Android Bootloader LittleKernel的两篇文章【转】

转自：http://blog.csdn.net/loongembedded/article/details/41747523

2014-12-05 14:37 3599人阅读评论(2) 收藏举报

分类：

Android Bootloader（68）

Android 开发之 ---- bootloader （LK）

LK是什么

LK 是 Little Kernel 它是 appsbl （Applications ARM Boot Loader）流程代码，little kernel 是小内核小操作系统。

LK 代码在 bootable/bootloadler/lk 目录下

LK 代码结构

+app // 应用相关

+arch // arm 体系

+dev // 设备相关

+include // 头文件

+kernel // lk系统相关

+platform // 相关驱动

+projiect // makefile文件

+scripts // Jtag 脚本

+target // 具体板子相关

LK 流程分析

在 bootable/bootloadler/lk/arch/arm/ssystem-onesegment.ld 连接文件中 ENTRY（_start）指定 LK 从_start 函数开始，_start 在 lk/arch/crt0.S中。crt0.S 主要做一些基本的 CPU 的初始化再通过 bl kmain ；跳转到 C 代码中。

kmain 在 lk/kernel/main.c 中

kmain()

kmain 主要做两件事：1、本身 lk 这个系统模块的初始化；2、boot 的启动初始化动作。

kmain 源码分析：

void kmain（）

{

1.初始化进程（lk 中的简单进程）相关结构体。

thread_init_early();

2.做一些如关闭 cache，使能 mmu 的 arm 相关工作。

arch_early_init();

3.相关平台的早期初始化

platform_early_init();

4.现在就一个函数跳转，初始化UART（板子相关）

target_early_init();

5.构造函数相关初始化

call_constructors();

6.lk系统相关的堆栈初始化

heap_init();

7.简短的初始化定时器对象

thread_init();

8.lk系统控制器初始化（相关事件初始化）

dpc_init();

9.初始化lk中的定时器

            timer_init();
           10.新建线程入口函数 bootstrap2 用于boot 工作（重点）
           thread_resume(thread_create("bootstrap2", &bootstrap2, NULL, DEFAULT_PRIORITY, DEFAULT_STACK_SIZE));

}

以上与 boot 启动初始化相关函数是 arch_early_init、 platform_early_init 、bootstrap2，这些是启动的重点，我们下面慢慢来看。

arch_early_init()

体系架构相关的初始化我们一般用的 ARM 体系

1.关闭cache

arch_disable_cache(UCACHE);

2.设置向量基地址（中断相关）

set_vector_base(MEMBASE);

3.初始化MMU

arm_mmu_init();

4.初始化MMU映射__平台相关

platform_init_mmu_mappings();

5.开启cache

arch_enable_cache(UCACHE)

6.使能 cp10 和 cp11

__asm__ volatile("mrc p15, 0, %0, c1, c0, 2" : "=r" (val));

val |= (3<<22)|(3<<20);

__asm__ volatile("mcr p15, 0, %0, c1, c0, 2" :: "r" (val));

7.设置使能 fpexc 位（中断相关）

__asm__ volatile("mrc p10, 7, %0, c8, c0, 0" : "=r" (val));

val |= (1<<30);

__asm__ volatile("mcr p10, 7, %0, c8, c0, 0" :: "r" (val));

8.使能循环计数寄存器

__asm__ volatile("mrc p15, 0, %0, c9, c12, 0" : "=r" (en));

en &= ~(1<<3); /*循环计算每个周期*/

en |= 1;

__asm__ volatile("mcr p15, 0, %0, c9, c12, 0" :: "r" (en));

9.使能循环计数器

en = (1<<31);
__asm__ volatile("mcr p15, 0, %0, c9, c12, 1" :: "r" (en));

platform_early_init()

平台相关初始化不同平台不同的初始化下面是msm7x30

1.初始化中断

platform_init_interrupts();

2.初始化定时器

platform_init_timer();

bootstrap2

bootstrap2 在kmain的末尾以线程方式开启。主要分三步：platform_init、target_init、apps_init。

1.platform_init

platform_init 中主要是函数 acpu_clock_init。

在 acpu_clock_init 对 arm11 进行系统时钟设置，超频

2.target_init

针对硬件平台进行设置。主要对 arm9 和 arm11 的分区表进行整合，初始化flash和读取FLASH信息

3.apps_init

apps_init 是关键，对 LK 中所谓 app 初始化并运行起来，而 aboot_init 就将在这里开始被运行，android Linux 内核的加载工作就在 aboot_init 中完成的。

aboot_init

1.设置NAND/ EMMC读取信息页面大小
if (target_is_emmc_boot())

{

page_size = 2048;

page_mask = page_size - 1;

}

else

{

page_size = flash_page_size();

page_mask = page_size - 1;

}

2.读取按键信息，判断是正常开机，还是进入 fastboot ,还是进入recovery 模式

。。。。。。。。。

通过一系列的 if (keys_get_state() == XXX) 判断

。。。。。。。。。

3.从 nand 中加载内核

boot_linux_from_flash();

partition_dump();

sz = target_get_max_flash_size();

fastboot_init(target_get_scratch_address(), sz);

udc_start(); // 开始 USB 协议

boot_linux_from_flash

主要是内核的加载过程，我们的 boot.img 包含：kernel 头、kernel、ramdisk、second stage（可以没有）。

1.读取boot 头部

flash_read(p, offset, raw_header, 2048)

           offset += 2048;
   2.读取内核
           memcmp(hdr->magic, BOOT_MAGIC, BOOT_MAGIC_SIZE)
           n = (hdr->kernel_size + (FLASH_PAGE_SIZE - 1)) & (~(FLASH_PAGE_SIZE - 1));

flash_read(p, offset, (void*) hdr->kernel_addr, n)

   offset += n;
           3.读取 ramdisk
   n = (hdr->ramdisk_size + (FLASH_PAGE_SIZE - 1)) & (~(FLASH_PAGE_SIZE - 1));

flash_read(p, offset, (void*) hdr->ramdisk_addr, n)

offset += n;

4.启动内核，

boot_linux()；//在boot_linux 中entry(0,machtype,tags);从kernel加载在内核中的地址开始运行了。

到这里LK的启动过程就结束了。

Android Kernel - Boot Loader

Android Boot loader 的 code 在 bootable/bootloader/lk 底下, LK 是 Little Kernel 的缩写, 是 andriod bootloader 的核心精神.

入口函数在 kernel/main.c 中的 kmain(), 以下就来读读这一段 code.

[c-sharp] view plain copy print ?

void kmain(void)
{
// get us into some sort of thread context
thread_init_early();
// early arch stuff
arch_early_init();
// do any super early platform initialization
platform_early_init();
// do any super early target initialization
target_early_init();
dprintf(INFO, "welcome to lk/n/n");
// deal with any static constructors
dprintf(SPEW, "calling constructors/n");
call_constructors();
// bring up the kernel heap
dprintf(SPEW, "initializing heap/n");
heap_init();
// initialize the threading system
dprintf(SPEW, "initializing threads/n");
thread_init();
// initialize the dpc system
dprintf(SPEW, "initializing dpc/n");
dpc_init();
// initialize kernel timers
dprintf(SPEW, "initializing timers/n");
timer_init();
#if (!ENABLE_NANDWRITE)
// create a thread to complete system initialization
dprintf(SPEW, "creating bootstrap completion thread/n");
thread_resume(thread_create("bootstrap2", &bootstrap2, NULL, DEFAULT_PRIORITY, DEFAULT_STACK_SIZE));
// enable interrupts
exit_critical_section();
// become the idle thread
thread_become_idle();
#else
bootstrap_nandwrite();
#endif
}

In include/debug.h: 我们可以看到 dprintf 的第一个参数是代表 debug level.

[c-sharp:nogutter] view plain copy print ?

/* debug levels */
#define CRITICAL 0
#define ALWAYS 0
#define INFO 1
#define SPEW 2

In include/debug.h:

[c-sharp:nogutter] view plain copy print ?

#define dprintf(level, x...) do { if ((level) <= DEBUGLEVEL) { _dprintf(x); } } while (0)

所以 dprintf 会依 DEBUGLEVEL 来判断是否输出信息.

来看第一个 call 的函数: thread_init_early, define in thread.c

[c-sharp:nogutter] view plain copy print ?

void thread_init_early(void)
{
int i;
/* initialize the run queues */
for (i=0; i < NUM_PRIORITIES; i++)
list_initialize(&run_queue[i]);
/* initialize the thread list */
list_initialize(&thread_list);
/* create a thread to cover the current running state */
thread_t *t = &bootstrap_thread;
init_thread_struct(t, "bootstrap");
/* half construct this thread, since we're already running */
t->priority = HIGHEST_PRIORITY;
t->state = THREAD_RUNNING;
t->saved_critical_section_count = 1;
list_add_head(&thread_list, &t->thread_list_node);
current_thread = t;
}

#define NUM_PRIORITIES 32 in include/kernel/thread.h

list_initialize() defined in include/list.h: initialized a list

[c-sharp:nogutter] view plain copy print ?

static inline void list_initialize(struct list_node *list)
{
list->prev = list->next = list;
}

run_queue 是 static struct list_node run_queue[NUM_PRIORITIES]

thread_list 是 static struct list_node thread_list

再来要 call 的函数是: arch_early_init() defined in arch/arm/arch.c

[c-sharp:nogutter] view plain copy print ?

void arch_early_init(void)
{
/* turn off the cache */
arch_disable_cache(UCACHE);
/* set the vector base to our exception vectors so we dont need to double map at 0 */
#if ARM_CPU_CORTEX_A8
set_vector_base(MEMBASE);
#endif
#if ARM_WITH_MMU
arm_mmu_init();
platform_init_mmu_mappings();
#endif
/* turn the cache back on */
arch_enable_cache(UCACHE);
#if ARM_WITH_NEON
/* enable cp10 and cp11 */
uint32_t val;
__asm__ volatile("mrc p15, 0, %0, c1, c0, 2" : "=r" (val));
val |= (3<<22)|(3<<20);
__asm__ volatile("mcr p15, 0, %0, c1, c0, 2" :: "r" (val));
/* set enable bit in fpexc */
val = (1<<30);
__asm__ volatile("mcr p10, 7, %0, c8, c0, 0" :: "r" (val));
#endif
}

现代操作系统普遍采用虚拟内存管理（Virtual Memory Management）机制，这需要处理器中的MMU（Memory Management Unit，

内存管理单元）提供支持。

CPU执行单元发出的内存地址将被MMU截获，从CPU到MMU的地址称为虚拟地址（Virtual Address，以下简称VA），而MMU将这个地

址翻译成另一个地址发到CPU芯片的外部地址引脚上，也就是将VA映射成PA

MMU将VA映射到PA是以页（Page）为单位的，32位处理器的页尺寸通常是4KB。例如，MMU可以通过一个映射项将VA的一页

0xb7001000~0xb7001fff映射到PA的一页0x2000~0x2fff，如果CPU执行单元要访问虚拟地址0xb7001008，则实际访问到的物理地

址是0x2008。物理内存中的页称为物理页面或者页帧（Page Frame）。虚拟内存的哪个页面映射到物理内存的哪个页帧是通过页

表（Page Table）来描述的，页表保存在物理内存中，MMU会查找页表来确定一个VA应该映射到什么PA。

操作系统和MMU是这样配合的：

1. 操作系统在初始化或分配、释放内存时会执行一些指令在物理内存中填写页表，然后用指令设置MMU，告诉MMU页表在物理内存中

的什么位置。

2. 设置好之后，CPU每次执行访问内存的指令都会自动引发MMU做查表和地址转换操作，地址转换操作由硬件自动完成，不需要用指令

控制MMU去做。

MMU除了做地址转换之外，还提供内存保护机制。各种体系结构都有用户模式（User Mode）和特权模式（Privileged Mode）之分，

操作系统可以在页表中设置每个内存页面的访问权限，有些页面不允许访问，有些页面只有在CPU处于特权模式时才允许访问，有些页面

在用户模式和特权模式都可以访问，访问权限又分为可读、可写和可执行三种。这样设定好之后，当CPU要访问一个VA时，MMU会检查

CPU当前处于用户模式还是特权模式，访问内存的目的是读数据、写数据还是取指令，如果和操作系统设定的页面权限相符，就允许访

问，把它转换成PA，否则不允许访问，产生一个异常（Exception）

常见的 segmentation fault 产生的原因:

用户程序要访问一段 VA, 经 MMU 检查后无权访问, MMU 会产生异常, CPU 从用户模式切换到特权模式, 跳转到内核代码中执行异常服务程序.

内核就会把这个异常解释为 segmentation fault, 将引发异常的程序终止.

简单的讲一下 NEON: NEON technology can accelerate multimedia and signal processing algorithms such as video encode/decode,

2D/3D graphics, gaming, audio and speech processing, image processing, telephony, and sound synthesis.

platform_early_init() defined in platform/<your-platform>/platform.c

[c-sharp:nogutter] view plain copy print ?

void platform_early_init(void)
{
uart_init();
platform_init_interrupts();
platform_init_timer();
}

uart_init.c defined in platform/<your-platform>/uart.c 所有用到的变数,也都定义在 uart.c

[c-sharp:nogutter] view plain copy print ?

void uart_init(void)
{
uwr(0x0A, UART_CR); /* disable TX and RX */
uwr(0x30, UART_CR); /* reset error status */
uwr(0x10, UART_CR); /* reset receiver */
uwr(0x20, UART_CR); /* reset transmitter */
#if PLATFORM_QSD8K
/* TCXO */
uwr(0x06, UART_MREG);
uwr(0xF1, UART_NREG);
uwr(0x0F, UART_DREG);
uwr(0x1A, UART_MNDREG);
#else
/* TCXO/4 */
uwr(0xC0, UART_MREG);
uwr(0xAF, UART_NREG);
uwr(0x80, UART_DREG);
uwr(0x19, UART_MNDREG);
#endif
uwr(0x10, UART_CR); /* reset RX */
uwr(0x20, UART_CR); /* reset TX */
uwr(0x30, UART_CR); /* reset error status */
uwr(0x40, UART_CR); /* reset RX break */
uwr(0x70, UART_CR); /* rest? */
uwr(0xD0, UART_CR); /* reset */
uwr(0x7BF, UART_IPR); /* stale timeout = 630 * bitrate */
uwr(0, UART_IMR);
uwr(115, UART_RFWR); /* RX watermark = 58 * 2 - 1 */
uwr(10, UART_TFWR); /* TX watermark */
uwr(0, UART_RFWR);
uwr(UART_CSR_115200, UART_CSR);
uwr(0, UART_IRDA);
uwr(0x1E, UART_HCR);
// uwr(0x7F4, UART_MR1); /* RFS/ CTS/ 500chr RFR */
uwr(16, UART_MR1);
uwr(0x34, UART_MR2); /* 8N1 */
uwr(0x05, UART_CR); /* enable TX & RX */
uart_ready = 1;
}

platform_init_interrupts: defined in platform/msm8x60/interrupts.c

[c-sharp:nogutter] view plain copy print ?

void platform_init_interrupts(void)
{
platform_gic_dist_init();
platform_gic_cpu_init();
}

GIC 指的是 Generic Interrupt Controller. The gic-cpu and gic-dist are two subcomponents of GIC.

Devices are wired to the Git-dist which is in charge of distributing interrupts to the gic-cpu (per cpu IRQ IF).

platform_init_timer(): defined in platform/<your-platform>/timer.c

[c-sharp:nogutter] view plain copy print ?

void platform_init_timer(void)
{
writel(0, DGT_ENABLE);
}

DGT: Digital Game Timer: presents the countdowns of two players, it is also called chess timer or chess clock.

target_early_init(): defined in target/init.c

[c-sharp:nogutter] view plain copy print ?

/*
* default implementations of these routines, if the target code
* chooses not to implement.
*/
__WEAK void target_early_init(void)
{
}
__WEAK void target_init(void)
{
}

call_constructors() is defined in kernel/main.c:

[c-sharp:nogutter] view plain copy print ?

static void call_constructors(void)
{
void **ctor;
ctor = &__ctor_list;
while(ctor != &__ctor_end) {
void (*func)(void);
func = (void (*)())*ctor;
func();
ctor++;
}
}

heap_init is defined in lib/heap/heap.c:

[c-sharp:nogutter] view plain copy print ?

void heap_init(void)
{
LTRACE_ENTRY;
// set the heap range
theheap.base = (void *)HEAP_START;
theheap.len = HEAP_LEN;
LTRACEF("base %p size %zd bytes/n", theheap.base, theheap.len);
// initialize the free list
list_initialize(&theheap.free_list);
// create an initial free chunk
heap_insert_free_chunk(heap_create_free_chunk(theheap.base, theheap.len));
// dump heap info
// heap_dump();
// dprintf(INFO, "running heap tests/n");
// heap_test();
}

thread_init is defined in kernel/thread.c but nothing coded, 先记下.

[c-sharp:nogutter] view plain copy print ?

void thread_init(void)
{
}

dpc_init() is defined in kernel/dpc.c:

[c-sharp:nogutter] view plain copy print ?

void dpc_init(void)
{
event_init(&dpc_event, false, 0);
thread_resume(thread_create("dpc", &dpc_thread_routine, NULL, DPC_PRIORITY, DEFAULT_STACK_SIZE));
}

dpc 为 Delayed Procedure Call 延迟过程调用的缩写.

timer_init() is defined in kernel/timer.c:

[c-sharp:nogutter] view plain copy print ?

void timer_init(void)
{
list_initialize(&timer_queue);
/* register for a periodic timer tick */
platform_set_periodic_timer(timer_tick, NULL, 10); /* 10ms */
}

执行 thread_resume 或是 bootstrap_nandwrite

[c-sharp:nogutter] view plain copy print ?

#if (!ENABLE_NANDWRITE)
// create a thread to complete system initialization
dprintf(SPEW, "creating bootstrap completion thread/n");
thread_resume(thread_create("bootstrap2", &bootstrap2, NULL, DEFAULT_PRIORITY, DEFAULT_STACK_SIZE));
// enable interrupts
exit_critical_section();
// become the idle thread
thread_become_idle();
#else
bootstrap_nandwrite();
#endif

In kermel/main.c:

[c-sharp:nogutter] view plain copy print ?

static int bootstrap2(void *arg)
{
dprintf(SPEW, "top of bootstrap2()/n");
arch_init();
// initialize the rest of the platform
dprintf(SPEW, "initializing platform/n");
platform_init();
// initialize the target
dprintf(SPEW, "initializing target/n");
target_init();
dprintf(SPEW, "calling apps_init()/n");
apps_init();
return 0;
}
#if (ENABLE_NANDWRITE)
void bootstrap_nandwrite(void)
{
dprintf(SPEW, "top of bootstrap2()/n");
arch_init();
// initialize the rest of the platform
dprintf(SPEW, "initializing platform/n");
platform_init();
// initialize the target
dprintf(SPEW, "initializing target/n");
target_init();
dprintf(SPEW, "calling nandwrite_init()/n");
nandwrite_init();
return 0;
}
#endif

continue to see apps_init(): app/app.c:void apps_init(void)

[c-sharp] view plain copy print ?

#include <app.h>
#include <kernel/thread.h>
extern const struct app_descriptor __apps_start;
extern const struct app_descriptor __apps_end;
static void start_app(const struct app_descriptor *app);
/* one time setup */
void apps_init(void)
{
const struct app_descriptor *app;
/* call all the init routines */
for (app = &__apps_start; app != &__apps_end; app++) {
if (app->init)
app->init(app);
}
/* start any that want to start on boot */
for (app = &__apps_start; app != &__apps_end; app++) {
if (app->entry && (app->flags & APP_FLAG_DONT_START_ON_BOOT) == 0) {
start_app(app);
}
}
}
static int app_thread_entry(void *arg)
{
const struct app_descriptor *app = (const struct app_descriptor *)arg;
app->entry(app, NULL);
return 0;
}
static void start_app(const struct app_descriptor *app)
{
printf("starting app %s/n", app->name);
thread_resume(thread_create(app->name, &app_thread_entry, (void *)app, DEFAULT_PRIORITY, DEFAULT_STACK_SIZE));
}

至于会有那些 app 被放入 boot thread section, 则定义在 include/app.h 中的 APP_START(appname)

[c-sharp:nogutter] view plain copy print ?

#define APP_START(appname) struct app_descriptor _app_##appname __SECTION(".apps") = { .name = #appname,
#define APP_END };

在 app 中只要像 app/aboot/aboot.c 指定就会在 bootloader bootup 时放入 thread section 中被执行.

[c-sharp:nogutter] view plain copy print ?

APP_START(aboot)
.init = aboot_init,
APP_END

在我的 bootloader 中有 app/aboot/aboot.c, app/tests/tests.c, app/shell/shell.c, 及 app/stringtests/string_tests.c 皆有此声明.

接下来关注: aboot.c 中的 aboot_init()

[c-sharp:nogutter] view plain copy print ?

void aboot_init(const struct app_descriptor *app)
{
unsigned reboot_mode = 0;
unsigned disp_init = 0;
unsigned usb_init = 0;
//test_ram();
/* Setup page size information for nand/emmc reads */
if (target_is_emmc_boot())
{
page_size = 2048;
page_mask = page_size - 1;
}
else
{
page_size = flash_page_size();
page_mask = page_size - 1;
}
/* Display splash screen if enabled */
#if DISPLAY_SPLASH_SCREEN
display_init();
dprintf(INFO, "Diplay initialized/n");
disp_init = 1;
diplay_image_on_screen();
#endif
/* Check if we should do something other than booting up */
if (keys_get_state(KEY_HOME) != 0)
boot_into_recovery = 1;
if (keys_get_state(KEY_BACK) != 0)
goto fastboot;
if (keys_get_state(KEY_CLEAR) != 0)
goto fastboot;
#if NO_KEYPAD_DRIVER
/* With no keypad implementation, check the status of USB connection. */
/* If USB is connected then go into fastboot mode. */
usb_init = 1;
udc_init(&surf_udc_device);
if (usb_cable_status())
goto fastboot;
#endif
init_vol_key();
if(voldown_press())
goto fastboot;
reboot_mode = check_reboot_mode();
if (reboot_mode == RECOVERY_MODE) {
boot_into_recovery = 1;
} else if(reboot_mode == FASTBOOT_MODE) {
goto fastboot;
}
if (target_is_emmc_boot())
{
boot_linux_from_mmc();
}
else
{
recovery_init();
boot_linux_from_flash();
}
dprintf(CRITICAL, "ERROR: Could not do normal boot. Reverting "
"to fastboot mode./n");
fastboot:
if(!usb_init)
udc_init(&surf_udc_device);
fastboot_register("boot", cmd_boot);
if (target_is_emmc_boot())
{
fastboot_register("flash:", cmd_flash_mmc);
fastboot_register("erase:", cmd_erase_mmc);
}
else
{
fastboot_register("flash:", cmd_flash);
fastboot_register("erase:", cmd_erase);
}
fastboot_register("continue", cmd_continue);
fastboot_register("reboot", cmd_reboot);
fastboot_register("reboot-bootloader", cmd_reboot_bootloader);
fastboot_publish("product", TARGET(BOARD));
fastboot_publish("kernel", "lk");
fastboot_init(target_get_scratch_address(), 120 * 1024 * 1024);
udc_start();
target_battery_charging_enable(1, 0);
}

target_is_emmc_boot() is defined in target/init.c: _EMMC_BOOT 是 compiler 时的 flags

[c-sharp:nogutter] view plain copy print ?

__WEAK int target_is_emmc_boot(void)
{
#if _EMMC_BOOT
return 1;
#else
return 0;
#endif
}

check_reboot_mode is defined in target/<your-platform>/init.c:

[c-sharp:nogutter] view plain copy print ?

unsigned check_reboot_mode(void)
{
unsigned restart_reason = 0;
void *restart_reason_addr = 0x401FFFFC;
/* Read reboot reason and scrub it */
restart_reason = readl(restart_reason_addr);
writel(0x00, restart_reason_addr);
return restart_reason;
}

reboot mode in bootloader:

[c-sharp:nogutter] view plain copy print ?

#define RECOVERY_MODE 0x77665502
#define FASTBOOT_MODE 0x77665500

再来就会执行 boot_linux_from_mmc():

[c-sharp:nogutter] view plain copy print ?

int boot_linux_from_mmc(void)
{
struct boot_img_hdr *hdr = (void*) buf;
struct boot_img_hdr *uhdr;
unsigned offset = 0;
unsigned long long ptn = 0;
unsigned n = 0;
const char *cmdline;
uhdr = (struct boot_img_hdr *)EMMC_BOOT_IMG_HEADER_ADDR;
if (!memcmp(uhdr->magic, BOOT_MAGIC, BOOT_MAGIC_SIZE)) {
dprintf(INFO, "Unified boot method!/n");
hdr = uhdr;
goto unified_boot;
}
if(!boot_into_recovery)
{
ptn = mmc_ptn_offset("boot");
if(ptn == 0) {
dprintf(CRITICAL, "ERROR: No boot partition found/n");
return -1;
}
}
else
{
ptn = mmc_ptn_offset("recovery");
if(ptn == 0) {
dprintf(CRITICAL, "ERROR: No recovery partition found/n");
return -1;
}
}
if (mmc_read(ptn + offset, (unsigned int *)buf, page_size)) {
dprintf(CRITICAL, "ERROR: Cannot read boot image header/n");
return -1;
}
if (memcmp(hdr->magic, BOOT_MAGIC, BOOT_MAGIC_SIZE)) {
dprintf(CRITICAL, "ERROR: Invaled boot image header/n");
return -1;
}
if (hdr->page_size && (hdr->page_size != page_size)) {
page_size = hdr->page_size;
page_mask = page_size - 1;
}
offset += page_size;
n = ROUND_TO_PAGE(hdr->kernel_size, page_mask);
if (mmc_read(ptn + offset, (void *)hdr->kernel_addr, n)) {
dprintf(CRITICAL, "ERROR: Cannot read kernel image/n");
return -1;
}
offset += n;
n = ROUND_TO_PAGE(hdr->ramdisk_size, page_mask);
if (mmc_read(ptn + offset, (void *)hdr->ramdisk_addr, n)) {
dprintf(CRITICAL, "ERROR: Cannot read ramdisk image/n");
return -1;
}
offset += n;
unified_boot:
dprintf(INFO, "/nkernel @ %x (%d bytes)/n", hdr->kernel_addr,
hdr->kernel_size);
dprintf(INFO, "ramdisk @ %x (%d bytes)/n", hdr->ramdisk_addr,
hdr->ramdisk_size);
if(hdr->cmdline[0]) {
cmdline = (char*) hdr->cmdline;
} else {
cmdline = DEFAULT_CMDLINE;
}
dprintf(INFO, "cmdline = '%s'/n", cmdline);
dprintf(INFO, "/nBooting Linux/n");
boot_linux((void *)hdr->kernel_addr, (void *)TAGS_ADDR,
(const char *)cmdline, board_machtype(),
(void *)hdr->ramdisk_addr, hdr->ramdisk_size);
return 0;
}

mmc_read() is defined in platform/<your-platform>/mmc.c:

[c-sharp:nogutter] view plain copy print ?

unsigned int mmc_read (unsigned long long data_addr, unsigned int* out, unsigned int data_len)
{
int val = 0;
val = mmc_boot_read_from_card( &mmc_host, &mmc_card, data_addr, data_len, out);
return val;
}