1 NOTE: ksymoops is useless on 2.6. Please use the Oops in its original format
2 (from dmesg, etc). Ignore any references in this or other docs to "decoding
3 the Oops" or "running it through ksymoops". If you post an Oops from 2.6 that
4 has been run through ksymoops, people will just tell you to repost it.
9 Find the Oops and send it to the maintainer of the kernel area that seems to be
10 involved with the problem. Don't worry too much about getting the wrong person.
11 If you are unsure send it to the person responsible for the code relevant to
12 what you were doing. If it occurs repeatably try and describe how to recreate
13 it. That's worth even more than the oops.
15 If you are totally stumped as to whom to send the report, send it to
16 linux-kernel@vger.kernel.org. Thanks for your help in making Linux as
17 stable as humanly possible.
20 ----------------------
22 Normally the Oops text is read from the kernel buffers by klogd and
23 handed to syslogd which writes it to a syslog file, typically
24 /var/log/messages (depends on /etc/syslog.conf). Sometimes klogd dies,
25 in which case you can run dmesg > file to read the data from the kernel
26 buffers and save it. Or you can cat /proc/kmsg > file, however you
27 have to break in to stop the transfer, kmsg is a "never ending file".
28 If the machine has crashed so badly that you cannot enter commands or
29 the disk is not available then you have three options :-
31 (1) Hand copy the text from the screen and type it in after the machine
32 has restarted. Messy but it is the only option if you have not
33 planned for a crash. Alternatively, you can take a picture of
34 the screen with a digital camera - not nice, but better than
35 nothing. If the messages scroll off the top of the console, you
36 may find that booting with a higher resolution (eg, vga=791)
37 will allow you to read more of the text. (Caveat: This needs vesafb,
38 so won't help for 'early' oopses)
40 (2) Boot with a serial console (see Documentation/serial-console.txt),
41 run a null modem to a second machine and capture the output there
42 using your favourite communication program. Minicom works well.
44 (3) Use Kdump (see Documentation/kdump/kdump.txt),
45 extract the kernel ring buffer from old memory with using dmesg
46 gdbmacro in Documentation/kdump/gdbmacros.txt.
52 NOTE: the message from Linus below applies to 2.4 kernel. I have preserved it
53 for historical reasons, and because some of the information in it still
54 applies. Especially, please ignore any references to ksymoops.
56 From: Linus Torvalds <torvalds@osdl.org>
58 How to track down an Oops.. [originally a mail to linux-kernel]
60 The main trick is having 5 years of experience with those pesky oops
63 Actually, there are things you can do that make this easier. I have two
66 gdb /usr/src/linux/vmlinux
67 gdb> disassemble <offending_function>
69 That's the easy way to find the problem, at least if the bug-report is
70 well made (like this one was - run through ksymoops to get the
71 information of which function and the offset in the function that it
74 Oh, it helps if the report happens on a kernel that is compiled with the
75 same compiler and similar setups.
77 The other thing to do is disassemble the "Code:" part of the bug report:
78 ksymoops will do this too with the correct tools, but if you don't have
79 the tools you can just do a silly program:
81 char str[] = "\xXX\xXX\xXX...";
84 and compile it with gcc -g and then do "disassemble str" (where the "XX"
85 stuff are the values reported by the Oops - you can just cut-and-paste
86 and do a replace of spaces to "\x" - that's what I do, as I'm too lazy
87 to write a program to automate this all).
89 Alternatively, you can use the shell script in scripts/decodecode.
90 Its usage is: decodecode < oops.txt
92 The hex bytes that follow "Code:" may (in some architectures) have a series
93 of bytes that precede the current instruction pointer as well as bytes at and
94 following the current instruction pointer. In some cases, one instruction
95 byte or word is surrounded by <> or (), as in "<86>" or "(f00d)". These
96 <> or () markings indicate the current instruction pointer. Example from
97 i386, split into multiple lines for readability:
99 Code: f9 0f 8d f9 00 00 00 8d 42 0c e8 dd 26 11 c7 a1 60 ea 2b f9 8b 50 08 a1
100 64 ea 2b f9 8d 34 82 8b 1e 85 db 74 6d 8b 15 60 ea 2b f9 <8b> 43 04 39 42 54
101 7e 04 40 89 42 54 8b 43 04 3b 05 00 f6 52 c0
103 Finally, if you want to see where the code comes from, you can do
106 make fs/buffer.s # or whatever file the bug happened in
108 and then you get a better idea of what happens than with the gdb
111 Now, the trick is just then to combine all the data you have: the C
112 sources (and general knowledge of what it _should_ do), the assembly
113 listing and the code disassembly (and additionally the register dump you
114 also get from the "oops" message - that can be useful to see _what_ the
115 corrupted pointers were, and when you have the assembler listing you can
116 also match the other registers to whatever C expressions they were used
119 Essentially, you just look at what doesn't match (in this case it was the
120 "Code" disassembly that didn't match with what the compiler generated).
121 Then you need to find out _why_ they don't match. Often it's simple - you
122 see that the code uses a NULL pointer and then you look at the code and
123 wonder how the NULL pointer got there, and if it's a valid thing to do
124 you just check against it..
126 Now, if somebody gets the idea that this is time-consuming and requires
127 some small amount of concentration, you're right. Which is why I will
128 mostly just ignore any panic reports that don't have the symbol table
129 info etc looked up: it simply gets too hard to look it up (I have some
130 programs to search for specific patterns in the kernel code segment, and
131 sometimes I have been able to look up those kinds of panics too, but
132 that really requires pretty good knowledge of the kernel just to be able
133 to pick out the right sequences etc..)
135 _Sometimes_ it happens that I just see the disassembled code sequence
136 from the panic, and I know immediately where it's coming from. That's when
137 I get worried that I've been doing this for too long ;-)
142 ---------------------------------------------------------------------------
143 Notes on Oops tracing with klogd:
145 In order to help Linus and the other kernel developers there has been
146 substantial support incorporated into klogd for processing protection
147 faults. In order to have full support for address resolution at least
148 version 1.3-pl3 of the sysklogd package should be used.
150 When a protection fault occurs the klogd daemon automatically
151 translates important addresses in the kernel log messages to their
152 symbolic equivalents. This translated kernel message is then
153 forwarded through whatever reporting mechanism klogd is using. The
154 protection fault message can be simply cut out of the message files
155 and forwarded to the kernel developers.
157 Two types of address resolution are performed by klogd. The first is
158 static translation and the second is dynamic translation. Static
159 translation uses the System.map file in much the same manner that
160 ksymoops does. In order to do static translation the klogd daemon
161 must be able to find a system map file at daemon initialization time.
162 See the klogd man page for information on how klogd searches for map
165 Dynamic address translation is important when kernel loadable modules
166 are being used. Since memory for kernel modules is allocated from the
167 kernel's dynamic memory pools there are no fixed locations for either
168 the start of the module or for functions and symbols in the module.
170 The kernel supports system calls which allow a program to determine
171 which modules are loaded and their location in memory. Using these
172 system calls the klogd daemon builds a symbol table which can be used
173 to debug a protection fault which occurs in a loadable kernel module.
175 At the very minimum klogd will provide the name of the module which
176 generated the protection fault. There may be additional symbolic
177 information available if the developer of the loadable module chose to
178 export symbol information from the module.
180 Since the kernel module environment can be dynamic there must be a
181 mechanism for notifying the klogd daemon when a change in module
182 environment occurs. There are command line options available which
183 allow klogd to signal the currently executing daemon that symbol
184 information should be refreshed. See the klogd manual page for more
187 A patch is included with the sysklogd distribution which modifies the
188 modules-2.0.0 package to automatically signal klogd whenever a module
189 is loaded or unloaded. Applying this patch provides essentially
190 seamless support for debugging protection faults which occur with
191 kernel loadable modules.
193 The following is an example of a protection fault in a loadable module
195 ---------------------------------------------------------------------------
196 Aug 29 09:51:01 blizard kernel: Unable to handle kernel paging request at virtual address f15e97cc
197 Aug 29 09:51:01 blizard kernel: current->tss.cr3 = 0062d000, %cr3 = 0062d000
198 Aug 29 09:51:01 blizard kernel: *pde = 00000000
199 Aug 29 09:51:01 blizard kernel: Oops: 0002
200 Aug 29 09:51:01 blizard kernel: CPU: 0
201 Aug 29 09:51:01 blizard kernel: EIP: 0010:[oops:_oops+16/3868]
202 Aug 29 09:51:01 blizard kernel: EFLAGS: 00010212
203 Aug 29 09:51:01 blizard kernel: eax: 315e97cc ebx: 003a6f80 ecx: 001be77b edx: 00237c0c
204 Aug 29 09:51:01 blizard kernel: esi: 00000000 edi: bffffdb3 ebp: 00589f90 esp: 00589f8c
205 Aug 29 09:51:01 blizard kernel: ds: 0018 es: 0018 fs: 002b gs: 002b ss: 0018
206 Aug 29 09:51:01 blizard kernel: Process oops_test (pid: 3374, process nr: 21, stackpage=00589000)
207 Aug 29 09:51:01 blizard kernel: Stack: 315e97cc 00589f98 0100b0b4 bffffed4 0012e38e 00240c64 003a6f80 00000001
208 Aug 29 09:51:01 blizard kernel: 00000000 00237810 bfffff00 0010a7fa 00000003 00000001 00000000 bfffff00
209 Aug 29 09:51:01 blizard kernel: bffffdb3 bffffed4 ffffffda 0000002b 0007002b 0000002b 0000002b 00000036
210 Aug 29 09:51:01 blizard kernel: Call Trace: [oops:_oops_ioctl+48/80] [_sys_ioctl+254/272] [_system_call+82/128]
211 Aug 29 09:51:01 blizard kernel: Code: c7 00 05 00 00 00 eb 08 90 90 90 90 90 90 90 90 89 ec 5d c3
212 ---------------------------------------------------------------------------
214 Dr. G.W. Wettstein Oncology Research Div. Computing Facility
215 Roger Maris Cancer Center INTERNET: greg@wind.rmcc.com
221 ---------------------------------------------------------------------------
224 Some oops reports contain the string 'Tainted: ' after the program
225 counter. This indicates that the kernel has been tainted by some
226 mechanism. The string is followed by a series of position-sensitive
227 characters, each representing a particular tainted value.
229 1: 'G' if all modules loaded have a GPL or compatible license, 'P' if
230 any proprietary module has been loaded. Modules without a
231 MODULE_LICENSE or with a MODULE_LICENSE that is not recognised by
232 insmod as GPL compatible are assumed to be proprietary.
234 2: 'F' if any module was force loaded by "insmod -f", ' ' if all
235 modules were loaded normally.
237 3: 'S' if the oops occurred on an SMP kernel running on hardware that
238 hasn't been certified as safe to run multiprocessor.
239 Currently this occurs only on various Athlons that are not
242 4: 'R' if a module was force unloaded by "rmmod -f", ' ' if all
243 modules were unloaded normally.
245 5: 'M' if any processor has reported a Machine Check Exception,
246 ' ' if no Machine Check Exceptions have occurred.
248 6: 'B' if a page-release function has found a bad page reference or
249 some unexpected page flags.
251 7: 'U' if a user or user application specifically requested that the
252 Tainted flag be set, ' ' otherwise.
254 8: 'D' if the kernel has died recently, i.e. there was an OOPS or BUG.
256 9: 'A' if the ACPI table has been overridden.
258 10: 'W' if a warning has previously been issued by the kernel.
259 (Though some warnings may set more specific taint flags.)
261 11: 'C' if a staging driver has been loaded.
263 12: 'I' if the kernel is working around a severe bug in the platform
264 firmware (BIOS or similar).
266 13: 'O' if an externally-built ("out-of-tree") module has been loaded.
268 The primary reason for the 'Tainted: ' string is to tell kernel
269 debuggers if this is a clean kernel or if anything unusual has
270 occurred. Tainting is permanent: even if an offending module is
271 unloaded, the tainted value remains to indicate that the kernel is not