blob: ec6998b1b6d04f3139ed6c066537cc059c89838d [file] [log] [blame]
Paul E. McKenney64db4cf2008-12-18 21:55:32 +01001CONFIG_RCU_TRACE debugfs Files and Formats
2
3
Paul E. McKenney8e79e1f2010-11-04 14:31:19 -07004The rcutree and rcutiny implementations of RCU provide debugfs trace
5output that summarizes counters and state. This information is useful for
6debugging RCU itself, and can sometimes also help to debug abuses of RCU.
7The following sections describe the debugfs files and formats, first
8for rcutree and next for rcutiny.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +01009
10
Pranith Kumar28f65692014-09-22 14:00:48 -040011CONFIG_TREE_RCU and CONFIG_PREEMPT_RCU debugfs Files and Formats
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010012
Paul E. McKenney40e80c42012-10-31 13:00:15 -070013These implementations of RCU provide several debugfs directories under the
Paul E. McKenney90e6ac32011-04-06 15:20:47 -070014top-level directory "rcu":
15
Paul E. McKenney40e80c42012-10-31 13:00:15 -070016rcu/rcu_bh
17rcu/rcu_preempt
18rcu/rcu_sched
19
20Each directory contains files for the corresponding flavor of RCU.
Pranith Kumar28f65692014-09-22 14:00:48 -040021Note that rcu/rcu_preempt is only present for CONFIG_PREEMPT_RCU.
Paul E. McKenney40e80c42012-10-31 13:00:15 -070022For CONFIG_TREE_RCU, the RCU flavor maps onto the RCU-sched flavor,
23so that activity for both appears in rcu/rcu_sched.
24
25In addition, the following file appears in the top-level directory:
26rcu/rcutorture. This file displays rcutorture test progress. The output
27of "cat rcu/rcutorture" looks as follows:
28
29rcutorture test sequence: 0 (test in progress)
30rcutorture update version number: 615
31
32The first line shows the number of rcutorture tests that have completed
33since boot. If a test is currently running, the "(test in progress)"
34string will appear as shown above. The second line shows the number of
35update cycles that the current test has started, or zero if there is
36no test in progress.
37
38
39Within each flavor directory (rcu/rcu_bh, rcu/rcu_sched, and possibly
40also rcu/rcu_preempt) the following files will be present:
41
42rcudata:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -070043 Displays fields in struct rcu_data.
Paul E. McKenneyd484a212012-10-31 13:22:08 -070044rcuexp:
45 Displays statistics for expedited grace periods.
Paul E. McKenney40e80c42012-10-31 13:00:15 -070046rcugp:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -070047 Displays grace-period counters.
Paul E. McKenney40e80c42012-10-31 13:00:15 -070048rcuhier:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -070049 Displays the struct rcu_node hierarchy.
Paul E. McKenney40e80c42012-10-31 13:00:15 -070050rcu_pending:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -070051 Displays counts of the reasons rcu_pending() decided that RCU had
52 work to do.
Paul E. McKenney40e80c42012-10-31 13:00:15 -070053rcuboost:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -070054 Displays RCU boosting statistics. Only present if
55 CONFIG_RCU_BOOST=y.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010056
Paul E. McKenney40e80c42012-10-31 13:00:15 -070057The output of "cat rcu/rcu_preempt/rcudata" looks as follows:
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010058
Paul E. McKenney0d43eb32015-08-06 14:17:29 -070059 0!c=30455 g=30456 cnq=1/0:1 dt=126535/140000000000000/0 df=2002 of=4 ql=0/0 qs=N... b=10 ci=74572 nci=0 co=1131 ca=716
60 1!c=30719 g=30720 cnq=1/0:0 dt=132007/140000000000000/0 df=1874 of=10 ql=0/0 qs=N... b=10 ci=123209 nci=0 co=685 ca=982
61 2!c=30150 g=30151 cnq=1/1:1 dt=138537/140000000000000/0 df=1707 of=8 ql=0/0 qs=N... b=10 ci=80132 nci=0 co=1328 ca=1458
62 3 c=31249 g=31250 cnq=1/1:0 dt=107255/140000000000000/0 df=1749 of=6 ql=0/450 qs=NRW. b=10 ci=151700 nci=0 co=509 ca=622
63 4!c=29502 g=29503 cnq=1/0:1 dt=83647/140000000000000/0 df=965 of=5 ql=0/0 qs=N... b=10 ci=65643 nci=0 co=1373 ca=1521
64 5 c=31201 g=31202 cnq=1/0:1 dt=70422/0/0 df=535 of=7 ql=0/0 qs=.... b=10 ci=58500 nci=0 co=764 ca=698
65 6!c=30253 g=30254 cnq=1/0:1 dt=95363/140000000000000/0 df=780 of=5 ql=0/0 qs=N... b=10 ci=100607 nci=0 co=1414 ca=1353
66 7 c=31178 g=31178 cnq=1/0:0 dt=91536/0/0 df=547 of=4 ql=0/0 qs=.... b=10 ci=109819 nci=0 co=1115 ca=969
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010067
Paul E. McKenney40e80c42012-10-31 13:00:15 -070068This file has one line per CPU, or eight for this 8-CPU system.
69The fields are as follows:
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010070
71o The number at the beginning of each line is the CPU number.
72 CPUs numbers followed by an exclamation mark are offline,
73 but have been online at least once since boot. There will be
74 no output for CPUs that have never been online, which can be
75 a good thing in the surprisingly common case where NR_CPUS is
76 substantially larger than the number of actual CPUs.
77
78o "c" is the count of grace periods that this CPU believes have
Paul E. McKenney40e80c42012-10-31 13:00:15 -070079 completed. Offlined CPUs and CPUs in dynticks idle mode may lag
80 quite a ways behind, for example, CPU 4 under "rcu_sched" above,
81 which has been offline through 16 RCU grace periods. It is not
82 unusual to see offline CPUs lagging by thousands of grace periods.
83 Note that although the grace-period number is an unsigned long,
84 it is printed out as a signed long to allow more human-friendly
85 representation near boot time.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010086
87o "g" is the count of grace periods that this CPU believes have
Paul E. McKenney2fa218d2011-03-27 21:37:58 -070088 started. Again, offlined CPUs and CPUs in dynticks idle mode
89 may lag behind. If the "c" and "g" values are equal, this CPU
90 has already reported a quiescent state for the last RCU grace
91 period that it is aware of, otherwise, the CPU believes that it
92 owes RCU a quiescent state.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +010093
94o "pq" indicates that this CPU has passed through a quiescent state
95 for the current grace period. It is possible for "pq" to be
96 "1" and "c" different than "g", which indicates that although
97 the CPU has passed through a quiescent state, either (1) this
98 CPU has not yet reported that fact, (2) some other CPU has not
99 yet reported for this grace period, or (3) both.
100
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100101o "qp" indicates that RCU still expects a quiescent state from
Paul E. McKenney2fa218d2011-03-27 21:37:58 -0700102 this CPU. Offlined CPUs and CPUs in dyntick idle mode might
103 well have qp=1, which is OK: RCU is still ignoring them.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100104
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100105o "dt" is the current value of the dyntick counter that is incremented
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700106 when entering or leaving idle, either due to a context switch or
107 due to an interrupt. This number is even if the CPU is in idle
108 from RCU's viewpoint and odd otherwise. The number after the
109 first "/" is the interrupt nesting depth when in idle state,
110 or a large number added to the interrupt-nesting depth when
111 running a non-idle task. Some architectures do not accurately
112 count interrupt nesting when running in non-idle kernel context,
113 which can result in interesting anomalies such as negative
114 interrupt-nesting levels. The number after the second "/"
115 is the NMI nesting depth.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100116
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100117o "df" is the number of times that some other CPU has forced a
118 quiescent state on behalf of this CPU due to this CPU being in
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700119 idle state.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100120
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100121o "of" is the number of times that some other CPU has forced a
122 quiescent state on behalf of this CPU due to this CPU being
Paul E. McKenney2fa218d2011-03-27 21:37:58 -0700123 offline. In a perfect world, this might never happen, but it
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100124 turns out that offlining and onlining a CPU can take several grace
125 periods, and so there is likely to be an extended period of time
126 when RCU believes that the CPU is online when it really is not.
127 Please note that erring in the other direction (RCU believing a
128 CPU is offline when it is really alive and kicking) is a fatal
129 error, so it makes sense to err conservatively.
130
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100131o "ql" is the number of RCU callbacks currently residing on
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700132 this CPU. The first number is the number of "lazy" callbacks
133 that are known to RCU to only be freeing memory, and the number
134 after the "/" is the total number of callbacks, lazy or not.
135 These counters count callbacks regardless of what phase of
136 grace-period processing that they are in (new, waiting for
137 grace period to start, waiting for grace period to end, ready
138 to invoke).
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100139
Paul E. McKenney0ac3d132011-03-28 15:47:07 -0700140o "qs" gives an indication of the state of the callback queue
141 with four characters:
142
143 "N" Indicates that there are callbacks queued that are not
144 ready to be handled by the next grace period, and thus
145 will be handled by the grace period following the next
146 one.
147
148 "R" Indicates that there are callbacks queued that are
149 ready to be handled by the next grace period.
150
151 "W" Indicates that there are callbacks queued that are
152 waiting on the current grace period.
153
154 "D" Indicates that there are callbacks queued that have
155 already been handled by a prior grace period, and are
156 thus waiting to be invoked. Note that callbacks in
157 the process of being invoked are not counted here.
158 Callbacks in the process of being invoked are those
159 that have been removed from the rcu_data structures
160 queues by rcu_do_batch(), but which have not yet been
161 invoked.
162
163 If there are no callbacks in a given one of the above states,
164 the corresponding character is replaced by ".".
165
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700166o "b" is the batch limit for this CPU. If more than this number
167 of RCU callbacks is ready to invoke, then the remainder will
168 be deferred.
169
170o "ci" is the number of RCU callbacks that have been invoked for
171 this CPU. Note that ci+nci+ql is the number of callbacks that have
172 been registered in absence of CPU-hotplug activity.
173
174o "nci" is the number of RCU callbacks that have been offloaded from
175 this CPU. This will always be zero unless the kernel was built
176 with CONFIG_RCU_NOCB_CPU=y and the "rcu_nocbs=" kernel boot
177 parameter was specified.
178
179o "co" is the number of RCU callbacks that have been orphaned due to
180 this CPU going offline. These orphaned callbacks have been moved
181 to an arbitrarily chosen online CPU.
182
183o "ca" is the number of RCU callbacks that have been adopted by this
184 CPU due to other CPUs going offline. Note that ci+co-ca+ql is
185 the number of RCU callbacks registered on this CPU.
186
187
188Kernels compiled with CONFIG_RCU_BOOST=y display the following from
189/debug/rcu/rcu_preempt/rcudata:
190
Paul E. McKenney0d43eb32015-08-06 14:17:29 -0700191 0!c=12865 g=12866 cnq=1/0:1 dt=83113/140000000000000/0 df=288 of=11 ql=0/0 qs=N... kt=0/O ktl=944 b=10 ci=60709 nci=0 co=748 ca=871
192 1 c=14407 g=14408 cnq=1/0:0 dt=100679/140000000000000/0 df=378 of=7 ql=0/119 qs=NRW. kt=0/W ktl=9b6 b=10 ci=109740 nci=0 co=589 ca=485
193 2 c=14407 g=14408 cnq=1/0:0 dt=105486/0/0 df=90 of=9 ql=0/89 qs=NRW. kt=0/W ktl=c0c b=10 ci=83113 nci=0 co=533 ca=490
194 3 c=14407 g=14408 cnq=1/0:0 dt=107138/0/0 df=142 of=8 ql=0/188 qs=NRW. kt=0/W ktl=b96 b=10 ci=121114 nci=0 co=426 ca=290
195 4 c=14405 g=14406 cnq=1/0:1 dt=50238/0/0 df=706 of=7 ql=0/0 qs=.... kt=0/W ktl=812 b=10 ci=34929 nci=0 co=643 ca=114
196 5!c=14168 g=14169 cnq=1/0:0 dt=45465/140000000000000/0 df=161 of=11 ql=0/0 qs=N... kt=0/O ktl=b4d b=10 ci=47712 nci=0 co=677 ca=722
197 6 c=14404 g=14405 cnq=1/0:0 dt=59454/0/0 df=94 of=6 ql=0/0 qs=.... kt=0/W ktl=e57 b=10 ci=55597 nci=0 co=701 ca=811
198 7 c=14407 g=14408 cnq=1/0:1 dt=68850/0/0 df=31 of=8 ql=0/0 qs=.... kt=0/W ktl=14bd b=10 ci=77475 nci=0 co=508 ca=1042
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700199
200This is similar to the output discussed above, but contains the following
201additional fields:
202
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700203o "kt" is the per-CPU kernel-thread state. The digit preceding
Paul E. McKenney15ba0ba2011-04-06 16:01:16 -0700204 the first slash is zero if there is no work pending and 1
Paul E. McKenney5ece5ba2011-04-22 18:08:51 -0700205 otherwise. The character between the first pair of slashes is
206 as follows:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700207
208 "S" The kernel thread is stopped, in other words, all
209 CPUs corresponding to this rcu_node structure are
210 offline.
211
212 "R" The kernel thread is running.
213
214 "W" The kernel thread is waiting because there is no work
215 for it to do.
216
Paul E. McKenney15ba0ba2011-04-06 16:01:16 -0700217 "O" The kernel thread is waiting because it has been
218 forced off of its designated CPU or because its
219 ->cpus_allowed mask permits it to run on other than
220 its designated CPU.
221
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700222 "Y" The kernel thread is yielding to avoid hogging CPU.
223
224 "?" Unknown value, indicates a bug.
225
Paul E. McKenney15ba0ba2011-04-06 16:01:16 -0700226 The number after the final slash is the CPU that the kthread
227 is actually running on.
228
Paul E. McKenneyd5988af2011-06-15 23:07:55 -0700229 This field is displayed only for CONFIG_RCU_BOOST kernels.
230
Paul E. McKenney5ece5ba2011-04-22 18:08:51 -0700231o "ktl" is the low-order 16 bits (in hexadecimal) of the count of
232 the number of times that this CPU's per-CPU kthread has gone
233 through its loop servicing invoke_rcu_cpu_kthread() requests.
234
Paul E. McKenneyd5988af2011-06-15 23:07:55 -0700235 This field is displayed only for CONFIG_RCU_BOOST kernels.
236
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100237
Paul E. McKenneyd484a212012-10-31 13:22:08 -0700238The output of "cat rcu/rcu_preempt/rcuexp" looks as follows:
239
Paul E. McKenneycdacbe12015-07-11 16:24:45 -0700240s=21872 wd0=0 wd1=0 wd2=0 wd3=5 n=0 enq=0 sc=21872
Paul E. McKenneyd484a212012-10-31 13:22:08 -0700241
242These fields are as follows:
243
Paul E. McKenneycdacbe12015-07-11 16:24:45 -0700244o "s" is the sequence number, with an odd number indicating that
245 an expedited grace period is in progress.
Paul E. McKenneyd484a212012-10-31 13:22:08 -0700246
Paul E. McKenneycdacbe12015-07-11 16:24:45 -0700247o "wd0", "wd1", "wd2", and "wd3" are the number of times that an
248 attempt to start an expedited grace period found that someone
249 else had completed an expedited grace period that satisfies the
Paul E. McKenneyd484a212012-10-31 13:22:08 -0700250 attempted request. "Our work is done."
251
Paul E. McKenneycdacbe12015-07-11 16:24:45 -0700252o "n" is number of times that a concurrent CPU-hotplug operation
253 forced a fallback to a normal grace period.
254
255o "enq" is the number of quiescent states still outstanding.
Paul E. McKenneyd484a212012-10-31 13:22:08 -0700256
257o "sc" is the number of times that the attempt to start a
258 new expedited grace period succeeded.
259
Paul E. McKenneyd484a212012-10-31 13:22:08 -0700260
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700261The output of "cat rcu/rcu_preempt/rcugp" looks as follows:
Paul E. McKenney269dcc12010-09-07 14:23:09 -0700262
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700263completed=31249 gpnum=31250 age=1 max=18
Paul E. McKenney269dcc12010-09-07 14:23:09 -0700264
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700265These fields are taken from the rcu_state structure, and are as follows:
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100266
267o "completed" is the number of grace periods that have completed.
268 It is comparable to the "c" field from rcu/rcudata in that a
269 CPU whose "c" field matches the value of "completed" is aware
270 that the corresponding RCU grace period has completed.
271
272o "gpnum" is the number of grace periods that have started. It is
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700273 similarly comparable to the "g" field from rcu/rcudata in that
274 a CPU whose "g" field matches the value of "gpnum" is aware that
275 the corresponding RCU grace period has started.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100276
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700277 If these two fields are equal, then there is no grace period
278 in progress, in other words, RCU is idle. On the other hand,
279 if the two fields differ (as they are above), then an RCU grace
280 period is in progress.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100281
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700282o "age" is the number of jiffies that the current grace period
283 has extended for, or zero if there is no grace period currently
284 in effect.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100285
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700286o "max" is the age in jiffies of the longest-duration grace period
287 thus far.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100288
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700289The output of "cat rcu/rcu_preempt/rcuhier" looks as follows:
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100290
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700291c=14407 g=14408 s=0 jfq=2 j=c863 nfqs=12040/nfqsng=0(12040) fqlh=1051 oqlen=0/0
2923/3 ..>. 0:7 ^0
293e/e ..>. 0:3 ^0 d/d ..>. 4:7 ^1
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100294
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700295The fields are as follows:
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100296
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700297o "c" is exactly the same as "completed" under rcu/rcu_preempt/rcugp.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100298
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700299o "g" is exactly the same as "gpnum" under rcu/rcu_preempt/rcugp.
300
301o "s" is the current state of the force_quiescent_state()
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100302 state machine.
303
304o "jfq" is the number of jiffies remaining for this grace period
305 before force_quiescent_state() is invoked to help push things
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700306 along. Note that CPUs in idle mode throughout the grace period
307 will not report on their own, but rather must be check by some
308 other CPU via force_quiescent_state().
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100309
310o "j" is the low-order four hex digits of the jiffies counter.
311 Yes, Paul did run into a number of problems that turned out to
312 be due to the jiffies counter no longer counting. Why do you ask?
313
314o "nfqs" is the number of calls to force_quiescent_state() since
315 boot.
316
317o "nfqsng" is the number of useless calls to force_quiescent_state(),
318 where there wasn't actually a grace period active. This can
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700319 no longer happen due to grace-period processing being pushed
320 into a kthread. The number in parentheses is the difference
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100321 between "nfqs" and "nfqsng", or the number of times that
322 force_quiescent_state() actually did some real work.
323
324o "fqlh" is the number of calls to force_quiescent_state() that
325 exited immediately (without even being counted in nfqs above)
326 due to contention on ->fqslock.
327
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700328o Each element of the form "3/3 ..>. 0:7 ^0" represents one rcu_node
329 structure. Each line represents one level of the hierarchy,
330 from root to leaves. It is best to think of the rcu_data
331 structures as forming yet another level after the leaves.
332 Note that there might be either one, two, three, or even four
333 levels of rcu_node structures, depending on the relationship
334 between CONFIG_RCU_FANOUT, CONFIG_RCU_FANOUT_LEAF (possibly
335 adjusted using the rcu_fanout_leaf kernel boot parameter), and
336 CONFIG_NR_CPUS (possibly adjusted using the nr_cpu_ids count of
337 possible CPUs for the booting hardware).
Paul E. McKenney0edf1a62009-10-14 10:15:59 -0700338
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100339 o The numbers separated by the "/" are the qsmask followed
340 by the qsmaskinit. The qsmask will have one bit
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700341 set for each entity in the next lower level that has
342 not yet checked in for the current grace period ("e"
343 indicating CPUs 5, 6, and 7 in the example above).
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100344 The qsmaskinit will have one bit for each entity that is
345 currently expected to check in during each grace period.
346 The value of qsmaskinit is assigned to that of qsmask
347 at the beginning of each grace period.
348
Paul E. McKenney0edf1a62009-10-14 10:15:59 -0700349 o The characters separated by the ">" indicate the state
Paul E. McKenney12f5f522010-11-29 21:56:39 -0800350 of the blocked-tasks lists. A "G" preceding the ">"
Paul E. McKenney0edf1a62009-10-14 10:15:59 -0700351 indicates that at least one task blocked in an RCU
352 read-side critical section blocks the current grace
Paul E. McKenney12f5f522010-11-29 21:56:39 -0800353 period, while a "E" preceding the ">" indicates that
354 at least one task blocked in an RCU read-side critical
355 section blocks the current expedited grace period.
356 A "T" character following the ">" indicates that at
357 least one task is blocked within an RCU read-side
358 critical section, regardless of whether any current
359 grace period (expedited or normal) is inconvenienced.
360 A "." character appears if the corresponding condition
361 does not hold, so that "..>." indicates that no tasks
362 are blocked. In contrast, "GE>T" indicates maximal
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700363 inconvenience from blocked tasks. CONFIG_TREE_RCU
364 builds of the kernel will always show "..>.".
Paul E. McKenney0edf1a62009-10-14 10:15:59 -0700365
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100366 o The numbers separated by the ":" are the range of CPUs
367 served by this struct rcu_node. This can be helpful
368 in working out how the hierarchy is wired together.
369
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700370 For example, the example rcu_node structure shown above
371 has "0:7", indicating that it covers CPUs 0 through 7.
Paul E. McKenney64db4cf2008-12-18 21:55:32 +0100372
373 o The number after the "^" indicates the bit in the
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700374 next higher level rcu_node structure that this rcu_node
375 structure corresponds to. For example, the "d/d ..>. 4:7
376 ^1" has a "1" in this position, indicating that it
377 corresponds to the "1" bit in the "3" shown in the
378 "3/3 ..>. 0:7 ^0" entry on the next level up.
Paul E. McKenney6fd9b3a2009-04-13 21:31:18 -0700379
380
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700381The output of "cat rcu/rcu_sched/rcu_pending" looks as follows:
Paul E. McKenney6fd9b3a2009-04-13 21:31:18 -0700382
Paul E. McKenney96d3fd02013-10-04 14:33:34 -0700383 0!np=26111 qsp=29 rpq=5386 cbr=1 cng=570 gpc=3674 gps=577 nn=15903 ndw=0
384 1!np=28913 qsp=35 rpq=6097 cbr=1 cng=448 gpc=3700 gps=554 nn=18113 ndw=0
385 2!np=32740 qsp=37 rpq=6202 cbr=0 cng=476 gpc=4627 gps=546 nn=20889 ndw=0
386 3 np=23679 qsp=22 rpq=5044 cbr=1 cng=415 gpc=3403 gps=347 nn=14469 ndw=0
387 4!np=30714 qsp=4 rpq=5574 cbr=0 cng=528 gpc=3931 gps=639 nn=20042 ndw=0
388 5 np=28910 qsp=2 rpq=5246 cbr=0 cng=428 gpc=4105 gps=709 nn=18422 ndw=0
389 6!np=38648 qsp=5 rpq=7076 cbr=0 cng=840 gpc=4072 gps=961 nn=25699 ndw=0
390 7 np=37275 qsp=2 rpq=6873 cbr=0 cng=868 gpc=3416 gps=971 nn=25147 ndw=0
Paul E. McKenney6fd9b3a2009-04-13 21:31:18 -0700391
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700392The fields are as follows:
393
394o The leading number is the CPU number, with "!" indicating
395 an offline CPU.
Paul E. McKenney6fd9b3a2009-04-13 21:31:18 -0700396
397o "np" is the number of times that __rcu_pending() has been invoked
398 for the corresponding flavor of RCU.
399
400o "qsp" is the number of times that the RCU was waiting for a
401 quiescent state from this CPU.
402
Paul E. McKenneyd21670a2010-04-14 17:39:26 -0700403o "rpq" is the number of times that the CPU had passed through
404 a quiescent state, but not yet reported it to RCU.
405
Paul E. McKenney6fd9b3a2009-04-13 21:31:18 -0700406o "cbr" is the number of times that this CPU had RCU callbacks
407 that had passed through a grace period, and were thus ready
408 to be invoked.
409
410o "cng" is the number of times that this CPU needed another
411 grace period while RCU was idle.
412
413o "gpc" is the number of times that an old grace period had
414 completed, but this CPU was not yet aware of it.
415
416o "gps" is the number of times that a new grace period had started,
417 but this CPU was not yet aware of it.
418
Paul E. McKenney96d3fd02013-10-04 14:33:34 -0700419o "ndw" is the number of times that a wakeup of an rcuo
420 callback-offload kthread had to be deferred in order to avoid
421 deadlock.
422
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700423o "nn" is the number of times that this CPU needed nothing.
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700424
425
426The output of "cat rcu/rcuboost" looks as follows:
427
Paul E. McKenney40e80c42012-10-31 13:00:15 -07004280:3 tasks=.... kt=W ntb=0 neb=0 nnb=0 j=c864 bt=c894
429 balk: nt=0 egt=4695 bt=0 nb=0 ny=56 nos=0
4304:7 tasks=.... kt=W ntb=0 neb=0 nnb=0 j=c864 bt=c894
431 balk: nt=0 egt=6541 bt=0 nb=0 ny=126 nos=0
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700432
433This information is output only for rcu_preempt. Each two-line entry
Paul E. McKenneyc9592ec2013-10-09 14:34:00 -0700434corresponds to a leaf rcu_node structure. The fields are as follows:
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700435
436o "n:m" is the CPU-number range for the corresponding two-line
437 entry. In the sample output above, the first entry covers
Paul E. McKenney40e80c42012-10-31 13:00:15 -0700438 CPUs zero through three and the second entry covers CPUs four
439 through seven.
Paul E. McKenney90e6ac32011-04-06 15:20:47 -0700440
441o "tasks=TNEB" gives the state of the various segments of the
442 rnp->blocked_tasks list:
443
444 "T" This indicates that there are some tasks that blocked
445 while running on one of the corresponding CPUs while
446 in an RCU read-side critical section.
447
448 "N" This indicates that some of the blocked tasks are preventing
449 the current normal (non-expedited) grace period from
450 completing.
451
452 "E" This indicates that some of the blocked tasks are preventing
453 the current expedited grace period from completing.
454
455 "B" This indicates that some of the blocked tasks are in
456 need of RCU priority boosting.
457
458 Each character is replaced with "." if the corresponding
459 condition does not hold.
460
461o "kt" is the state of the RCU priority-boosting kernel
462 thread associated with the corresponding rcu_node structure.
463 The state can be one of the following:
464
465 "S" The kernel thread is stopped, in other words, all
466 CPUs corresponding to this rcu_node structure are
467 offline.
468
469 "R" The kernel thread is running.
470
471 "W" The kernel thread is waiting because there is no work
472 for it to do.
473
474 "Y" The kernel thread is yielding to avoid hogging CPU.
475
476 "?" Unknown value, indicates a bug.
477
478o "ntb" is the number of tasks boosted.
479
480o "neb" is the number of tasks boosted in order to complete an
481 expedited grace period.
482
483o "nnb" is the number of tasks boosted in order to complete a
484 normal (non-expedited) grace period. When boosting a task
485 that was blocking both an expedited and a normal grace period,
486 it is counted against the expedited total above.
487
488o "j" is the low-order 16 bits of the jiffies counter in
489 hexadecimal.
490
491o "bt" is the low-order 16 bits of the value that the jiffies
492 counter will have when we next start boosting, assuming that
493 the current grace period does not end beforehand. This is
494 also in hexadecimal.
495
496o "balk: nt" counts the number of times we didn't boost (in
497 other words, we balked) even though it was time to boost because
498 there were no blocked tasks to boost. This situation occurs
499 when there is one blocked task on one rcu_node structure and
500 none on some other rcu_node structure.
501
502o "egt" counts the number of times we balked because although
503 there were blocked tasks, none of them were blocking the
504 current grace period, whether expedited or otherwise.
505
506o "bt" counts the number of times we balked because boosting
507 had already been initiated for the current grace period.
508
509o "nb" counts the number of times we balked because there
510 was at least one task blocking the current non-expedited grace
511 period that never had blocked. If it is already running, it
512 just won't help to boost its priority!
513
514o "ny" counts the number of times we balked because it was
515 not yet time to start boosting.
516
517o "nos" counts the number of times we balked for other
518 reasons, e.g., the grace period ended first.
519
520
Paul E. McKenney7807acd2013-03-27 11:32:12 -0700521CONFIG_TINY_RCU debugfs Files and Formats
Paul E. McKenney8e79e1f2010-11-04 14:31:19 -0700522
523These implementations of RCU provides a single debugfs file under the
524top-level directory RCU, namely rcu/rcudata, which displays fields in
Paul E. McKenney7807acd2013-03-27 11:32:12 -0700525rcu_bh_ctrlblk and rcu_sched_ctrlblk.
Paul E. McKenney8e79e1f2010-11-04 14:31:19 -0700526
527The output of "cat rcu/rcudata" is as follows:
528
Paul E. McKenney8e79e1f2010-11-04 14:31:19 -0700529rcu_sched: qlen: 0
530rcu_bh: qlen: 0
531
Paul E. McKenney7807acd2013-03-27 11:32:12 -0700532This is split into rcu_sched and rcu_bh sections. The field is as
533follows:
Paul E. McKenney8e79e1f2010-11-04 14:31:19 -0700534
535o "qlen" is the number of RCU callbacks currently waiting either
536 for an RCU grace period or waiting to be invoked. This is the
537 only field present for rcu_sched and rcu_bh, due to the
538 short-circuiting of grace period in those two cases.