branch [ref]</filename>:
                Creates a new branch relative to the current branch
                (typically <filename>${KTYPE}</filename>) using
                the currently checked-out branch, or "ref" if specified.
                </para></listitem>
            <listitem><para><filename>define</filename>:
                Defines variables, such as <filename>KMACHINE</filename>,
                <filename>KTYPE</filename>, <filename>KARCH</filename>,
                and <filename>KFEATURE_DESCRIPTION</filename>.</para></listitem>
            <listitem><para><filename>include SCC_FILE</filename>:
                Includes an SCC file in the current file.
                The file is parsed as if you had inserted it inline.
                </para></listitem>
            <listitem><para><filename>kconf [hardware|non-hardware] CFG_FILE</filename>:
                Queues a configuration fragment for merging into the final
                Linux <filename>.config</filename> file.</para></listitem>
            <listitem><para><filename>git merge GIT_BRANCH</filename>:
                Merges the feature branch into the current branch.
                </para></listitem>
            <listitem><para><filename>patch PATCH_FILE</filename>:
                Applies the patch to the current Git branch.</para></listitem>
        </itemizedlist>
    </para>
</section>

</chapter>
<!--
vim: expandtab tw=80 ts=4
-->