1 RCU Torture Test Operation
4 CONFIG_RCU_TORTURE_TEST
6 The CONFIG_RCU_TORTURE_TEST config option is available for all RCU
7 implementations. It creates an rcutorture kernel module that can
8 be loaded to run a torture test. The test periodically outputs
9 status messages via printk(), which can be examined via the dmesg
10 command (perhaps grepping for "torture"). The test is started
11 when the module is loaded, and stops when the module is unloaded.
13 However, actually setting this config option to "y" results in the system
14 running the test immediately upon boot, and ending only when the system
15 is taken down. Normally, one will instead want to build the system
16 with CONFIG_RCU_TORTURE_TEST=m and to use modprobe and rmmod to control
17 the test, perhaps using a script similar to the one shown at the end of
18 this document. Note that you will need CONFIG_MODULE_UNLOAD in order
19 to be able to end the test.
24 This module has the following parameters:
26 nreaders This is the number of RCU reading threads supported.
27 The default is twice the number of CPUs. Why twice?
28 To properly exercise RCU implementations with preemptible
29 read-side critical sections.
31 nfakewriters This is the number of RCU fake writer threads to run. Fake
32 writer threads repeatedly use the synchronous "wait for
33 current readers" function of the interface selected by
34 torture_type, with a delay between calls to allow for various
35 different numbers of writers running in parallel.
36 nfakewriters defaults to 4, which provides enough parallelism
37 to trigger special cases caused by multiple writers, such as
38 the synchronize_srcu() early return optimization.
40 stat_interval The number of seconds between output of torture
41 statistics (via printk()). Regardless of the interval,
42 statistics are printed when the module is unloaded.
43 Setting the interval to zero causes the statistics to
44 be printed -only- when the module is unloaded, and this
48 The number of seconds to keep the test threads affinitied
49 to a particular subset of the CPUs. Used in conjunction
52 test_no_idle_hz Whether or not to test the ability of RCU to operate in
53 a kernel that disables the scheduling-clock interrupt to
54 idle CPUs. Boolean parameter, "1" to test, "0" otherwise.
56 torture_type The type of RCU to test: "rcu" for the rcu_read_lock() API,
57 "rcu_sync" for rcu_read_lock() with synchronous reclamation,
58 "rcu_bh" for the rcu_read_lock_bh() API, "rcu_bh_sync" for
59 rcu_read_lock_bh() with synchronous reclamation, "srcu" for
60 the "srcu_read_lock()" API, and "sched" for the use of
61 preempt_disable() together with synchronize_sched().
63 verbose Enable debug printk()s. Default is disabled.
68 The statistics output is as follows:
70 rcu-torture: --- Start of test: nreaders=16 stat_interval=0 verbose=0
71 rcu-torture: rtc: 0000000000000000 ver: 1916 tfle: 0 rta: 1916 rtaf: 0 rtf: 1915
72 rcu-torture: Reader Pipe: 1466408 9747 0 0 0 0 0 0 0 0 0
73 rcu-torture: Reader Batch: 1464477 11678 0 0 0 0 0 0 0 0
74 rcu-torture: Free-Block Circulation: 1915 1915 1915 1915 1915 1915 1915 1915 1915 1915 0
75 rcu-torture: --- End of test
77 The command "dmesg | grep torture:" will extract this information on
78 most systems. On more esoteric configurations, it may be necessary to
79 use other commands to access the output of the printk()s used by
80 the RCU torture test. The printk()s use KERN_ALERT, so they should
83 The entries are as follows:
85 o "ggp": The number of counter flips (or batches) since boot.
87 o "rtc": The hexadecimal address of the structure currently visible
90 o "ver": The number of times since boot that the rcutw writer task
91 has changed the structure visible to readers.
93 o "tfle": If non-zero, indicates that the "torture freelist"
94 containing structure to be placed into the "rtc" area is empty.
95 This condition is important, since it can fool you into thinking
96 that RCU is working when it is not. :-/
98 o "rta": Number of structures allocated from the torture freelist.
100 o "rtaf": Number of allocations from the torture freelist that have
101 failed due to the list being empty.
103 o "rtf": Number of frees into the torture freelist.
105 o "Reader Pipe": Histogram of "ages" of structures seen by readers.
106 If any entries past the first two are non-zero, RCU is broken.
107 And rcutorture prints the error flag string "!!!" to make sure
108 you notice. The age of a newly allocated structure is zero,
109 it becomes one when removed from reader visibility, and is
110 incremented once per grace period subsequently -- and is freed
111 after passing through (RCU_TORTURE_PIPE_LEN-2) grace periods.
113 The output displayed above was taken from a correctly working
114 RCU. If you want to see what it looks like when broken, break
117 o "Reader Batch": Another histogram of "ages" of structures seen
118 by readers, but in terms of counter flips (or batches) rather
119 than in terms of grace periods. The legal number of non-zero
120 entries is again two. The reason for this separate view is
121 that it is easier to get the third entry to show up in the
122 "Reader Batch" list than in the "Reader Pipe" list.
124 o "Free-Block Circulation": Shows the number of torture structures
125 that have reached a given point in the pipeline. The first element
126 should closely correspond to the number of structures allocated,
127 the second to the number that have been removed from reader view,
128 and all but the last remaining to the corresponding number of
129 passes through a grace period. The last entry should be zero,
130 as it is only incremented if a torture structure's counter
131 somehow gets incremented farther than it should.
133 Different implementations of RCU can provide implementation-specific
134 additional information. For example, SRCU provides the following:
136 srcu-torture: rtc: f8cf46a8 ver: 355 tfle: 0 rta: 356 rtaf: 0 rtf: 346 rtmbe: 0
137 srcu-torture: Reader Pipe: 559738 939 0 0 0 0 0 0 0 0 0
138 srcu-torture: Reader Batch: 560434 243 0 0 0 0 0 0 0 0
139 srcu-torture: Free-Block Circulation: 355 354 353 352 351 350 349 348 347 346 0
140 srcu-torture: per-CPU(idx=1): 0(0,1) 1(0,1) 2(0,0) 3(0,1)
142 The first four lines are similar to those for RCU. The last line shows
143 the per-CPU counter state. The numbers in parentheses are the values
144 of the "old" and "current" counters for the corresponding CPU. The
145 "idx" value maps the "old" and "current" values to the underlying array,
146 and is useful for debugging.
151 The following script may be used to torture RCU:
158 dmesg | grep torture:
160 The output can be manually inspected for the error flag of "!!!".
161 One could of course create a more elaborate script that automatically
162 checked for such errors. The "rmmod" command forces a "SUCCESS" or
163 "FAILURE" indication to be printk()ed.