blob: 9dba3bb90e60a4710222d864124453d0d861b8c6 [file] [log] [blame]
Paul E. McKenneya241ec62005-10-30 15:03:12 -08001RCU Torture Test Operation
2
3
4CONFIG_RCU_TORTURE_TEST
5
6The CONFIG_RCU_TORTURE_TEST config option is available for all RCU
7implementations. It creates an rcutorture kernel module that can
8be loaded to run a torture test. The test periodically outputs
9status messages via printk(), which can be examined via the dmesg
Paul E. McKenney72e9bb52006-06-27 02:54:03 -070010command (perhaps grepping for "torture"). The test is started
Paul E. McKenneya241ec62005-10-30 15:03:12 -080011when the module is loaded, and stops when the module is unloaded.
12
Paul E. McKenney31a72bc2008-06-18 09:26:49 -070013CONFIG_RCU_TORTURE_TEST_RUNNABLE
14
15It is also possible to specify CONFIG_RCU_TORTURE_TEST=y, which will
16result in the tests being loaded into the base kernel. In this case,
17the CONFIG_RCU_TORTURE_TEST_RUNNABLE config option is used to specify
18whether the RCU torture tests are to be started immediately during
19boot or whether the /proc/sys/kernel/rcutorture_runnable file is used
20to enable them. This /proc file can be used to repeatedly pause and
21restart the tests, regardless of the initial state specified by the
22CONFIG_RCU_TORTURE_TEST_RUNNABLE config option.
23
24You will normally -not- want to start the RCU torture tests during boot
25(and thus the default is CONFIG_RCU_TORTURE_TEST_RUNNABLE=n), but doing
26this can sometimes be useful in finding boot-time bugs.
Paul E. McKenneya241ec62005-10-30 15:03:12 -080027
28
29MODULE PARAMETERS
30
31This module has the following parameters:
32
Paul E. McKenney0729fbf2008-06-25 12:24:52 -070033irqreaders Says to invoke RCU readers from irq level. This is currently
34 done via timers. Defaults to "1" for variants of RCU that
35 permit this. (Or, more accurately, variants of RCU that do
36 -not- permit this know to ignore this variable.)
Paul E. McKenneya241ec62005-10-30 15:03:12 -080037
Josh Triplettb772e1d2006-10-04 02:17:13 -070038nfakewriters This is the number of RCU fake writer threads to run. Fake
39 writer threads repeatedly use the synchronous "wait for
40 current readers" function of the interface selected by
41 torture_type, with a delay between calls to allow for various
42 different numbers of writers running in parallel.
43 nfakewriters defaults to 4, which provides enough parallelism
44 to trigger special cases caused by multiple writers, such as
45 the synchronize_srcu() early return optimization.
46
Paul E. McKenney0729fbf2008-06-25 12:24:52 -070047nreaders This is the number of RCU reading threads supported.
48 The default is twice the number of CPUs. Why twice?
49 To properly exercise RCU implementations with preemptible
50 read-side critical sections.
51
52shuffle_interval
53 The number of seconds to keep the test threads affinitied
54 to a particular subset of the CPUs, defaults to 3 seconds.
55 Used in conjunction with test_no_idle_hz.
56
Paul E. McKenneya241ec62005-10-30 15:03:12 -080057stat_interval The number of seconds between output of torture
58 statistics (via printk()). Regardless of the interval,
59 statistics are printed when the module is unloaded.
60 Setting the interval to zero causes the statistics to
61 be printed -only- when the module is unloaded, and this
62 is the default.
63
Paul E. McKenneyd120f652008-06-18 05:21:44 -070064stutter The length of time to run the test before pausing for this
65 same period of time. Defaults to "stutter=5", so as
66 to run and pause for (roughly) five-second intervals.
67 Specifying "stutter=0" causes the test to run continuously
68 without pausing, which is the old default behavior.
69
Paul E. McKenney29766f12006-06-27 02:54:02 -070070test_no_idle_hz Whether or not to test the ability of RCU to operate in
71 a kernel that disables the scheduling-clock interrupt to
72 idle CPUs. Boolean parameter, "1" to test, "0" otherwise.
Paul E. McKenneyf85d6c72008-01-25 21:08:25 +010073 Defaults to omitting this test.
Paul E. McKenney29766f12006-06-27 02:54:02 -070074
Josh Triplett20d2e422006-10-04 02:17:15 -070075torture_type The type of RCU to test: "rcu" for the rcu_read_lock() API,
76 "rcu_sync" for rcu_read_lock() with synchronous reclamation,
Josh Triplett11a14702006-10-04 02:17:16 -070077 "rcu_bh" for the rcu_read_lock_bh() API, "rcu_bh_sync" for
Josh Triplett4b6c2cc2006-10-04 02:17:16 -070078 rcu_read_lock_bh() with synchronous reclamation, "srcu" for
Paul E. McKenney240ebbf2009-06-25 09:08:18 -070079 the "srcu_read_lock()" API, "sched" for the use of
80 preempt_disable() together with synchronize_sched(),
81 and "sched_expedited" for the use of preempt_disable()
82 with synchronize_sched_expedited().
Paul E. McKenney72e9bb52006-06-27 02:54:03 -070083
Paul E. McKenneya241ec62005-10-30 15:03:12 -080084verbose Enable debug printk()s. Default is disabled.
85
86
87OUTPUT
88
89The statistics output is as follows:
90
Paul E. McKenney72e9bb52006-06-27 02:54:03 -070091 rcu-torture: --- Start of test: nreaders=16 stat_interval=0 verbose=0
92 rcu-torture: rtc: 0000000000000000 ver: 1916 tfle: 0 rta: 1916 rtaf: 0 rtf: 1915
93 rcu-torture: Reader Pipe: 1466408 9747 0 0 0 0 0 0 0 0 0
94 rcu-torture: Reader Batch: 1464477 11678 0 0 0 0 0 0 0 0
95 rcu-torture: Free-Block Circulation: 1915 1915 1915 1915 1915 1915 1915 1915 1915 1915 0
96 rcu-torture: --- End of test
Paul E. McKenneya241ec62005-10-30 15:03:12 -080097
Paul E. McKenney72e9bb52006-06-27 02:54:03 -070098The command "dmesg | grep torture:" will extract this information on
Paul E. McKenneya241ec62005-10-30 15:03:12 -080099most systems. On more esoteric configurations, it may be necessary to
100use other commands to access the output of the printk()s used by
101the RCU torture test. The printk()s use KERN_ALERT, so they should
102be evident. ;-)
103
104The entries are as follows:
105
Paul E. McKenneya241ec62005-10-30 15:03:12 -0800106o "rtc": The hexadecimal address of the structure currently visible
107 to readers.
108
109o "ver": The number of times since boot that the rcutw writer task
110 has changed the structure visible to readers.
111
112o "tfle": If non-zero, indicates that the "torture freelist"
113 containing structure to be placed into the "rtc" area is empty.
114 This condition is important, since it can fool you into thinking
115 that RCU is working when it is not. :-/
116
117o "rta": Number of structures allocated from the torture freelist.
118
119o "rtaf": Number of allocations from the torture freelist that have
120 failed due to the list being empty.
121
122o "rtf": Number of frees into the torture freelist.
123
124o "Reader Pipe": Histogram of "ages" of structures seen by readers.
125 If any entries past the first two are non-zero, RCU is broken.
126 And rcutorture prints the error flag string "!!!" to make sure
127 you notice. The age of a newly allocated structure is zero,
128 it becomes one when removed from reader visibility, and is
129 incremented once per grace period subsequently -- and is freed
130 after passing through (RCU_TORTURE_PIPE_LEN-2) grace periods.
131
132 The output displayed above was taken from a correctly working
133 RCU. If you want to see what it looks like when broken, break
134 it yourself. ;-)
135
136o "Reader Batch": Another histogram of "ages" of structures seen
137 by readers, but in terms of counter flips (or batches) rather
138 than in terms of grace periods. The legal number of non-zero
Paul E. McKenneyf85d6c72008-01-25 21:08:25 +0100139 entries is again two. The reason for this separate view is that
140 it is sometimes easier to get the third entry to show up in the
Paul E. McKenneya241ec62005-10-30 15:03:12 -0800141 "Reader Batch" list than in the "Reader Pipe" list.
142
143o "Free-Block Circulation": Shows the number of torture structures
144 that have reached a given point in the pipeline. The first element
145 should closely correspond to the number of structures allocated,
146 the second to the number that have been removed from reader view,
147 and all but the last remaining to the corresponding number of
148 passes through a grace period. The last entry should be zero,
149 as it is only incremented if a torture structure's counter
150 somehow gets incremented farther than it should.
151
Paul E. McKenneyb2896d22006-10-04 02:17:03 -0700152Different implementations of RCU can provide implementation-specific
153additional information. For example, SRCU provides the following:
154
155 srcu-torture: rtc: f8cf46a8 ver: 355 tfle: 0 rta: 356 rtaf: 0 rtf: 346 rtmbe: 0
156 srcu-torture: Reader Pipe: 559738 939 0 0 0 0 0 0 0 0 0
157 srcu-torture: Reader Batch: 560434 243 0 0 0 0 0 0 0 0
158 srcu-torture: Free-Block Circulation: 355 354 353 352 351 350 349 348 347 346 0
159 srcu-torture: per-CPU(idx=1): 0(0,1) 1(0,1) 2(0,0) 3(0,1)
160
161The first four lines are similar to those for RCU. The last line shows
162the per-CPU counter state. The numbers in parentheses are the values
163of the "old" and "current" counters for the corresponding CPU. The
164"idx" value maps the "old" and "current" values to the underlying array,
165and is useful for debugging.
166
Paul E. McKenney240ebbf2009-06-25 09:08:18 -0700167Similarly, sched_expedited RCU provides the following:
168
169 sched_expedited-torture: rtc: d0000000016c1880 ver: 1090796 tfle: 0 rta: 1090796 rtaf: 0 rtf: 1090787 rtmbe: 0 nt: 27713319
170 sched_expedited-torture: Reader Pipe: 12660320201 95875 0 0 0 0 0 0 0 0 0
171 sched_expedited-torture: Reader Batch: 12660424885 0 0 0 0 0 0 0 0 0 0
172 sched_expedited-torture: Free-Block Circulation: 1090795 1090795 1090794 1090793 1090792 1090791 1090790 1090789 1090788 1090787 0
173 state: -1 / 0:0 3:0 4:0
174
175As before, the first four lines are similar to those for RCU.
176The last line shows the task-migration state. The first number is
177-1 if synchronize_sched_expedited() is idle, -2 if in the process of
178posting wakeups to the migration kthreads, and N when waiting on CPU N.
179Each of the colon-separated fields following the "/" is a CPU:state pair.
180Valid states are "0" for idle, "1" for waiting for quiescent state,
181"2" for passed through quiescent state, and "3" when a race with a
182CPU-hotplug event forces use of the synchronize_sched() primitive.
183
Paul E. McKenneya241ec62005-10-30 15:03:12 -0800184
185USAGE
186
187The following script may be used to torture RCU:
188
189 #!/bin/sh
190
191 modprobe rcutorture
192 sleep 100
193 rmmod rcutorture
Paul E. McKenney72e9bb52006-06-27 02:54:03 -0700194 dmesg | grep torture:
Paul E. McKenneya241ec62005-10-30 15:03:12 -0800195
196The output can be manually inspected for the error flag of "!!!".
197One could of course create a more elaborate script that automatically
Paul E. McKenney29766f12006-06-27 02:54:02 -0700198checked for such errors. The "rmmod" command forces a "SUCCESS" or
199"FAILURE" indication to be printk()ed.