内核异常分析方法

1.通过sysdump/ramdump或者串口抓取内核日记

展讯平台sysdump解析_Android开发-CSDN博客_sysdump

高通平台抓取ramdump并用qcap解析_Android开发-CSDN博客_高通ramdump

mtk平台使用GAT工具分析aee db文件

对于深度休眠后，出现死机情况，测试前需打开串口打印(默认休眠后会关闭串口)

echo 0 > /sys/module/printk/parameters/console_suspend或者直接修改程序

一般还需要调整默认的串口打印等级echo 8 > /proc/sys/kernel/printk

bool console_suspend_enabled = true;
EXPORT_SYMBOL(console_suspend_enabled);

static int __init console_suspend_disable(char *str)
{
	console_suspend_enabled = false;
	return 1;
}
__setup("no_console_suspend", console_suspend_disable);
module_param_named(console_suspend, console_suspend_enabled,
		bool, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(console_suspend, "suspend console during suspend"
	" and hibernate operations");

 出现深度休眠后死机的情况，有如下的情况

晶体类型配置出错，导致深度休眠后，器件掉电

ddr一致性差，需要降频

空指针问题(驱动没适配好)

2.通过工具分析出错原因

写的一个空指针demo

运行代码,通过串口或者sysdump抓取内核日记

2.1通过arm-eabi-addr2line(将地址转换为行号)

从上面可以看出PC的地址是0xc04e8394,获取出错行号

./arm-eabi-addr2line -e vmlinux 0xc04e8394
/home/w/AndroidP/kernel/drivers/misc/suspend_lock.c:15

2.2通过arm-eabi-objdump反汇编

./arm-eabi-objdump -S vmlinux > vmlinux.txt

grep -rin c04e8394 vmlinux.txt 
1315671:c04e8394:	e5843000 	str	r3, [r4]

2.3通过arm-eabi-gdb定位出错位置

w@server:/home/w/AndroidP$ ./arm-eabi-gdb out/target/product/sp9850ka_1h10_go/obj/KERNEL/vmlinux
GNU gdb (GDB) 7.6
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "--host=x86_64-linux-gnu --target=arm-linux-android".
For bug reporting instructions, please see:
<http://source.android.com/source/report-bugs.html>...
Reading symbols from /home/w/AndroidP/out/target/product/sp9850ka_1h10_go/obj/KERNEL/vmlinux...done.
(gdb) list *0xc04e8394
0xc04e8394 is in null_pointer (/home/w/AndroidP/kernel/drivers/misc/suspend_lock.c:15).
10	static struct miscdevice poll_dev;
11	
12	ssize_t null_pointer(struct device *dev, struct device_attribute *attr,char *buf)
13	{
14		char *p=NULL;
15		memcpy(p, "test", 4);
16		printk("%s\n",__FUNCTION__);
17		return 0;
18	}

如不在代码上运行，结果如下

./arm-eabi-gdb vmlinux
(gdb) list *0xc04e8394
0xc04e8394 is in null_pointer (/home/w/AndroidP/kernel/drivers/misc/suspend_lock.c:15).
10	/home/w/AndroidP/kernel/drivers/misc/suspend_lock.c: 没有那个文件或目录

 不初始化,然后使用互斥锁mutex

 运行后,定屏,触摸没反应(cat /proc/interrupts没看到tp的中断),adb能连上手机,dmesg查看内核缓冲区log,如下图

 根据pc位置调到指定的代码

(gdb) list *0xc0953844
0xc0953844 is in _raw_spin_lock (/home/w/AndroidP/kernel/arch/arm/include/asm/spinlock.h:77).
72		: "r" (&lock->slock), "I" (1 << TICKET_SHIFT)
73		: "cc");
74	
75		while (lockval.tickets.next != lockval.tickets.owner) {
76			wfe();
77			lockval.tickets.owner = ACCESS_ONCE(lock->tickets.owner);
78		}
79	
80		smp_mb();
81	}

3.测试DDR稳定性

Andoird中使用stressapptest测试DDR稳定性_Android开发-CSDN博客_android ddr测试，实际效果不好

或者使用相应平台的DDR稳定性测试工具。

4.更换器件重新测试

对应器件的虚焊，可采取重焊测试。如果还不行，也可以更换器件测试。