1 Kernel Lock Torture Test Operation
3 CONFIG_LOCK_TORTURE_TEST
5 The CONFIG LOCK_TORTURE_TEST config option provides a kernel module
6 that runs torture tests on core kernel locking primitives. The kernel
7 module, 'locktorture', may be built after the fact on the running
8 kernel to be tested, if desired. The tests periodically output status
9 messages via printk(), which can be examined via the dmesg (perhaps
10 grepping for "torture"). The test is started when the module is loaded,
11 and stops when the module is unloaded. This program is based on how RCU
12 is tortured, via rcutorture.
14 This torture test consists of creating a number of kernel threads which
15 acquire the lock and hold it for specific amount of time, thus simulating
16 different critical region behaviors. The amount of contention on the lock
17 can be simulated by either enlarging this critical region hold time and/or
18 creating more kthreads.
23 This module has the following parameters:
26 ** Locktorture-specific **
28 nwriters_stress Number of kernel threads that will stress exclusive lock
29 ownership (writers). The default value is twice the number
32 nreaders_stress Number of kernel threads that will stress shared lock
33 ownership (readers). The default is the same amount of writer
34 locks. If the user did not specify nwriters_stress, then
35 both readers and writers be the amount of online CPUs.
37 torture_type Type of lock to torture. By default, only spinlocks will
38 be tortured. This module can torture the following locks,
39 with string values as follows:
41 o "lock_busted": Simulates a buggy lock implementation.
43 o "spin_lock": spin_lock() and spin_unlock() pairs.
45 o "spin_lock_irq": spin_lock_irq() and spin_unlock_irq()
48 o "rw_lock": read/write lock() and unlock() rwlock pairs.
50 o "rw_lock_irq": read/write lock_irq() and unlock_irq()
53 o "mutex_lock": mutex_lock() and mutex_unlock() pairs.
55 o "rtmutex_lock": rtmutex_lock() and rtmutex_unlock()
56 pairs. Kernel must have CONFIG_RT_MUTEX=y.
58 o "rwsem_lock": read/write down() and up() semaphore pairs.
60 torture_runnable Start locktorture at boot time in the case where the
61 module is built into the kernel, otherwise wait for
62 torture_runnable to be set via sysfs before starting.
63 By default it will begin once the module is loaded.
66 ** Torture-framework (RCU + locking) **
68 shutdown_secs The number of seconds to run the test before terminating
69 the test and powering off the system. The default is
70 zero, which disables test termination and system shutdown.
71 This capability is useful for automated testing.
73 onoff_interval The number of seconds between each attempt to execute a
74 randomly selected CPU-hotplug operation. Defaults
75 to zero, which disables CPU hotplugging. In
76 CONFIG_HOTPLUG_CPU=n kernels, locktorture will silently
77 refuse to do any CPU-hotplug operations regardless of
78 what value is specified for onoff_interval.
80 onoff_holdoff The number of seconds to wait until starting CPU-hotplug
81 operations. This would normally only be used when
82 locktorture was built into the kernel and started
83 automatically at boot time, in which case it is useful
84 in order to avoid confusing boot-time code with CPUs
85 coming and going. This parameter is only useful if
86 CONFIG_HOTPLUG_CPU is enabled.
88 stat_interval Number of seconds between statistics-related printk()s.
89 By default, locktorture will report stats every 60 seconds.
90 Setting the interval to zero causes the statistics to
91 be printed -only- when the module is unloaded, and this
94 stutter The length of time to run the test before pausing for this
95 same period of time. Defaults to "stutter=5", so as
96 to run and pause for (roughly) five-second intervals.
97 Specifying "stutter=0" causes the test to run continuously
98 without pausing, which is the old default behavior.
100 shuffle_interval The number of seconds to keep the test threads affinitied
101 to a particular subset of the CPUs, defaults to 3 seconds.
102 Used in conjunction with test_no_idle_hz.
104 verbose Enable verbose debugging printing, via printk(). Enabled
105 by default. This extra information is mostly related to
106 high-level errors and reports from the main 'torture'
112 Statistics are printed in the following format:
114 spin_lock-torture: Writes: Total: 93746064 Max/Min: 0/0 Fail: 0
117 (A): Lock type that is being tortured -- torture_type parameter.
119 (B): Number of writer lock acquisitions. If dealing with a read/write primitive
120 a second "Reads" statistics line is printed.
122 (C): Number of times the lock was acquired.
124 (D): Min and max number of times threads failed to acquire the lock.
126 (E): true/false values if there were errors acquiring the lock. This should
127 -only- be positive if there is a bug in the locking primitive's
128 implementation. Otherwise a lock should never fail (i.e., spin_lock()).
129 Of course, the same applies for (C), above. A dummy example of this is
130 the "lock_busted" type.
134 The following script may be used to torture locks:
141 dmesg | grep torture:
143 The output can be manually inspected for the error flag of "!!!".
144 One could of course create a more elaborate script that automatically
145 checked for such errors. The "rmmod" command forces a "SUCCESS",
146 "FAILURE", or "RCU_HOTPLUG" indication to be printk()ed. The first
147 two are self-explanatory, while the last indicates that while there
148 were no locking failures, CPU-hotplug problems were detected.
150 Also see: Documentation/RCU/torture.txt