Documentation/RCU/torture.txt

   1 RCU Torture Test Operation
   2
   3
   4 CONFIG_RCU_TORTURE_TEST
   5
   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.
  12
  13 CONFIG_RCU_TORTURE_TEST_RUNNABLE
  14
  15 It is also possible to specify CONFIG_RCU_TORTURE_TEST=y, which will
  16 result in the tests being loaded into the base kernel.  In this case,
  17 the CONFIG_RCU_TORTURE_TEST_RUNNABLE config option is used to specify
  18 whether the RCU torture tests are to be started immediately during
  19 boot or whether the /proc/sys/kernel/rcutorture_runnable file is used
  20 to enable them.  This /proc file can be used to repeatedly pause and
  21 restart the tests, regardless of the initial state specified by the
  22 CONFIG_RCU_TORTURE_TEST_RUNNABLE config option.
  23
  24 You 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
  26 this can sometimes be useful in finding boot-time bugs.
  27
  28
  29 MODULE PARAMETERS
  30
  31 This module has the following parameters:
  32
  33 nreaders        This is the number of RCU reading threads supported.
  34                 The default is twice the number of CPUs.  Why twice?
  35                 To properly exercise RCU implementations with preemptible
  36                 read-side critical sections.
  37
  38 nfakewriters    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
  47 stat_interval   The number of seconds between output of torture
  48                 statistics (via printk()).  Regardless of the interval,
  49                 statistics are printed when the module is unloaded.
  50                 Setting the interval to zero causes the statistics to
  51                 be printed -only- when the module is unloaded, and this
  52                 is the default.
  53
  54 shuffle_interval
  55                 The number of seconds to keep the test threads affinitied
  56                 to a particular subset of the CPUs, defaults to 3 seconds.
  57                 Used in conjunction with test_no_idle_hz.
  58
  59 stutter         The length of time to run the test before pausing for this
  60                 same period of time.  Defaults to "stutter=5", so as
  61                 to run and pause for (roughly) five-second intervals.
  62                 Specifying "stutter=0" causes the test to run continuously
  63                 without pausing, which is the old default behavior.
  64
  65 test_no_idle_hz Whether or not to test the ability of RCU to operate in
  66                 a kernel that disables the scheduling-clock interrupt to
  67                 idle CPUs.  Boolean parameter, "1" to test, "0" otherwise.
  68                 Defaults to omitting this test.
  69
  70 torture_type    The type of RCU to test: "rcu" for the rcu_read_lock() API,
  71                 "rcu_sync" for rcu_read_lock() with synchronous reclamation,
  72                 "rcu_bh" for the rcu_read_lock_bh() API, "rcu_bh_sync" for
  73                 rcu_read_lock_bh() with synchronous reclamation, "srcu" for
  74                 the "srcu_read_lock()" API, and "sched" for the use of
  75                 preempt_disable() together with synchronize_sched().
  76
  77 verbose         Enable debug printk()s.  Default is disabled.
  78
  79
  80 OUTPUT
  81
  82 The statistics output is as follows:
  83
  84         rcu-torture: --- Start of test: nreaders=16 stat_interval=0 verbose=0
  85         rcu-torture: rtc: 0000000000000000 ver: 1916 tfle: 0 rta: 1916 rtaf: 0 rtf: 1915
  86         rcu-torture: Reader Pipe:  1466408 9747 0 0 0 0 0 0 0 0 0
  87         rcu-torture: Reader Batch:  1464477 11678 0 0 0 0 0 0 0 0
  88         rcu-torture: Free-Block Circulation:  1915 1915 1915 1915 1915 1915 1915 1915 1915 1915 0
  89         rcu-torture: --- End of test
  90
  91 The command "dmesg | grep torture:" will extract this information on
  92 most systems.  On more esoteric configurations, it may be necessary to
  93 use other commands to access the output of the printk()s used by
  94 the RCU torture test.  The printk()s use KERN_ALERT, so they should
  95 be evident.  ;-)
  96
  97 The entries are as follows:
  98
  99 o       "rtc": The hexadecimal address of the structure currently visible
 100         to readers.
 101
 102 o       "ver": The number of times since boot that the rcutw writer task
 103         has changed the structure visible to readers.
 104
 105 o       "tfle": If non-zero, indicates that the "torture freelist"
 106         containing structure to be placed into the "rtc" area is empty.
 107         This condition is important, since it can fool you into thinking
 108         that RCU is working when it is not.  :-/
 109
 110 o       "rta": Number of structures allocated from the torture freelist.
 111
 112 o       "rtaf": Number of allocations from the torture freelist that have
 113         failed due to the list being empty.
 114
 115 o       "rtf": Number of frees into the torture freelist.
 116
 117 o       "Reader Pipe": Histogram of "ages" of structures seen by readers.
 118         If any entries past the first two are non-zero, RCU is broken.
 119         And rcutorture prints the error flag string "!!!" to make sure
 120         you notice.  The age of a newly allocated structure is zero,
 121         it becomes one when removed from reader visibility, and is
 122         incremented once per grace period subsequently -- and is freed
 123         after passing through (RCU_TORTURE_PIPE_LEN-2) grace periods.
 124
 125         The output displayed above was taken from a correctly working
 126         RCU.  If you want to see what it looks like when broken, break
 127         it yourself.  ;-)
 128
 129 o       "Reader Batch": Another histogram of "ages" of structures seen
 130         by readers, but in terms of counter flips (or batches) rather
 131         than in terms of grace periods.  The legal number of non-zero
 132         entries is again two.  The reason for this separate view is that
 133         it is sometimes easier to get the third entry to show up in the
 134         "Reader Batch" list than in the "Reader Pipe" list.
 135
 136 o       "Free-Block Circulation": Shows the number of torture structures
 137         that have reached a given point in the pipeline.  The first element
 138         should closely correspond to the number of structures allocated,
 139         the second to the number that have been removed from reader view,
 140         and all but the last remaining to the corresponding number of
 141         passes through a grace period.  The last entry should be zero,
 142         as it is only incremented if a torture structure's counter
 143         somehow gets incremented farther than it should.
 144
 145 Different implementations of RCU can provide implementation-specific
 146 additional information.  For example, SRCU provides the following:
 147
 148         srcu-torture: rtc: f8cf46a8 ver: 355 tfle: 0 rta: 356 rtaf: 0 rtf: 346 rtmbe: 0
 149         srcu-torture: Reader Pipe:  559738 939 0 0 0 0 0 0 0 0 0
 150         srcu-torture: Reader Batch:  560434 243 0 0 0 0 0 0 0 0
 151         srcu-torture: Free-Block Circulation:  355 354 353 352 351 350 349 348 347 346 0
 152         srcu-torture: per-CPU(idx=1): 0(0,1) 1(0,1) 2(0,0) 3(0,1)
 153
 154 The first four lines are similar to those for RCU.  The last line shows
 155 the per-CPU counter state.  The numbers in parentheses are the values
 156 of the "old" and "current" counters for the corresponding CPU.  The
 157 "idx" value maps the "old" and "current" values to the underlying array,
 158 and is useful for debugging.
 159
 160
 161 USAGE
 162
 163 The following script may be used to torture RCU:
 164
 165         #!/bin/sh
 166
 167         modprobe rcutorture
 168         sleep 100
 169         rmmod rcutorture
 170         dmesg | grep torture:
 171
 172 The output can be manually inspected for the error flag of "!!!".
 173 One could of course create a more elaborate script that automatically
 174 checked for such errors.  The "rmmod" command forces a "SUCCESS" or
 175 "FAILURE" indication to be printk()ed.