Gitiles
Code Review Sign In
gerrit-public.fairphone.software / kernel / msm-4.9 / 4e3fd7a06dc20b2d8ec6892233ad2012968fe7b6 / . / Documentation / lockstat.txt
blob: cef00d42ed5be272fdeffb89e1eede3c931ef25b [file] [log] [blame]
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]1
2LOCK STATISTICS
3
4- WHAT
5
6As the name suggests, it provides statistics on locks.
7
8- WHY
9
10Because things like lock contention can severely impact performance.
11
12- HOW
13
14Lockdep already has hooks in the lock functions and maps lock instances to
Juri Lellif62508f2011-05-19 12:53:53 +0200[diff] [blame]15lock classes. We build on that (see Documentation/lockdep-design.txt).
16The graph below shows the relation between the lock functions and the various
17hooks therein.
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]18
19 __acquire
20 |
21 lock _____
22 | \
23 | __contended
24 | |
25 | <wait>
26 | _______/
27 |/
28 |
29 __acquired
30 |
31 .
32 <hold>
33 .
34 |
35 __release
36 |
37 unlock
38
39lock, unlock - the regular lock functions
40__* - the hooks
41<> - states
42
43With these hooks we provide the following statistics:
44
45 con-bounces - number of lock contention that involved x-cpu data
46 contentions - number of lock acquisitions that had to wait
47 wait time min - shortest (non-0) time we ever had to wait for a lock
48 max - longest time we ever had to wait for a lock
49 total - total time we spend waiting on this lock
50 acq-bounces - number of lock acquisitions that involved x-cpu data
51 acquisitions - number of times we took the lock
52 hold time min - shortest (non-0) time we ever held the lock
53 max - longest time we ever held the lock
54 total - total time this lock was held
55
56From these number various other statistics can be derived, such as:
57
58 hold time average = hold time total / acquisitions
59
60These numbers are gathered per lock class, per read/write state (when
61applicable).
62
63It also tracks 4 contention points per class. A contention point is a call site
64that had to wait on lock acquisition.
65
Frank Rowandf5754bf2009-11-19 13:54:05 -0800[diff] [blame]66 - CONFIGURATION
67
68Lock statistics are enabled via CONFIG_LOCK_STATS.
69
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]70 - USAGE
71
Frank Rowandf5754bf2009-11-19 13:54:05 -0800[diff] [blame]72Enable collection of statistics:
73
74# echo 1 >/proc/sys/kernel/lock_stat
75
76Disable collection of statistics:
77
78# echo 0 >/proc/sys/kernel/lock_stat
79
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]80Look at the current lock statistics:
81
82( line numbers not part of actual output, done for clarity in the explanation
83 below )
84
85# less /proc/lock_stat
86
Peter Zijlstrac7e78cf2008-10-16 23:17:09 +0200[diff] [blame]8701 lock_stat version 0.3
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]8802 -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
8903 class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total
9004 -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
9105
Peter Zijlstrac7e78cf2008-10-16 23:17:09 +0200[diff] [blame]9206 &mm->mmap_sem-W: 233 538 18446744073708 22924.27 607243.51 1342 45806 1.71 8595.89 1180582.34
9307 &mm->mmap_sem-R: 205 587 18446744073708 28403.36 731975.00 1940 412426 0.58 187825.45 6307502.88
9408 ---------------
9509 &mm->mmap_sem 487 [<ffffffff8053491f>] do_page_fault+0x466/0x928
9610 &mm->mmap_sem 179 [<ffffffff802a6200>] sys_mprotect+0xcd/0x21d
9711 &mm->mmap_sem 279 [<ffffffff80210a57>] sys_mmap+0x75/0xce
9812 &mm->mmap_sem 76 [<ffffffff802a490b>] sys_munmap+0x32/0x59
9913 ---------------
10014 &mm->mmap_sem 270 [<ffffffff80210a57>] sys_mmap+0x75/0xce
10115 &mm->mmap_sem 431 [<ffffffff8053491f>] do_page_fault+0x466/0x928
10216 &mm->mmap_sem 138 [<ffffffff802a490b>] sys_munmap+0x32/0x59
10317 &mm->mmap_sem 145 [<ffffffff802a6200>] sys_mprotect+0xcd/0x21d
10418
10519 ...............................................................................................................................................................................................
10620
10721 dcache_lock: 621 623 0.52 118.26 1053.02 6745 91930 0.29 316.29 118423.41
10822 -----------
10923 dcache_lock 179 [<ffffffff80378274>] _atomic_dec_and_lock+0x34/0x54
11024 dcache_lock 113 [<ffffffff802cc17b>] d_alloc+0x19a/0x1eb
11125 dcache_lock 99 [<ffffffff802ca0dc>] d_rehash+0x1b/0x44
11226 dcache_lock 104 [<ffffffff802cbca0>] d_instantiate+0x36/0x8a
11327 -----------
11428 dcache_lock 192 [<ffffffff80378274>] _atomic_dec_and_lock+0x34/0x54
11529 dcache_lock 98 [<ffffffff802ca0dc>] d_rehash+0x1b/0x44
11630 dcache_lock 72 [<ffffffff802cc17b>] d_alloc+0x19a/0x1eb
11731 dcache_lock 112 [<ffffffff802cbca0>] d_instantiate+0x36/0x8a
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]118
119This excerpt shows the first two lock class statistics. Line 01 shows the
120output version - each time the format changes this will be updated. Line 02-04
Peter Zijlstrac7e78cf2008-10-16 23:17:09 +0200[diff] [blame]121show the header with column descriptions. Lines 05-18 and 20-31 show the actual
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]122statistics. These statistics come in two parts; the actual stats separated by a
Peter Zijlstrac7e78cf2008-10-16 23:17:09 +0200[diff] [blame]123short separator (line 08, 13) from the contention points.
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]124
Peter Zijlstrac7e78cf2008-10-16 23:17:09 +0200[diff] [blame]125The first lock (05-18) is a read/write lock, and shows two lines above the
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]126short separator. The contention points don't match the column descriptors,
Peter Zijlstrac7e78cf2008-10-16 23:17:09 +0200[diff] [blame]127they have two: contentions and [<IP>] symbol. The second set of contention
128points are the points we're contending with.
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]129
Peter Zijlstracdad7222008-09-03 16:00:57 +0200[diff] [blame]130The integer part of the time values is in us.
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]131
Juri Lellif62508f2011-05-19 12:53:53 +0200[diff] [blame]132Dealing with nested locks, subclasses may appear:
133
13432...............................................................................................................................................................................................
13533
13634 &rq->lock: 13128 13128 0.43 190.53 103881.26 97454 3453404 0.00 401.11 13224683.11
13735 ---------
13836 &rq->lock 645 [<ffffffff8103bfc4>] task_rq_lock+0x43/0x75
13937 &rq->lock 297 [<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
14038 &rq->lock 360 [<ffffffff8103c4c5>] select_task_rq_fair+0x1f0/0x74a
14139 &rq->lock 428 [<ffffffff81045f98>] scheduler_tick+0x46/0x1fb
14240 ---------
14341 &rq->lock 77 [<ffffffff8103bfc4>] task_rq_lock+0x43/0x75
14442 &rq->lock 174 [<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
14543 &rq->lock 4715 [<ffffffff8103ed4b>] double_rq_lock+0x42/0x54
14644 &rq->lock 893 [<ffffffff81340524>] schedule+0x157/0x7b8
14745
14846...............................................................................................................................................................................................
14947
15048 &rq->lock/1: 11526 11488 0.33 388.73 136294.31 21461 38404 0.00 37.93 109388.53
15149 -----------
15250 &rq->lock/1 11526 [<ffffffff8103ed58>] double_rq_lock+0x4f/0x54
15351 -----------
15452 &rq->lock/1 5645 [<ffffffff8103ed4b>] double_rq_lock+0x42/0x54
15553 &rq->lock/1 1224 [<ffffffff81340524>] schedule+0x157/0x7b8
15654 &rq->lock/1 4336 [<ffffffff8103ed58>] double_rq_lock+0x4f/0x54
15755 &rq->lock/1 181 [<ffffffff8104ba65>] try_to_wake_up+0x127/0x25a
158
159Line 48 shows statistics for the second subclass (/1) of &rq->lock class
160(subclass starts from 0), since in this case, as line 50 suggests,
161double_rq_lock actually acquires a nested lock of two spinlocks.
162
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]163View the top contending locks:
164
165# grep : /proc/lock_stat | head
166 &inode->i_data.tree_lock-W: 15 21657 0.18 1093295.30 11547131054.85 58 10415 0.16 87.51 6387.60
167 &inode->i_data.tree_lock-R: 0 0 0.00 0.00 0.00 23302 231198 0.25 8.45 98023.38
168 dcache_lock: 1037 1161 0.38 45.32 774.51 6611 243371 0.15 306.48 77387.24
169 &inode->i_mutex: 161 286 18446744073709 62882.54 1244614.55 3653 20598 18446744073709 62318.60 1693822.74
170 &zone->lru_lock: 94 94 0.53 7.33 92.10 4366 32690 0.29 59.81 16350.06
Peter Zijlstra3d48ae42011-05-24 17:12:06 -0700[diff] [blame]171 &inode->i_data.i_mmap_mutex: 79 79 0.40 3.77 53.03 11779 87755 0.28 116.93 29898.44
Peter Zijlstraa560aa42007-10-07 00:24:33 -0700[diff] [blame]172 &q->__queue_lock: 48 50 0.52 31.62 86.31 774 13131 0.17 113.08 12277.52
173 &rq->rq_lock_key: 43 47 0.74 68.50 170.63 3706 33929 0.22 107.99 17460.62
174 &rq->rq_lock_key#2: 39 46 0.75 6.68 49.03 2979 32292 0.17 125.17 17137.63
175 tasklist_lock-W: 15 15 1.45 10.87 32.70 1201 7390 0.58 62.55 13648.47
176
177Clear the statistics:
178
179# echo 0 > /proc/lock_stat