| blob | 5ad30a6ba99f8c521663f94ad1d80c1bdea0b9e8 |
1 '\" te
2 .\" Copyright (c) 2004, Sun Microsystems, Inc. All Rights Reserved.
3 .\" The contents of this file are subject to the terms of the Common Development and Distribution License (the "License"). You may not use this file except in compliance with the License.
4 .\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE or http://www.opensolaris.org/os/licensing. See the License for the specific language governing permissions and limitations under the License.
5 .\" When distributing Covered Code, include this CDDL HEADER in each file and include the License file at usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your own identifying information: Portions Copyright [yyyy] [name of copyright owner]
6 .TH TRAPSTAT 1M "April 9, 2016"
7 .SH NAME
8 trapstat \- report trap statistics
9 .SH SYNOPSIS
10 .LP
11 .nf
12 \fB/usr/sbin/trapstat\fR [\fB-t\fR | \fB-T\fR | \fB-e\fR \fIentry\fR]
13 [\fB-C\fR \fIprocessor_set_id\fR | \fB-c\fR \fIcpulist\fR] [\fB-P\fR] [\fB-a\fR]
14 [\fB-r\fR \fIrate\fR] [ [\fIinterval\fR [\fIcount\fR]] | \fIcommand\fR | [\fIargs\fR]]
15 .fi
17 .LP
18 .nf
19 \fB/usr/sbin/trapstat\fR \fB-l\fR
20 .fi
22 .SH DESCRIPTION
23 .LP
24 The \fBtrapstat\fR utility gathers and displays run-time trap statistics on
25 UltraSPARC-based systems. The default output is a table of trap types and
26 \fBCPU\fR \fBID\fRs, with each row of the table denoting a trap type and each
27 column of the table denoting a \fBCPU\fR. If standard output is a terminal, the
28 table contains as many columns of data as can fit within the terminal width; if
29 standard output is not a terminal, the table contains at most six columns of
30 data. By default, data is gathered and displayed for all \fBCPU\fRs; if the
31 data cannot fit in a single table, it is printed across multiple tables. The
32 set of \fBCPU\fRs for which data is gathered and displayed can be optionally
33 specified with the \fB-c\fR or \fB-C\fR option.
34 .sp
35 .LP
36 Unless the \fB-r\fR option or the \fB-a\fR option is specified, the value
37 displayed in each entry of the table corresponds to the number of traps per
38 second. If the \fB-r\fR option is specified, the value corresponds to the
39 number of traps over the interval implied by the specified sampling rate; if
40 the \fB-a\fR option is specified, the value corresponds to the accumulated
41 number of traps since the invocation of \fBtrapstat\fR.
42 .sp
43 .LP
44 By default, \fBtrapstat\fR displays data once per second, and runs
45 indefinitely; both of these behaviors can be optionally controlled with the
46 \fIinterval\fR and \fIcount\fR parameters, respectively. The \fIinterval\fR is
47 specified in seconds; the \fIcount\fR indicates the number of intervals to be
48 executed before exiting. Alternatively, \fBcommand\fR can be specified, in
49 which case \fBtrapstat\fR executes the provided command and continues to run
50 until the command exits. A positive integer is assumed to be an \fIinterval\fR;
51 if the desired \fB\fIcommand\fR\fR cannot be distinguished from an integer, the
52 full path of \fIcommand\fR must be specified.
53 .sp
54 .LP
55 UltraSPARC I (obsolete), II, and III handle translation lookaside buffer (TLB)
56 misses by trapping to the operating system. TLB miss traps can be a significant
57 component of overall system performance for some workloads; the \fB-t\fR option
58 provides in-depth information on these traps. When run with this option,
59 \fBtrapstat\fR displays both the rate of TLB miss traps \fI\fR and the
60 percentage of time spent processing those traps. Additionally, TLB misses that
61 hit in the translation storage buffer (TSB) are differentiated from TLB misses
62 that further miss in the TSB. (The TSB is a software structure used as a
63 translation entry cache to allow the TLB to be quickly filled; it is discussed
64 in detail in the \fIUltraSPARC II User's Manual\fR.) The TLB and TSB miss
65 information is further broken down into user- and kernel-mode misses.
66 .sp
67 .LP
68 Workloads with working sets that exceed the TLB reach may spend a significant
69 amount of time missing in the TLB. To accommodate such workloads, the operating
70 system supports multiple page sizes: larger page sizes increase the effective
71 TLB reach and thereby reduce the number of TLB misses. To provide insight into
72 the relationship between page size and TLB miss rate, \fBtrapstat\fR optionally
73 provides in-depth TLB miss information broken down by page size using the
74 \fB-T\fR option. The information provided by the \fB-T\fR option is a superset
75 of that provided by the \fB-t\fR option; only one of \fB-t\fR and \fB-T\fR can
76 be specified.
77 .SH OPTIONS
78 .LP
79 The following options are supported:
80 .sp
81 .ne 2
82 .na
83 \fB\fB-a\fR\fR
84 .ad
85 .RS 24n
86 Displays the number of traps as accumulating, monotonically increasing values
87 instead of per-second or per-interval rates.
88 .RE
90 .sp
91 .ne 2
92 .na
93 \fB\fB-c\fR \fIcpulist\fR \fR
94 .ad
95 .RS 24n
96 Enables \fBtrapstat\fR only on the \fBCPU\fRs specified by \fIcpulist\fR.
97 .sp
98 \fIcpulist\fR can be a single processor \fBID\fR (for example, \fB4\fR), a
99 range of processor\fB ID\fRs (for example, \fB4-6\fR), or a comma separated
100 list of processor\fB ID\fRs or processor \fBID\fR ranges (for example,
101 \fB4,5,6\fR or \fB4,6-8\fR).
102 .RE
104 .sp
105 .ne 2
106 .na
107 \fB\fB-C\fR \fIprocessor_set_id\fR \fR
108 .ad
109 .RS 24n
110 Enables \fBtrapstat\fR only on the \fBCPU\fRs in the processor set specified by
111 \fIprocessor_set_id\fR.
112 .sp
113 \fBtrapstat\fR modifies its output to always reflect the \fBCPU\fRs in the
114 specified processor set. If a \fBCPU\fR is added to the set, \fBtrapstat\fR
115 modifies its output to include the added \fBCPU\fR; if a \fBCPU\fR is removed
116 from the set, \fBtrapstat\fR modifies its output to exclude the removed CPU. At
117 most one processor set can be specified.
118 .RE
120 .sp
121 .ne 2
122 .na
123 \fB\fB-e\fR \fIentrylist\fR \fR
124 .ad
125 .RS 24n
126 Enables \fBtrapstat\fR only for the trap table entry or entries specified by
127 \fIentrylist\fR. A trap table entry can be specified by trap number or by trap
128 name (for example, the level-10 trap can be specified as \fB74\fR, \fB0x4A\fR,
129 \fB0x4a\fR, or \fBlevel-10\fR).
130 .sp
131 \fIentrylist\fR can be a single trap table entry or a comma separated list of
132 trap table entries. If the specified trap table entry is not valid,
133 \fBtrapstat\fR prints a table of all valid trap table entries and values. A
134 list of valid trap table entries is also found in \fIThe SPARC Architecture
135 Manual, Version 9\fR and the \fISun Microelectronics UltraSPARC II User's
136 Manual.\fR If the parsable option (\fB-P\fR) is specified in addition to the
137 \fB-e\fR option, the format of the data is as follows:
138 .sp
140 .sp
141 .TS
142 c c
143 l l .
144 Field Contents
145 1 Timestamp (nanoseconds since start)
146 2 CPU ID
147 3 Trap number (in hexadecimal)
148 4 Trap name
149 5 Trap rate per interval
150 .TE
152 Each field is separated with whitespace. If the format is modified, it will be
153 modified by adding potentially new fields beginning with field 6; exant fields
154 will remain unchanged.
155 .RE
157 .sp
158 .ne 2
159 .na
160 \fB\fB-l\fR \fR
161 .ad
162 .RS 24n
163 Lists trap table entries. By default, a table is displayed containing all valid
164 trap numbers, their names and a brief description. The trap name is used in
165 both the default output and in the \fIentrylist\fR parameter for the \fB-e\fR
166 argument. If the parsable option (\fB-P\fR) is specified in addition to the
167 \fB-l\fR option, the format of the data is as follows:
168 .sp
170 .sp
171 .TS
172 c c
173 l l .
174 Field Contents
175 1 Trap number in hexadecimal
176 2 Trap number in decimal
177 3 Trap name
178 Remaining Trap description
179 .TE
181 .RE
183 .sp
184 .ne 2
185 .na
186 \fB\fB-P\fR \fR
187 .ad
188 .RS 24n
189 Generates parsable output. When run without other data gathering modifying
190 options (that is, \fB-e\fR, \fB-t\fR or \fB-T\fR), \fBtrapstat\fR's the
191 parsable output has the following format:
192 .sp
194 .sp
195 .TS
196 c c
197 l l .
198 Field Contents
199 1 Timestamp (nanoseconds since start)
200 2 CPU ID
201 3 Trap number (in hexadecimal)
202 4 Trap name
203 5 Trap rate per interval
204 .TE
206 Each field is separated with whitespace. If the format is modified, it will be
207 modified by adding potentially new fields beginning with field 6; extant fields
208 will remain unchanged.
209 .RE
211 .sp
212 .ne 2
213 .na
214 \fB\fB-r\fR \fIrate\fR \fR
215 .ad
216 .RS 24n
217 Explicitly sets the sampling rate to be \fIrate\fR samples per second. If this
218 option is specified, \fBtrapstat\fR's output changes from a traps-per-second to
219 traps-per-sampling-interval.
220 .RE
222 .sp
223 .ne 2
224 .na
225 \fB\fB-t\fR\fR
226 .ad
227 .RS 24n
228 Enables TLB statistics.
229 .sp
230 A table is displayed with four principal columns of data: \fIitlb-miss\fR,
231 \fIitsb-miss\fR, \fIdtlb-miss\fR, and \fIdtsb-miss\fR. The columns contain both
232 the rate of the corresponding event and the percentage of \fBCPU\fR time spent
233 processing the event. The percentage of \fBCPU\fR time is given only in terms
234 of a single \fBCPU\fR. The rows of the table correspond to \fBCPU\fRs, with
235 each \fBCPU\fR consuming two rows: one row for user-mode events (denoted with
236 \fBu\fR) and one row for kernel-mode events (denoted with \fBk\fR). For each
237 row, the percentage of \fBCPU\fR time is totalled and displayed in the
238 rightmost column. The \fBCPU\fRs are delineated with a solid line. If the
239 parsable option (\fB-P\fR) is specified in addition to the \fB-t\fR option, the
240 format of the data is as follows:
241 .sp
243 .sp
244 .TS
245 c c
246 l l .
247 Field Contents
248 1 Timestamp (nanoseconds since start)
249 2 CPU ID
250 3 Mode (\fBk\fR denotes kernel, \fBu\fR denotes user)
251 4 I-TLB misses
252 5 Percentage of time in I-TLB miss handler
253 6 I-TSB misses
254 7 Percentage of time in I-TSB miss handler
255 8 D-TLB misses
256 9 Percentage of time in D-TLB miss handler
257 10 D-TSB misses
258 11 Percentage of time in D-TSB miss handler
259 .TE
261 Each field is separated with whitespace. If the format is modified, it will be
262 modified by adding potentially new fields beginning with field 12; extant
263 fields will remain unchanged.
264 .RE
266 .sp
267 .ne 2
268 .na
269 \fB\fB-T\fR \fR
270 .ad
271 .RS 24n
272 Enables \fBTLB\fR statistics, with page size information. As with the \fB-t\fR
273 option, a table is displayed with four principal columns of data:
274 \fIitlb-miss\fR, \fIitsb-miss\fR, \fIdtlb-miss\fR, and \fIdtsb-miss\fR. The
275 columns contain both the absolute number of the corresponding event, and the
276 percentage of \fBCPU\fR time spent processing the event. The percentage of
277 \fBCPU\fR time is given only in terms of a single \fBCPU\fR. The rows of the
278 table correspond to \fBCPU\fRs, with each \fBCPU\fR consuming two sets of rows:
279 one set for user-level events (denoted with \fBu\fR) and one set for
280 kernel-level events (denoted with \fBk\fR). Each set, in turn, contains as many
281 rows as there are page sizes supported (see \fBgetpagesizes\fR(3C)). For each
282 row, the percentage of \fBCPU\fR time is totalled and displayed in the
283 right-most column. The two sets are delineated with a dashed line; CPUs are
284 delineated with a solid line. If the parsable option (\fB-P\fR) is specified in
285 addition to the \fB-T\fR option, the format of the data is as follows:
286 .sp
288 .sp
289 .TS
290 c c
291 l l .
292 Field Contents
293 1 Timestamp (nanoseconds since start)
294 2 CPU ID
295 3 Mode \fBk\fR denotes kernel, \fBu\fR denotes user)
296 4 Page size, in decimal
297 5 I-TLB misses
298 6 Percentage of time in I-TLB miss handler
299 7 I-TSB misses
300 8 Percentage of time in I-TSB miss handler
301 9 D-TLB misses
302 10 Percentage of time in D-TLB miss handler
303 11 D-TSB misses
304 12 Percentage of time in D-TSB miss handler
305 .TE
307 Each field is separated with whitespace. If the format is modified, it will be
308 modified by adding potentially new fields beginning with field 13; extant
309 fields will remain unchanged.
310 .RE
312 .SH EXAMPLES
313 .LP
314 \fBExample 1 \fRUsing \fBtrapstat\fR Without Options
315 .sp
316 .LP
317 When run without options, \fBtrapstat\fR displays a table of trap types and
318 CPUs. At most six columns can fit in the default terminal width; if (as in this
319 example) there are more than six CPUs, multiple tables are displayed:
321 .sp
322 .in +2
323 .nf
324 example# \fBtrapstat\fR
325 vct name | cpu0 cpu1 cpu4 cpu5 cpu8 cpu9
326 ------------------------+------------------------------------------------------
327 24 cleanwin | 6446 4837 6368 2153 2623 1321
328 41 level-1 | 100 0 0 0 1 0
329 44 level-4 | 0 1 1 1 0 0
330 45 level-5 | 0 0 0 0 0 0
331 47 level-7 | 0 0 0 0 9 0
332 49 level-9 | 100 100 100 100 100 100
333 4a level-10 | 100 0 0 0 0 0
334 4d level-13 | 6 10 7 16 13 11
335 4e level-14 | 100 0 0 0 1 0
336 60 int-vec | 2607 2740 2642 2922 2920 3033
337 64 itlb-miss | 3129 2475 3167 1037 1200 569
338 68 dtlb-miss | 121061 86162 109838 37386 45639 20269
339 6c dtlb-prot | 997 847 1061 379 406 184
340 84 spill-user-32 | 2809 2133 2739 200806 332776 454504
341 88 spill-user-64 | 45819 207856 93487 228529 68373 77590
342 8c spill-user-32-cln | 784 561 767 274 353 215
343 90 spill-user-64-cln | 9 37 17 39 12 13
344 98 spill-kern-64 | 62913 50145 63869 21916 28431 11738
345 a4 spill-asuser-32 | 1327 947 1288 460 572 335
346 a8 spill-asuser-64 | 26 48 18 54 10 14
347 ac spill-asuser-32-cln | 4580 3599 4555 1538 1978 857
348 b0 spill-asuser-64-cln | 26 0 0 2 0 0
349 c4 fill-user-32 | 2862 2161 2798 191746 318115 435850
350 c8 fill-user-64 | 45813 197781 89179 217668 63905 74281
351 cc fill-user-32-cln | 3802 2833 3733 10153 16419 19475
352 d0 fill-user-64-cln | 329 10105 4873 10603 4235 3649
353 d8 fill-kern-64 | 62519 49943 63611 21824 28328 11693
354 108 syscall-32 | 2285 1634 2278 737 957 383
355 126 self-xcall | 100 0 0 0 0 0
357 vct name | cpu12 cpu13 cpu14 cpu15
358 ------------------------+------------------------------------
359 24 cleanwin | 5435 4232 6302 6104
360 41 level-1 | 0 0 0 0
361 44 level-4 | 2 0 0 1
362 45 level-5 | 0 0 0 0
363 47 level-7 | 0 0 0 0
364 49 level-9 | 100 100 100 100
365 4a level-10 | 0 0 0 0
366 4d level-13 | 15 11 22 11
367 4e level-14 | 0 0 0 0
368 60 int-vec | 2813 2833 2738 2714
369 64 itlb-miss | 2636 1925 3133 3029
370 68 dtlb-miss | 90528 70639 107786 103425
371 6c dtlb-prot | 819 675 988 954
372 84 spill-user-32 | 175768 39933 2811 2742
373 88 spill-user-64 | 0 241348 96907 118298
374 8c spill-user-32-cln | 681 513 753 730
375 90 spill-user-64-cln | 0 42 16 20
376 98 spill-kern-64 | 52158 40914 62305 60141
377 a4 spill-asuser-32 | 1113 856 1251 1208
378 a8 spill-asuser-64 | 0 64 16 24
379 ac spill-asuser-32-cln | 3816 2942 4515 4381
380 b0 spill-asuser-64-cln | 0 0 0 0
381 c4 fill-user-32 | 170744 38444 2876 2784
382 c8 fill-user-64 | 0 230381 92941 111694
383 cc fill-user-32-cln | 8550 3790 3612 3553
384 d0 fill-user-64-cln | 0 10726 4495 5845
385 d8 fill-kern-64 | 51968 40760 62053 59922
386 108 syscall-32 | 1839 1495 2144 2083
387 126 self-xcall | 0 0 0 0
388 .fi
389 .in -2
390 .sp
392 .LP
393 \fBExample 2 \fRUsing \fBtrapset\fR with CPU Filtering
394 .sp
395 .LP
396 The \fB-c\fR option can be used to limit the \fBCPU\fRs on which
397 \fBtrapstat\fR is enabled. This example limits \fBCPU 1\fR and \fBCPU\fRs
398 \fB12\fR through \fB15\fR.
400 .sp
401 .in +2
402 .nf
403 example# \fBtrapstat -c 1,12-15\fR
406 vct name | cpu1 cpu12 cpu13 cpu14 cpu15
407 ------------------------+---------------------------------------------
408 24 cleanwin | 6923 3072 2500 3518 2261
409 44 level-4 | 3 0 0 1 1
410 49 level-9 | 100 100 100 100 100
411 4d level-13 | 23 8 14 19 14
412 60 int-vec | 2559 2699 2752 2688 2792
413 64 itlb-miss | 3296 1548 1174 1698 1087
414 68 dtlb-miss | 114788 54313 43040 58336 38057
415 6c dtlb-prot | 1046 549 417 545 370
416 84 spill-user-32 | 66551 29480 301588 26522 213032
417 88 spill-user-64 | 0 318652 111239 299829 221716
418 8c spill-user-32-cln | 856 347 331 416 293
419 90 spill-user-64-cln | 0 55 21 59 39
420 98 spill-kern-64 | 66464 31803 24758 34004 22277
421 a4 spill-asuser-32 | 1423 569 560 698 483
422 a8 spill-asuser-64 | 0 74 32 98 46
423 ac spill-asuser-32-cln | 4875 2250 1728 2384 1584
424 b0 spill-asuser-64-cln | 0 2 0 1 0
425 c4 fill-user-32 | 64193 28418 287516 27055 202093
426 c8 fill-user-64 | 0 305016 106692 288542 210654
427 cc fill-user-32-cln | 6733 3520 15185 2396 12035
428 d0 fill-user-64-cln | 0 13226 3506 12933 11032
429 d8 fill-kern-64 | 66220 31680 24674 33892 22196
430 108 syscall-32 | 2446 967 817 1196 755
431 .fi
432 .in -2
433 .sp
435 .LP
436 \fBExample 3 \fRUsing \fBtrapstat\fR with TLB Statistics
437 .sp
438 .LP
439 The \fB-t\fR option displays in-depth \fBTLB\fR statistics, including the
440 amount of time spent performing \fBTLB\fR miss processing. The following
441 example shows that the machine is spending 14.1 percent of its time just
442 handling D-TLB misses:
444 .sp
445 .in +2
446 .nf
447 example# \fBtrapstat -t\fR
448 cpu m| itlb-miss %tim itsb-miss %tim | dtlb-miss %tim dtsb-miss %tim |%tim
449 -----+-------------------------------+-------------------------------+----
450 0 u| 2571 0.3 0 0.0 | 10802 1.3 0 0.0 | 1.6
451 0 k| 0 0.0 0 0.0 | 106420 13.4 184 0.1 |13.6
452 -----+-------------------------------+-------------------------------+----
453 1 u| 3069 0.3 0 0.0 | 10983 1.2 100 0.0 | 1.6
454 1 k| 27 0.0 0 0.0 | 106974 12.6 19 0.0 |12.7
455 -----+-------------------------------+-------------------------------+----
456 2 u| 3033 0.3 0 0.0 | 11045 1.2 105 0.0 | 1.6
457 2 k| 43 0.0 0 0.0 | 107842 12.7 108 0.0 |12.8
458 -----+-------------------------------+-------------------------------+----
459 3 u| 2924 0.3 0 0.0 | 10380 1.2 121 0.0 | 1.6
460 3 k| 54 0.0 0 0.0 | 102682 12.2 16 0.0 |12.2
461 -----+-------------------------------+-------------------------------+----
462 4 u| 3064 0.3 0 0.0 | 10832 1.2 120 0.0 | 1.6
463 4 k| 31 0.0 0 0.0 | 107977 13.0 236 0.1 |13.1
464 =====+===============================+===============================+====
465 ttl | 14816 0.3 0 0.0 | 585937 14.1 1009 0.0 |14.5
466 .fi
467 .in -2
468 .sp
470 .LP
471 \fBExample 4 \fRUsing \fBtrapstat\fR with TLB Statistics and Page Size
472 Information
473 .sp
474 .LP
475 By specifying the \fB-T\fR option, \fBtrapstat\fR shows \fBTLB\fR misses broken
476 down by page size. In this example, CPU 0 is spending 7.9 percent of its time
477 handling user-mode TLB misses on 8K pages, and another 2.3 percent of its time
478 handling user-mode TLB misses on 64K pages.
480 .sp
481 .in +2
482 .nf
483 example# \fBtrapstat -T -c 0\fR
484 cpu m size| itlb-miss %tim itsb-miss %tim | dtlb-miss %tim dtsb-miss %tim |%tim
485 ----------+-------------------------------+-------------------------------+----
486 0 u 8k| 1300 0.1 15 0.0 | 104897 7.9 90 0.0 | 8.0
487 0 u 64k| 0 0.0 0 0.0 | 29935 2.3 7 0.0 | 2.3
488 0 u 512k| 0 0.0 0 0.0 | 3569 0.2 2 0.0 | 0.2
489 0 u 4m| 0 0.0 0 0.0 | 233 0.0 2 0.0 | 0.0
490 - - - - - + - - - - - - - - - - - - - - - + - - - - - - - - - - - - - - - + - -
491 0 k 8k| 13 0.0 0 0.0 | 71733 6.5 110 0.0 | 6.5
492 0 k 64k| 0 0.0 0 0.0 | 0 0.0 0 0.0 | 0.0
493 0 k 512k| 0 0.0 0 0.0 | 0 0.0 206 0.1 | 0.1
494 0 k 4m| 0 0.0 0 0.0 | 0 0.0 0 0.0 | 0.0
495 ==========+===============================+===============================+====
496 ttl | 1313 0.1 15 0.0 | 210367 17.1 417 0.2 |17.5
497 .fi
498 .in -2
499 .sp
501 .LP
502 \fBExample 5 \fRUsing \fBtrapstat\fR with Entry Filtering
503 .sp
504 .LP
505 By specifying the \fB-e\fR option, \fBtrapstat\fR displays statistics for only
506 specific trap types. Using this option minimizes the probe effect when seeking
507 specific data. This example yields statistics for only the \fIdtlb-prot\fR and
508 \fIsyscall-32\fR traps on CPUs 12 through 15:
510 .sp
511 .in +2
512 .nf
513 example# \fBtrapstat -e dtlb-prot,syscall-32 -c 12-15\fR
514 vct name | cpu12 cpu13 cpu14 cpu15
515 ------------------------+------------------------------------
516 6c dtlb-prot | 817 754 1018 560
517 108 syscall-32 | 1426 1647 2186 1142
519 vct name | cpu12 cpu13 cpu14 cpu15
520 ------------------------+------------------------------------
521 6c dtlb-prot | 1085 996 800 707
522 108 syscall-32 | 2578 2167 1638 1452
523 .fi
524 .in -2
525 .sp
527 .LP
528 \fBExample 6 \fRUsing \fBtrapstat\fR with a Higher Sampling Rate
529 .sp
530 .LP
531 The following example uses the \fB-r\fR option to specify a sampling rate of
532 1000 samples per second, and filter only for the level-10 trap. Additionally,
533 specifying the \fB-P\fR option yields parsable output.
535 .sp
536 .LP
537 Notice the timestamp difference between the level-10 events: 9,998,000
538 nanoseconds and 10,007,000 nanoseconds. These level-10 events correspond to the
539 system clock, which by default ticks at 100 hertz (that is, every 10,000,000
540 nanoseconds).
542 .sp
543 .in +2
544 .nf
545 example# \fBtrapstat -e level-10 -P -r 1000\fR
546 1070400 0 4a level-10 0
547 2048600 0 4a level-10 0
548 3030400 0 4a level-10 1
549 4035800 0 4a level-10 0
550 5027200 0 4a level-10 0
551 6027200 0 4a level-10 0
552 7027400 0 4a level-10 0
553 8028200 0 4a level-10 0
554 9026400 0 4a level-10 0
555 10029600 0 4a level-10 0
556 11028600 0 4a level-10 0
557 12024000 0 4a level-10 0
558 13028400 0 4a level-10 1
559 14031200 0 4a level-10 0
560 15027200 0 4a level-10 0
561 16027600 0 4a level-10 0
562 17025000 0 4a level-10 0
563 18026000 0 4a level-10 0
564 19027800 0 4a level-10 0
565 20025600 0 4a level-10 0
566 21025200 0 4a level-10 0
567 22025000 0 4a level-10 0
568 23035400 0 4a level-10 1
569 24027400 0 4a level-10 0
570 25026000 0 4a level-10 0
571 26027000 0 4a level-10 0
572 .fi
573 .in -2
574 .sp
576 .SH ATTRIBUTES
577 .LP
578 See \fBattributes\fR(5) for descriptions of the following attributes:
579 .sp
581 .sp
582 .TS
583 box;
584 c c
585 l l .
586 ATTRIBUTE TYPE ATTRIBUTE VALUE
587 Interface Stability
588 Human Readable Output Unstable
589 Parsable Output Evolving
590 .TE
592 .SH SEE ALSO
593 .LP
594 \fBlockstat\fR(1M), \fBpmap\fR(1), \fBpsrset\fR(1M), \fBpsrinfo\fR(1M),
595 \fBpbind\fR(1M), \fBppgsz\fR(1), \fBgetpagesizes\fR(3C)
596 .sp
597 .LP
598 \fISun Microelectronics UltraSPARC II User's Manual,\fR January 1997, STP1031,
599 .sp
600 .LP
601 \fIThe SPARC Architecture Manual, Version 9,\fR 1994, Prentice-Hall.
602 .SH NOTES
603 .LP
604 When enabled, \fBtrapstat\fR induces a varying probe effect, depending on the
605 type of information collected. While the precise probe effect depends upon the
606 specifics of the hardware, the following table can be used as a rough guide:
607 .sp
609 .sp
610 .TS
611 c c
612 l l .
613 Option Approximate probe effect
614 default 3-5% per trap
615 \fB-e\fR 3-5% per specified trap
616 \fB-t\fR, \fB-T\fR T{
617 40-45% per TLB miss trap hitting in the TSB, 25-30% per TLB miss trap missing in the TSB
618 T}
619 .TE
621 .sp
622 .LP
623 These probe effects are \fIper trap\fR not for the system as a whole. For
624 example, running \fBtrapstat\fR with the default options on a system that
625 spends 7% of total time handling traps induces a performance degradation of
626 less than one half of one percent; running \fBtrapstat\fR with the \fB-t\fR or
627 \fB-T\fR option on a system spending 5% of total time processing TLB misses
628 induce a performance degradation of no more than 2.5%.
629 .sp
630 .LP
631 When run with the \fB-t\fR or \fB-T\fR option, \fBtrapstat\fR accounts for its
632 probe effect when calculating the \fI%tim\fR fields. This assures that the
633 \fI%tim\fR fields are a reasonably accurate indicator of the time a given
634 workload is spending handling TLB misses \(em regardless of the perturbing
635 presence of \fBtrapstat\fR.
636 .sp
637 .LP
638 While the \fI%tim\fR fields include the explicit cost of executing the TLB miss
639 handler, they do \fInot\fR include the implicit costs of TLB miss traps (for
640 example, pipeline effects, cache pollution, etc). These implicit costs become
641 more significant as the trap rate grows; if high \fI%tim\fR values are reported
642 (greater than 50%), you can accurately infer that much of the balance of time
643 is being spent on the implicit costs of the TLB miss traps.
644 .sp
645 .LP
646 Due to the potential system wide degradation induced, only the super-user can
647 run \fBtrapstat\fR.
648 .sp
649 .LP
650 Due to the limitation of the underlying statistics gathering methodology, only
651 one instance of \fBtrapstat\fR can run at a time.