全面解析PowerPC架构下的扁平设备树FDT(ZT)

Sailor_forever  sailing_9806#163.com

(本原创文章发表于Sailor_forever 的个人blog，未经本人许可，不得用于商业用途。任何个人、媒体、其他网站不得私自抄袭；网络媒体转载请注明出处，增加原文链接，否则属于侵权行为。如有任何问题，请留言或者发邮件给[e-mail])
http://blog.csdn.net/sailor_8318/archive/2009/12/26/5078959.aspx



【摘要】本文以MPC8378处理器、Linux2.6.25内核及U-boot1.3.4为例，讲述了如何在PowerPC架构下使用FDT。首先介绍了引入FDT的背景，接着详细介绍了FDT的组成及制作。最后介绍了U-boot及内核如何支持FDT。


【关键字】PowerPC，MPC8378，DTS，DTB，FDT，device node，property，compatible


1    背景    2
2    设备树的描述方式    2
2.1    root Node    3
2.2    chosen    3
2.3    cpus Node    3
2.4    System Memory    5
2.5    Devices    5
2.5.1    Compatible属性    6
2.5.2    Addressing    6
2.6    Interrupts and Interrupt Controllers    7
3    如何制作设备树映像    8
3.1    输入    8
3.2    输出    9
3.3    命令格式    9
4    设备树的传递途径    9
4.1    U-boot对FDT的支持    10
4.2    如何配置FDT    10
4.3    如何传递设备树    10
5    内核如何解析设备树    12
6    设备树对驱动设计产生的影响    15



1    背景
通常情况下，桌面机和服务器可以兼容大部分软件。最好的结果就是当添加新硬件时，无需重新编译Linux内核。标准的固件接口可以保证bootloader将正确的参数传递给内核。PC可以采用bios，PowerPC and Sparc采用Open-Firmware接口。但是对于嵌入式系统，软件差别太大，内核本身也是定制的，bootloader只需要传递很少的参数，因为大部分信息都是硬编码在系统的配置中的。这样同一个内核映像很难同时使用在多个不同的平台上。

早期的PowerPC平台采用特定的数据结构bd_info来传递参数，其定义在include/asm-ppc/ppcboot.h，#define来定义特定平台的数据域，但并没有什么信息来说明当前采用的是那个bd_info结构，所以必须保证在内核和bootloader中同时更新，以便保持一致。

合并32-bit (arch/ppc) 和 64-bit (arch/ppc64) PowerPC的同时，决定重新整理固件接口，建立新的目录arch/powerpc，这里所有的平台必须向内核提供Open Firmware风格的设备树，以便内核启动时可以获得当前平台的硬件配置。

2    设备树的描述方式
简单的说设备书是一种描述硬件配置信息的数据结构，包括CPU，内存，总线及相关外设。内核启动时可以解析这些信息，以此决定如何配置内核及加载那些驱动。该数据结构有一个单一的根节点“/”。每个节点有个名字并可以包含多个子节点。数据的格式遵循IEEE standard 1275。

Device tree source (.dts)采用一种易编辑的文本方式来表达设备树，device tree compiler tool (dtc)将.dts转换成binary device tree blob(.dtb)。设备树并不是控制系统设备的唯一方法，比如内核对USB和PCI已经有非常方便的检测机制。

/ { // the root node
    an-empty-property;
   
    a-child-node {
    array-prop = <0x100 32>;
    string-prop = "hello, world";
    };
   
    another-child-node {
    binary-prop = [0102CAFE];
    string-list = "yes","no","maybe";
    };
};
Figure 1: Simple example of the .dts file format

2.1    root Node
设备树的起点是根节点，Model和compatible属性指明了当前平台的名字，格式为<mfg>,<board>：
Mfg是vendor，board是板子模型

Compatible属性不一定非得要，但是当两个系统在硬件配置上基本一致时，这个参数可以用于辨别当前系统。
/ {
    model = "fsl,mpc8377rdb";
    compatible = "fsl,mpc8377rdb";
    #address-cells = <1>;
    #size-cells = <1>;

    aliases {
        ethernet0 = &enet0;
        ethernet1 = &enet1;
        serial0 = &serial0;
        serial1 = &serial1;
        //pci0 = &pci0;
    };
    // Child nodes go here
};
Figure 3: Example system root node

2.2    chosen
此节点并不真正代表设备节点，而是一些虚拟的由bootloader传递给内核的一些参数，包括bootargs(cmdline)和initrd等。一般由bootloader在启动内核时添加此节点。

2.3    cpus Node
cpus节点是root节点的子节点，对于多核CPU系统，每个CPU有一个子节点。Cpus节点并不需要特别的特性，但是通常习惯指定#address-cells = <1>和#size-cells = <0>，这指定了各个CPU节点的reg属性的格式，其用于编码物理CPU号。

CPU节点的格式为cpu@x，model属性描述CPU类型，其他的是时钟频率及cache 相关属性。
cpus {
    #cpus = <1>;
    #address-cells = <1>;
    #size-cells = <0>;

    PowerPC,8377@0 {
        device_type = "cpu";
        model = "PowerPC, 8377";
        reg = <0x0>;
        d-cache-line-size = <32>;
        i-cache-line-size = <32>;
        d-cache-size = <32768>;
        i-cache-size = <32768>;
        timebase-frequency = <0>;
        bus-frequency = <0>;
        clock-frequency = <0>;
    };
};
Figure 4: cpus node

cpus {
    #cpus = <2>;
    #address-cells = <1>;
    #size-cells = <0>;

    PowerPC,8641@0 {
        device_type = "cpu";
        reg = <0>;
        d-cache-line-size = <20>;    // 32 bytes
        i-cache-line-size = <20>;    // 32 bytes
        d-cache-size = <8000>;        // L1, 32K
        i-cache-size = <8000>;        // L1, 32K
        timebase-frequency = <0>;    // 33 MHz, from uboot
        bus-frequency = <0>;        // From uboot
        clock-frequency = <0>;        // From uboot
        32-bit;
        linux,boot-cpu;
    };
    PowerPC,8641@1 {
        device_type = "cpu";
        reg = <1>;
        d-cache-line-size = <20>;    // 32 bytes
        i-cache-line-size = <20>;    // 32 bytes
        d-cache-size = <8000>;        // L1, 32K
        i-cache-size = <8000>;        // L1, 32K
        timebase-frequency = <0>;    // 33 MHz, from uboot
        bus-frequency = <0>;        // From uboot
        clock-frequency = <0>;        // From uboot
        32-bit;
    };
};
2.4    System Memory
描述系统内存的节点成为memory node，其为root节点的子节点，通常只用一个memory节点描述系统所有的内存范围，reg属性用来定义当前可用的各个memory范围。
memory {
    device_type = "memory";
    reg = <0x00000000 0x40000000>;    // 256MB at 0
};
Figure 5: Memory node

2.5    Devices
一系列节点用于描述系统总线及设备，每个总线及设备在设备树种都有自己的节点。处理器的local bus通常直接作为根节点的子节点，附着在local bus上的Devices and bridges将作为其子节点。
下图显示的PLB bus上的设备包括interrupt controller, an Ethernet device,
及 OPB bridge，OPB总线上有serial devices and a Flash device
plb {
    compatible = "simple-bus";
    #address-cells = <1>;
    #size-cells = <1>;
    ranges;
    UIC0: interrupt-controller {
    compatible = "ibm,uic-440gp",
    "ibm,uic";
    interrupt-controller;
    #interrupt-cells = <2>;
    };
    ethernet@20000 {
    compatible = "ibm,emac-440gp";
    reg = <0x20000 0x70>;
    interrupt-parent = <&UIC0>;
    interrupts = <0 4>;
    };
    opb {
        compatible = "simple-bus";
        #address-cells = <1>;
        #size-cells = <1>;
        ranges = <0x0 0xe0000000
        0x20000000>;
        serial@0 {
        compatible = "ns16550";
        reg = <0x0 0x10>;
        interrupt-parent = <&UIC0>;
        interrupts = <1 4>;
        };
        serial@10000 {
        compatible = "ns16550";
        reg = <0x10000 0x10>;
        interrupt-parent = <&UIC0>;
        interrupts = <2 4>;
        };
        flash@1ff00000 {
        compatible = "amd,s29gl256n",
        "cfi-flash";
        reg = <0x1ff00000 0x100000>;
        };
    };
};
Figure 6: Simple System Device Hierarchy

2.5.1    Compatible属性
几乎每个设备都有compatible属性，OS利用此关键字来确定node所描述的设备，通常compatible字符串的格式如下：
<manufacturer>,<part-num>
对于每个特定的compatible值，需要为该设备定义一个device tree binding。
有时候compatible是一系列字符串，如果某个设备在寄存器级别和某个旧设备兼容，则可以同时指定多个字串，这样OS就知道这两个驱动是兼容的。通常该设备的compatible字串在前，然后是兼容的旧设备的字串。

2.5.2    Addressing
设备地址由reg属性指定，其为一系列cell单元。格式如下：
reg = <base1 size1 [base2 size2 [...]]>;
每个reg的实际大小有父节点的#address-cells and #size-cells属性决定，#address-cells是用来指定base address基地址的cells个数，#size-cells是用来指定region size的cells个数。Reg所使用的cells个数必须是（#address-cells + #size-cells）的倍数。

Reg定义的是bus address，而非system address，bus address是设备依赖的总线上的相对地址，或者更专业的说bus address是相对于父节点的。

Ranges属性可以将bus address映射到父节点一级，格式如下：
ranges = <addr1 parent1 size1 [...]>;
addr为总线地址，宽度为#address-cells，parent是父节点总线上的地址，宽度为父节点的#address-cells，size宽度为父节点的#size-cells。但是当总线地址和父节点地址映射关系为1：1时，可以简化映射关系：
ranges;

在本示例中，Flash在OPB总线上的地址为0x1ff00000，但OPB总线中，PLB bus address 0xe0000000 映射到了0x0000000 on the OPB bus，因此Flash设备的地址为0xfff00000。

2.6    Interrupts and Interrupt Controllers
设备树的自然布局很方便描述设备间的简单关系，但是中断系统是个比较复杂的例子。可以将serial device描述为OPB总线的子节点，但也可以说其是interrupt controller设备的子节点，那么如何描述呢？目前的规范是，自然树的结构适用于描述那些寻址和控制设备的主要接口，次要连接可以通过phandle属性来描述相互之间的关系，其为节点中的一个指针，指向另一个节点。

对于中断连接，设备节点利用interrupt-parent and interrupts属性来描述到interrupt controller的连接。interrupt-parent是指向描述interrupt controller节点的指针，interrupts是interrupt controller可以触发的一系列中断信号。

Interrupt controller节点必须定义空属性interrupt-controller，同时定义#interrupt-cells，确定几个cells描述一个中断信号。由于Interrupt controller节点在设备树种被其他节点链接，因此必须定义属性linux,phandle = <xx>。

对于大部分SOC系统，通常只有一个interrupt controller,，但是多个interrupt controller之间可以级联。interrupt controller和设备之间的关系就形成了interrupt tree。

对于serial device node，interrupt-parent属性定义了其在中断树中与其父节点的关系。Interrupts属性定义了特定的中断标识，其格式取决于中断树中父节点的#interrupt-cells，通常#interrupt-cells为2，这样第一个值表示interrupt controller中的硬件中断编号，第二个值表示中断触发方式：电平触发或者边沿触发。

/* IPIC
* interrupts cell = <intr #, sense>
* sense values match linux IORESOURCE_IRQ_* defines:
* sense == 8: Level, low assertion
* sense == 2: Edge, high-to-low change
*/
pic@700 {
    linux,phandle = <700>;
    interrupt-controller;
    #address-cells = <0>;
    #interrupt-cells = <2>;
    reg = <700 100>;
    built-in;
    device_type = "ipic";
};

serial@4500 {
    device_type = "serial";
    compatible = "ns16550";
    reg = <4500 100>;
    clock-frequency = <0>;
    interrupts = <9 8>;
    interrupt-parent = <700>;
};

3    如何制作设备树映像
Device tree compiler(dtc)负责将文本格式的设备树转换成OS可以识别的格式。
3.1    输入
Dtc接受三种输入格式：
源文件，即device tree source；
Blob (dtb)，flattened tree format，主要用于检查现有的DTB映像；
FS文件系统，/proc/device-tree下面的文件树目录，主要用于从当前运行的内核中获得设备树映像。

-sh-3.1# ls -al /proc/device-tree
ls -l /proc/device-tree/
-r--r--r--    1 root     root            4 Jan  1 00:05 #address-cells
-r--r--r--    1 root     root            4 Jan  1 00:05 #size-cells
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 aliases
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 chosen
-r--r--r--    1 root     root           15 Jan  1 00:05 compatible
dr-xr-xr-x    3 root     root            0 Jan  1 00:05 cpus
dr-xr-xr-x   15 root     root            0 Jan  1 00:05 immr@e0000000
dr-xr-xr-x    4 root     root            0 Jan  1 00:05 localbus@e0005000
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 memory
-r--r--r--    1 root     root            1 Jan  1 00:05 name
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 pci@e000a000
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 redbox-fpga-card0@F0000000
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 redbox-fpga-card1@F0000000
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 redbox-fpga-dbstate@F0000000
dr-xr-xr-x    2 root     root            0 Jan  1 00:05 redbox-fpga-misc@F0000000

ls -l /proc/device-tree/immr/@e0000000/
-r--r--r--    1 root     root            4 Jan  1 00:06 #address-cells
-r--r--r--    1 root     root            4 Jan  1 00:06 #size-cells
-r--r--r--    1 root     root            4 Jan  1 00:06 bus-frequency
-r--r--r--    1 root     root           11 Jan  1 00:06 compatible
-r--r--r--    1 root     root            4 Jan  1 00:06 device_type
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 ethernet@24000
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 ethernet@25000
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 i2c@3000
dr-xr-xr-x    3 root     root            0 Jan  1 00:06 i2c@3100
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 interrupt-controller@700
dr-xr-xr-x    3 root     root            0 Jan  1 00:06 mdio@24520
-r--r--r--    1 root     root            5 Jan  1 00:06 name
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 power@b00
-r--r--r--    1 root     root           12 Jan  1 00:06 ranges
-r--r--r--    1 root     root            8 Jan  1 00:06 reg
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 serial@4500
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 serial@4600
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 spi@7000
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 timer@500
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 usb@23000
dr-xr-xr-x    2 root     root            0 Jan  1 00:06 wdt@200

3.2    输出
Blob (dtb)，主要用于从DTS获得设备树映像；
source (dts), 当输入参数为Blob (dtb)，可以“反汇编”出设备树源文件；
assembler source (asm)，其最终可以编译成.O文件，可以链接到bootloader中或者frrmware image。

3.3    命令格式
dtc [-I <input -format > ] [-O <output -format > ] [-o output-filename] [-V output_version] input_filename

dtc -I dts -O dtb -S 0x3000 -o obj_name.dtb source_name.dts
-S 指定的是生成的dtb文件的大小，需要适当地扩大以供u-boot 创建/choose节点时使用

4    设备树的传递途径
对于Open Firmware (OF)系统，prom_init.c中的代码负责解析设备树，并将其转化为blob印象。

对于无Open Firmware (OF)的系统，内核可以直接从入口启动，并接受外部传递的flattened device tree参数。对于嵌入式系统，此参数由bootloader提供，或者封装过的zImage映像提供。

4.1    U-boot对FDT的支持
U-boot为了支持FDT，专门添加了新的代码，如下：
/Libfdt目录
fdt.h
libfdt.h
fdt_support.h
fdt_support.c

4.2    如何配置FDT
通常在板子配置头文件定义相关宏，以支持FDT
/* Pass open firmware flat tree */
#define CONFIG_OF_LIBFDT    1
#define CONFIG_OF_BOARD_SETUP    1
#define CONFIG_OF_STDOUT_VIA_ALIAS 1

CONFIG_OF_BOARD_SETUP宏表示会对设备树中的部分参数进行调整，主要是timebase-frequency，bus-frequency，clock-frequency等参数，在设备树配置文件中，这些参数可能为0，即采用U-boot中的参数。

4.3    如何传递设备树
Lib_ppc/board.c中do_bootm_linux负责启动Linux内核。
#if defined(CONFIG_OF_LIBFDT)
#include <fdt.h>
#include <libfdt.h>
#include <fdt_support.h>

static void fdt_error (const char *msg);
static int boot_get_fdt (cmd_tbl_t *cmdtp, int flag, int argc, char *argv[],
        bootm_headers_t *images, char **of_flat_tree, ulong *of_size);
static int boot_relocate_fdt (struct lmb *lmb, ulong bootmap_base,
        cmd_tbl_t *cmdtp, int flag, int argc, char *argv[],
        char **of_flat_tree, ulong *of_size);
#endif

相关流程如下：


在该流程中，主要从启动参数中找出设备树，然后在设备树中添加chosen 节点， 并将initrd 的地址地址，结束地址，bootargs,cmd_line 等参数保存到chosen 节点中，最后根据是否支持扁平设备树选择不同的内核启动方式。

#if defined(CONFIG_OF_LIBFDT)
    if (of_flat_tree) {    /* device tree; boot new style */
        /*
         * Linux Kernel Parameters (passing device tree):
         *   r3: pointer to the fdt, followed by the board info data
         *   r4: physical pointer to the kernel itself
         *   r5: NULL
         *   r6: NULL
         *   r7: NULL
         */
        debug ("   Booting using OF flat tree.../n");
        (*kernel) ((bd_t *)of_flat_tree, (ulong)kernel, 0, 0, 0);
        /* does not return */
    } else
#endif
    {
        /*
         * Linux Kernel Parameters (passing board info data):
         *   r3: ptr to board info data
         *   r4: initrd_start or 0 if no initrd
         *   r5: initrd_end - unused if r4 is 0
         *   r6: Start of command line string
         *   r7: End   of command line string
         */
        debug ("   Booting using board info.../n");
        (*kernel) (kbd, initrd_start, initrd_end, cmd_start, cmd_end);
        /* does not return */
    }

如果未定义CONFIG_OF_LIBFDT或者当前bootm命令没有FDT参数时则采用传统的方式启动内核。

启动命令格式如下：
Bootm kernel_addr  ramdisk_addr/-  fdt_addr
当不采用ramdisk时，第二个参数为“-”

5    内核如何解析设备树
1）首先将从u-boot 传递过来的映像基地址和dtb 文件映像基地址保存通用寄存器r30,r31；
2）通过调用machine_init()、early_init_devtree()函数来获取内核前期初始化所需的bootargs,cmd_line等系统引导参数；

3）调用start_kernel()、setup_arch()、unflatten_device_tree（）函数来解析dtb 文件，构建一个由device_node 结构连接而成的单项链表，并使用全局变量allnodes 指针来保存这个链表的头指针；
4）内核调用OF 提供的API 函数获取allnodes链表信息来初始化内核其他子系统、设备等。


Head_32.S
/*
* This is where the main kernel code starts.
*/
start_here:
。。。。
/*
* Do early platform-specific initialization,
* and set up the MMU.
*/
    mr    r3,r31
    mr    r4,r30
    bl    machine_init

/*
* Find out what kind of machine we're on and save any data we need
* from the early boot process (devtree is copied on pmac by prom_init()).
* This is called very early on the boot process, after a minimal
* MMU environment has been set up but before MMU_init is called.
*/
void __init machine_init(unsigned long dt_ptr, unsigned long phys)
{
    /* If btext is enabled, we might have a BAT setup for early display,
     * thus we do enable some very basic udbg output
     */
#ifdef CONFIG_BOOTX_TEXT
    udbg_putc = btext_drawchar;
#endif

    /* Do some early initialization based on the flat device tree */
    early_init_devtree(__va(dt_ptr));
}

/* Warning, IO base is not yet inited */
void __init setup_arch(char **cmdline_p)
{
    *cmdline_p = cmd_line;

    /* so udelay does something sensible, assume <= 1000 bogomips */
    loops_per_jiffy = 500000000 / HZ;

    unflatten_device_tree();
…..
}

/**
* unflattens the device-tree passed by the firmware, creating the
* tree of struct device_node. It also fills the "name" and "type"
* pointers of the nodes so the normal device-tree walking functions
* can be used (this used to be done by finish_device_tree)
*/
void __init unflatten_device_tree(void)
{
    unsigned long start, mem, size;
    struct device_node **allnextp = &allnodes;

    DBG(" -> unflatten_device_tree()/n");

    /* First pass, scan for size */
    start = ((unsigned long)initial_boot_params) +
        initial_boot_params->off_dt_struct;
    size = unflatten_dt_node(0, &start, NULL, NULL, 0);
    size = (size | 3) + 1;

    DBG("  size is %lx, allocating.../n", size);

    /* Allocate memory for the expanded device tree */
    mem = lmb_alloc(size + 4, __alignof__(struct device_node));
    mem = (unsigned long) __va(mem);

    ((u32 *)mem)[size / 4] = 0xdeadbeef;

    DBG("  unflattening %lx.../n", mem);

    /* Second pass, do actual unflattening */
    start = ((unsigned long)initial_boot_params) +
        initial_boot_params->off_dt_struct;
    unflatten_dt_node(mem, &start, NULL, &allnextp, 0);
    if (*((u32 *)start) != OF_DT_END)
        printk(KERN_WARNING "Weird tag at end of tree: %08x/n", *((u32 *)start));
    if (((u32 *)mem)[size / 4] != 0xdeadbeef)
        printk(KERN_WARNING "End of tree marker overwritten: %08x/n",
               ((u32 *)mem)[size / 4] );
    *allnextp = NULL;

    /* Get pointer to OF "/chosen" node for use everywhere */
    of_chosen = of_find_node_by_path("/chosen");
    if (of_chosen == NULL)
        of_chosen = of_find_node_by_path("/chosen@0");

    DBG(" <- unflatten_device_tree()/n");
}

6    设备树对驱动设计产生的影响
TBD

========================================================================

linux\Documentation\devicetree\booting-without-of.txt
           Booting the Linux/ppc kernel without Open Firmware
           --------------------------------------------------

(c) 2005 Benjamin Herrenschmidt <benh at kernel.crashing.org>,
    IBM Corp.
(c) 2005 Becky Bruce <becky.bruce at freescale.com>,
    Freescale Semiconductor, FSL SOC and 32-bit additions
(c) 2006 MontaVista Software, Inc.
    Flash chip node definition

Table of Contents
=================

  I - Introduction
    1) Entry point for arch/arm
    2) Entry point for arch/powerpc
    3) Entry point for arch/x86

  II - The DT block format
    1) Header
    2) Device tree generalities
    3) Device tree "structure" block
    4) Device tree "strings" block

  III - Required content of the device tree
    1) Note about cells and address representation
    2) Note about "compatible" properties
    3) Note about "name" properties
    4) Note about node and property names and character set
    5) Required nodes and properties
      a) The root node
      b) The /cpus node
      c) The /cpus/* nodes
      d) the /memory node(s)
      e) The /chosen node
      f) the /soc<SOCname> node

  IV - "dtc", the device tree compiler

  V - Recommendations for a bootloader

  VI - System-on-a-chip devices and nodes
    1) Defining child nodes of an SOC
    2) Representing devices without a current OF specification

  VII - Specifying interrupt information for devices
    1) interrupts property
    2) interrupt-parent property
    3) OpenPIC Interrupt Controllers
    4) ISA Interrupt Controllers

  VIII - Specifying device power management information (sleep property)

  Appendix A - Sample SOC node for MPC8540


Revision Information
====================

   May 18, 2005: Rev 0.1 - Initial draft, no chapter III yet.

   May 19, 2005: Rev 0.2 - Add chapter III and bits & pieces here or
                           clarifies the fact that a lot of things are
                           optional, the kernel only requires a very
                           small device tree, though it is encouraged
                           to provide an as complete one as possible.

   May 24, 2005: Rev 0.3 - Precise that DT block has to be in RAM
- Misc fixes
- Define version 3 and new format version 16
   for the DT block (version 16 needs kernel
   patches, will be fwd separately).
   String block now has a size, and full path
   is replaced by unit name for more
   compactness.
   linux,phandle is made optional, only nodes
   that are referenced by other nodes need it.
   "name" property is now automatically
   deduced from the unit name

   June 1, 2005: Rev 0.4 - Correct confusion between OF_DT_END and
                           OF_DT_END_NODE in structure definition.
                         - Change version 16 format to always align
                           property data to 4 bytes. Since tokens are
                           already aligned, that means no specific
                           required alignment between property size
                           and property data. The old style variable
                           alignment would make it impossible to do
                           "simple" insertion of properties using
                           memmove (thanks Milton for
                           noticing). Updated kernel patch as well
- Correct a few more alignment constraints
- Add a chapter about the device-tree
                           compiler and the textural representation of
                           the tree that can be "compiled" by dtc.

   November 21, 2005: Rev 0.5
- Additions/generalizations for 32-bit
- Changed to reflect the new arch/powerpc
   structure
- Added chapter VI


ToDo:
- Add some definitions of interrupt tree (simple/complex)
- Add some definitions for PCI host bridges
- Add some common address format examples
- Add definitions for standard properties and "compatible"
  names for cells that are not already defined by the existing
  OF spec.
- Compare FSL SOC use of PCI to standard and make sure no new
  node definition required.
- Add more information about node definitions for SOC devices
    that currently have no standard, like the FSL CPM.


I - Introduction
================

During the development of the Linux/ppc64 kernel, and more
specifically, the addition of new platform types outside of the old
IBM pSeries/iSeries pair, it was decided to enforce some strict rules
regarding the kernel entry and bootloader <-> kernel interfaces, in
order to avoid the degeneration that had become the ppc32 kernel entry
point and the way a new platform should be added to the kernel. The
legacy iSeries platform breaks those rules as it predates this scheme,
but no new board support will be accepted in the main tree that
doesn't follow them properly.  In addition, since the advent of the
arch/powerpc merged architecture for ppc32 and ppc64, new 32-bit
platforms and 32-bit platforms which move into arch/powerpc will be
required to use these rules as well.

The main requirement that will be defined in more detail below is
the presence of a device-tree whose format is defined after Open
Firmware specification. However, in order to make life easier
to embedded board vendors, the kernel doesn't require the device-tree
to represent every device in the system and only requires some nodes
and properties to be present. This will be described in detail in
section III, but, for example, the kernel does not require you to
create a node for every PCI device in the system. It is a requirement
to have a node for PCI host bridges in order to provide interrupt
routing information and memory/IO ranges, among others. It is also
recommended to define nodes for on chip devices and other buses that
don't specifically fit in an existing OF specification. This creates a
great flexibility in the way the kernel can then probe those and match
drivers to device, without having to hard code all sorts of tables. It
also makes it more flexible for board vendors to do minor hardware
upgrades without significantly impacting the kernel code or cluttering
it with special cases.


1) Entry point for arch/arm
---------------------------

   There is one single entry point to the kernel, at the start
   of the kernel image. That entry point supports two calling
   conventions.  A summary of the interface is described here.  A full
   description of the boot requirements is documented in
   Documentation/arm/Booting

        a) ATAGS interface.  Minimal information is passed from firmware
        to the kernel with a tagged list of predefined parameters.

                r0 : 0

                r1 : Machine type number

                r2 : Physical address of tagged list in system RAM

        b) Entry with a flattened device-tree block.  Firmware loads the
        physical address of the flattened device tree block (dtb) into r2,
        r1 is not used, but it is considered good practise to use a valid
        machine number as described in Documentation/arm/Booting.

                r0 : 0

                r1 : Valid machine type number.  When using a device tree,
                a single machine type number will often be assigned to
                represent a class or family of SoCs.

                r2 : physical pointer to the device-tree block
                (defined in chapter II) in RAM.  Device tree can be located
                anywhere in system RAM, but it should be aligned on a 64 bit
                boundary.

   The kernel will differentiate between ATAGS and device tree booting by
   reading the memory pointed to by r2 and looking for either the flattened
   device tree block magic value (0xd00dfeed) or the ATAG_CORE value at
   offset 0x4 from r2 (0x54410001).

2) Entry point for arch/powerpc
-------------------------------

   There is one single entry point to the kernel, at the start
   of the kernel image. That entry point supports two calling
   conventions:

        a) Boot from Open Firmware. If your firmware is compatible
        with Open Firmware (IEEE 1275) or provides an OF compatible
        client interface API (support for "interpret" callback of
        forth words isn't required), you can enter the kernel with:

              r5 : OF callback pointer as defined by IEEE 1275
              bindings to powerpc. Only the 32-bit client interface
              is currently supported

              r3, r4 : address & length of an initrd if any or 0

              The MMU is either on or off; the kernel will run the
              trampoline located in arch/powerpc/kernel/prom_init.c to
              extract the device-tree and other information from open
              firmware and build a flattened device-tree as described
              in b). prom_init() will then re-enter the kernel using
              the second method. This trampoline code runs in the
              context of the firmware, which is supposed to handle all
              exceptions during that time.

        b) Direct entry with a flattened device-tree block. This entry
        point is called by a) after the OF trampoline and can also be
        called directly by a bootloader that does not support the Open
        Firmware client interface. It is also used by "kexec" to
        implement "hot" booting of a new kernel from a previous
        running one. This method is what I will describe in more
        details in this document, as method a) is simply standard Open
        Firmware, and thus should be implemented according to the
        various standard documents defining it and its binding to the
        PowerPC platform. The entry point definition then becomes:

                r3 : physical pointer to the device-tree block
                (defined in chapter II) in RAM

                r4 : physical pointer to the kernel itself. This is
                used by the assembly code to properly disable the MMU
                in case you are entering the kernel with MMU enabled
                and a non-1:1 mapping.

                r5 : NULL (as to differentiate with method a)

        Note about SMP entry: Either your firmware puts your other
        CPUs in some sleep loop or spin loop in ROM where you can get
        them out via a soft reset or some other means, in which case
        you don't need to care, or you'll have to enter the kernel
        with all CPUs. The way to do that with method b) will be
        described in a later revision of this document.

   Board supports (platforms) are not exclusive config options. An
   arbitrary set of board supports can be built in a single kernel
   image. The kernel will "know" what set of functions to use for a
   given platform based on the content of the device-tree. Thus, you
   should:

        a) add your platform support as a _boolean_ option in
        arch/powerpc/Kconfig, following the example of PPC_PSERIES,
        PPC_PMAC and PPC_MAPLE. The later is probably a good
        example of a board support to start from.

        b) create your main platform file as
        "arch/powerpc/platforms/myplatform/myboard_setup.c" and add it
        to the Makefile under the condition of your CONFIG_
        option. This file will define a structure of type "ppc_md"
        containing the various callbacks that the generic code will
        use to get to your platform specific code

  A kernel image may support multiple platforms, but only if the
  platforms feature the same core architecture.  A single kernel build
  cannot support both configurations with Book E and configurations
  with classic Powerpc architectures.

3) Entry point for arch/x86
-------------------------------

  There is one single 32bit entry point to the kernel at code32_start,
  the decompressor (the real mode entry point goes to the same  32bit
  entry point once it switched into protected mode). That entry point
  supports one calling convention which is documented in
  Documentation/x86/boot.txt
  The physical pointer to the device-tree block (defined in chapter II)
  is passed via setup_data which requires at least boot protocol 2.09.
  The type filed is defined as

  #define SETUP_DTB                      2

  This device-tree is used as an extension to the "boot page". As such it
  does not parse / consider data which is already covered by the boot
  page. This includes memory size, reserved ranges, command line arguments
  or initrd address. It simply holds information which can not be retrieved
  otherwise like interrupt routing or a list of devices behind an I2C bus.

II - The DT block format
========================


This chapter defines the actual format of the flattened device-tree
passed to the kernel. The actual content of it and kernel requirements
are described later. You can find example of code manipulating that
format in various places, including arch/powerpc/kernel/prom_init.c
which will generate a flattened device-tree from the Open Firmware
representation, or the fs2dt utility which is part of the kexec tools
which will generate one from a filesystem representation. It is
expected that a bootloader like uboot provides a bit more support,
that will be discussed later as well.

Note: The block has to be in main memory. It has to be accessible in
both real mode and virtual mode with no mapping other than main
memory. If you are writing a simple flash bootloader, it should copy
the block to RAM before passing it to the kernel.


1) Header
---------

   The kernel is passed the physical address pointing to an area of memory
   that is roughly described in include/linux/of_fdt.h by the structure
   boot_param_header:

struct boot_param_header {
        u32     magic;                  /* magic word OF_DT_HEADER */
        u32     totalsize;              /* total size of DT block */
        u32     off_dt_struct;          /* offset to structure */
        u32     off_dt_strings;         /* offset to strings */
        u32     off_mem_rsvmap;         /* offset to memory reserve map
                                           */
        u32     version;                /* format version */
        u32     last_comp_version;      /* last compatible version */

        /* version 2 fields below */
        u32     boot_cpuid_phys;        /* Which physical CPU id we're
                                           booting on */
        /* version 3 fields below */
        u32     size_dt_strings;        /* size of the strings block */

        /* version 17 fields below */
        u32 size_dt_struct; /* size of the DT structure block */
};

   Along with the constants:

/* Definitions used by the flattened device tree */
#define OF_DT_HEADER            0xd00dfeed      /* 4: version,
   4: total size */
#define OF_DT_BEGIN_NODE        0x1             /* Start node: full name
   */
#define OF_DT_END_NODE          0x2             /* End node */
#define OF_DT_PROP              0x3             /* Property: name off,
                                                   size, content */
#define OF_DT_END               0x9

   All values in this header are in big endian format, the various
   fields in this header are defined more precisely below. All
   "offset" values are in bytes from the start of the header; that is
   from the physical base address of the device tree block.

   - magic

     This is a magic value that "marks" the beginning of the
     device-tree block header. It contains the value 0xd00dfeed and is
     defined by the constant OF_DT_HEADER

   - totalsize

     This is the total size of the DT block including the header. The
     "DT" block should enclose all data structures defined in this
     chapter (who are pointed to by offsets in this header). That is,
     the device-tree structure, strings, and the memory reserve map.

   - off_dt_struct

     This is an offset from the beginning of the header to the start
     of the "structure" part the device tree. (see 2) device tree)

   - off_dt_strings

     This is an offset from the beginning of the header to the start
     of the "strings" part of the device-tree

   - off_mem_rsvmap

     This is an offset from the beginning of the header to the start
     of the reserved memory map. This map is a list of pairs of 64-
     bit integers. Each pair is a physical address and a size. The
     list is terminated by an entry of size 0. This map provides the
     kernel with a list of physical memory areas that are "reserved"
     and thus not to be used for memory allocations, especially during
     early initialization. The kernel needs to allocate memory during
     boot for things like un-flattening the device-tree, allocating an
     MMU hash table, etc... Those allocations must be done in such a
     way to avoid overriding critical things like, on Open Firmware
     capable machines, the RTAS instance, or on some pSeries, the TCE
     tables used for the iommu. Typically, the reserve map should
     contain _at least_ this DT block itself (header,total_size). If
     you are passing an initrd to the kernel, you should reserve it as
     well. You do not need to reserve the kernel image itself. The map
     should be 64-bit aligned.

   - version

     This is the version of this structure. Version 1 stops
     here. Version 2 adds an additional field boot_cpuid_phys.
     Version 3 adds the size of the strings block, allowing the kernel
     to reallocate it easily at boot and free up the unused flattened
     structure after expansion. Version 16 introduces a new more
     "compact" format for the tree itself that is however not backward
     compatible. Version 17 adds an additional field, size_dt_struct,
     allowing it to be reallocated or moved more easily (this is
     particularly useful for bootloaders which need to make
     adjustments to a device tree based on probed information). You
     should always generate a structure of the highest version defined
     at the time of your implementation. Currently that is version 17,
     unless you explicitly aim at being backward compatible.

   - last_comp_version

     Last compatible version. This indicates down to what version of
     the DT block you are backward compatible. For example, version 2
     is backward compatible with version 1 (that is, a kernel build
     for version 1 will be able to boot with a version 2 format). You
     should put a 1 in this field if you generate a device tree of
     version 1 to 3, or 16 if you generate a tree of version 16 or 17
     using the new unit name format.

   - boot_cpuid_phys

     This field only exist on version 2 headers. It indicate which
     physical CPU ID is calling the kernel entry point. This is used,
     among others, by kexec. If you are on an SMP system, this value
     should match the content of the "reg" property of the CPU node in
     the device-tree corresponding to the CPU calling the kernel entry
     point (see further chapters for more information on the required
     device-tree contents)

   - size_dt_strings

     This field only exists on version 3 and later headers.  It
     gives the size of the "strings" section of the device tree (which
     starts at the offset given by off_dt_strings).

   - size_dt_struct

     This field only exists on version 17 and later headers.  It gives
     the size of the "structure" section of the device tree (which
     starts at the offset given by off_dt_struct).

   So the typical layout of a DT block (though the various parts don't
   need to be in that order) looks like this (addresses go from top to
   bottom):


             ------------------------------
     base -> |  struct boot_param_header  |
             ------------------------------
             |      (alignment gap) (*)   |
             ------------------------------
             |      memory reserve map    |
             ------------------------------
             |      (alignment gap)       |
             ------------------------------
             |                            |
             |    device-tree structure   |
             |                            |
             ------------------------------
             |      (alignment gap)       |
             ------------------------------
             |                            |
             |     device-tree strings    |
             |                            |
      -----> ------------------------------
      |
      |
      --- (base + totalsize)

  (*) The alignment gaps are not necessarily present; their presence
      and size are dependent on the various alignment requirements of
      the individual data blocks.


2) Device tree generalities
---------------------------

This device-tree itself is separated in two different blocks, a
structure block and a strings block. Both need to be aligned to a 4
byte boundary.

First, let's quickly describe the device-tree concept before detailing
the storage format. This chapter does _not_ describe the detail of the
required types of nodes & properties for the kernel, this is done
later in chapter III.

The device-tree layout is strongly inherited from the definition of
the Open Firmware IEEE 1275 device-tree. It's basically a tree of
nodes, each node having two or more named properties. A property can
have a value or not.

It is a tree, so each node has one and only one parent except for the
root node who has no parent.

A node has 2 names. The actual node name is generally contained in a
property of type "name" in the node property list whose value is a
zero terminated string and is mandatory for version 1 to 3 of the
format definition (as it is in Open Firmware). Version 16 makes it
optional as it can generate it from the unit name defined below.

There is also a "unit name" that is used to differentiate nodes with
the same name at the same level, it is usually made of the node
names, the "@" sign, and a "unit address", which definition is
specific to the bus type the node sits on.

The unit name doesn't exist as a property per-se but is included in
the device-tree structure. It is typically used to represent "path" in
the device-tree. More details about the actual format of these will be
below.

The kernel generic code does not make any formal use of the
unit address (though some board support code may do) so the only real
requirement here for the unit address is to ensure uniqueness of
the node unit name at a given level of the tree. Nodes with no notion
of address and no possible sibling of the same name (like /memory or
/cpus) may omit the unit address in the context of this specification,
or use the "@0" default unit address. The unit name is used to define
a node "full path", which is the concatenation of all parent node
unit names separated with "/".

The root node doesn't have a defined name, and isn't required to have
a name property either if you are using version 3 or earlier of the
format. It also has no unit address (no @ symbol followed by a unit
address). The root node unit name is thus an empty string. The full
path to the root node is "/".

Every node which actually represents an actual device (that is, a node
which isn't only a virtual "container" for more nodes, like "/cpus"
is) is also required to have a "compatible" property indicating the
specific hardware and an optional list of devices it is fully
backwards compatible with.

Finally, every node that can be referenced from a property in another
node is required to have either a "phandle" or a "linux,phandle"
property. Real Open Firmware implementations provide a unique
"phandle" value for every node that the "prom_init()" trampoline code
turns into "linux,phandle" properties. However, this is made optional
if the flattened device tree is used directly. An example of a node
referencing another node via "phandle" is when laying out the
interrupt tree which will be described in a further version of this
document.

The "phandle" property is a 32-bit value that uniquely
identifies a node. You are free to use whatever values or system of
values, internal pointers, or whatever to generate these, the only
requirement is that every node for which you provide that property has
a unique value for it.

Here is an example of a simple device-tree. In this example, an "o"
designates a node followed by the node unit name. Properties are
presented with their name followed by their content. "content"
represents an ASCII string (zero terminated) value, while <content>
represents a 32-bit hexadecimal value. The various nodes in this
example will be discussed in a later chapter. At this point, it is
only meant to give you a idea of what a device-tree looks like. I have
purposefully kept the "name" and "linux,phandle" properties which
aren't necessary in order to give you a better idea of what the tree
looks like in practice.

  / o device-tree
      |- name = "device-tree"
      |- model = "MyBoardName"
      |- compatible = "MyBoardFamilyName"
      |- #address-cells = <2>
      |- #size-cells = <2>
      |- linux,phandle = <0>
      |
      o cpus
      | | - name = "cpus"
      | | - linux,phandle = <1>
      | | - #address-cells = <1>
      | | - #size-cells = <0>
      | |
      | o PowerPC,970@0
      |   |- name = "PowerPC,970"
      |   |- device_type = "cpu"
      |   |- reg = <0>
      |   |- clock-frequency = <5f5e1000>
      |   |- 64-bit
      |   |- linux,phandle = <2>
      |
      o memory@0
      | |- name = "memory"
      | |- device_type = "memory"
      | |- reg = <00000000 00000000 00000000 20000000>
      | |- linux,phandle = <3>
      |
      o chosen
        |- name = "chosen"
        |- bootargs = "root=/dev/sda2"
        |- linux,phandle = <4>

This tree is almost a minimal tree. It pretty much contains the
minimal set of required nodes and properties to boot a linux kernel;
that is, some basic model information at the root, the CPUs, and the
physical memory layout.  It also includes misc information passed
through /chosen, like in this example, the platform type (mandatory)
and the kernel command line arguments (optional).

The /cpus/PowerPC,970@0/64-bit property is an example of a
property without a value. All other properties have a value. The
significance of the #address-cells and #size-cells properties will be
explained in chapter IV which defines precisely the required nodes and
properties and their content.


3) Device tree "structure" block

The structure of the device tree is a linearized tree structure. The
"OF_DT_BEGIN_NODE" token starts a new node, and the "OF_DT_END_NODE"
ends that node definition. Child nodes are simply defined before
"OF_DT_END_NODE" (that is nodes within the node). A 'token' is a 32
bit value. The tree has to be "finished" with a OF_DT_END token

Here's the basic structure of a single node:

     * token OF_DT_BEGIN_NODE (that is 0x00000001)
     * for version 1 to 3, this is the node full path as a zero
       terminated string, starting with "/". For version 16 and later,
       this is the node unit name only (or an empty string for the
       root node)
     * [align gap to next 4 bytes boundary]
     * for each property:
        * token OF_DT_PROP (that is 0x00000003)
        * 32-bit value of property value size in bytes (or 0 if no
          value)
        * 32-bit value of offset in string block of property name
        * property value data if any
        * [align gap to next 4 bytes boundary]
     * [child nodes if any]
     * token OF_DT_END_NODE (that is 0x00000002)

So the node content can be summarized as a start token, a full path,
a list of properties, a list of child nodes, and an end token. Every
child node is a full node structure itself as defined above.

NOTE: The above definition requires that all property definitions for
a particular node MUST precede any subnode definitions for that node.
Although the structure would not be ambiguous if properties and
subnodes were intermingled, the kernel parser requires that the
properties come first (up until at least 2.6.22).  Any tools
manipulating a flattened tree must take care to preserve this
constraint.

4) Device tree "strings" block

In order to save space, property names, which are generally redundant,
are stored separately in the "strings" block. This block is simply the
whole bunch of zero terminated strings for all property names
concatenated together. The device-tree property definitions in the
structure block will contain offset values from the beginning of the
strings block.


III - Required content of the device tree
=========================================

WARNING: All "linux,*" properties defined in this document apply only
to a flattened device-tree. If your platform uses a real
implementation of Open Firmware or an implementation compatible with
the Open Firmware client interface, those properties will be created
by the trampoline code in the kernel's prom_init() file. For example,
that's where you'll have to add code to detect your board model and
set the platform number. However, when using the flattened device-tree
entry point, there is no prom_init() pass, and thus you have to
provide those properties yourself.


1) Note about cells and address representation
----------------------------------------------

The general rule is documented in the various Open Firmware
documentations. If you choose to describe a bus with the device-tree
and there exist an OF bus binding, then you should follow the
specification. However, the kernel does not require every single
device or bus to be described by the device tree.

In general, the format of an address for a device is defined by the
parent bus type, based on the #address-cells and #size-cells
properties.  Note that the parent's parent definitions of #address-cells
and #size-cells are not inherited so every node with children must specify
them.  The kernel requires the root node to have those properties defining
addresses format for devices directly mapped on the processor bus.

Those 2 properties define 'cells' for representing an address and a
size. A "cell" is a 32-bit number. For example, if both contain 2
like the example tree given above, then an address and a size are both
composed of 2 cells, and each is a 64-bit number (cells are
concatenated and expected to be in big endian format). Another example
is the way Apple firmware defines them, with 2 cells for an address
and one cell for a size.  Most 32-bit implementations should define
#address-cells and #size-cells to 1, which represents a 32-bit value.
Some 32-bit processors allow for physical addresses greater than 32
bits; these processors should define #address-cells as 2.

"reg" properties are always a tuple of the type "address size" where
the number of cells of address and size is specified by the bus
#address-cells and #size-cells. When a bus supports various address
spaces and other flags relative to a given address allocation (like
prefetchable, etc...) those flags are usually added to the top level
bits of the physical address. For example, a PCI physical address is
made of 3 cells, the bottom two containing the actual address itself
while the top cell contains address space indication, flags, and pci
bus & device numbers.

For buses that support dynamic allocation, it's the accepted practice
to then not provide the address in "reg" (keep it 0) though while
providing a flag indicating the address is dynamically allocated, and
then, to provide a separate "assigned-addresses" property that
contains the fully allocated addresses. See the PCI OF bindings for
details.

In general, a simple bus with no address space bits and no dynamic
allocation is preferred if it reflects your hardware, as the existing
kernel address parsing functions will work out of the box. If you
define a bus type with a more complex address format, including things
like address space bits, you'll have to add a bus translator to the
prom_parse.c file of the recent kernels for your bus type.

The "reg" property only defines addresses and sizes (if #size-cells is
non-0) within a given bus. In order to translate addresses upward
(that is into parent bus addresses, and possibly into CPU physical
addresses), all buses must contain a "ranges" property. If the
"ranges" property is missing at a given level, it's assumed that
translation isn't possible, i.e., the registers are not visible on the
parent bus.  The format of the "ranges" property for a bus is a list
of:

bus address, parent bus address, size

"bus address" is in the format of the bus this bus node is defining,
that is, for a PCI bridge, it would be a PCI address. Thus, (bus
address, size) defines a range of addresses for child devices. "parent
bus address" is in the format of the parent bus of this bus. For
example, for a PCI host controller, that would be a CPU address. For a
PCI<->ISA bridge, that would be a PCI address. It defines the base
address in the parent bus where the beginning of that range is mapped.

For new 64-bit board support, I recommend either the 2/2 format or
Apple's 2/1 format which is slightly more compact since sizes usually
fit in a single 32-bit word.   New 32-bit board support should use a
1/1 format, unless the processor supports physical addresses greater
than 32-bits, in which case a 2/1 format is recommended.

Alternatively, the "ranges" property may be empty, indicating that the
registers are visible on the parent bus using an identity mapping
translation.  In other words, the parent bus address space is the same
as the child bus address space.

2) Note about "compatible" properties
-------------------------------------

These properties are optional, but recommended in devices and the root
node. The format of a "compatible" property is a list of concatenated
zero terminated strings. They allow a device to express its
compatibility with a family of similar devices, in some cases,
allowing a single driver to match against several devices regardless
of their actual names.

3) Note about "name" properties
-------------------------------

While earlier users of Open Firmware like OldWorld macintoshes tended
to use the actual device name for the "name" property, it's nowadays
considered a good practice to use a name that is closer to the device
class (often equal to device_type). For example, nowadays, Ethernet
controllers are named "ethernet", an additional "model" property
defining precisely the chip type/model, and "compatible" property
defining the family in case a single driver can driver more than one
of these chips. However, the kernel doesn't generally put any
restriction on the "name" property; it is simply considered good
practice to follow the standard and its evolutions as closely as
possible.

Note also that the new format version 16 makes the "name" property
optional. If it's absent for a node, then the node's unit name is then
used to reconstruct the name. That is, the part of the unit name
before the "@" sign is used (or the entire unit name if no "@" sign
is present).

4) Note about node and property names and character set
-------------------------------------------------------

While Open Firmware provides more flexible usage of 8859-1, this
specification enforces more strict rules. Nodes and properties should
be comprised only of ASCII characters 'a' to 'z', '0' to
'9', ',', '.', '_', '+', '#', '?', and '-'. Node names additionally
allow uppercase characters 'A' to 'Z' (property names should be
lowercase. The fact that vendors like Apple don't respect this rule is
irrelevant here). Additionally, node and property names should always
begin with a character in the range 'a' to 'z' (or 'A' to 'Z' for node
names).

The maximum number of characters for both nodes and property names
is 31. In the case of node names, this is only the leftmost part of
a unit name (the pure "name" property), it doesn't include the unit
address which can extend beyond that limit.


5) Required nodes and properties
--------------------------------
  These are all that are currently required. However, it is strongly
  recommended that you expose PCI host bridges as documented in the
  PCI binding to Open Firmware, and your interrupt tree as documented
  in OF interrupt tree specification.

  a) The root node

  The root node requires some properties to be present:

    - model : this is your board name/model
    - #address-cells : address representation for "root" devices
    - #size-cells: the size representation for "root" devices
    - compatible : the board "family" generally finds its way here,
      for example, if you have 2 board models with a similar layout,
      that typically get driven by the same platform code in the
      kernel, you would specify the exact board model in the
      compatible property followed by an entry that represents the SoC
      model.

  The root node is also generally where you add additional properties
  specific to your board like the serial number if any, that sort of
  thing. It is recommended that if you add any "custom" property whose
  name may clash with standard defined ones, you prefix them with your
  vendor name and a comma.

  b) The /cpus node

  This node is the parent of all individual CPU nodes. It doesn't
  have any specific requirements, though it's generally good practice
  to have at least:

               #address-cells = <00000001>
               #size-cells    = <00000000>

  This defines that the "address" for a CPU is a single cell, and has
  no meaningful size. This is not necessary but the kernel will assume
  that format when reading the "reg" properties of a CPU node, see
  below

  c) The /cpus/* nodes

  So under /cpus, you are supposed to create a node for every CPU on
  the machine. There is no specific restriction on the name of the
  CPU, though it's common to call it <architecture>,<core>. For
  example, Apple uses PowerPC,G5 while IBM uses PowerPC,970FX.
  However, the Generic Names convention suggests that it would be
  better to simply use 'cpu' for each cpu node and use the compatible
  property to identify the specific cpu core.

  Required properties:

    - device_type : has to be "cpu"
    - reg : This is the physical CPU number, it's a single 32-bit cell
      and is also used as-is as the unit number for constructing the
      unit name in the full path. For example, with 2 CPUs, you would
      have the full path:
        /cpus/PowerPC,970FX@0
        /cpus/PowerPC,970FX@1
      (unit addresses do not require leading zeroes)
    - d-cache-block-size : one cell, L1 data cache block size in bytes (*)
    - i-cache-block-size : one cell, L1 instruction cache block size in
      bytes
    - d-cache-size : one cell, size of L1 data cache in bytes
    - i-cache-size : one cell, size of L1 instruction cache in bytes

(*) The cache "block" size is the size on which the cache management
instructions operate. Historically, this document used the cache
"line" size here which is incorrect. The kernel will prefer the cache
block size and will fallback to cache line size for backward
compatibility.

  Recommended properties:

    - timebase-frequency : a cell indicating the frequency of the
      timebase in Hz. This is not directly used by the generic code,
      but you are welcome to copy/paste the pSeries code for setting
      the kernel timebase/decrementer calibration based on this
      value.
    - clock-frequency : a cell indicating the CPU core clock frequency
      in Hz. A new property will be defined for 64-bit values, but if
      your frequency is < 4Ghz, one cell is enough. Here as well as
      for the above, the common code doesn't use that property, but
      you are welcome to re-use the pSeries or Maple one. A future
      kernel version might provide a common function for this.
    - d-cache-line-size : one cell, L1 data cache line size in bytes
      if different from the block size
    - i-cache-line-size : one cell, L1 instruction cache line size in
      bytes if different from the block size

  You are welcome to add any property you find relevant to your board,
  like some information about the mechanism used to soft-reset the
  CPUs. For example, Apple puts the GPIO number for CPU soft reset
  lines in there as a "soft-reset" property since they start secondary
  CPUs by soft-resetting them.


  d) the /memory node(s)

  To define the physical memory layout of your board, you should
  create one or more memory node(s). You can either create a single
  node with all memory ranges in its reg property, or you can create
  several nodes, as you wish. The unit address (@ part) used for the
  full path is the address of the first range of memory defined by a
  given node. If you use a single memory node, this will typically be
  @0.

  Required properties:

    - device_type : has to be "memory"
    - reg : This property contains all the physical memory ranges of
      your board. It's a list of addresses/sizes concatenated
      together, with the number of cells of each defined by the
      #address-cells and #size-cells of the root node. For example,
      with both of these properties being 2 like in the example given
      earlier, a 970 based machine with 6Gb of RAM could typically
      have a "reg" property here that looks like:

      00000000 00000000 00000000 80000000
      00000001 00000000 00000001 00000000

      That is a range starting at 0 of 0x80000000 bytes and a range
      starting at 0x100000000 and of 0x100000000 bytes. You can see
      that there is no memory covering the IO hole between 2Gb and
      4Gb. Some vendors prefer splitting those ranges into smaller
      segments, but the kernel doesn't care.

  e) The /chosen node

  This node is a bit "special". Normally, that's where Open Firmware
  puts some variable environment information, like the arguments, or
  the default input/output devices.

  This specification makes a few of these mandatory, but also defines
  some linux-specific properties that would be normally constructed by
  the prom_init() trampoline when booting with an OF client interface,
  but that you have to provide yourself when using the flattened format.

  Recommended properties:

    - bootargs : This zero-terminated string is passed as the kernel
      command line
    - linux,stdout-path : This is the full path to your standard
      console device if any. Typically, if you have serial devices on
      your board, you may want to put the full path to the one set as
      the default console in the firmware here, for the kernel to pick
      it up as its own default console.

  Note that u-boot creates and fills in the chosen node for platforms
  that use it.

  (Note: a practice that is now obsolete was to include a property
  under /chosen called interrupt-controller which had a phandle value
  that pointed to the main interrupt controller)

  f) the /soc<SOCname> node

  This node is used to represent a system-on-a-chip (SoC) and must be
  present if the processor is a SoC. The top-level soc node contains
  information that is global to all devices on the SoC. The node name
  should contain a unit address for the SoC, which is the base address
  of the memory-mapped register set for the SoC. The name of an SoC
  node should start with "soc", and the remainder of the name should
  represent the part number for the soc.  For example, the MPC8540's
  soc node would be called "soc8540".

  Required properties:

    - ranges : Should be defined as specified in 1) to describe the
      translation of SoC addresses for memory mapped SoC registers.
    - bus-frequency: Contains the bus frequency for the SoC node.
      Typically, the value of this field is filled in by the boot
      loader.
    - compatible : Exact model of the SoC


  Recommended properties:

    - reg : This property defines the address and size of the
      memory-mapped registers that are used for the SOC node itself.
      It does not include the child device registers - these will be
      defined inside each child node.  The address specified in the
      "reg" property should match the unit address of the SOC node.
    - #address-cells : Address representation for "soc" devices.  The
      format of this field may vary depending on whether or not the
      device registers are memory mapped.  For memory mapped
      registers, this field represents the number of cells needed to
      represent the address of the registers.  For SOCs that do not
      use MMIO, a special address format should be defined that
      contains enough cells to represent the required information.
      See 1) above for more details on defining #address-cells.
    - #size-cells : Size representation for "soc" devices
    - #interrupt-cells : Defines the width of cells used to represent
       interrupts.  Typically this value is <2>, which includes a
       32-bit number that represents the interrupt number, and a
       32-bit number that represents the interrupt sense and level.
       This field is only needed if the SOC contains an interrupt
       controller.

  The SOC node may contain child nodes for each SOC device that the
  platform uses.  Nodes should not be created for devices which exist
  on the SOC but are not used by a particular platform. See chapter VI
  for more information on how to specify devices that are part of a SOC.

  Example SOC node for the MPC8540:

soc8540@e0000000 {
#address-cells = <1>;
#size-cells = <1>;
#interrupt-cells = <2>;
device_type = "soc";
ranges = <00000000 e0000000 00100000>
reg = <e0000000 00003000>;
bus-frequency = <0>;
}



IV - "dtc", the device tree compiler
====================================


dtc source code can be found at
<http://git.jdl.com/gitweb/?p=dtc.git>

WARNING: This version is still in early development stage; the
resulting device-tree "blobs" have not yet been validated with the
kernel. The current generated block lacks a useful reserve map (it will
be fixed to generate an empty one, it's up to the bootloader to fill
it up) among others. The error handling needs work, bugs are lurking,
etc...

dtc basically takes a device-tree in a given format and outputs a
device-tree in another format. The currently supported formats are:

  Input formats:
  -------------

     - "dtb": "blob" format, that is a flattened device-tree block
       with
        header all in a binary blob.
     - "dts": "source" format. This is a text file containing a
       "source" for a device-tree. The format is defined later in this
        chapter.
     - "fs" format. This is a representation equivalent to the
        output of /proc/device-tree, that is nodes are directories and
properties are files

Output formats:
---------------

     - "dtb": "blob" format
     - "dts": "source" format
     - "asm": assembly language file. This is a file that can be
       sourced by gas to generate a device-tree "blob". That file can
       then simply be added to your Makefile. Additionally, the
       assembly file exports some symbols that can be used.


The syntax of the dtc tool is

    dtc [-I <input-format>] [-O <output-format>]
        [-o output-filename] [-V output_version] input_filename


The "output_version" defines what version of the "blob" format will be
generated. Supported versions are 1,2,3 and 16. The default is
currently version 3 but that may change in the future to version 16.

Additionally, dtc performs various sanity checks on the tree, like the
uniqueness of linux, phandle properties, validity of strings, etc...

The format of the .dts "source" file is "C" like, supports C and C++
style comments.

/ {
}

The above is the "device-tree" definition. It's the only statement
supported currently at the toplevel.

/ {
  property1 = "string_value"; /* define a property containing a 0
                                 * terminated string
*/

  property2 = <1234abcd>; /* define a property containing a
                                 * numerical 32-bit value (hexadecimal)
*/

  property3 = <12345678 12345678 deadbeef>;
                                /* define a property containing 3
                                 * numerical 32-bit values (cells) in
                                 * hexadecimal
*/
  property4 = [0a 0b 0c 0d de ea ad be ef];
                                /* define a property whose content is
                                 * an arbitrary array of bytes
                                 */

  childnode@address { /* define a child node named "childnode"
                                 * whose unit name is "childnode at
* address"
                                 */

    childprop = "hello\n";      /* define a property "childprop" of
                                 * childnode (in this case, a string)
                                 */
  };
};

Nodes can contain other nodes etc... thus defining the hierarchical
structure of the tree.

Strings support common escape sequences from C: "\n", "\t", "\r",
"\(octal value)", "\x(hex value)".

It is also suggested that you pipe your source file through cpp (gcc
preprocessor) so you can use #include's, #define for constants, etc...

Finally, various options are planned but not yet implemented, like
automatic generation of phandles, labels (exported to the asm file so
you can point to a property content and change it easily from whatever
you link the device-tree with), label or path instead of numeric value
in some cells to "point" to a node (replaced by a phandle at compile
time), export of reserve map address to the asm file, ability to
specify reserve map content at compile time, etc...

We may provide a .h include file with common definitions of that
proves useful for some properties (like building PCI properties or
interrupt maps) though it may be better to add a notion of struct
definitions to the compiler...


V - Recommendations for a bootloader
====================================


Here are some various ideas/recommendations that have been proposed
while all this has been defined and implemented.

  - The bootloader may want to be able to use the device-tree itself
    and may want to manipulate it (to add/edit some properties,
    like physical memory size or kernel arguments). At this point, 2
    choices can be made. Either the bootloader works directly on the
    flattened format, or the bootloader has its own internal tree
    representation with pointers (similar to the kernel one) and
    re-flattens the tree when booting the kernel. The former is a bit
    more difficult to edit/modify, the later requires probably a bit
    more code to handle the tree structure. Note that the structure
    format has been designed so it's relatively easy to "insert"
    properties or nodes or delete them by just memmoving things
    around. It contains no internal offsets or pointers for this
    purpose.

  - An example of code for iterating nodes & retrieving properties
    directly from the flattened tree format can be found in the kernel
    file drivers/of/fdt.c.  Look at the of_scan_flat_dt() function,
    its usage in early_init_devtree(), and the corresponding various
    early_init_dt_scan_*() callbacks. That code can be re-used in a
    GPL bootloader, and as the author of that code, I would be happy
    to discuss possible free licensing to any vendor who wishes to
    integrate all or part of this code into a non-GPL bootloader.
    (reference needed; who is 'I' here? ---gcl Jan 31, 2011)



VI - System-on-a-chip devices and nodes
=======================================

Many companies are now starting to develop system-on-a-chip
processors, where the processor core (CPU) and many peripheral devices
exist on a single piece of silicon.  For these SOCs, an SOC node
should be used that defines child nodes for the devices that make
up the SOC. While platforms are not required to use this model in
order to boot the kernel, it is highly encouraged that all SOC
implementations define as complete a flat-device-tree as possible to
describe the devices on the SOC.  This will allow for the
genericization of much of the kernel code.


1) Defining child nodes of an SOC
---------------------------------

Each device that is part of an SOC may have its own node entry inside
the SOC node.  For each device that is included in the SOC, the unit
address property represents the address offset for this device's
memory-mapped registers in the parent's address space.  The parent's
address space is defined by the "ranges" property in the top-level soc
node. The "reg" property for each node that exists directly under the
SOC node should contain the address mapping from the child address space
to the parent SOC address space and the size of the device's
memory-mapped register file.

For many devices that may exist inside an SOC, there are predefined
specifications for the format of the device tree node.  All SOC child
nodes should follow these specifications, except where noted in this
document.

See appendix A for an example partial SOC node definition for the
MPC8540.


2) Representing devices without a current OF specification
----------------------------------------------------------

Currently, there are many devices on SoCs that do not have a standard
representation defined as part of the Open Firmware specifications,
mainly because the boards that contain these SoCs are not currently
booted using Open Firmware.  Binding documentation for new devices
should be added to the Documentation/devicetree/bindings directory.
That directory will expand as device tree support is added to more and
more SoCs.


VII - Specifying interrupt information for devices
===================================================

The device tree represents the buses and devices of a hardware
system in a form similar to the physical bus topology of the
hardware.

In addition, a logical 'interrupt tree' exists which represents the
hierarchy and routing of interrupts in the hardware.

The interrupt tree model is fully described in the
document "Open Firmware Recommended Practice: Interrupt
Mapping Version 0.9".  The document is available at:
<http://playground.sun.com/1275/practice>.

1) interrupts property
----------------------

Devices that generate interrupts to a single interrupt controller
should use the conventional OF representation described in the
OF interrupt mapping documentation.

Each device which generates interrupts must have an 'interrupt'
property.  The interrupt property value is an arbitrary number of
of 'interrupt specifier' values which describe the interrupt or
interrupts for the device.

The encoding of an interrupt specifier is determined by the
interrupt domain in which the device is located in the
interrupt tree.  The root of an interrupt domain specifies in
its #interrupt-cells property the number of 32-bit cells
required to encode an interrupt specifier.  See the OF interrupt
mapping documentation for a detailed description of domains.

For example, the binding for the OpenPIC interrupt controller
specifies  an #interrupt-cells value of 2 to encode the interrupt
number and level/sense information. All interrupt children in an
OpenPIC interrupt domain use 2 cells per interrupt in their interrupts
property.

The PCI bus binding specifies a #interrupt-cell value of 1 to encode
which interrupt pin (INTA,INTB,INTC,INTD) is used.

2) interrupt-parent property
----------------------------

The interrupt-parent property is specified to define an explicit
link between a device node and its interrupt parent in
the interrupt tree.  The value of interrupt-parent is the
phandle of the parent node.

If the interrupt-parent property is not defined for a node, its
interrupt parent is assumed to be an ancestor in the node's
_device tree_ hierarchy.

3) OpenPIC Interrupt Controllers
--------------------------------

OpenPIC interrupt controllers require 2 cells to encode
interrupt information.  The first cell defines the interrupt
number.  The second cell defines the sense and level
information.

Sense and level information should be encoded as follows:

0 = low to high edge sensitive type enabled
1 = active low level sensitive type enabled
2 = active high level sensitive type enabled
3 = high to low edge sensitive type enabled

4) ISA Interrupt Controllers
----------------------------

ISA PIC interrupt controllers require 2 cells to encode
interrupt information.  The first cell defines the interrupt
number.  The second cell defines the sense and level
information.

ISA PIC interrupt controllers should adhere to the ISA PIC
encodings listed below:

0 =  active low level sensitive type enabled
1 =  active high level sensitive type enabled
2 =  high to low edge sensitive type enabled
3 =  low to high edge sensitive type enabled

VIII - Specifying Device Power Management Information (sleep property)
===================================================================

Devices on SOCs often have mechanisms for placing devices into low-power
states that are decoupled from the devices' own register blocks.  Sometimes,
this information is more complicated than a cell-index property can
reasonably describe.  Thus, each device controlled in such a manner
may contain a "sleep" property which describes these connections.

The sleep property consists of one or more sleep resources, each of
which consists of a phandle to a sleep controller, followed by a
controller-specific sleep specifier of zero or more cells.

The semantics of what type of low power modes are possible are defined
by the sleep controller.  Some examples of the types of low power modes
that may be supported are:

- Dynamic: The device may be disabled or enabled at any time.
- System Suspend: The device may request to be disabled or remain
   awake during system suspend, but will not be disabled until then.
- Permanent: The device is disabled permanently (until the next hard
   reset).

Some devices may share a clock domain with each other, such that they should
only be suspended when none of the devices are in use.  Where reasonable,
such nodes should be placed on a virtual bus, where the bus has the sleep
property.  If the clock domain is shared among devices that cannot be
reasonably grouped in this manner, then create a virtual sleep controller
(similar to an interrupt nexus, except that defining a standardized
sleep-map should wait until its necessity is demonstrated).

Appendix A - Sample SOC node for MPC8540
========================================

soc@e0000000 {
#address-cells = <1>;
#size-cells = <1>;
compatible = "fsl,mpc8540-ccsr", "simple-bus";
device_type = "soc";
ranges = <0x00000000 0xe0000000 0x00100000>
bus-frequency = <0>;
interrupt-parent = <&pic>;

ethernet@24000 {
#address-cells = <1>;
#size-cells = <1>;
device_type = "network";
model = "TSEC";
compatible = "gianfar", "simple-bus";
reg = <0x24000 0x1000>;
local-mac-address = [ 00 E0 0C 00 73 00 ];
interrupts = <29 2 30 2 34 2>;
phy-handle = <&phy0>;
sleep = <&pmc 00000080>;
ranges;

mdio@24520 {
reg = <0x24520 0x20>;
compatible = "fsl,gianfar-mdio";

phy0: ethernet-phy@0 {
interrupts = <5 1>;
reg = <0>;
device_type = "ethernet-phy";
};

phy1: ethernet-phy@1 {
interrupts = <5 1>;
reg = <1>;
device_type = "ethernet-phy";
};

phy3: ethernet-phy@3 {
interrupts = <7 1>;
reg = <3>;
device_type = "ethernet-phy";
};
};
};

ethernet@25000 {
device_type = "network";
model = "TSEC";
compatible = "gianfar";
reg = <0x25000 0x1000>;
local-mac-address = [ 00 E0 0C 00 73 01 ];
interrupts = <13 2 14 2 18 2>;
phy-handle = <&phy1>;
sleep = <&pmc 00000040>;
};

ethernet@26000 {
device_type = "network";
model = "FEC";
compatible = "gianfar";
reg = <0x26000 0x1000>;
local-mac-address = [ 00 E0 0C 00 73 02 ];
interrupts = <41 2>;
phy-handle = <&phy3>;
sleep = <&pmc 00000020>;
};

serial@4500 {
#address-cells = <1>;
#size-cells = <1>;
compatible = "fsl,mpc8540-duart", "simple-bus";
sleep = <&pmc 00000002>;
ranges;

serial@4500 {
device_type = "serial";
compatible = "ns16550";
reg = <0x4500 0x100>;
clock-frequency = <0>;
interrupts = <42 2>;
};

serial@4600 {
device_type = "serial";
compatible = "ns16550";
reg = <0x4600 0x100>;
clock-frequency = <0>;
interrupts = <42 2>;
};
};

pic: pic@40000 {
interrupt-controller;
#address-cells = <0>;
#interrupt-cells = <2>;
reg = <0x40000 0x40000>;
compatible = "chrp,open-pic";
device_type = "open-pic";
};

i2c@3000 {
interrupts = <43 2>;
reg = <0x3000 0x100>;
compatible  = "fsl-i2c";
dfsrr;
sleep = <&pmc 00000004>;
};

pmc: power@e0070 {
compatible = "fsl,mpc8540-pmc", "fsl,mpc8548-pmc";
reg = <0xe0070 0x20>;
};
};