问题描述:
写了一个linux驱动,在probe函数中创建了一个线程,然后线程会一直等待一个全局变量标志位,如果变量为1,就将变量重新置零。然后进行相关的业务逻辑操作。到时在加载完驱动或者是卸载完驱动后,会报一下错误,但是程序没有core dump。
Unable to handle kernel paging request at virtual address bf933218
pgd = c0004000
[bf933218] *pgd=26f20811, *pte=00000000, *ppte=00000000
Internal error: Oops: 80000007 [#1] PREEMPT SMP THUMB2
Modules linked in: xxx_touch(O+) mdrv_crypto sd_mod usb_storage ehci_hcd ntfs vfat msdos fat kdrv_sdmmc mmc_block mmc_core nfsv2 nfs lockd sunrpc grace nls_utf8 cifs [last unloaded: xxx_touch]
CPU: 1 PID: 734 Comm: usb_touch_threa Tainted: G O 4.9.84 #18
Hardware name: SStar Soc (Flattened Device Tree)
task: c721c500 task.stack: c6eca000
PC is at 0xbf933218
LR is at _raw_spin_unlock_irqrestore+0x9/0x2e
pc : [<bf933218>] lr : [<c024e229>] psr: 80080133
sp : c6ecbea8 ip : c6eca03c fp : 00000000
r10: 00000000 r9 : 00000000 r8 : bf93374c
r7 : bf933764 r6 : bf9336ff r5 : c03d5bcc r4 : c6fa0400
r3 : 00001000 r2 : 00000000 r1 : c721c500 r0 : 00000000
Flags: Nzcv IRQs on FIQs on Mode SVC_32 ISA Thumb Segment none
Control: 50c5387d Table: 26a0006a DAC: 00000051
Process usb_touch_threa (pid: 734, stack limit = 0xc6eca218)
Stack: (0xc6ecbea8 to 0xc6ecc000)
bea0: c6ecbeb0 000003e8 00000040 00616161 00000000 00000000
bec0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
bee0: 00000000 00000000 00000000 00000000 00000000 00737373 00000000 00000000
bf00: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
bf20: 00000000 00000000 00000000 00000000 00000000 00040954 c6eca000 c6a6c2c0
bf40: 00000000 c6fa0400 bf93311d c002bdc1 c721c500 00000000 c78cf6c0 c6fa0400
bf60: 00000000 00000000 dead4ead ffffffff ffffffff c6ecbf74 c6ecbf74 00000000
bf80: 00000000 dead4ead ffffffff ffffffff c6ecbf90 c6ecbf90 c6a6c2c0 c002bd21
bfa0: 00000000 00000000 00000000 c000d581 00000000 00000000 00000000 00000000
bfc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
bfe0: 00000000 00000000 00000000 00000000 00000013 00000000 00180104 180c0f00
[<c024e229>] (_raw_spin_unlock_irqrestore) from [<000003e8>] (0x3e8)
Code: bad PC value
---[ end trace c1c4aae0adc938f4 ]---
解决方法
因为是加了一个内核线程之后,才出现这个问题的,经过查看代码,发现在线程中有任务切换的代码,如下代码:
if(flag == 1){
flag = 0;
......
}else{
msleep(1);
schedule();
}
注释掉else分支的代码,不让进行任务切换就好了。
cat /proc/ksyms或者cat /proc/kallsyms 可以查看核心符号表,对与core dump的程序可以通过这个命令查看程序崩溃时的函数调用