weixin_39716971

linux内核oops错误码说明,Linux Kernel Oops异常分析

0．linux内核异常常用分析方法

异常地址是否在0附近，确认是否是空指针解引用问题

异常地址是否在iomem映射区，确认是否是设备访问总线异常问题，如PCI异常导致的地址访问异常

异常地址是否在stack附近，如果相邻，要考虑是否被踩

比较delay reset/nmi watchdog等多种机制打印的栈信息，看看pc是否在动，确定是否是死锁

用SysRq判断是真死还是假死

通过反汇编获得发生异常的C代码段和函数，查找开源社区是否已有补丁修复

下面分别通过PowerPC和Mips64的2个异常例子详细讲解分析过程。

1．PowerPC小系统内核异常分析

1.1 异常打印

Unable to handle kernel paging request for data at address 0x36fef31e

Faulting instruction address: 0xc0088b8c

Oops: Kernel access of bad area, sig: 11 [#1]

PREEMPT SMP NR_CPUS=2

Modules linked in: ossmod tipc ohci_hcd ehci_hcd cmm uart1655x bcm334 bootflash mtdchar bsp_flash_init boardctrl 85xx_debug util

NIP: C0088B8C LR: C0088CF8 CTR: 00000000

REGS: ce283e20 TRAP: 0300 Not tainted (2.6.21.7-EMBSYS-CGEL-3.04.10.P6.F5)

MSR: 00021000 CR: 22004222 XER: 00000000

DAR: 36FEF31E, DSISR: 00800000

TASK = cffdf180[26] 'events/1' THREAD: ce282000 CPU: 1

GPR00: 00100100 CE283ED0 CFFDF180 CF528000 C09EA500 EFFEAD20 CF5188A0 00000000

GPR08: CF5188BC 00200200 36FEF31E D1FD7F9E 22004222 1010DA44 00000290 00000000

GPR16: 1011C858 100147F4 BF9BC9C4 10100000 00000001 C0460000 C06454CC 00000000

GPR24: C0640000 CE282000 C0640000 00000005 00000000 00000000 EFFE8EC0 CFFED958

NIP [C0088B8C] free_block+0xc4/0x16c

LR [C0088CF8] drain_array+0xc4/0x100

Pass 2: Checking directory structure

Pass 3: Checking directory connectivity

Pass 4: Checking reference counts

Call Trace:

[CE283ED0] [C06ABEC0] 0xc06abec0(unreliable)

[CE283EF0] [C0088CF8] drain_array+0xc4/0x100

[CE283F10] [C008A70C] cache_reap+0x94/0x13c

[CE283F30] [C003DA2C] run_workqueue+0xc4/0x198

[CE283F60] [C003E6D4] worker_thread+0x130/0x154

[CE283FB0] [C0042E80] kthread+0xd4/0x110

[CE283FF0] [C0011A70] original_kernel_thread+0x44/0x60

Instruction dump:

5400cffe 0f000000 80c4001c 7d1cf214 3c000010 3d200020 80a8001c 60000100

81660000 61290200 81460004 3906001c <916a0000> 914b0004 90060000 91260004

------------[ cut here ]------------

Badness at c0011e4c [verbose debug info unavailable]

Call Trace:

[CE283C50] [C00080BC] show_stack+0x3c/0x1a0 (unreliable)

[CE283C80] [C018EA28] report_bug+0xb0/0xb8

[CE283C90] [C000EC94] program_check_exception+0xcc/0x4f8

[CE283CD0] [C0010BE4] ret_from_except_full+0x0/0x4c

[CE283D90] [C0640000] 0xc0640000

[CE283DD0] [C000E61C] die+0x1f0/0x27c

[CE283E00] [C0014B18] bad_page_fault+0x98/0xe8

[CE283E10] [C0010A88] handle_page_fault+0x7c/0x80

[CE283ED0] [C06ABEC0] 0xc06abec0

[CE283EF0] [C0088CF8] drain_array+0xc4/0x100

[CE283F10] [C008A70C] cache_reap+0x94/0x13c

[CE283F30] [C003DA2C] run_workqueue+0xc4/0x198

[CE283F60] [C003E6D4] worker_thread+0x130/0x154

[CE283FB0] [C0042E80] kthread+0xd4/0x110

[CE283FF0] [C0011A70] original_kernel_thread+0x44/0x60

1.2 Oops分析

Oops: Kernel access of bad area, sig: 11 [#1]

异常分类

Oops：内核态指令异常；

BUG：内核检测到逻辑异常(类似于assert)，会影响内核的后续运行；

WARNING：类似于BUG，但是不会影响内核的后续运行；

PANIC：类似于BUG，系统不能继续运行，直接挂起或重启；

SOFTLOCK：长时间任务得不到调度；

异常信号

Signal

Code

Default Action

Description

SIGABRT

Process abort signal

SIGALRM

Alarm clock

SIGBUS

Access to an undefined portion of a memory object

SIGCHLD

I - Ignore the Signal

Child process terminated, stopped,

SIGCONT

C - Continue the process

Continue executing, if stopped.

SIGFPE

Erroneous arithmetic operation.

SIGHUP

Hangup.

SIGILL

Illegal instruction.

SIGINT

Terminal interrupt signal.

SIGKILL

Kill (cannot be caught or ignored).

SIGPIPE

T - Abnormal termination of the process

Write on a pipe with no one to read it.

SIGQUIT

A - Abnormal termination of the process

Terminal quit signal.

SIGSEGV

Invalid memory reference.

SIGSTOP

S - Stop the process

Stop executing (cannot be caught or ignored).

SIGTERM

Termination signal.

SIGTSTP

Terminal stop signal.

SIGTTIN

Background process attempting read.

SIGTTOU

Background process attempting write.

SIGUSR1

User-defined signal 1.

SIGUSR2

User-defined signal 2.

SIGPOLL

Pollable event.

SIGPROF

Profiling timer expired.

SIGSYS

Bad system call.

SIGTRAP

Trace/breakpoint trap.

SIGURG

High bandwidth data is available at a socket.

SIGVTALRM

Virtual timer expired.

SIGXCPU

CPU time limit exceeded.

SIGXFSZ

File size limit exceeded

Default Actions:

T - Abnormal termination of the process. The process is terminated with all the consequences of _exit() except that the status made available to wait() and waitpid() indicates abnormal termination by the specified signal.

A - Abnormal termination of the process. Additionally, implementation-defined abnormal termination actions, such as creation of a core file, may occur.

I - Ignore the signal.

S - Stop the process.

C - Continue the process, if it is stopped; otherwise, ignore the signal.

具体针对powerpc e500内核，异常与信号的对应关系如下：

所以有进程访问了超出其虚拟地址空间的地址，内核报SIGSEGV(segment fault)信号。

那是什么进程呢？

其他

#1，die_counter，表示Oops发生的次数，一般来说，如果有多条Oops，看第一条Oops信息，因为后面的Oops可能是第一条Oops的错误传播导致的。

1.3 寄存器分析

NIP: C0088B8C LR: C0088CF8 CTR: 00000000

NIP是next instruction pointer，值就是当前指令的地址。这里列出了3个寄存器的值。

LR是link register其值为上一条指令的地址。

CTR是count register，其值用于循环指令。

REGS: ce283e20 TRAP: 0300 Not tainted (2.6.21.7-EMBSYS-CGEL-3.04.10.P6.F5)

TRAP ：异常处理函数入口地址；REGS ：系统栈pt_regs的基址。pt_regs这个结构封装了需要在内核入口中保存的最少的状态信息。比如说每一次的系统调用、中断、陷阱、故障。

0x100: "(System Reset)"

0x200: "(Machine Check)"

0x300: "(Data Access)"

0x380: "(Data SLB Access)"

0x400: "(Instruction Access)"

0x480: "(Instruction SLB Access)"

0x500: "(Hardware Interrupt)"

0x600: "(Alignment)"

0x700: "(Program Check)"

0x800: "(FPU Unavailable)"

0x900: "(Decrementer)"

0xc00: "(System Call)"

0xd00: "(Single Step)"

0xf00: "(Performance Monitor)"

0xf20: "(Altivec Unavailable)"

0x1300: "(Instruction Breakpoint)"

详细解释见《PowerPC™ e500 Core Family Reference Manual》“5.7 Interrupt Definitions”。

tainted ：内核错误信息，由add_taint设置，解释如下：

* 'P' - Proprietary module has been loaded.

* 'F' - Module has been forcibly loaded.

* 'S' - SMP with CPUs not designed for SMP.

* 'R' - User forced a module unload.

* 'M' - System experienced a machine check exception.

* 'B' - System has hit bad_page.

* 'U' - Userspace-defined naughtiness.

* 'D' - Kernel has oopsed before

* 'A' - ACPI table overridden.

* 'W' - Taint on warning.

* 'C' - modules from drivers/staging are loaded.

MSR: 00021000 CR: 22004222 XER: 00000000

DAR: 36FEF31E, DSISR: 00800000

MSR是machine state register；

CR是condition register；

XER为Integer Exception Register

DAR为data address register，其值为造成了内存访问异常的地址。E500中为Data Exception Address Register (DEAR)

DSISR为Data Storage Interrupt Status Register，是存储着发生内存访问异常原因的寄存器。E500中为Exception Syndrome Register (ESR)。0x00800000表示Store operation中的Alignment, data storage, data TLB error异常。

TASK = cffdf180[26] 'events/1' THREAD: ce282000 CPU: 1

cffdf180：进程task_struct结构体的地址；

26：进程号；

events/1：进程名；

THREAD：进程的内核栈起始地址；

CPU：当前CPU；

当前进程也就是'events/1进程，出现SIGSEGV异常了。

GPR00: 00100100 CE283ED0 CFFDF180 CF528000 C09EA500 EFFEAD20 CF5188A0 00000000

GPR08: CF5188BC 00200200 36FEF31E D1FD7F9E 22004222 1010DA44 00000290 00000000

GPR16: 1011C858 100147F4 BF9BC9C4 10100000 00000001 C0460000 C06454CC 00000000

GPR24: C0640000 CE282000 C0640000 00000005 00000000 00000000 EFFE8EC0 CFFED958

PowerPC的ABI规定的寄存器的使用规则如下：

(1)GPR0：属于易失性寄存器，ABI规定普通用户不能使用此寄存器。GCC编译器用此寄存器来保存LR寄存器，Linux PowerPC用此寄存器来传递系统调用号码。

(2)GPR1：属于专用寄存器，ABI规定用次寄存器来保存堆栈的栈顶指针。

(3)GPR2：属于专用寄存器，ABI规定普通用户不使用才寄存器，Linux PowerPC用此寄存器来保存当前进程的进程描述符地址。

(4)GPR3-GPR4：属于易失性寄存器，ABI使用这两个寄存器来保存函数的返回值，或者用来传递参数。

(5)GPR5-GPR10：也属于易失性寄存器，加上GPR3和GPR4共8个寄存器用来传递函数的参数。当函数的参数超过八个时使用堆栈来传递。

(6)GPR11-GPR12：属于易失性寄存器，ABI规定普通用户不使用该寄存器，Linux PowerPC有时用这两个寄存器来存放临时变量，但是GCC编译器没有使用这两个寄存器。

(7)GPR13：属于专用寄存器，ABI规定该寄存器sdata段的基地址指针。Linux PowerPC在系统初始化时使用该寄存器来存放临时变量。GCC有时会根据某些规则将一些常用的数据放入sdata或者sbss段中。应用程序对sdata或者sbss段数据的访问与对data和bss段数据的访问机制不同，访问sdata段的数据速度更快。

(8)GPR14-GPR31：属于非易失性寄存器。ABI使用这些寄存器来存放一些临时变量，在应用程序中可以自由使用这些变量。

1.4 调用栈分析

调用链

NIP [C0088B8C] free_block+0xc4/0x16c

LR [C0088CF8] drain_array+0xc4/0x100

Call Trace:

[CE283ED0] [C06ABEC0] 0xc06abec0(unreliable)

[CE283EF0] [C0088CF8] drain_array+0xc4/0x100

[CE283F10] [C008A70C] cache_reap+0x94/0x13c

[CE283F30] [C003DA2C] run_workqueue+0xc4/0x198

[CE283F60] [C003E6D4] worker_thread+0x130/0x154

[CE283FB0] [C0042E80] kthread+0xd4/0x110

[CE283FF0] [C0011A70] original_kernel_thread+0x44/0x60

Instruction dump:

5400cffe 0f000000 80c4001c 7d1cf214 3c000010 3d200020 80a8001c 60000100

81660000 61290200 81460004 3906001c <916a0000> 914b0004 90060000 91260004

[CE283FB0] [C0042E80] kthread+0xd4/0x110

CE283FB0：栈地址；

C0042E80：栈上保存的LR值，即函数返回地址。

kthread：函数名；

0xd4/0x110：异常指令偏移/调用函数长度。

static void free_block(struct kmem_cache *cachep, void **objpp, int nr_objects, int node)

从调用栈上看，内核在drain_array中调用free_block出现异常，查看free_block原型，对比入栈参数(CF528000 C09EA500 EFFEAD20 CF5188A0)，可以发现int nr_objects, int node明显异常，可能推断调用栈可能已经被踩。

指令码

Instruction dump:

5400cffe 0f000000 80c4001c 7d1cf214 3c000010 3d200020 80a8001c 60000100

81660000 61290200 81460004 3906001c <916a0000> 914b0004 90060000 91260004

Instruction dump打印出NIP附近的指令字节码。其中<916a0000>为NIP的指令码。

反汇编定位

objump -dS vmlinux > /tmp/kernel.s

通过查找<916a0000>对应的C代码，确定具体那句C代码出现异常。

其中vmlinux为已打开调试信息的，与故障相同版本的内核镜像。

2．MIPS小系统内核异常分析

2.1 异常打印

0:Oops[#1]:

0:Cpu 0

0:Show thread info from vcpu 0

0: VCPU Stack bottom Task Ti at

0: 0 c000000595057fe0 swapper c000000595054000

0:Thread info( c000000595054000 ):

0: Process swapper (pid: 1)

0: exec_domain ffffffffc0f299b0

0: flags 100000

0: tp_value 0

0: cpu 0

0: preempt_count 2

0: regs (null)

0:STACK_END_MAGIC at va( c000000595054068 ): 57AC6E9D( =? 57AC6E9D)

0:$ 0 : 0: 0000000000000000 0: 0000000000000000 0: 0000000000000000 0: 0000000000000001 0:

0:$ 4 : 0: 0000000000000000 0: 0000000000000000 0: ffffffffffffffff 0: 0000000000002976 0:

0:$ 8 : 0: 0000000000007fff 0: 000000000000000a 0: 5f73746172747570 0: 000000000000006c 0:

0:$12 : 0: 0000000000000068 0: 000000000000004c 0: ffffffffc10bc384 0: c000000593338000 0:

0:$16 : 0: 0000000000000000 0: ffffffffc10e42b8 0: ffffffffc10e0000 0: ffffffffc10e0000 0:

0:$20 : 0: 0000000000000000 0: 0000000000000080 0: 0000000000000080 0: 0000000000000000 0:

0:$24 : 0: 0000000000000006 0: ffffffffc06501a8 0: 0: 0:

0:$28 : 0: c000000595054000 0: c000000595057c88 0: 0000000000000000 0: ffffffffc087bf40 0:

0:Hi : 0000000000000000

0:Lo : 0000000000000000

0:epc : ffffffffc087c4b4 _bcore_cleanup+0x34/0x190

0: Not tainted

0:ra : ffffffffc087bf40 _init+0x3e8/0x480

0:Status: 5400ffe3 0:KX 0:SX 0:UX 0:KERNEL 0:EXL 0:IE 0:

0:Cause : 00800008

0:BadVA : 0000000000000008

0:PrId : 000c1102 (XLP316 A2 )

0:Modules linked in: 0:

0:Process swapper (pid: 1, threadinfo=c000000595054000, task=c000000595053898, tls=0000000000000000)

0:Stack : 0: ffffffffffffffff 0: ffffffffc10e0000 0: c000000595193240 0: 0000000000000001 0:

0: ffffffffc104365c 0: ffffffffc087bf40 0: 000001fac104365c 0: ffffffffc087cb30 0:

0: ffffffffc087c3a8 0: 0000000000000000 0: ffffffffc0f4a778 0: c000000595193000 0:

0: c000000595193240 0: 0000000000000001 0: ffffffffc10e0000 0: c000000595193240 0:

0: 0000000000000001 0: ffffffffc104365c 0: 0000000000000000 0: 0000000000000080 0:

0: 0000000000000080 0: ffffffffc1043c44 0: 00008a17bc300000 0: ffffffffc10e0000 0:

0: c00000059333dd40 0: 0000000000000000 0: 3800000000000000 0: 0000000000000000 0:

0: 000000009333dd40 0: ffffffffc1043638 0: 000000005400ffe0 0: ffffffffbfff00fe 0:

0: ffffffffc1070000 0: ffffffffc1063200 0: 0000000000000001 0: ffffffffc104365c 0:

0: 0000000000000000 0: 0000000000000080 0: 0000000000000080 0: 0000000000000000 0:

0: ... 0:

0:Call Trace: [jiffies: 0xfffff79f]

0:[] _bcore_cleanup+0x34/0x190

0:[] _init+0x3e8/0x480

0:[] bcmxgs_init_module+0x5e8/0xc00

0:[] do_one_initcall+0x3c/0x1a0

0:[] kernel_init+0x220/0x2b8

0:[] kernel_thread_helper+0x10/0x20

Code: 0: ffbf0028 0: 0000802d 0: 663142b8 0:0: 0040f809 0: 00000000 0: 0202102a 0: 1040001d 0: 00000000

0:<4>Disabling lock debugging due to kernel taint

2.2 异常信号

异常与信号之间的关系：

2.3 线程信息分析

0:Cpu 0：这2个0为当前CPU核ID；

0:Show thread info from vcpu 0

0: VCPU Stack bottom Task Ti at

0: 0 c000000595057fe0 swapper c000000595054000

VCPU：CPU核；

Stack bottom：栈底指针；

Task：线程名；

Ti at：线程thread_info结构体指针；

0:Thread info( c000000595054000 ):

0: Process swapper (pid: 1)

0: exec_domain ffffffffc0f299b0

0: flags 100000

0: tp_value 0

0: cpu 0

0: preempt_count 2

0: regs (null)

0:STACK_END_MAGIC at va( c000000595054068 ): 57AC6E9D( =? 57AC6E9D)

flags ：线程标志位，具体标记如下表。此时值为TIF_FIXADE，表示有address errors。Thread info( c000000595054000 )：产生异常的线程信息；下面的字段为thread_info结构体中的字段信息。其中，

preempt_count：为抢占计数。为0时，内核可以安全的执行抢占此线程。不为0，表示当前进程持有锁不能释放CPU控制权(不能被抢占)。

STACK_END_MAGIC：栈底部的魔幻数，可以辅助判断栈是否被踩。

#define TIF_SIGPENDING1/* signal pending */

#define TIF_NEED_RESCHED2/* rescheduling necessary */

#define TIF_SYSCALL_AUDIT3/* syscall auditing active */

#define TIF_SECCOMP4/* secure computing */

#define TIF_NOTIFY_RESUME5/* callback before returning to user */

#define TIF_RESTORE_SIGMASK9/* restore signal mask in do_signal() */

#define TIF_USEDFPU16/* FPU was used by this task this quantum (SMP) */

#define TIF_POLLING_NRFLAG17/* true if poll_idle() is polling TIF_NEED_RESCHED */

#define TIF_MEMDIE18

#define TIF_FREEZE19

#define TIF_FIXADE20/* Fix address errors in software */

#define TIF_LOGADE21/* Log address errors to syslog */

#define TIF_32BIT_REGS22/* also implies 16/32 fprs */

#define TIF_32BIT_ADDR23/* 32-bit address space (o32/n32) */

#define TIF_FPUBOUND24/* thread bound to FPU-full CPU set */

#define TIF_LOAD_WATCH25/* If set, load watch registers */

#define TIF_XKPHYS_MEM_EN26

#define TIF_XKPHYS_IO_EN27

#define TIF_SYSCALL_TRACE31/* syscall trace active */

2.4 寄存器分析

0:$ 0 : 0: 0000000000000000 0: 0000000000000000 0: 0000000000000000 0: 0000000000000001 0:

0:$ 4 : 0: 0000000000000000 0: 0000000000000000 0: ffffffffffffffff 0: 0000000000002976 0:

0:$ 8 : 0: 0000000000007fff 0: 000000000000000a 0: 5f73746172747570 0: 000000000000006c 0:

0:$12 : 0: 0000000000000068 0: 000000000000004c 0: ffffffffc10bc384 0: c000000593338000 0:

0:$16 : 0: 0000000000000000 0: ffffffffc10e42b8 0: ffffffffc10e0000 0: ffffffffc10e0000 0:

0:$20 : 0: 0000000000000000 0: 0000000000000080 0: 0000000000000080 0: 0000000000000000 0:

0:$24 : 0: 0000000000000006 0: ffffffffc06501a8 0: 0: 0:

0:$28 : 0: c000000595054000 0: c000000595057c88 0: 0000000000000000 0: ffffffffc087bf40 0:

0:Hi : 0000000000000000

0:Lo : 0000000000000000

0:epc : ffffffffc087c4b4 _bcore_cleanup+0x34/0x190

0: Not tainted

0:ra : ffffffffc087bf40 _init+0x3e8/0x480

0:Status: 5400ffe3 0:KX 0:SX 0:UX 0:KERNEL 0:EXL 0:IE 0:

0:Cause : 00800008

0:BadVA : 0000000000000008

0:PrId : 000c1102 (XLP316 A2 )

Mips核心寄存器组有4组，分别是GP, COP0, COP1, COP2。

其中COP0几个重要的寄存器解释如下：

Status：c0p0状态cp0_status。其中EXL标示在异常模式中，具体解释请参照《参考资料6.7 第193页》

Cause：00800008，标示 TLB exception(load or instruction fetch)

BadVA：产生异常的虚拟地址，如地址错误、无效的TLB，TLB modified等等。

2.5 调用栈分析