]>
Commit | Line | Data |
---|---|---|
de1ba09b AM |
1 | Fault injection capabilities infrastructure |
2 | =========================================== | |
3 | ||
4 | See also drivers/md/faulty.c and "every_nth" module option for scsi_debug. | |
5 | ||
6 | ||
7 | Available fault injection capabilities | |
8 | -------------------------------------- | |
9 | ||
10 | o failslab | |
11 | ||
12 | injects slab allocation failures. (kmalloc(), kmem_cache_alloc(), ...) | |
13 | ||
14 | o fail_page_alloc | |
15 | ||
16 | injects page allocation failures. (alloc_pages(), get_free_pages(), ...) | |
17 | ||
ab51fbab DB |
18 | o fail_futex |
19 | ||
20 | injects futex deadlock and uaddr fault errors. | |
21 | ||
de1ba09b AM |
22 | o fail_make_request |
23 | ||
5d0ffa2b | 24 | injects disk IO errors on devices permitted by setting |
de1ba09b AM |
25 | /sys/block/<device>/make-it-fail or |
26 | /sys/block/<device>/<partition>/make-it-fail. (generic_make_request()) | |
27 | ||
1e4cb22b PF |
28 | o fail_mmc_request |
29 | ||
30 | injects MMC data errors on devices permitted by setting | |
31 | debugfs entries under /sys/kernel/debug/mmc0/fail_mmc_request | |
32 | ||
de1ba09b AM |
33 | Configure fault-injection capabilities behavior |
34 | ----------------------------------------------- | |
35 | ||
36 | o debugfs entries | |
37 | ||
38 | fault-inject-debugfs kernel module provides some debugfs entries for runtime | |
39 | configuration of fault-injection capabilities. | |
40 | ||
156f5a78 | 41 | - /sys/kernel/debug/fail*/probability: |
de1ba09b AM |
42 | |
43 | likelihood of failure injection, in percent. | |
44 | Format: <percent> | |
45 | ||
5d0ffa2b DM |
46 | Note that one-failure-per-hundred is a very high error rate |
47 | for some testcases. Consider setting probability=100 and configure | |
156f5a78 | 48 | /sys/kernel/debug/fail*/interval for such testcases. |
de1ba09b | 49 | |
156f5a78 | 50 | - /sys/kernel/debug/fail*/interval: |
de1ba09b AM |
51 | |
52 | specifies the interval between failures, for calls to | |
53 | should_fail() that pass all the other tests. | |
54 | ||
55 | Note that if you enable this, by setting interval>1, you will | |
56 | probably want to set probability=100. | |
57 | ||
156f5a78 | 58 | - /sys/kernel/debug/fail*/times: |
de1ba09b AM |
59 | |
60 | specifies how many times failures may happen at most. | |
61 | A value of -1 means "no limit". | |
62 | ||
156f5a78 | 63 | - /sys/kernel/debug/fail*/space: |
de1ba09b AM |
64 | |
65 | specifies an initial resource "budget", decremented by "size" | |
66 | on each call to should_fail(,size). Failure injection is | |
67 | suppressed until "space" reaches zero. | |
68 | ||
156f5a78 | 69 | - /sys/kernel/debug/fail*/verbose |
de1ba09b AM |
70 | |
71 | Format: { 0 | 1 | 2 } | |
5d0ffa2b DM |
72 | specifies the verbosity of the messages when failure is |
73 | injected. '0' means no messages; '1' will print only a single | |
74 | log line per failure; '2' will print a call trace too -- useful | |
75 | to debug the problems revealed by fault injection. | |
de1ba09b | 76 | |
156f5a78 | 77 | - /sys/kernel/debug/fail*/task-filter: |
de1ba09b | 78 | |
5d0ffa2b DM |
79 | Format: { 'Y' | 'N' } |
80 | A value of 'N' disables filtering by process (default). | |
de1ba09b AM |
81 | Any positive value limits failures to only processes indicated by |
82 | /proc/<pid>/make-it-fail==1. | |
83 | ||
156f5a78 GL |
84 | - /sys/kernel/debug/fail*/require-start: |
85 | - /sys/kernel/debug/fail*/require-end: | |
86 | - /sys/kernel/debug/fail*/reject-start: | |
87 | - /sys/kernel/debug/fail*/reject-end: | |
de1ba09b AM |
88 | |
89 | specifies the range of virtual addresses tested during | |
90 | stacktrace walking. Failure is injected only if some caller | |
329409ae AM |
91 | in the walked stacktrace lies within the required range, and |
92 | none lies within the rejected range. | |
93 | Default required range is [0,ULONG_MAX) (whole of virtual address space). | |
94 | Default rejected range is [0,0). | |
de1ba09b | 95 | |
156f5a78 | 96 | - /sys/kernel/debug/fail*/stacktrace-depth: |
de1ba09b AM |
97 | |
98 | specifies the maximum stacktrace depth walked during search | |
5d0ffa2b DM |
99 | for a caller within [require-start,require-end) OR |
100 | [reject-start,reject-end). | |
de1ba09b | 101 | |
156f5a78 | 102 | - /sys/kernel/debug/fail_page_alloc/ignore-gfp-highmem: |
de1ba09b | 103 | |
5d0ffa2b DM |
104 | Format: { 'Y' | 'N' } |
105 | default is 'N', setting it to 'Y' won't inject failures into | |
de1ba09b AM |
106 | highmem/user allocations. |
107 | ||
156f5a78 GL |
108 | - /sys/kernel/debug/failslab/ignore-gfp-wait: |
109 | - /sys/kernel/debug/fail_page_alloc/ignore-gfp-wait: | |
de1ba09b | 110 | |
5d0ffa2b DM |
111 | Format: { 'Y' | 'N' } |
112 | default is 'N', setting it to 'Y' will inject failures | |
de1ba09b AM |
113 | only into non-sleep allocations (GFP_ATOMIC allocations). |
114 | ||
156f5a78 | 115 | - /sys/kernel/debug/fail_page_alloc/min-order: |
54114994 AM |
116 | |
117 | specifies the minimum page allocation order to be injected | |
118 | failures. | |
119 | ||
ab51fbab DB |
120 | - /sys/kernel/debug/fail_futex/ignore-private: |
121 | ||
122 | Format: { 'Y' | 'N' } | |
123 | default is 'N', setting it to 'Y' will disable failure injections | |
124 | when dealing with private (address space) futexes. | |
125 | ||
de1ba09b AM |
126 | o Boot option |
127 | ||
128 | In order to inject faults while debugfs is not available (early boot time), | |
129 | use the boot option: | |
130 | ||
131 | failslab= | |
132 | fail_page_alloc= | |
1e4cb22b | 133 | fail_make_request= |
ab51fbab | 134 | fail_futex= |
199e3f4b | 135 | mmc_core.fail_request=<interval>,<probability>,<space>,<times> |
de1ba09b | 136 | |
e41d5818 DV |
137 | o proc entries |
138 | ||
168c42bc AM |
139 | - /proc/<pid>/fail-nth: |
140 | - /proc/self/task/<tid>/fail-nth: | |
e41d5818 | 141 | |
9049f2f6 | 142 | Write to this file of integer N makes N-th call in the task fail. |
bfc74093 AM |
143 | Read from this file returns a integer value. A value of '0' indicates |
144 | that the fault setup with a previous write to this file was injected. | |
145 | A positive integer N indicates that the fault wasn't yet injected. | |
e41d5818 DV |
146 | Note that this file enables all types of faults (slab, futex, etc). |
147 | This setting takes precedence over all other generic debugfs settings | |
148 | like probability, interval, times, etc. But per-capability settings | |
149 | (e.g. fail_futex/ignore-private) take precedence over it. | |
150 | ||
151 | This feature is intended for systematic testing of faults in a single | |
152 | system call. See an example below. | |
153 | ||
de1ba09b AM |
154 | How to add new fault injection capability |
155 | ----------------------------------------- | |
156 | ||
157 | o #include <linux/fault-inject.h> | |
158 | ||
159 | o define the fault attributes | |
160 | ||
161 | DECLARE_FAULT_INJECTION(name); | |
162 | ||
163 | Please see the definition of struct fault_attr in fault-inject.h | |
164 | for details. | |
165 | ||
5d0ffa2b | 166 | o provide a way to configure fault attributes |
de1ba09b AM |
167 | |
168 | - boot option | |
169 | ||
170 | If you need to enable the fault injection capability from boot time, you can | |
5d0ffa2b | 171 | provide boot option to configure it. There is a helper function for it: |
de1ba09b | 172 | |
5d0ffa2b | 173 | setup_fault_attr(attr, str); |
de1ba09b AM |
174 | |
175 | - debugfs entries | |
176 | ||
177 | failslab, fail_page_alloc, and fail_make_request use this way. | |
5d0ffa2b | 178 | Helper functions: |
de1ba09b | 179 | |
dd48c085 | 180 | fault_create_debugfs_attr(name, parent, attr); |
de1ba09b AM |
181 | |
182 | - module parameters | |
183 | ||
184 | If the scope of the fault injection capability is limited to a | |
185 | single kernel module, it is better to provide module parameters to | |
186 | configure the fault attributes. | |
187 | ||
188 | o add a hook to insert failures | |
189 | ||
5d0ffa2b | 190 | Upon should_fail() returning true, client code should inject a failure. |
de1ba09b | 191 | |
5d0ffa2b | 192 | should_fail(attr, size); |
de1ba09b AM |
193 | |
194 | Application Examples | |
195 | -------------------- | |
196 | ||
18584870 | 197 | o Inject slab allocation failures into module init/exit code |
de1ba09b | 198 | |
de1ba09b AM |
199 | #!/bin/bash |
200 | ||
18584870 | 201 | FAILTYPE=failslab |
156f5a78 GL |
202 | echo Y > /sys/kernel/debug/$FAILTYPE/task-filter |
203 | echo 10 > /sys/kernel/debug/$FAILTYPE/probability | |
204 | echo 100 > /sys/kernel/debug/$FAILTYPE/interval | |
205 | echo -1 > /sys/kernel/debug/$FAILTYPE/times | |
206 | echo 0 > /sys/kernel/debug/$FAILTYPE/space | |
207 | echo 2 > /sys/kernel/debug/$FAILTYPE/verbose | |
208 | echo 1 > /sys/kernel/debug/$FAILTYPE/ignore-gfp-wait | |
de1ba09b | 209 | |
18584870 | 210 | faulty_system() |
de1ba09b | 211 | { |
18584870 | 212 | bash -c "echo 1 > /proc/self/make-it-fail && exec $*" |
de1ba09b AM |
213 | } |
214 | ||
18584870 AM |
215 | if [ $# -eq 0 ] |
216 | then | |
217 | echo "Usage: $0 modulename [ modulename ... ]" | |
218 | exit 1 | |
219 | fi | |
220 | ||
221 | for m in $* | |
222 | do | |
223 | echo inserting $m... | |
224 | faulty_system modprobe $m | |
de1ba09b | 225 | |
18584870 AM |
226 | echo removing $m... |
227 | faulty_system modprobe -r $m | |
228 | done | |
de1ba09b AM |
229 | |
230 | ------------------------------------------------------------------------------ | |
231 | ||
18584870 | 232 | o Inject page allocation failures only for a specific module |
de1ba09b | 233 | |
de1ba09b AM |
234 | #!/bin/bash |
235 | ||
18584870 AM |
236 | FAILTYPE=fail_page_alloc |
237 | module=$1 | |
de1ba09b | 238 | |
18584870 AM |
239 | if [ -z $module ] |
240 | then | |
241 | echo "Usage: $0 <modulename>" | |
242 | exit 1 | |
243 | fi | |
de1ba09b | 244 | |
18584870 | 245 | modprobe $module |
de1ba09b | 246 | |
18584870 AM |
247 | if [ ! -d /sys/module/$module/sections ] |
248 | then | |
249 | echo Module $module is not loaded | |
250 | exit 1 | |
251 | fi | |
252 | ||
156f5a78 GL |
253 | cat /sys/module/$module/sections/.text > /sys/kernel/debug/$FAILTYPE/require-start |
254 | cat /sys/module/$module/sections/.data > /sys/kernel/debug/$FAILTYPE/require-end | |
18584870 | 255 | |
156f5a78 GL |
256 | echo N > /sys/kernel/debug/$FAILTYPE/task-filter |
257 | echo 10 > /sys/kernel/debug/$FAILTYPE/probability | |
258 | echo 100 > /sys/kernel/debug/$FAILTYPE/interval | |
259 | echo -1 > /sys/kernel/debug/$FAILTYPE/times | |
260 | echo 0 > /sys/kernel/debug/$FAILTYPE/space | |
261 | echo 2 > /sys/kernel/debug/$FAILTYPE/verbose | |
262 | echo 1 > /sys/kernel/debug/$FAILTYPE/ignore-gfp-wait | |
263 | echo 1 > /sys/kernel/debug/$FAILTYPE/ignore-gfp-highmem | |
264 | echo 10 > /sys/kernel/debug/$FAILTYPE/stacktrace-depth | |
18584870 | 265 | |
156f5a78 | 266 | trap "echo 0 > /sys/kernel/debug/$FAILTYPE/probability" SIGINT SIGTERM EXIT |
18584870 AM |
267 | |
268 | echo "Injecting errors into the module $module... (interrupt to stop)" | |
269 | sleep 1000000 | |
de1ba09b | 270 | |
c24aa64d AM |
271 | Tool to run command with failslab or fail_page_alloc |
272 | ---------------------------------------------------- | |
273 | In order to make it easier to accomplish the tasks mentioned above, we can use | |
274 | tools/testing/fault-injection/failcmd.sh. Please run a command | |
275 | "./tools/testing/fault-injection/failcmd.sh --help" for more information and | |
276 | see the following examples. | |
277 | ||
278 | Examples: | |
279 | ||
280 | Run a command "make -C tools/testing/selftests/ run_tests" with injecting slab | |
281 | allocation failure. | |
282 | ||
283 | # ./tools/testing/fault-injection/failcmd.sh \ | |
284 | -- make -C tools/testing/selftests/ run_tests | |
285 | ||
286 | Same as above except to specify 100 times failures at most instead of one time | |
287 | at most by default. | |
288 | ||
289 | # ./tools/testing/fault-injection/failcmd.sh --times=100 \ | |
290 | -- make -C tools/testing/selftests/ run_tests | |
291 | ||
292 | Same as above except to inject page allocation failure instead of slab | |
293 | allocation failure. | |
294 | ||
295 | # env FAILCMD_TYPE=fail_page_alloc \ | |
296 | ./tools/testing/fault-injection/failcmd.sh --times=100 \ | |
297 | -- make -C tools/testing/selftests/ run_tests | |
e41d5818 DV |
298 | |
299 | Systematic faults using fail-nth | |
300 | --------------------------------- | |
301 | ||
302 | The following code systematically faults 0-th, 1-st, 2-nd and so on | |
303 | capabilities in the socketpair() system call. | |
304 | ||
305 | #include <sys/types.h> | |
306 | #include <sys/stat.h> | |
307 | #include <sys/socket.h> | |
308 | #include <sys/syscall.h> | |
309 | #include <fcntl.h> | |
310 | #include <unistd.h> | |
311 | #include <string.h> | |
312 | #include <stdlib.h> | |
313 | #include <stdio.h> | |
314 | #include <errno.h> | |
315 | ||
316 | int main() | |
317 | { | |
318 | int i, err, res, fail_nth, fds[2]; | |
319 | char buf[128]; | |
320 | ||
321 | system("echo N > /sys/kernel/debug/failslab/ignore-gfp-wait"); | |
322 | sprintf(buf, "/proc/self/task/%ld/fail-nth", syscall(SYS_gettid)); | |
323 | fail_nth = open(buf, O_RDWR); | |
9049f2f6 | 324 | for (i = 1;; i++) { |
e41d5818 DV |
325 | sprintf(buf, "%d", i); |
326 | write(fail_nth, buf, strlen(buf)); | |
327 | res = socketpair(AF_LOCAL, SOCK_STREAM, 0, fds); | |
328 | err = errno; | |
bfc74093 | 329 | pread(fail_nth, buf, sizeof(buf), 0); |
e41d5818 DV |
330 | if (res == 0) { |
331 | close(fds[0]); | |
332 | close(fds[1]); | |
333 | } | |
bfc74093 AM |
334 | printf("%d-th fault %c: res=%d/%d\n", i, atoi(buf) ? 'N' : 'Y', |
335 | res, err); | |
336 | if (atoi(buf)) | |
e41d5818 DV |
337 | break; |
338 | } | |
339 | return 0; | |
340 | } | |
341 | ||
342 | An example output: | |
343 | ||
e41d5818 DV |
344 | 1-th fault Y: res=-1/23 |
345 | 2-th fault Y: res=-1/23 | |
346 | 3-th fault Y: res=-1/12 | |
347 | 4-th fault Y: res=-1/12 | |
348 | 5-th fault Y: res=-1/23 | |
349 | 6-th fault Y: res=-1/23 | |
350 | 7-th fault Y: res=-1/23 | |
351 | 8-th fault Y: res=-1/12 | |
352 | 9-th fault Y: res=-1/12 | |
353 | 10-th fault Y: res=-1/12 | |
354 | 11-th fault Y: res=-1/12 | |
355 | 12-th fault Y: res=-1/12 | |
356 | 13-th fault Y: res=-1/12 | |
357 | 14-th fault Y: res=-1/12 | |
358 | 15-th fault Y: res=-1/12 | |
359 | 16-th fault N: res=0/12 |