Linux Kernel oops

資料來源: Linux Kernel oops

每次kernel掛掉總是會出現一堆16進制的東西
但我都沒有認真看過
因為….真的看不懂
不過上星期上完課之後
終於知道這些數字代表什麼

例如main.c裡面有個faulty_write function, 故意讓0的位置寫0

ssize_t faulty_write (struct file *filp, const char __user *buf, size_t count, loff_t *pos) { /* make a simple fault by dereferencing a NULL pointer */ *(int *)0 = 0; return 0; }

當我去執行的時候kernel掛掉

$ echo “1″ > /dev/aaa Unable to handle kernel NULL pointer dereference at virtual address 00000000 pgd = c3a70000 [00000000] *pgd=33a7e031, *pte=00000000, *ppte=00000000 Internal error: Oops: 817 [#3] Modules linked in: faulty CPU: 0 Tainted: G D (2.6.29.2 #2)PC is at faulty_write+0×10/0×18 [faulty] LR is at vfs_write+0xc4/0×148 pc : [<bf00009c>] lr : [<c00a5f6c>] psr: a0000013 sp : c3a69f44 ip : c3a69f54 fp : c3a69f50 r10: 400d0948 r9 : c3a68000 r8 : 00000000 r7 : 00000002 r6 : c3a69f78 r5 : 400d3704 r4 : c382cea0 r3 : c3a69f78 r2 : 00000002 r1 : 400d3704 r0 : 00000000 Flags: NzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment user Control: 0000717f Table: 33a70000 DAC: 00000015 Process echo (pid: 882, stack limit = 0xc3a68260) Stack: (0xc3a69f44 to 0xc3a6a000) 9f40: c3a69f74 c3a69f54 c00a5f6c bf00009c 00000000 c382cec0 c382cea0 9f60: c3a69f78 00000000 c3a69fa4 c3a69f78 c00a60b0 c00a5eb8 00000000 00000000 9f80: 00000000 400d1188 00000002 400d3704 00000004 c0035fa4 00000000 c3a69fa8 9fa0: c0035e00 c00a6074 400d1188 00000002 00000001 400d3704 00000002 4009f9b4 9fc0: 400d1188 00000002 400d3704 00000002 000a8374 00000002 400d0948 00000000 9fe0: 400a491c bec33d74 4009f008 4008accc 20000010 00000001 fffff7f7 fff5ffdf Backtrace: [<bf00008c>] (faulty_write+0×0/0×18 [faulty]) from [<c00a5f6c>] (vfs_write+0xc4/0×148) [<c00a5ea8>] (vfs_write+0×0/0×148) from [<c00a60b0>] (sys_write+0×4c/0×74) r7:00000000 r6:c3a69f78 r5:c382cea0 r4:c382cec0 [<c00a6064>] (sys_write+0×0/0×74) from [<c0035e00>] (ret_fast_syscall+0×0/0×2c) r8:c0035fa4 r7:00000004 r6:400d3704 r5:00000002 r4:400d1188Code: e1a0c00d e92dd800 e24cb004 e3a00000 (e5800000)  —[ end trace 0f24d785ce6e74e6 ]— Segmentation fault / $

注意看裡面

PC is at faulty_write+0×10/0×18 [faulty] Code: e1a0c00d e92dd800 e24cb004 e3a00000 (e5800000)

因為instruction length是4 bytefaulty_write+0×10/0×18 = > 4 / 6 (6行的第四行錯)
在位置e3a00000 的地方

執行mipsel-linux-objdump -S main.o 可以看到組語和程式碼的對應

ssize_t faulty_write (struct file *filp, const char __user *buf, size_t count, loff_t *pos) { 8c:     e1a0c00d mov ip, sp 90:     e92dd800 stmdb sp!, {fp, ip, lr, pc} 94:     e24cb004 sub fp, ip, #4 ; 0×4 /* make a simple fault by dereferencing a NULL pointer */ *(int *)0 = 0; 98:     e3a00000 mov r0, #0 ; 0×0 9c:     e5800000 str r0, [r0] return 0; }

所以這樣就可以找到哪一個錯誤!