1 ==============================================
2 Documentation for Kdump - The kexec-based Cras
3 ==============================================
4
5 This document includes overview, setup, instal
6 information.
7
8 Overview
9 ========
10
11 Kdump uses kexec to quickly boot to a dump-cap
12 dump of the system kernel's memory needs to be
13 the system panics). The system kernel's memory
14 the reboot and is accessible to the dump-captu
15
16 You can use common commands, such as cp, scp o
17 the memory image to a dump file on the local d
18 to a remote system.
19
20 Kdump and kexec are currently supported on the
21 s390x, arm and arm64 architectures.
22
23 When the system kernel boots, it reserves a sm
24 the dump-capture kernel. This ensures that ong
25 (DMA) from the system kernel does not corrupt
26 The kexec -p command loads the dump-capture ke
27 memory.
28
29 On x86 machines, the first 640 KB of physical
30 regardless of where the kernel loads. For simp
31 low 1M is reserved to avoid any later kernel o
32 data into this area. Like this, the low 1M can
33 by kdump kernel without extra handling.
34
35 On PPC64 machines first 32KB of physical memor
36 regardless of where the kernel is loaded and t
37 kexec backs up the first 64KB memory.
38
39 For s390x, when kdump is triggered, the crashk
40 with the region [0, crashkernel region size] a
41 runs in [0, crashkernel region size]. Therefor
42 needed for s390x.
43
44 All of the necessary information about the sys
45 encoded in the ELF format, and stored in a res
46 before a crash. The physical address of the st
47 passed to the dump-capture kernel through the
48 parameter. Optionally the size of the ELF head
49 when using the elfcorehdr=[size[KMG]@]offset[K
50
51 With the dump-capture kernel, you can access t
52 /proc/vmcore. This exports the dump as an ELF-
53 write out using file copy commands such as cp
54 makedumpfile utility to analyze and write out
55 options, e.g with '-d 31' it will only write o
56 you can use analysis tools such as the GNU Deb
57 tool to debug the dump file. This method ensur
58 correctly ordered.
59
60 Setup and Installation
61 ======================
62
63 Install kexec-tools
64 -------------------
65
66 1) Login as the root user.
67
68 2) Download the kexec-tools user-space package
69
70 http://kernel.org/pub/linux/utils/kernel/kexec
71
72 This is a symlink to the latest version.
73
74 The latest kexec-tools git tree is available a
75
76 - git://git.kernel.org/pub/scm/utils/kernel/ke
77 - http://www.kernel.org/pub/scm/utils/kernel/k
78
79 There is also a gitweb interface available at
80 http://www.kernel.org/git/?p=utils/kernel/kexe
81
82 More information about kexec-tools can be foun
83 http://horms.net/projects/kexec/
84
85 3) Unpack the tarball with the tar command, as
86
87 tar xvpzf kexec-tools.tar.gz
88
89 4) Change to the kexec-tools directory, as fol
90
91 cd kexec-tools-VERSION
92
93 5) Configure the package, as follows::
94
95 ./configure
96
97 6) Compile the package, as follows::
98
99 make
100
101 7) Install the package, as follows::
102
103 make install
104
105
106 Build the system and dump-capture kernels
107 -----------------------------------------
108 There are two possible methods of using Kdump.
109
110 1) Build a separate custom dump-capture kernel
111 kernel core dump.
112
113 2) Or use the system kernel binary itself as d
114 no need to build a separate dump-capture ke
115 only with the architectures which support a
116 of today, i386, x86_64, ppc64, arm and arm6
117 relocatable kernel.
118
119 Building a relocatable kernel is advantageous
120 one does not have to build a second kernel for
121 at the same time one might want to build a cus
122 suitable to his needs.
123
124 Following are the configuration setting requir
125 dump-capture kernels for enabling kdump suppor
126
127 System kernel config options
128 ----------------------------
129
130 1) Enable "kexec system call" or "kexec file b
131 "Processor type and features."::
132
133 CONFIG_KEXEC=y or CONFIG_KEXEC_FILE=y
134
135 And both of them will select KEXEC_CORE::
136
137 CONFIG_KEXEC_CORE=y
138
139 2) Enable "sysfs file system support" in "File
140 filesystems." This is usually enabled by de
141
142 CONFIG_SYSFS=y
143
144 Note that "sysfs file system support" might
145 filesystems" menu if "Configure standard ke
146 is not enabled in "General Setup." In this
147 itself to ensure that sysfs is turned on, a
148
149 grep 'CONFIG_SYSFS' .config
150
151 3) Enable "Compile the kernel with debug info"
152
153 CONFIG_DEBUG_INFO=Y
154
155 This causes the kernel to be built with deb
156 analysis tools require a vmlinux with debug
157 and analyze a dump file.
158
159 Dump-capture kernel config options (Arch Indep
160 ----------------------------------------------
161
162 1) Enable "kernel crash dumps" support under "
163 features"::
164
165 CONFIG_CRASH_DUMP=y
166
167 And this will select VMCORE_INFO and CRASH_
168 CONFIG_VMCORE_INFO=y
169 CONFIG_CRASH_RESERVE=y
170
171 2) Enable "/proc/vmcore support" under "Filesy
172
173 CONFIG_PROC_VMCORE=y
174
175 (CONFIG_PROC_VMCORE is set by default when
176
177 Dump-capture kernel config options (Arch Depen
178 ----------------------------------------------
179
180 1) On i386, enable high memory support under "
181 features"::
182
183 CONFIG_HIGHMEM64G=y
184
185 or::
186
187 CONFIG_HIGHMEM4G
188
189 2) With CONFIG_SMP=y, usually nr_cpus=1 need s
190 command line when loading the dump-capture
191 CPU is enough for kdump kernel to dump vmco
192
193 However, you can also specify nr_cpus=X to
194 in kdump kernel.
195
196 With CONFIG_SMP=n, the above things are not
197
198 3) A relocatable kernel is suggested to be bui
199 enable "Build a relocatable kernel" support
200 features"::
201
202 CONFIG_RELOCATABLE=y
203
204 4) Use a suitable value for "Physical address
205 loaded" (under "Processor type and features
206 "kernel crash dumps" is enabled. A suitable
207 whether kernel is relocatable or not.
208
209 If you are using a relocatable kernel use C
210 This will compile the kernel for physical a
211 kernel is relocatable, it can be run from a
212 kexec boot loader will load it in memory re
213 kernel.
214
215 Otherwise it should be the start of memory
216 second kernel using boot parameter "crashke
217 start of memory region reserved for dump-ca
218 Generally X is 16MB (0x1000000). So you can
219 CONFIG_PHYSICAL_START=0x1000000
220
221 5) Make and install the kernel and its modules
222 to the boot loader configuration files.
223
224 Dump-capture kernel config options (Arch Depen
225 ----------------------------------------------
226
227 1) Enable "Build a kdump crash kernel" support
228
229 CONFIG_CRASH_DUMP=y
230
231 2) Enable "Build a relocatable kernel" suppo
232
233 CONFIG_RELOCATABLE=y
234
235 Make and install the kernel and its modules
236
237 Dump-capture kernel config options (Arch Depen
238 ----------------------------------------------
239
240 - To use a relocatable kernel,
241 Enable "AUTO_ZRELADDR" support under "Boot
242
243 AUTO_ZRELADDR=y
244
245 Dump-capture kernel config options (Arch Depen
246 ----------------------------------------------
247
248 - Please note that kvm of the dump-capture ker
249 on non-VHE systems even if it is configured.
250 will not be reset to EL2 on panic.
251
252 crashkernel syntax
253 ===========================
254 1) crashkernel=size@offset
255
256 Here 'size' specifies how much memory to re
257 and 'offset' specifies the beginning of thi
258 "crashkernel=64M@16M" tells the system kern
259 starting at physical address 0x01000000 (16
260
261 The crashkernel region can be automatically
262 kernel at run time. This is done by specify
263 or omitting it all together::
264
265 crashkernel=256M@0
266
267 or::
268
269 crashkernel=256M
270
271 If the start address is specified, note tha
272 kernel will be aligned to a value (which is
273 start address is not then any space below t
274 wasted.
275
276 2) range1:size1[,range2:size2,...][@offset]
277
278 While the "crashkernel=size[@offset]" synta
279 configurations, sometimes it's handy to hav
280 on the value of System RAM -- that's mostly
281 the kernel command line to avoid a unbootab
282 been removed from the machine.
283
284 The syntax is::
285
286 crashkernel=<range1>:<size1>[,<range2>:
287 range=start-[end]
288
289 For example::
290
291 crashkernel=512M-2G:64M,2G-:128M
292
293 This would mean:
294
295 1) if the RAM is smaller than 512M, the
296 (this is the "rescue" case)
297 2) if the RAM size is between 512M and
298 3) if the RAM size is larger than 2G, t
299
300 3) crashkernel=size,high and crashkernel=size,
301
302 If memory above 4G is preferred, crashkerne
303 fulfill that. With it, physical memory is a
304 so could be above 4G if system has more tha
305 memory region will be allocated below 4G if
306
307 When crashkernel=X,high is passed, kernel c
308 region above 4G, low memory under 4G is nee
309 three ways to get low memory:
310
311 1) Kernel will allocate at least 256M me
312 if crashkernel=Y,low is not specified
313 2) Let user specify low memory size inst
314 3) Specified value 0 will disable low me
315
316 crashkernel=0,low
317
318 Boot into System Kernel
319 -----------------------
320 1) Update the boot loader (such as grub, yaboo
321 files as necessary.
322
323 2) Boot the system kernel with the boot parame
324
325 On x86 and x86_64, use "crashkernel=Y[@X]".
326 start address 'X' is not necessary, kernel
327 area. Unless an explicit start address is e
328
329 On ppc64, use "crashkernel=128M@32M".
330
331 On s390x, typically use "crashkernel=xxM".
332 on the memory consumption of the kdump syst
333 dependent on the memory size of the product
334
335 On arm, the use of "crashkernel=Y@X" is no
336 kernel will automatically locate the crash
337 first 512MB of RAM if X is not given.
338
339 On arm64, use "crashkernel=Y[@X]". Note th
340 the kernel, X if explicitly specified, must
341
342 Load the Dump-capture Kernel
343 ============================
344
345 After booting to the system kernel, dump-captu
346 loaded.
347
348 Based on the architecture and type of image (r
349 can choose to load the uncompressed vmlinux or
350 of dump-capture kernel. Following is the summa
351
352 For i386 and x86_64:
353
354 - Use bzImage/vmlinuz if kernel is rel
355 - Use vmlinux if kernel is not relocat
356
357 For ppc64:
358
359 - Use vmlinux
360
361 For s390x:
362
363 - Use image or bzImage
364
365 For arm:
366
367 - Use zImage
368
369 For arm64:
370
371 - Use vmlinux or Image
372
373 If you are using an uncompressed vmlinux image
374 to load dump-capture kernel::
375
376 kexec -p <dump-capture-kernel-vmlinux-image
377 --initrd=<initrd-for-dump-capture-kernel> -
378 --append="root=<root-dev> <arch-specific-op
379
380 If you are using a compressed bzImage/vmlinuz,
381 to load dump-capture kernel::
382
383 kexec -p <dump-capture-kernel-bzImage> \
384 --initrd=<initrd-for-dump-capture-kernel> \
385 --append="root=<root-dev> <arch-specific-op
386
387 If you are using a compressed zImage, then use
388 to load dump-capture kernel::
389
390 kexec --type zImage -p <dump-capture-kernel
391 --initrd=<initrd-for-dump-capture-kernel> \
392 --dtb=<dtb-for-dump-capture-kernel> \
393 --append="root=<root-dev> <arch-specific-op
394
395 If you are using an uncompressed Image, then u
396 to load dump-capture kernel::
397
398 kexec -p <dump-capture-kernel-Image> \
399 --initrd=<initrd-for-dump-capture-kernel> \
400 --append="root=<root-dev> <arch-specific-op
401
402 Following are the arch specific command line o
403 loading dump-capture kernel.
404
405 For i386 and x86_64:
406
407 "1 irqpoll nr_cpus=1 reset_devices"
408
409 For ppc64:
410
411 "1 maxcpus=1 noirqdistrib reset_device
412
413 For s390x:
414
415 "1 nr_cpus=1 cgroup_disable=memory"
416
417 For arm:
418
419 "1 maxcpus=1 reset_devices"
420
421 For arm64:
422
423 "1 nr_cpus=1 reset_devices"
424
425 Notes on loading the dump-capture kernel:
426
427 * By default, the ELF headers are stored in EL
428 systems with more than 4GB memory. On i386,
429 the physical RAM size exceeds the 4 GB limit
430 So, on non-PAE systems, ELF32 is always used
431
432 The --elf32-core-headers option can be used
433 headers. This is necessary because GDB curre
434 with ELF64 headers on 32-bit systems.
435
436 * The "irqpoll" boot parameter reduces driver
437 due to shared interrupts in the dump-capture
438
439 * You must specify <root-dev> in the format co
440 device name in the output of mount command.
441
442 * Boot parameter "1" boots the dump-capture ke
443 mode without networking. If you want network
444
445 * We generally don't have to bring up a SMP ke
446 dump. Hence generally it is useful either to
447 kernel or specify maxcpus=1 option while loa
448 Note, though maxcpus always works, you had b
449 nr_cpus to save memory if supported by the c
450
451 * You should enable multi-cpu support in dump-
452 to use multi-thread programs with it, such a
453 makedumpfile. Otherwise, the multi-thread pr
454 performance degradation. To enable multi-cpu
455 SMP dump-capture kernel and specify maxcpus/
456
457 * For s390x there are two kdump modes: If a EL
458 the elfcorehdr= kernel parameter, it is used
459 is done on all other architectures. If no el
460 specified, the s390x kdump kernel dynamicall
461 second mode has the advantage that for CPU a
462 not to be reloaded with kexec_load().
463
464 * For s390x systems with many attached devices
465 parameter should be used for the kdump kerne
466 of kernel memory for devices that are not re
467 applies to systems that use SCSI/FCP devices
468 "allow_lun_scan" zfcp module parameter shoul
469 setting FCP devices online.
470
471 Kernel Panic
472 ============
473
474 After successfully loading the dump-capture ke
475 described, the system will reboot into the dum
476 system crash is triggered. Trigger points are
477 die(), die_nmi() and in the sysrq handler (ALT
478
479 The following conditions will execute a crash
480
481 If a hard lockup is detected and "NMI watchdog
482 will boot into the dump-capture kernel ( die_n
483
484 If die() is called, and it happens to be a thr
485 is called inside interrupt context or die() is
486 the system will boot into the dump-capture ker
487
488 On powerpc systems when a soft-reset is genera
489 and the system will boot into the dump-capture
490
491 For testing purposes, you can trigger a crash
492 "echo c > /proc/sysrq-trigger" or write a modu
493
494 Write Out the Dump File
495 =======================
496
497 After the dump-capture kernel is booted, write
498 the following command::
499
500 cp /proc/vmcore <dump-file>
501
502 or use scp to write out the dump file between
503
504 scp /proc/vmcore remote_username@remote_ip:
505
506 You can also use makedumpfile utility to write
507 with specified options to filter out unwanted
508
509 makedumpfile -l --message-level 1 -d 31 /pr
510
511 Analysis
512 ========
513
514 Before analyzing the dump image, you should re
515
516 You can do limited analysis using GDB on the d
517 /proc/vmcore. Use the debug vmlinux built with
518 command::
519
520 gdb vmlinux <dump-file>
521
522 Stack trace for the task on processor 0, regis
523 display work fine.
524
525 Note: GDB cannot analyze core files generated
526 On systems with a maximum of 4GB of memory, yo
527 ELF32-format headers using the --elf32-core-he
528 dump kernel.
529
530 You can also use the Crash utility to analyze
531 format. Crash is available at the following UR
532
533 https://github.com/crash-utility/crash
534
535 Crash document can be found at:
536 https://crash-utility.github.io/
537
538 Trigger Kdump on WARN()
539 =======================
540
541 The kernel parameter, panic_on_warn, calls pan
542 will cause a kdump to occur at the panic() cal
543 to specify this during runtime, /proc/sys/kern
544 to achieve the same behaviour.
545
546 Trigger Kdump on add_taint()
547 ============================
548
549 The kernel parameter panic_on_taint facilitate
550 from within add_taint() whenever the value set
551 bit flag being set by add_taint().
552 This will cause a kdump to occur at the add_ta
553
554 Contact
555 =======
556
557 - kexec@lists.infradead.org
558
559 GDB macros
560 ==========
561
562 .. include:: gdbmacros.txt
563 :literal:
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.