diff options
author | Davidlohr Bueso <dave@stgolabs.net> | 2014-09-11 23:40:17 -0400 |
---|---|---|
committer | Paul E. McKenney <paulmck@linux.vnet.ibm.com> | 2014-09-16 16:39:13 -0400 |
commit | cdf26bb10bcb50161d452b16eb3cf2901645d625 (patch) | |
tree | ba7e69bb55b3836ae06c8bb82bf9aae72a376bc5 /Documentation/locking | |
parent | 23a8e5c2d2a481fcf382490369c27b405a650212 (diff) |
locktorture: Add documentation
Just like Documentation/RCU/torture.txt, begin a document for the
locktorture module. This module is still pretty green, so I have
just added some specific sections to the doc (general desc, params,
usage, etc.). Further development should update the file.
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
[ paulmck: Apply Randy Dunlap review comments. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Diffstat (limited to 'Documentation/locking')
-rw-r--r-- | Documentation/locking/locktorture.txt | 130 |
1 files changed, 130 insertions, 0 deletions
diff --git a/Documentation/locking/locktorture.txt b/Documentation/locking/locktorture.txt new file mode 100644 index 000000000000..3eb9b81454d0 --- /dev/null +++ b/Documentation/locking/locktorture.txt | |||
@@ -0,0 +1,130 @@ | |||
1 | Kernel Lock Torture Test Operation | ||
2 | |||
3 | CONFIG_LOCK_TORTURE_TEST | ||
4 | |||
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. | ||
13 | |||
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. | ||
19 | |||
20 | |||
21 | MODULE PARAMETERS | ||
22 | |||
23 | This module has the following parameters: | ||
24 | |||
25 | |||
26 | ** Locktorture-specific ** | ||
27 | |||
28 | nwriters_stress Number of kernel threads that will stress exclusive lock | ||
29 | ownership (writers). The default value is twice the number | ||
30 | of online CPUs. | ||
31 | |||
32 | torture_type Type of lock to torture. By default, only spinlocks will | ||
33 | be tortured. This module can torture the following locks, | ||
34 | with string values as follows: | ||
35 | |||
36 | o "lock_busted": Simulates a buggy lock implementation. | ||
37 | |||
38 | o "spin_lock": spin_lock() and spin_unlock() pairs. | ||
39 | |||
40 | o "spin_lock_irq": spin_lock_irq() and spin_unlock_irq() | ||
41 | pairs. | ||
42 | |||
43 | torture_runnable Start locktorture at boot time in the case where the | ||
44 | module is built into the kernel, otherwise wait for | ||
45 | torture_runnable to be set via sysfs before starting. | ||
46 | By default it will begin once the module is loaded. | ||
47 | |||
48 | |||
49 | ** Torture-framework (RCU + locking) ** | ||
50 | |||
51 | shutdown_secs The number of seconds to run the test before terminating | ||
52 | the test and powering off the system. The default is | ||
53 | zero, which disables test termination and system shutdown. | ||
54 | This capability is useful for automated testing. | ||
55 | |||
56 | onoff_interval The number of seconds between each attempt to execute a | ||
57 | randomly selected CPU-hotplug operation. Defaults | ||
58 | to zero, which disables CPU hotplugging. In | ||
59 | CONFIG_HOTPLUG_CPU=n kernels, locktorture will silently | ||
60 | refuse to do any CPU-hotplug operations regardless of | ||
61 | what value is specified for onoff_interval. | ||
62 | |||
63 | onoff_holdoff The number of seconds to wait until starting CPU-hotplug | ||
64 | operations. This would normally only be used when | ||
65 | locktorture was built into the kernel and started | ||
66 | automatically at boot time, in which case it is useful | ||
67 | in order to avoid confusing boot-time code with CPUs | ||
68 | coming and going. This parameter is only useful if | ||
69 | CONFIG_HOTPLUG_CPU is enabled. | ||
70 | |||
71 | stat_interval Number of seconds between statistics-related printk()s. | ||
72 | By default, locktorture will report stats every 60 seconds. | ||
73 | Setting the interval to zero causes the statistics to | ||
74 | be printed -only- when the module is unloaded, and this | ||
75 | is the default. | ||
76 | |||
77 | stutter The length of time to run the test before pausing for this | ||
78 | same period of time. Defaults to "stutter=5", so as | ||
79 | to run and pause for (roughly) five-second intervals. | ||
80 | Specifying "stutter=0" causes the test to run continuously | ||
81 | without pausing, which is the old default behavior. | ||
82 | |||
83 | shuffle_interval The number of seconds to keep the test threads affinitied | ||
84 | to a particular subset of the CPUs, defaults to 3 seconds. | ||
85 | Used in conjunction with test_no_idle_hz. | ||
86 | |||
87 | verbose Enable verbose debugging printing, via printk(). Enabled | ||
88 | by default. This extra information is mostly related to | ||
89 | high-level errors and reports from the main 'torture' | ||
90 | framework. | ||
91 | |||
92 | |||
93 | STATISTICS | ||
94 | |||
95 | Statistics are printed in the following format: | ||
96 | |||
97 | spin_lock-torture: Writes: Total: 93746064 Max/Min: 0/0 Fail: 0 | ||
98 | (A) (B) (C) (D) | ||
99 | |||
100 | (A): Lock type that is being tortured -- torture_type parameter. | ||
101 | |||
102 | (B): Number of times the lock was acquired. | ||
103 | |||
104 | (C): Min and max number of times threads failed to acquire the lock. | ||
105 | |||
106 | (D): true/false values if there were errors acquiring the lock. This should | ||
107 | -only- be positive if there is a bug in the locking primitive's | ||
108 | implementation. Otherwise a lock should never fail (i.e., spin_lock()). | ||
109 | Of course, the same applies for (C), above. A dummy example of this is | ||
110 | the "lock_busted" type. | ||
111 | |||
112 | USAGE | ||
113 | |||
114 | The following script may be used to torture locks: | ||
115 | |||
116 | #!/bin/sh | ||
117 | |||
118 | modprobe locktorture | ||
119 | sleep 3600 | ||
120 | rmmod locktorture | ||
121 | dmesg | grep torture: | ||
122 | |||
123 | The output can be manually inspected for the error flag of "!!!". | ||
124 | One could of course create a more elaborate script that automatically | ||
125 | checked for such errors. The "rmmod" command forces a "SUCCESS", | ||
126 | "FAILURE", or "RCU_HOTPLUG" indication to be printk()ed. The first | ||
127 | two are self-explanatory, while the last indicates that while there | ||
128 | were no locking failures, CPU-hotplug problems were detected. | ||
129 | |||
130 | Also see: Documentation/RCU/torture.txt | ||