Root/Documentation/trace/tracepoint-analysis.txt

1        Notes on Analysing Behaviour Using Events and Tracepoints
2
3            Documentation written by Mel Gorman
4        PCL information heavily based on email from Ingo Molnar
5
61. Introduction
7===============
8
9Tracepoints (see Documentation/trace/tracepoints.txt) can be used without
10creating custom kernel modules to register probe functions using the event
11tracing infrastructure.
12
13Simplistically, tracepoints represent important events that can be
14taken in conjunction with other tracepoints to build a "Big Picture" of
15what is going on within the system. There are a large number of methods for
16gathering and interpreting these events. Lacking any current Best Practises,
17this document describes some of the methods that can be used.
18
19This document assumes that debugfs is mounted on /sys/kernel/debug and that
20the appropriate tracing options have been configured into the kernel. It is
21assumed that the PCL tool tools/perf has been installed and is in your path.
22
232. Listing Available Events
24===========================
25
262.1 Standard Utilities
27----------------------
28
29All possible events are visible from /sys/kernel/debug/tracing/events. Simply
30calling
31
32  $ find /sys/kernel/debug/tracing/events -type d
33
34will give a fair indication of the number of events available.
35
362.2 PCL (Performance Counters for Linux)
37-------
38
39Discovery and enumeration of all counters and events, including tracepoints,
40are available with the perf tool. Getting a list of available events is a
41simple case of:
42
43  $ perf list 2>&1 | grep Tracepoint
44  ext4:ext4_free_inode [Tracepoint event]
45  ext4:ext4_request_inode [Tracepoint event]
46  ext4:ext4_allocate_inode [Tracepoint event]
47  ext4:ext4_write_begin [Tracepoint event]
48  ext4:ext4_ordered_write_end [Tracepoint event]
49  [ .... remaining output snipped .... ]
50
51
523. Enabling Events
53==================
54
553.1 System-Wide Event Enabling
56------------------------------
57
58See Documentation/trace/events.txt for a proper description on how events
59can be enabled system-wide. A short example of enabling all events related
60to page allocation would look something like:
61
62  $ for i in `find /sys/kernel/debug/tracing/events -name "enable" | grep mm_`; do echo 1 > $i; done
63
643.2 System-Wide Event Enabling with SystemTap
65---------------------------------------------
66
67In SystemTap, tracepoints are accessible using the kernel.trace() function
68call. The following is an example that reports every 5 seconds what processes
69were allocating the pages.
70
71  global page_allocs
72
73  probe kernel.trace("mm_page_alloc") {
74      page_allocs[execname()]++
75  }
76
77  function print_count() {
78      printf ("%-25s %-s\n", "#Pages Allocated", "Process Name")
79      foreach (proc in page_allocs-)
80          printf("%-25d %s\n", page_allocs[proc], proc)
81      printf ("\n")
82      delete page_allocs
83  }
84
85  probe timer.s(5) {
86          print_count()
87  }
88
893.3 System-Wide Event Enabling with PCL
90---------------------------------------
91
92By specifying the -a switch and analysing sleep, the system-wide events
93for a duration of time can be examined.
94
95 $ perf stat -a \
96    -e kmem:mm_page_alloc -e kmem:mm_page_free_direct \
97    -e kmem:mm_pagevec_free \
98    sleep 10
99 Performance counter stats for 'sleep 10':
100
101           9630 kmem:mm_page_alloc
102           2143 kmem:mm_page_free_direct
103           7424 kmem:mm_pagevec_free
104
105   10.002577764 seconds time elapsed
106
107Similarly, one could execute a shell and exit it as desired to get a report
108at that point.
109
1103.4 Local Event Enabling
111------------------------
112
113Documentation/trace/ftrace.txt describes how to enable events on a per-thread
114basis using set_ftrace_pid.
115
1163.5 Local Event Enablement with PCL
117-----------------------------------
118
119Events can be activated and tracked for the duration of a process on a local
120basis using PCL such as follows.
121
122  $ perf stat -e kmem:mm_page_alloc -e kmem:mm_page_free_direct \
123         -e kmem:mm_pagevec_free ./hackbench 10
124  Time: 0.909
125
126    Performance counter stats for './hackbench 10':
127
128          17803 kmem:mm_page_alloc
129          12398 kmem:mm_page_free_direct
130           4827 kmem:mm_pagevec_free
131
132    0.973913387 seconds time elapsed
133
1344. Event Filtering
135==================
136
137Documentation/trace/ftrace.txt covers in-depth how to filter events in
138ftrace. Obviously using grep and awk of trace_pipe is an option as well
139as any script reading trace_pipe.
140
1415. Analysing Event Variances with PCL
142=====================================
143
144Any workload can exhibit variances between runs and it can be important
145to know what the standard deviation is. By and large, this is left to the
146performance analyst to do it by hand. In the event that the discrete event
147occurrences are useful to the performance analyst, then perf can be used.
148
149  $ perf stat --repeat 5 -e kmem:mm_page_alloc -e kmem:mm_page_free_direct
150            -e kmem:mm_pagevec_free ./hackbench 10
151  Time: 0.890
152  Time: 0.895
153  Time: 0.915
154  Time: 1.001
155  Time: 0.899
156
157   Performance counter stats for './hackbench 10' (5 runs):
158
159          16630 kmem:mm_page_alloc ( +- 3.542% )
160          11486 kmem:mm_page_free_direct ( +- 4.771% )
161           4730 kmem:mm_pagevec_free ( +- 2.325% )
162
163    0.982653002 seconds time elapsed ( +- 1.448% )
164
165In the event that some higher-level event is required that depends on some
166aggregation of discrete events, then a script would need to be developed.
167
168Using --repeat, it is also possible to view how events are fluctuating over
169time on a system-wide basis using -a and sleep.
170
171  $ perf stat -e kmem:mm_page_alloc -e kmem:mm_page_free_direct \
172        -e kmem:mm_pagevec_free \
173        -a --repeat 10 \
174        sleep 1
175  Performance counter stats for 'sleep 1' (10 runs):
176
177           1066 kmem:mm_page_alloc ( +- 26.148% )
178            182 kmem:mm_page_free_direct ( +- 5.464% )
179            890 kmem:mm_pagevec_free ( +- 30.079% )
180
181    1.002251757 seconds time elapsed ( +- 0.005% )
182
1836. Higher-Level Analysis with Helper Scripts
184============================================
185
186When events are enabled the events that are triggering can be read from
187/sys/kernel/debug/tracing/trace_pipe in human-readable format although binary
188options exist as well. By post-processing the output, further information can
189be gathered on-line as appropriate. Examples of post-processing might include
190
191  o Reading information from /proc for the PID that triggered the event
192  o Deriving a higher-level event from a series of lower-level events.
193  o Calculating latencies between two events
194
195Documentation/trace/postprocess/trace-pagealloc-postprocess.pl is an example
196script that can read trace_pipe from STDIN or a copy of a trace. When used
197on-line, it can be interrupted once to generate a report without exiting
198and twice to exit.
199
200Simplistically, the script just reads STDIN and counts up events but it
201also can do more such as
202
203  o Derive high-level events from many low-level events. If a number of pages
204    are freed to the main allocator from the per-CPU lists, it recognises
205    that as one per-CPU drain even though there is no specific tracepoint
206    for that event
207  o It can aggregate based on PID or individual process number
208  o In the event memory is getting externally fragmented, it reports
209    on whether the fragmentation event was severe or moderate.
210  o When receiving an event about a PID, it can record who the parent was so
211    that if large numbers of events are coming from very short-lived
212    processes, the parent process responsible for creating all the helpers
213    can be identified
214
2157. Lower-Level Analysis with PCL
216================================
217
218There may also be a requirement to identify what functions within a program
219were generating events within the kernel. To begin this sort of analysis, the
220data must be recorded. At the time of writing, this required root:
221
222  $ perf record -c 1 \
223    -e kmem:mm_page_alloc -e kmem:mm_page_free_direct \
224    -e kmem:mm_pagevec_free \
225    ./hackbench 10
226  Time: 0.894
227  [ perf record: Captured and wrote 0.733 MB perf.data (~32010 samples) ]
228
229Note the use of '-c 1' to set the event period to sample. The default sample
230period is quite high to minimise overhead but the information collected can be
231very coarse as a result.
232
233This record outputted a file called perf.data which can be analysed using
234perf report.
235
236  $ perf report
237  # Samples: 30922
238  #
239  # Overhead Command Shared Object
240  # ........ ......... ................................
241  #
242      87.27% hackbench [vdso]
243       6.85% hackbench /lib/i686/cmov/libc-2.9.so
244       2.62% hackbench /lib/ld-2.9.so
245       1.52% perf [vdso]
246       1.22% hackbench ./hackbench
247       0.48% hackbench [kernel]
248       0.02% perf /lib/i686/cmov/libc-2.9.so
249       0.01% perf /usr/bin/perf
250       0.01% perf /lib/ld-2.9.so
251       0.00% hackbench /lib/i686/cmov/libpthread-2.9.so
252  #
253  # (For more details, try: perf report --sort comm,dso,symbol)
254  #
255
256According to this, the vast majority of events triggered on events
257within the VDSO. With simple binaries, this will often be the case so let's
258take a slightly different example. In the course of writing this, it was
259noticed that X was generating an insane amount of page allocations so let's look
260at it:
261
262  $ perf record -c 1 -f \
263        -e kmem:mm_page_alloc -e kmem:mm_page_free_direct \
264        -e kmem:mm_pagevec_free \
265        -p `pidof X`
266
267This was interrupted after a few seconds and
268
269  $ perf report
270  # Samples: 27666
271  #
272  # Overhead Command Shared Object
273  # ........ ....... .......................................
274  #
275      51.95% Xorg [vdso]
276      47.95% Xorg /opt/gfx-test/lib/libpixman-1.so.0.13.1
277       0.09% Xorg /lib/i686/cmov/libc-2.9.so
278       0.01% Xorg [kernel]
279  #
280  # (For more details, try: perf report --sort comm,dso,symbol)
281  #
282
283So, almost half of the events are occurring in a library. To get an idea which
284symbol:
285
286  $ perf report --sort comm,dso,symbol
287  # Samples: 27666
288  #
289  # Overhead Command Shared Object Symbol
290  # ........ ....... ....................................... ......
291  #
292      51.95% Xorg [vdso] [.] 0x000000ffffe424
293      47.93% Xorg /opt/gfx-test/lib/libpixman-1.so.0.13.1 [.] pixmanFillsse2
294       0.09% Xorg /lib/i686/cmov/libc-2.9.so [.] _int_malloc
295       0.01% Xorg /opt/gfx-test/lib/libpixman-1.so.0.13.1 [.] pixman_region32_copy_f
296       0.01% Xorg [kernel] [k] read_hpet
297       0.01% Xorg /opt/gfx-test/lib/libpixman-1.so.0.13.1 [.] get_fast_path
298       0.00% Xorg [kernel] [k] ftrace_trace_userstack
299
300To see where within the function pixmanFillsse2 things are going wrong:
301
302  $ perf annotate pixmanFillsse2
303  [ ... ]
304    0.00 : 34eeb: 0f 18 08 prefetcht0 (%eax)
305         : }
306         :
307         : extern __inline void __attribute__((__gnu_inline__, __always_inline__, _
308         : _mm_store_si128 (__m128i *__P, __m128i __B) : {
309         : *__P = __B;
310   12.40 : 34eee: 66 0f 7f 80 40 ff ff movdqa %xmm0,-0xc0(%eax)
311    0.00 : 34ef5: ff
312   12.40 : 34ef6: 66 0f 7f 80 50 ff ff movdqa %xmm0,-0xb0(%eax)
313    0.00 : 34efd: ff
314   12.39 : 34efe: 66 0f 7f 80 60 ff ff movdqa %xmm0,-0xa0(%eax)
315    0.00 : 34f05: ff
316   12.67 : 34f06: 66 0f 7f 80 70 ff ff movdqa %xmm0,-0x90(%eax)
317    0.00 : 34f0d: ff
318   12.58 : 34f0e: 66 0f 7f 40 80 movdqa %xmm0,-0x80(%eax)
319   12.31 : 34f13: 66 0f 7f 40 90 movdqa %xmm0,-0x70(%eax)
320   12.40 : 34f18: 66 0f 7f 40 a0 movdqa %xmm0,-0x60(%eax)
321   12.31 : 34f1d: 66 0f 7f 40 b0 movdqa %xmm0,-0x50(%eax)
322
323At a glance, it looks like the time is being spent copying pixmaps to
324the card. Further investigation would be needed to determine why pixmaps
325are being copied around so much but a starting point would be to take an
326ancient build of libpixmap out of the library path where it was totally
327forgotten about from months ago!
328

Archive Download this file



interactive