本文深入探讨了设备树(DTB)如何将参数传递给内核，内核如何解析设备树并转化为设备节点结构。在启动过程中，U-Boot将设备树信息传递给内核，内核通过初始化流程获取设备树地址，并使用`setup_machine_fdt`等函数将设备树转换为设备节点。设备树中的`/chosen`节点用于传递运行时参数，`/memreserve`用于预留内存。内核通过`of_platform_populate`遍历设备树，为每个节点创建platform_device，并根据`compatible`属性匹配总线。整个过程展示了设备树在Linux内核设备驱动模型中的关键作用。
uboot将一些参数，设备树文件传给内核，那么内核如何处理这些设备树文件呢？本章就kernel解析设备树的过程和原理，本章的主要内容以Device Tree相关的数据流分析为索引，对ARM linux kernel的代码进行解析。主要的数据流包括：
- 设备树对于内核的意义
- 从u-boot传递dtb开始，kernel初始化流程，如何将dtb并将其转换成Device Tree Structure
- 传递运行时参数传递以及platform的识别流程分析
- 如何将Device Tree Structure并入linux kernel的设备驱动模型。
1. 设备树的作用
由前面几章已经大致可以得出设备树对于内核的作用
| 作用 | 详细描述 |
|---|---|
| 平台标识 | 告诉内核dtb支持哪些平台 ； 用DT 来标识特定的machine ； root 节点的compatible 字段，匹配machine_desc的dt_compat |
| 运行时配置 | chosen节点的属性 |
| 设备信息集合 | 传递各种设备信息 |
2. 初始化流程
从上一章我们已经知道fdt的地址是作为参数传递到kernel。下面看一下kernel阶段怎么获取这个地址值的。bootloader启动内核时,会设置r0,r1,r2三个寄存器,
r0一般设置为0；
r1一般设置为machine id (在使用设备树时该参数没有被使用)；
r2一般设置ATAGS或DTB的开始地址；
对于启动的流程代码如下：
ENTRY(stext)
ARM_BE8(setend be ) @ ensure we are in BE8 mode
THUMB( adr r9, BSYM(1f) ) @ Kernel is always entered in ARM.
THUMB( bx r9 ) @ If this is a Thumb-2 kernel,
THUMB( .thumb ) @ switch to Thumb now.
THUMB(1: )
#ifdef CONFIG_ARM_VIRT_EXT
bl __hyp_stub_install
#endif
@ ensure svc mode and all interrupts masked
safe_svcmode_maskall r9
mrc p15<span class="token punctuation">,</span> <span class="token number">0</span><span class="token punctuation">,</span> r9<span class="token punctuation">,</span> c0<span class="token punctuation">,</span> c0 @ get processor id
bl __lookup_processor_type @ r5<span class="token operator">=</span>procinfo r9<span class="token operator">=</span>cpuid
movs r10<span class="token punctuation">,</span> r5 @ invalid processor <span class="token punctuation">(</span>r5<span class="token operator">=</span><span class="token number">0</span><span class="token punctuation">)</span><span class="token operator">?</span>
THUMB( it eq ) @ force fixup-able long branch encoding
beq __error_p @ yes, error ‘p’
#ifdef CONFIG_ARM_LPAE
mrc p15, 0, r3, c0, c1, 4 @ read ID_MMFR0
and r3, r3, #0xf @ extract VMSA support
cmp r3, #5 @ long-descriptor translation table format?
THUMB( it lo ) @ force fixup-able long branch encoding
blo __error_lpae @ only classic page table format
#endif
#ifndef CONFIG_XIP_KERNEL
adr r3, 2f
ldmia r3, {
r4, r8}
sub r4, r3, r4 @ (PHYS_OFFSET - PAGE_OFFSET)
add r8, r8, r4 @ PHYS_OFFSET
#else
ldr r8, =PLAT_PHYS_OFFSET @ always constant in this case
#endif
<span class="token comment">/*
* r1 = machine no, r2 = atags or dtb,
* r8 = phys_offset, r9 = cpuid, r10 = procinfo
*/</span>
bl __vet_atags
#ifdef CONFIG_SMP_ON_UP
bl __fixup_smp
#endif
#ifdef CONFIG_ARM_PATCH_PHYS_VIRT
bl __fixup_pv_table
#endif
bl __create_page_tables
<span class="token comment">/*
* The following calls CPU specific code in a position independent
* manner. See arch/arm/mm/proc-*.S for details. r10 = base of
* xxx_proc_info structure selected by __lookup_processor_type
* above. On return, the CPU will be ready for the MMU to be
* turned on, and r0 will hold the CPU control register value.
*/</span>
ldr r13<span class="token punctuation">,</span> <span class="token operator">=</span>__mmap_switched @ address to jump to after
@ mmu has been enabled
adr lr<span class="token punctuation">,</span> <span class="token function">BSYM</span><span class="token punctuation">(</span><span class="token number">1f</span><span class="token punctuation">)</span> @ <span class="token keyword">return</span> <span class="token punctuation">(</span>PIC<span class="token punctuation">)</span> address
mov r8<span class="token punctuation">,</span> r4 @ set TTBR1 to swapper_pg_dir
ARM( add pc, r10, #PROCINFO_INITFUNC )
THUMB( add r12, r10, #PROCINFO_INITFUNC )
THUMB( ret r12 )
1: b __enable_mmu
ENDPROC(stext)
.ltorg
#ifndef CONFIG_XIP_KERNEL
2: .long .
.long PAGE_OFFSET
#endif
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
- 35
- 36
- 37
- 38
- 39
- 40
- 41
- 42
- 43
- 44
- 45
- 46
- 47
- 48
- 49
- 50
- 51
- 52
- 53
- 54
- 55
- 56
- 57
- 58
- 59
- 60
- 61
- 62
- 63
- 64
- 65
- 66
- 67
- 68
- 69
- 70
- 71
- __lookup_processor_type : 使用汇编指令读取CPU ID, 根据该ID找到对应的proc_info_list结构体(里面含有这类CPU的初始化函数、信息)
- __vet_atags : 判断是否存在可用的ATAGS或DTB
在汇编的阶段，大概可以看出来用变量__atags_pointer指向FDT的首地址，执行完汇编的阶段就会调到C代码的流程里面了
3. 平台信息的处理(machine_desc)
进入到start_kernel的处理流程中
asmlinkage void __init start_kernel(void)
{
...
setup_arch(&command_line); //设置架构相关的内容
...
}
- 1
- 2
- 3
- 4
- 5
由于涉及的知识点内容实在太多，那么我们只是重点的关注fdt的处理，直接进到setup_ arch()函数。
void __init setup_arch(char **cmdline_p)
{
...
mdesc = setup_machine_fdt(__atags_pointer);
if (!mdesc)
mdesc = setup_machine_tags(__atags_pointer, __machine_arch_type);
machine_desc = mdesc;
machine_name = mdesc->name;
...
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
首先通过set_machine_fdt来set_machine描述符，如果返回值是NULL，那么就采用传统的方式，如果u-boot传递了，就采用设备树方式
- 传统方式：对于如何确定mdesc，旧的方法是静态定义若干的machine描述符(struct machine_desc)，在系统启动的时候，通过machine type ID作为索引，在这些静态定义的machine描述符中，找到对应哪个ID匹配的描述符。
- 设备树：通过__atags_pointer来找到对应的machine_desc设备描述符
首先我们来看看struct machine_desc的定义方式：
struct machine_desc {
...
unsigned int nr; /* architecture number */
const char *name; /* architecture name */
unsigned long atag_offset; /* tagged list (relative) */
const char *const *dt_compat; /* array of device tree
...
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
nr成员就是过去使用的machine type ID。内核machine描述符的table有若干个entry，每个都有自己的ID。bootloader传递了machine type ID，指明使用哪一个machine描述符。而dtb方式中目前匹配machine描述符使用compatible strings，也就是dt_compat成员，这是一个string list，定义了这个machine所支持的列表。
const struct machine_desc * __init setup_machine_fdt(unsigned int dt_phys) { const struct machine_desc *mdesc, *mdesc_best = NULL; if (!dt_phys || !early_init_dt_verify(phys_to_virt(dt_phys))) return NULL;mdesc <span class="token operator">=</span> <span class="token function">of_flat_dt_match_machine</span><span class="token punctuation">(</span>mdesc_best<span class="token punctuation">,</span> arch_get_next_mach<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span>mdesc<span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token keyword">const</span> <span class="token keyword">char</span> <span class="token operator">*</span>prop<span class="token punctuation">;</span> <span class="token keyword">int</span> size<span class="token punctuation">;</span> <span class="token keyword">unsigned</span> <span class="token keyword">long</span> dt_root<span class="token punctuation">;</span> <span class="token function">early_print</span><span class="token punctuation">(</span><span class="token string">"\nError: unrecognized/unsupported "</span> <span class="token string">"device tree compatible list:\n[ "</span><span class="token punctuation">)</span><span class="token punctuation">;</span> dt_root <span class="token operator">=</span> <span class="token function">of_get_flat_dt_root</span><span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">;</span> prop <span class="token operator">=</span> <span class="token function">of_get_flat_dt_prop</span><span class="token punctuation">(</span>dt_root<span class="token punctuation">,</span> <span class="token string">"compatible"</span><span class="token punctuation">,</span> <span class="token operator">&</span>size<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">while</span> <span class="token punctuation">(</span>size <span class="token operator">></span> <span class="token number">0</span><span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token function">early_print</span><span class="token punctuation">(</span><span class="token string">"'%s' "</span><span class="token punctuation">,</span> prop<span class="token punctuation">)</span><span class="token punctuation">;</span> size <span class="token operator">-</span><span class="token operator">=</span> <span class="token function">strlen</span><span class="token punctuation">(</span>prop<span class="token punctuation">)</span> <span class="token operator">+</span> <span class="token number">1</span><span class="token punctuation">;</span> prop <span class="token operator">+</span><span class="token operator">=</span> <span class="token function">strlen</span><span class="token punctuation">(</span>prop<span class="token punctuation">)</span> <span class="token operator">+</span> <span class="token number">1</span><span class="token punctuation">;</span> <span class="token punctuation">}</span> <span class="token function">early_print</span><span class="token punctuation">(</span><span class="token string">"]\n\n"</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token function">dump_machine_table</span><span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token comment">/* does not return */</span> <span class="token punctuation">}</span> <span class="token comment">/* We really don't want to do this, but sometimes firmware provides buggy data */</span> <span class="token keyword">if</span> <span class="token punctuation">(</span>mdesc<span class="token operator">-></span>dt_fixup<span class="token punctuation">)</span> mdesc<span class="token operator">-></span><span class="token function">dt_fixup</span><span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token function">early_init_dt_scan_nodes</span><span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token comment">/* Change machine number to match the mdesc we're using */</span> __machine_arch_type <span class="token operator">=</span> mdesc<span class="token operator">-></span>nr<span class="token punctuation">;</span> <span class="token keyword">return</span> mdesc<span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
- 35
- 36
- 37
- 38
setup_machine_fdt函数的功能就是根据Device Tree的信息，找到最适合的machine描述符。其主要做了下面几件事情
- 传进来的fdt地址是物理地址，所以用phys_to_virt()函数转换为虚拟地址，同时进行合法检测
- 在machine描述符的列表中scan，找到最合适的那个machine描述符。和传统的方法类似，也是静态定义的。DT_MACHINE_START和MACHINE_END用来定义一个machine描述符。编译的时候，compiler会把这些machine descriptor放到一个特殊的段中（.arch.info.init），形成machine描述符的列表。
- of_get_flat_dt_prop(dt_root, “compatible”, &size)使用compatile属性的值, 跟’’‘每一个machine_desc.dt_compat’’'比较,
成绩为"吻合的compatile属性值的位置",成绩越低越匹配, 对应的machine_desc即被选中
static const void * __init arch_get_next_mach(const char *const **match) { static const struct machine_desc *mdesc = __arch_info_begin; const struct machine_desc *m = mdesc;<span class="token keyword">if</span> <span class="token punctuation">(</span>m <span class="token operator">>=</span> __arch_info_end<span class="token punctuation">)</span> <span class="token keyword">return</span> <span class="token constant">NULL</span><span class="token punctuation">;</span> mdesc<span class="token operator">++</span><span class="token punctuation">;</span> <span class="token operator">*</span>match <span class="token operator">=</span> m<span class="token operator">-></span>dt_compat<span class="token punctuation">;</span> <span class="token keyword">return</span> m<span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
_arch_info_begin指向machine描述符列表第一个entry。通过mdesc++不断的移动machine描述符指针（Note：mdesc是static的）。match返回了该machine描述符的compatible string list。具体匹配的算法倒是很简单，就是比较字符串而已，最终找到对应的machine type。 从该流程可以知道，内核是可以支持很多种不同类型的设备，只要在bootloader传递的时候，传递对应不同的dtb表即可。
4. 运行时参数传递
设备树只是起一个信息传递的作用，对这些信息配置的处理，也比较简单，即从设备树的DTB文件中，把这些设备信息提取出来赋给内核中的某个变量即可。那么在系统初始化的过程中，我们需要将DTB转换成节点是device_node的树状结构，以便后续方便操作。紧接上章的函数继续分析
void __init early_init_dt_scan_nodes(void) { /* Retrieve various information from the /chosen node */ of_scan_flat_dt(early_init_dt_scan_chosen, boot_command_line);<span class="token comment">/* Initialize {size,address}-cells info */</span> <span class="token function">of_scan_flat_dt</span><span class="token punctuation">(</span>early_init_dt_scan_root<span class="token punctuation">,</span> <span class="token constant">NULL</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token comment">/* Setup memory, calling early_init_dt_add_memory_arch */</span> <span class="token function">of_scan_flat_dt</span><span class="token punctuation">(</span>early_init_dt_scan_memory<span class="token punctuation">,</span> <span class="token constant">NULL</span><span class="token punctuation">)</span><span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
该函数主要完成3个工作：
- 扫描 /chosen node，保存运行时参数（bootargs）到boot_command_line，此外，还通过early_init_dt_check_for_initrd处理initrd相关的property，并保存在initrd_start和initrd_end这两个全局变量中 。其中主要是解析dts的配置为
chosen {
bootargs = "earlycon=sprd_serial,0x70100000,115200n8 loglevel=8 console=ttyS1,115200n8 init=/init root=/dev/ram0 rw androidboot.hardware=sc9830";
linux,initrd-start = <0x85500000>;
linux,initrd-end = <0x855a3212>;
};
- 1
- 2
- 3
- 4
bootargs属性就是内核启动的命令行参数，它里面可以指定根文件系统在哪里，第一个运行的应用程序是哪一个，指定内核的打印信息从哪个设备里打印出来
- 扫描根节点，获取 {size,address}-cells信息，并保存在dt_root_size_cells和dt_root_addr_cells全局变量中 ,memory中的reg属性的地址是32位还是64位，大小是用一个32位表示，还是两个32位表示
- 扫描DTB中的memory node，并把相关信息保存在meminfo中，全局变量meminfo通过memblock_add保存了系统内存相关的信息
5. dtb解析成device node
uboot把设备树DTB文件随便放到内存的某一个地方就可以使用，为什么内核运行中，他不会去覆盖DTB所占用的那块内存呢？在设备树文件中，可以使用/memreserve/指定一块内存，这块内存就是保留的内存，内核不会占用它。即使你没有指定这块内存，当我们内核启动时，他也会把设备树所占用的区域保留下来。内核在arm_memblock_init中会使用early_init_fdt_scan_reserved_mem来配置fdt的内存，通知也回对memreserve指定内存进行保留操作。
void __init early_init_fdt_scan_reserved_mem(void) { int n; u64 base, size;<span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span>initial_boot_params<span class="token punctuation">)</span> <span class="token keyword">return</span><span class="token punctuation">;</span> <span class="token comment">/* Reserve the dtb region */</span> <span class="token function">early_init_dt_reserve_memory_arch</span><span class="token punctuation">(</span><span class="token function">__pa</span><span class="token punctuation">(</span>initial_boot_params<span class="token punctuation">)</span><span class="token punctuation">,</span> <span class="token function">fdt_totalsize</span><span class="token punctuation">(</span>initial_boot_params<span class="token punctuation">)</span><span class="token punctuation">,</span> <span class="token number">0</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token comment">/* Process header /memreserve/ fields */</span> <span class="token keyword">for</span> <span class="token punctuation">(</span>n <span class="token operator">=</span> <span class="token number">0</span><span class="token punctuation">;</span> <span class="token punctuation">;</span> n<span class="token operator">++</span><span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token function">fdt_get_mem_rsv</span><span class="token punctuation">(</span>initial_boot_params<span class="token punctuation">,</span> n<span class="token punctuation">,</span> <span class="token operator">&</span>base<span class="token punctuation">,</span> <span class="token operator">&</span>size<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span>size<span class="token punctuation">)</span> <span class="token keyword">break</span><span class="token punctuation">;</span> <span class="token function">early_init_dt_reserve_memory_arch</span><span class="token punctuation">(</span>base<span class="token punctuation">,</span> size<span class="token punctuation">,</span> <span class="token number">0</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token punctuation">}</span> <span class="token function">of_scan_flat_dt</span><span class="token punctuation">(</span>__fdt_scan_reserved_mem<span class="token punctuation">,</span> <span class="token constant">NULL</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token function">fdt_init_reserved_mem</span><span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- initial_boot_params实际上是dtb的虚拟地址，在early_init_dt_verify初始化的时候设定，首先进来就判断dtb是否存在，如果存在就将dtb的空间进行保留
- 对fdt中的每一个节点调用__fdt_scan_reserved_mem函数，进行reserved-memory节点的扫描，之后调用fdt_init_reserved_mem函数进行内存预留的动作
说完了dtb对于内存的流程，那么来到这节的重点，dtb解析成device node，首先来看看下面的代码
void __init unflatten_device_tree(void) { __unflatten_device_tree(initial_boot_params, &of_allnodes, early_init_dt_alloc_memory_arch);<span class="token comment">/* Get pointer to "/chosen" and "/aliases" nodes for use everywhere */</span> <span class="token function">of_alias_scan</span><span class="token punctuation">(</span>early_init_dt_alloc_memory_arch<span class="token punctuation">)</span><span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
分析以上代码，在unflatten_device_tree()中，调用函数__unflatten_device_tree()，参数initial_boot_params指向Device Tree在内存中的首地址，of_root在经过该函数处理之后，会指向根节点，early_init_dt_alloc_memory_arch是一个函数指针，为struct device_node和struct property结构体分配内存的回调函数（callback）。在__unflatten_device_tree()函数中，两次调用unflatten_dt_node()函数，第一次是为了得到Device Tree转换成struct device_node和struct property结构体需要分配的内存大小，第二次调用才是具体填充每一个struct device_node和struct property结构体。那么Device Tree中的每一个node节点经过kernel处理都会生成一个struct device_node的结构体，struct device_node最终一般会被挂接到具体的struct device结构体。struct device_node结构体描述如下：
struct device_node { const char *name; //device node name const char *type; //对应device_type的属性 phandle phandle; //对应该节点的phandle属性 const char *full_name; //从“/”开始的，表示该node的full path<span class="token keyword">struct</span> property <span class="token operator">*</span>properties<span class="token punctuation">;</span> <span class="token comment">//该节点的属性列表 </span> <span class="token keyword">struct</span> property <span class="token operator">*</span>deadprops<span class="token punctuation">;</span> <span class="token comment">//如果需要删除某些属性，kernel并非真的删除，而是挂入到deadprops的列表 </span> <span class="token keyword">struct</span> device_node <span class="token operator">*</span>parent<span class="token punctuation">;</span> <span class="token comment">//parent、child以及sibling将所有的device node连接起来</span> <span class="token keyword">struct</span> device_node <span class="token operator">*</span>child<span class="token punctuation">;</span> <span class="token keyword">struct</span> device_node <span class="token operator">*</span>sibling<span class="token punctuation">;</span> <span class="token keyword">struct</span> device_node <span class="token operator">*</span>next<span class="token punctuation">;</span> <span class="token comment">//通过该指针可以获取相同类型的下一个node </span> <span class="token keyword">struct</span> device_node <span class="token operator">*</span>allnext<span class="token punctuation">;</span> <span class="token comment">//通过该指针可以获取node global list下一个node</span> <span class="token keyword">struct</span> kobject kobj<span class="token punctuation">;</span> <span class="token keyword">unsigned</span> <span class="token keyword">long</span> _flags<span class="token punctuation">;</span> <span class="token keyword">void</span> <span class="token operator">*</span>data<span class="token punctuation">;</span>
#if defined(CONFIG_SPARC)
const char path_component_name;
unsigned int unique_id;
struct of_irq_controller irq_trans;
#endif
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
对于dtb最后都存储在，其结构图如下
6. linux kernel的设备驱动模型
在linux kernel引入统一设备模型之后，bus、driver和device形成了设备模型中的铁三角。在驱动初始化的时候会将代表该driver的一个数据结构挂入bus上的driver链表，device的数据结构挂入bus上的devie链表，那么如何让device遇到“对”的那个driver呢？那么就要靠缘分了，也就是bus的match函数来完成。在传统的方式中，代码中会定义一个static struct platform_device *xxx_devices的静态数组，在初始化的时候调用platform_add_devices。这些静态定义的platform_device往往又需要静态定义各种resource，那么对于设备树，也就是需要根据device_node的树状结构（root是of_allnodes）将一个个的device node挂入到相应的总线device链表中即可。
static int __init customize_machine(void)
{
/*
* customizes platform devices, or adds new ones
* On DT based machines, we fall back to populating the
* machine from the device tree, if no callback is provided,
* otherwise we would always need an init_machine callback.
*/
if (machine_desc->init_machine)
machine_desc->init_machine();
#ifdef CONFIG_OF
else
of_platform_populate(NULL, of_default_bus_match_table,
NULL, NULL);
#endif
return 0;
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
那么Linux系统是怎么知道哪些device node要注册为platform_device，哪些是用于i2c_client，哪些是用于spi_device？不知道你有没有注意到调用of_platform_populate的时候给它传递了一个参数of_default_bus_match_table
const struct of_device_id of_default_bus_match_table[] = {
{ .compatible = "simple-bus", },
#ifdef CONFIG_ARM_AMBA
{ .compatible = "arm,amba-bus", },
#endif /* CONFIG_ARM_AMBA */
{} /* Empty terminated list */
};
- 1
- 2
- 3
- 4
- 5
- 6
那么在dts文件会也会对对应的驱动进行配置
ap-apb { compatible = "simple-bus"; #address-cells = <1>; #size-cells = <1>; ranges;uart0: serial@70000000 { compatible = "sprd,sc9836-uart"; reg = <0x70000000 0x100>; interrupts = <GIC_SPI 2 IRQ_TYPE_LEVEL_HIGH>; clock-names = "uart", "source","enable"; clocks = <&clk_uart0>, <&ext_26m>, <&clk_ap_apb_gates 13>; status = "disabled"; };
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
如果某个device node的compatible属性的值与数组of_default_bus_match_table中的任意一个元素的compatible的值match，那么这个device node的child device node（device_node的child成员变量指向的是这个device node的子节点，也是一个链表）仍旧会被注册为platform_device。下面来看看重点的解析过程
int of_platform_populate(struct device_node *root, const struct of_device_id *matches, const struct of_dev_auxdata *lookup, struct device *parent) { struct device_node *child; int rc = 0;root <span class="token operator">=</span> root <span class="token operator">?</span> <span class="token function">of_node_get</span><span class="token punctuation">(</span>root<span class="token punctuation">)</span> <span class="token punctuation">:</span> <span class="token function">of_find_node_by_path</span><span class="token punctuation">(</span><span class="token string">"/"</span><span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span>root<span class="token punctuation">)</span> <span class="token keyword">return</span> <span class="token operator">-</span>EINVAL<span class="token punctuation">;</span> <span class="token function">for_each_child_of_node</span><span class="token punctuation">(</span>root<span class="token punctuation">,</span> child<span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> rc <span class="token operator">=</span> <span class="token function">of_platform_bus_create</span><span class="token punctuation">(</span>child<span class="token punctuation">,</span> matches<span class="token punctuation">,</span> lookup<span class="token punctuation">,</span> parent<span class="token punctuation">,</span> true<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span>rc<span class="token punctuation">)</span> <span class="token keyword">break</span><span class="token punctuation">;</span> <span class="token punctuation">}</span> <span class="token function">of_node_put</span><span class="token punctuation">(</span>root<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">return</span> rc<span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
该函数主要完成完成：
- 获取根节点，如果传递进来的参数root为NULL，那么需要通过of_find_node_by_path函数找到device tree中的根节点。
- 得到根节点之后，就可以通过这个根节点来遍历device tree中的节点了。得到一个子节点之后，调用of_platform_bus_create函数为每一个节点创建platform_device结构体
static int of_platform_bus_create(struct device_node *bus, const struct of_device_id *matches, const struct of_dev_auxdata *lookup, struct device *parent, bool strict) { const struct of_dev_auxdata *auxdata; struct device_node *child; struct platform_device *dev; const char *bus_id = NULL; void *platform_data = NULL; int rc = 0;<span class="token comment">/* Make sure it has a compatible property */</span> <span class="token keyword">if</span> <span class="token punctuation">(</span>strict <span class="token operator">&&</span> <span class="token punctuation">(</span><span class="token operator">!</span><span class="token function">of_get_property</span><span class="token punctuation">(</span>bus<span class="token punctuation">,</span> <span class="token string">"compatible"</span><span class="token punctuation">,</span> <span class="token constant">NULL</span><span class="token punctuation">)</span><span class="token punctuation">)</span><span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token function">pr_debug</span><span class="token punctuation">(</span><span class="token string">"%s() - skipping %s, no compatible prop\n"</span><span class="token punctuation">,</span> <span class="token constant">__func__</span><span class="token punctuation">,</span> bus<span class="token operator">-></span>full_name<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">return</span> <span class="token number">0</span><span class="token punctuation">;</span> <span class="token punctuation">}</span> auxdata <span class="token operator">=</span> <span class="token function">of_dev_lookup</span><span class="token punctuation">(</span>lookup<span class="token punctuation">,</span> bus<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span>auxdata<span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> bus_id <span class="token operator">=</span> auxdata<span class="token operator">-></span>name<span class="token punctuation">;</span> platform_data <span class="token operator">=</span> auxdata<span class="token operator">-></span>platform_data<span class="token punctuation">;</span> <span class="token punctuation">}</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token function">of_device_is_compatible</span><span class="token punctuation">(</span>bus<span class="token punctuation">,</span> <span class="token string">"arm,primecell"</span><span class="token punctuation">)</span><span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token comment">/* * Don't return an error here to keep compatibility with older * device tree files. */</span> <span class="token function">of_amba_device_create</span><span class="token punctuation">(</span>bus<span class="token punctuation">,</span> bus_id<span class="token punctuation">,</span> platform_data<span class="token punctuation">,</span> parent<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">return</span> <span class="token number">0</span><span class="token punctuation">;</span> <span class="token punctuation">}</span> dev <span class="token operator">=</span> <span class="token function">of_platform_device_create_pdata</span><span class="token punctuation">(</span>bus<span class="token punctuation">,</span> bus_id<span class="token punctuation">,</span> platform_data<span class="token punctuation">,</span> parent<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span>dev <span class="token operator">||</span> <span class="token operator">!</span><span class="token function">of_match_node</span><span class="token punctuation">(</span>matches<span class="token punctuation">,</span> bus<span class="token punctuation">)</span><span class="token punctuation">)</span> <span class="token keyword">return</span> <span class="token number">0</span><span class="token punctuation">;</span> <span class="token function">for_each_child_of_node</span><span class="token punctuation">(</span>bus<span class="token punctuation">,</span> child<span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token function">pr_debug</span><span class="token punctuation">(</span><span class="token string">" create child: %s\n"</span><span class="token punctuation">,</span> child<span class="token operator">-></span>full_name<span class="token punctuation">)</span><span class="token punctuation">;</span> rc <span class="token operator">=</span> <span class="token function">of_platform_bus_create</span><span class="token punctuation">(</span>child<span class="token punctuation">,</span> matches<span class="token punctuation">,</span> lookup<span class="token punctuation">,</span> <span class="token operator">&</span>dev<span class="token operator">-></span>dev<span class="token punctuation">,</span> strict<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span>rc<span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token function">of_node_put</span><span class="token punctuation">(</span>child<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">break</span><span class="token punctuation">;</span> <span class="token punctuation">}</span> <span class="token punctuation">}</span> <span class="token function">of_node_set_flag</span><span class="token punctuation">(</span>bus<span class="token punctuation">,</span> OF_POPULATED_BUS<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">return</span> rc<span class="token punctuation">;</span>
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
- 35
- 36
- 37
- 38
- 39
- 40
- 41
- 42
- 43
- 44
- 45
- 46
- 47
- 48
- 需要确定节点是否有"compatible"属性，如果没有"compatible"属性，则直接返回，即不会创建platform设备的。
- 如果"compatible"属性值有"arm,primecell"，则会调用of_amba_device_create函数去创建amba_device，它设计了AMBA的总线来连接SOC内的各个block。符合这个总线标准的SOC上的外设叫做ARM Primecell Peripherals
- 如果不是ARM Primecell Peripherals，那么我们就需要向platform bus上增加一个platform device了，of_platform_device_create_pdata才是真正的platform_device
- 一个device node可能是一个桥设备，因此要重复调用of_platform_bus_create来把所有的device node处理掉
static struct platform_device *of_platform_device_create_pdata( struct device_node *np, const char *bus_id, void *platform_data, struct device *parent) { struct platform_device *dev;<span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span><span class="token function">of_device_is_available</span><span class="token punctuation">(</span>np<span class="token punctuation">)</span> <span class="token operator">||</span> <span class="token function">of_node_test_and_set_flag</span><span class="token punctuation">(</span>np<span class="token punctuation">,</span> OF_POPULATED<span class="token punctuation">)</span><span class="token punctuation">)</span> <span class="token keyword">return</span> <span class="token constant">NULL</span><span class="token punctuation">;</span> dev <span class="token operator">=</span> <span class="token function">of_device_alloc</span><span class="token punctuation">(</span>np<span class="token punctuation">,</span> bus_id<span class="token punctuation">,</span> parent<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token operator">!</span>dev<span class="token punctuation">)</span> <span class="token keyword">goto</span> err_clear_flag<span class="token punctuation">;</span> <span class="token function">of_dma_configure</span><span class="token punctuation">(</span><span class="token operator">&</span>dev<span class="token operator">-></span>dev<span class="token punctuation">)</span><span class="token punctuation">;</span> dev<span class="token operator">-></span>dev<span class="token punctuation">.</span>bus <span class="token operator">=</span> <span class="token operator">&</span>platform_bus_type<span class="token punctuation">;</span> dev<span class="token operator">-></span>dev<span class="token punctuation">.</span>platform_data <span class="token operator">=</span> platform_data<span class="token punctuation">;</span> <span class="token comment">/* We do not fill the DMA ops for platform devices by default. * This is currently the responsibility of the platform code * to do such, possibly using a device notifier */</span> <span class="token keyword">if</span> <span class="token punctuation">(</span><span class="token function">of_device_add</span><span class="token punctuation">(</span>dev<span class="token punctuation">)</span> <span class="token operator">!=</span> <span class="token number">0</span><span class="token punctuation">)</span> <span class="token punctuation">{<!-- --></span> <span class="token function">platform_device_put</span><span class="token punctuation">(</span>dev<span class="token punctuation">)</span><span class="token punctuation">;</span> <span class="token keyword">goto</span> err_clear_flag<span class="token punctuation">;</span> <span class="token punctuation">}</span> <span class="token keyword">return</span> dev<span class="token punctuation">;</span>
err_clear_flag:
of_node_clear_flag(np, OF_POPULATED);
return NULL;
}
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
- 13
- 14
- 15
- 16
- 17
- 18
- 19
- 20
- 21
- 22
- 23
- 24
- 25
- 26
- 27
- 28
- 29
- 30
- 31
- 32
- 33
- 34
- 35
- of_device_is_available函数，这个函数主要是用于检测"status"属性的，如果没有"status"属性，那还好说直接返回true。如果有"status"属性，而它的值又不是"okay"或"ok"，那么不好意思，返回false，否则还是返回true。所以"status"属性就是用来检测是否可用，是否需要创建platform_node
- of_device_alloc除了分配struct platform_device的内存，还分配了该platform device需要的resource的内存。当然，这就需要解析该device node的interrupt资源以及memory address资源。
- 回到of_platform_device_create_pdata函数中，平台设备已经申请好了，然后对平台设备继续进行赋值操作，例如平台设备的总线赋值为平台总线，平台设备的私有数据赋值为platform_data，最终会调用of_device_add函数将平台设备注册到内核中。
也就是说当of_platform_populate()函数执行完毕，kernel就为DTB中所有包含compatible属性名的第一级node创建platform_device结构体，并向平台设备总线注册设备信息。如果第一级node的compatible属性值等于“simple-bus”、“simple-mfd”或者"arm,amba-bus"的话，kernel会继续为当前node的第二级包含compatible属性的node创建platform_device结构体，并注册设备。Linux系统下的设备大多都是挂载在平台总线下的，因此在平台总线被注册后，会根据of_root节点的树结构，去寻找该总线的子节点，所有的子节点将被作为设备注册到该总线上。
7 参考文档
http://www.wowotech.net/device_model/dt-code-analysis.html
https://blog.youkuaiyun.com/thisway_diy/article/details/84336817
扫一扫
808
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
