diff options
author | Maxin B. John <maxin.john@enea.com> | 2015-03-20 18:07:56 +0100 |
---|---|---|
committer | Richard Purdie <richard.purdie@linuxfoundation.org> | 2015-04-08 10:53:11 +0100 |
commit | 12f3850485094b97f4405c5160b44ad399149403 (patch) | |
tree | 981e5b5a681e0cb786e983b136ef66e636452e35 /meta/recipes-rt/rt-tests | |
parent | b559d7f81f0793e70bdfcf6c1258530ba07c0217 (diff) | |
download | poky-12f3850485094b97f4405c5160b44ad399149403.tar.gz |
rt-tests: bump version and enable ptest support
1. Upgrade version to 0.91
2. Add ptest support
[YOCTO #1569]
(From OE-Core rev: d3c3b49384d034e84cf197c3bb165709a2088cbf)
Signed-off-by: Thomas Lundström <thomas.lundstrom@enea.com>
Signed-off-by: Maxin B. John <maxin.john@enea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Diffstat (limited to 'meta/recipes-rt/rt-tests')
-rwxr-xr-x | meta/recipes-rt/rt-tests/files/rt_bmark.py | 412 | ||||
-rw-r--r-- | meta/recipes-rt/rt-tests/files/run-ptest | 3 | ||||
-rw-r--r-- | meta/recipes-rt/rt-tests/rt-tests.inc | 6 | ||||
-rw-r--r-- | meta/recipes-rt/rt-tests/rt-tests_0.91.bb (renamed from meta/recipes-rt/rt-tests/rt-tests_0.89.bb) | 10 |
4 files changed, 428 insertions, 3 deletions
diff --git a/meta/recipes-rt/rt-tests/files/rt_bmark.py b/meta/recipes-rt/rt-tests/files/rt_bmark.py new file mode 100755 index 0000000000..e2280e43e2 --- /dev/null +++ b/meta/recipes-rt/rt-tests/files/rt_bmark.py | |||
@@ -0,0 +1,412 @@ | |||
1 | #!/usr/bin/python | ||
2 | # -*- coding: utf-8 -*- | ||
3 | |||
4 | # SPDX-License-Identifier: BSD-3-Clause | ||
5 | # Copyright (C) 2015 Enea Software AB | ||
6 | # Author: Thomas Lundström <thomas.lundstrom@enea.com> | ||
7 | |||
8 | # The script measures interrupt latency together with different types of | ||
9 | # system load. This is done using the programs cyclictest and stress. | ||
10 | # | ||
11 | # The main output is: | ||
12 | # | ||
13 | # Best case (min) latency | ||
14 | # This has very limited value, but is presented since it can be done | ||
15 | # easily | ||
16 | # | ||
17 | # Average latency | ||
18 | # This value is of interrest for throughput oriented systems. Limited | ||
19 | # value for a real-time system. Also presented because it is easy to do. | ||
20 | # | ||
21 | # Worst case (max) latency | ||
22 | # This is the interesting number for a real-time system. The number | ||
23 | # presented is (of cource) the largest number observed. The challenge | ||
24 | # is to know how the observed worst case relates to the actual worst case. | ||
25 | # | ||
26 | # To get an indication of the confidence, the following method is used: | ||
27 | # 1) Instead of one long run, the measurement is made as a set of shorter | ||
28 | # runs. The number of runs must be a power of 2 for reasons that will | ||
29 | # shorlty be obvious | ||
30 | # | ||
31 | # 2) First, a list of the max values are created. | ||
32 | # | ||
33 | # 3) The smallest value in that list is recorded. | ||
34 | # | ||
35 | # 4) Then a new list is create by taking the max value of each pair of | ||
36 | # values in the original list. In this list the smallest value is | ||
37 | # recorded. | ||
38 | # | ||
39 | # 5) Step 3 is repeated until there is only one value in the list. See | ||
40 | # example below: | ||
41 | # | ||
42 | # Samples: | ||
43 | # | 44 | | | | | | ||
44 | # | 77 | 77 | | | | | ||
45 | # | 118 | | | | | | ||
46 | # | 119 | 119 | 119 | | | | ||
47 | # | 138 | | | | | | ||
48 | # | 57 | 138 | | | | | ||
49 | # | 175 | | | | | | ||
50 | # | 130 | 175 | 175 | 175 | | | ||
51 | # | 54 | | | | | | ||
52 | # | 150 | 150 | | | | | ||
53 | # | 47 | | | | | | ||
54 | # | 59 | 59 | 150 | | | | ||
55 | # | 199 | | | | | | ||
56 | # | 115 | 199 | | | | | ||
57 | # | 177 | | | | | | ||
58 | # | 129 | 177 | 199 | 199 | 199 | | ||
59 | # | ||
60 | # Smallest value: | ||
61 | # | 44 | 59 | 119 | 175 | 199 | | ||
62 | # | ||
63 | # 6) The generated list of smallest values is analyzed. In this case, it | ||
64 | # can be observed that the values are increasing significantly through | ||
65 | # the entire list, which leads to the conclusion that the number of | ||
66 | # samples is too small. | ||
67 | # If instead the list had been (167, 191, 196, 199, 199), there had | ||
68 | # been a very small, or no, increase at the end of the list. We might | ||
69 | # then suspect that the number of samples is probably large enough. | ||
70 | # There is however no guarantee for that. | ||
71 | # | ||
72 | # Steps 1-2 are done in run_cyclictest_suite | ||
73 | # Steps 3-5 are done in gen_minmax_list. | ||
74 | # Step 6 needs to be done manually since there is (yet) no well defined | ||
75 | # FAIL criterion and a theoretically solid PASS criterion may never be | ||
76 | # available. | ||
77 | |||
78 | import multiprocessing | ||
79 | import os | ||
80 | import re | ||
81 | import signal | ||
82 | import subprocess | ||
83 | import time | ||
84 | import traceback | ||
85 | |||
86 | # See comment on the function set_hung_tmo | ||
87 | has_hung_task_detection = True | ||
88 | |||
89 | #------------------------------------------------------------------------------- | ||
90 | |||
91 | class TestFail(Exception): | ||
92 | def __init__(self, msg): | ||
93 | self.msg = msg | ||
94 | |||
95 | def __str__(self): | ||
96 | return "Test failure: (" + self.msg + ")" | ||
97 | |||
98 | #------------------------------------------------------------------------------- | ||
99 | |||
100 | def tc_name(sub_name): | ||
101 | return "rt_bmark.intlat." + sub_name | ||
102 | |||
103 | #------------------------------------------------------------------------------- | ||
104 | # log() does the same job as print except that a '#' is added at the beginning | ||
105 | # of each line. This causes TEFEL to ignore it | ||
106 | |||
107 | def log(*msg): | ||
108 | tmp = "".join(map(str, msg)) # 'map(str, ...' allows numbers | ||
109 | for line in tmp.split("\n"): | ||
110 | print "#", line | ||
111 | |||
112 | #------------------------------------------------------------------------------- | ||
113 | # Like log(), but with a timestamp added | ||
114 | |||
115 | def log_ts(*msg): | ||
116 | ts = time.localtime() | ||
117 | stamp = "%2d:%02d:%02d: " % (ts.tm_hour, ts.tm_min, ts.tm_sec) | ||
118 | log(stamp, *msg) | ||
119 | |||
120 | #------------------------------------------------------------------------------- | ||
121 | |||
122 | def log_test_header(seq_no, nr_of_tests, name): | ||
123 | log("=" * 78) | ||
124 | log() | ||
125 | log(" Test case (%d/%d): %s" % (seq_no, nr_of_tests, tc_name(name))) | ||
126 | log() | ||
127 | log("." * 78) | ||
128 | log() | ||
129 | |||
130 | #------------------------------------------------------------------------------- | ||
131 | |||
132 | def start_stress(*args): | ||
133 | stress_cmd = [ "stress" ] | ||
134 | added_stress_types = [] | ||
135 | req_stress_types = set(args) | ||
136 | cpu_cnt = str(multiprocessing.cpu_count()) | ||
137 | |||
138 | # The function cond_add_stress appends the options to the stress | ||
139 | # command if the stress type is in the set of requested stress types | ||
140 | |||
141 | def cond_add_stress(stress_type, options): | ||
142 | if stress_type in req_stress_types: | ||
143 | req_stress_types.remove(stress_type) | ||
144 | added_stress_types.append(stress_type) | ||
145 | stress_cmd.extend(options) | ||
146 | |||
147 | #---------- | ||
148 | |||
149 | cond_add_stress("io", ["-i", cpu_cnt]) | ||
150 | cond_add_stress("cpu", ["-c", cpu_cnt]) | ||
151 | cond_add_stress("hdd", ["-d", cpu_cnt, "--hdd-bytes", "20M"]) | ||
152 | cond_add_stress("vm", ["-m", cpu_cnt, "--vm-bytes", "10M"]) | ||
153 | |||
154 | unknown = ", ".join(req_stress_types) | ||
155 | if unknown != "": | ||
156 | raise TestFail("Unknown stress type(s): %s" % unknown) | ||
157 | |||
158 | if not added_stress_types: | ||
159 | log("No stress requested") | ||
160 | return None | ||
161 | |||
162 | added = "+".join(added_stress_types) | ||
163 | stress_cmd_str = " ".join(stress_cmd) | ||
164 | |||
165 | log("Starting stress(", added, ")") | ||
166 | log(" Command: '", stress_cmd_str, "'") | ||
167 | log() | ||
168 | |||
169 | # preexec_fn=os.setsid causes stress to be executed in a separate | ||
170 | # session, => it gets a new process group (incl. children). It | ||
171 | # can then be terminated using os.killpg in end_stress without | ||
172 | # terminating this script. | ||
173 | |||
174 | p = subprocess.Popen(stress_cmd, preexec_fn=os.setsid) | ||
175 | |||
176 | return p | ||
177 | |||
178 | #------------------------------------------------------------------------------- | ||
179 | |||
180 | def end_stress(p): | ||
181 | if p is None: | ||
182 | # The value None indicates that no stress scenario was started | ||
183 | return | ||
184 | |||
185 | if p.poll() is not None: | ||
186 | raise TestFail("stress prematurely terminated.") | ||
187 | |||
188 | os.killpg(os.getpgid(p.pid), signal.SIGTERM) | ||
189 | log("Terminated stress") | ||
190 | |||
191 | #------------------------------------------------------------------------------- | ||
192 | |||
193 | def us2hms_str(us): | ||
194 | s = (us+500000) // 1000000 # Round microseconds to s | ||
195 | m = s//60 | ||
196 | s -= 60*m; | ||
197 | h = m//60 | ||
198 | m -= 60*h | ||
199 | |||
200 | return "%d:%02d:%02d" % (h, m, s) | ||
201 | |||
202 | #------------------------------------------------------------------------------- | ||
203 | # Sometime the hung task supervision is triggered during execution of | ||
204 | # cyclictest (cyclictest starves stress). To avoid that, the supervision | ||
205 | # is temporarily disabled | ||
206 | |||
207 | def set_hung_tmo(new_tmo): | ||
208 | global has_hung_task_detection | ||
209 | |||
210 | tmo_file = "/proc/sys/kernel/hung_task_timeout_secs" | ||
211 | |||
212 | if not has_hung_task_detection: | ||
213 | return | ||
214 | |||
215 | if not os.access(tmo_file, os.W_OK): | ||
216 | log("Hung task detection not supported") | ||
217 | log(" (File ", tmo_file, " not found)") | ||
218 | has_hung_task_detection = False | ||
219 | return | ||
220 | |||
221 | orig_tmo = int(subprocess.check_output(["cat", tmo_file]).strip()) | ||
222 | if new_tmo != orig_tmo: | ||
223 | cmd = ( "echo " + str(new_tmo) + " > " + tmo_file ) | ||
224 | subprocess.check_output(cmd, shell=True) | ||
225 | log("Changed timeout for detection of hung tasks: ", | ||
226 | orig_tmo, " -> ", new_tmo) | ||
227 | |||
228 | return orig_tmo | ||
229 | |||
230 | #------------------------------------------------------------------------------- | ||
231 | |||
232 | def gen_minmax_list(max_list): | ||
233 | res = [min(max_list)] | ||
234 | |||
235 | while True: | ||
236 | tmp = max_list | ||
237 | max_list = [] | ||
238 | while tmp: | ||
239 | max_list.append(max(tmp.pop(0), tmp.pop(0))) | ||
240 | |||
241 | res.append(min(max_list)) | ||
242 | |||
243 | if len(max_list) < 2: | ||
244 | return res | ||
245 | |||
246 | #------------------------------------------------------------------------------- | ||
247 | # Parameters for cyclictest: | ||
248 | # | ||
249 | # On the -S option (from cyclictest.c): | ||
250 | # -S implies options -a -t -n and same priority of all threads | ||
251 | # -a: One thread per core | ||
252 | # -n: use clock_nanosleep instead of posix interval timers | ||
253 | # -t: (without argument) Set number of threads to the number | ||
254 | # of cpus | ||
255 | |||
256 | interval_core_0 = 100 # Timer interval on core 0 [us] | ||
257 | interval_delta = 20 # Interval increment for each core [us] | ||
258 | loop_count = 30000 # Number of loops (on core 0). | ||
259 | |||
260 | cmd = ("cyclictest", | ||
261 | "-S", # Standard SMP testing. See below | ||
262 | "-p", "99", # RT priority 99 | ||
263 | "-q", # Quiet mode, i.e. print only a summary | ||
264 | "-i", str(interval_core_0), | ||
265 | "-d", str(interval_delta), | ||
266 | "-l", str(loop_count) | ||
267 | ) | ||
268 | rex = re.compile("C:\s*(\d+).*Min:\s*(\d+).*Avg:\s*(\d+).*Max:\s*(\d+)") | ||
269 | |||
270 | def run_cyclictest_once(): | ||
271 | res = subprocess.check_output(cmd) | ||
272 | |||
273 | # minlist and maxlist are lists with the extremes for each core | ||
274 | # avg_cnt is the sum of cycles for all cores | ||
275 | # avg_sum is the sum of (cycle count*average) for each core | ||
276 | # Since cyclictest runs different number of cycles on | ||
277 | # different cores, avg_sum/avg_cnt gives a more accurate | ||
278 | # value of the overall average than just taking the average | ||
279 | # of each core's averages | ||
280 | |||
281 | minlist = [] | ||
282 | maxlist = [] | ||
283 | avg_sum = 0 | ||
284 | avg_cnt = 0 | ||
285 | |||
286 | for line in res.split("\n"): | ||
287 | m = rex.search(line) | ||
288 | if m is not None: | ||
289 | minlist.append(int(m.group(2))) | ||
290 | maxlist.append(int(m.group(4))) | ||
291 | n = int(m.group(1)) | ||
292 | avg_sum += n * int(m.group(3)) | ||
293 | avg_cnt += n | ||
294 | |||
295 | return min(minlist), [avg_sum, avg_cnt], max(maxlist) | ||
296 | |||
297 | #------------------------------------------------------------------------------- | ||
298 | # A precondition for the tracking of min-max values is that | ||
299 | # the suite size os a power of 2. | ||
300 | |||
301 | N = 5 | ||
302 | suite_size = 2**N | ||
303 | |||
304 | est_exec_time_once = interval_core_0 * loop_count | ||
305 | est_exec_time_suite = suite_size * est_exec_time_once | ||
306 | |||
307 | def run_cyclictest_suite(): | ||
308 | log("Starting cyclictest") | ||
309 | log(" Command : ", " ".join(cmd)) | ||
310 | log(" Number of cycles : ", loop_count*suite_size, | ||
311 | " (", suite_size, " sets of ", loop_count, " cycles)") | ||
312 | log(" Exec. time (est) : ", us2hms_str(est_exec_time_suite)) | ||
313 | log() | ||
314 | |||
315 | orig_tmo = set_hung_tmo(0) # 0 => Disable | ||
316 | |||
317 | # float('inf') emulates infinity. At least in the sense that it is | ||
318 | # guaranteed to be larger than any actual number. | ||
319 | ack_min = float('inf') | ||
320 | ack_avg = [0, 0] | ||
321 | |||
322 | log() | ||
323 | log_ts("Start of execution") | ||
324 | t = time.time() | ||
325 | max_list = [] | ||
326 | |||
327 | for i in xrange(0, suite_size): | ||
328 | tmp_min, tmp_avg, tmp_max = run_cyclictest_once() | ||
329 | |||
330 | msg = "%2d/%2d:" % (i+1, suite_size) | ||
331 | msg += " min: %4d" % tmp_min | ||
332 | msg += " avg: %5.1f" % (float(tmp_avg[0])/tmp_avg[1]) | ||
333 | msg += " max: %4d" % tmp_max | ||
334 | log_ts(msg) | ||
335 | |||
336 | # Track minimum value | ||
337 | if tmp_min < ack_min: | ||
338 | ack_min = tmp_min | ||
339 | |||
340 | # Track smallest max value | ||
341 | max_list.append(tmp_max) | ||
342 | |||
343 | ack_avg[0] += tmp_avg[0] | ||
344 | ack_avg[1] += tmp_avg[1] | ||
345 | |||
346 | t = time.time()-t | ||
347 | log_ts("Cyclictest completed. Actual execution time:", | ||
348 | us2hms_str(t*1000000)) | ||
349 | log() | ||
350 | set_hung_tmo(orig_tmo) | ||
351 | |||
352 | return ack_min, float(ack_avg[0])/ack_avg[1], gen_minmax_list(max_list) | ||
353 | |||
354 | #------------------------------------------------------------------------------- | ||
355 | |||
356 | class cyclictest_runner: | ||
357 | def run_test(self, seq_no, nr_of_tests, name, stressparams): | ||
358 | |||
359 | try: | ||
360 | log_test_header(seq_no, nr_of_tests, name) | ||
361 | |||
362 | p = start_stress(*stressparams) | ||
363 | |||
364 | bm_min, bm_avg, bm_max_list = run_cyclictest_suite() | ||
365 | |||
366 | end_stress(p) | ||
367 | |||
368 | bm_max = bm_max_list[-1] | ||
369 | |||
370 | log() | ||
371 | log("Min: %d us" % bm_min) | ||
372 | log("Avg: %.1f us" % bm_avg) | ||
373 | log("Max: %d us" % bm_max) | ||
374 | log() | ||
375 | log("Max list: ", bm_max_list) | ||
376 | log() | ||
377 | log("PASS") | ||
378 | |||
379 | |||
380 | print tc_name(name), "[Min/us,Avg/us,Max/us]:", | ||
381 | print "%d,%.1f,%d" % (bm_min,bm_avg, bm_max) | ||
382 | print "PASS:", tc_name(name) | ||
383 | |||
384 | |||
385 | except Exception: | ||
386 | log() | ||
387 | log("Exception!") | ||
388 | log() | ||
389 | log("Traceback:", traceback.format_exc()) | ||
390 | log() | ||
391 | log("WD: ", os.getcwd()) | ||
392 | log() | ||
393 | log("FAIL") | ||
394 | |||
395 | print "FAIL:", tc_name(name) | ||
396 | |||
397 | |||
398 | #------------------------------------------------------------------------------- | ||
399 | |||
400 | runner = cyclictest_runner() | ||
401 | |||
402 | tests = (("no_stress", []), | ||
403 | ("cpu", ["cpu"]), | ||
404 | ("hdd", ["hdd"]), | ||
405 | ("io", ["io"]), | ||
406 | ("vm", ["vm"]), | ||
407 | ("full", ["io", "cpu", "hdd", "vm"]), | ||
408 | ) | ||
409 | |||
410 | nr_of_tests = len(tests) | ||
411 | for seq_no, params in enumerate(tests, start=1): | ||
412 | runner.run_test(seq_no, nr_of_tests, *params) | ||
diff --git a/meta/recipes-rt/rt-tests/files/run-ptest b/meta/recipes-rt/rt-tests/files/run-ptest new file mode 100644 index 0000000000..8e6d3e3588 --- /dev/null +++ b/meta/recipes-rt/rt-tests/files/run-ptest | |||
@@ -0,0 +1,3 @@ | |||
1 | #!/bin/sh | ||
2 | |||
3 | python ./rt_bmark.py | ||
diff --git a/meta/recipes-rt/rt-tests/rt-tests.inc b/meta/recipes-rt/rt-tests/rt-tests.inc index 4942701531..68a972ab4e 100644 --- a/meta/recipes-rt/rt-tests/rt-tests.inc +++ b/meta/recipes-rt/rt-tests/rt-tests.inc | |||
@@ -1,6 +1,6 @@ | |||
1 | # Version v0.89 | 1 | # Version v0.91 |
2 | PV = "0.89" | 2 | PV = "0.91" |
3 | SRCREV = "9164ec4db2645f52ffdd20e6d1b72bb0411ab40e" | 3 | SRCREV = "3fed00ff8d2c0cdbc5ba078cbd60ce3cfcee615f" |
4 | 4 | ||
5 | SRC_URI = "git://git.kernel.org/pub/scm/linux/kernel/git/clrkwllms/rt-tests.git \ | 5 | SRC_URI = "git://git.kernel.org/pub/scm/linux/kernel/git/clrkwllms/rt-tests.git \ |
6 | file://0001-rt-tests-Allow-for-user-specified-PYLIB.patch \ | 6 | file://0001-rt-tests-Allow-for-user-specified-PYLIB.patch \ |
diff --git a/meta/recipes-rt/rt-tests/rt-tests_0.89.bb b/meta/recipes-rt/rt-tests/rt-tests_0.91.bb index df646de8ed..fbe2f669e8 100644 --- a/meta/recipes-rt/rt-tests/rt-tests_0.89.bb +++ b/meta/recipes-rt/rt-tests/rt-tests_0.91.bb | |||
@@ -8,7 +8,11 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=751419260aa954499f7abaabaa882bbe \ | |||
8 | file://src/pi_tests/pi_stress.c;beginline=6;endline=19;md5=bd426a634a43ec612e9fbf125dfcc949" | 8 | file://src/pi_tests/pi_stress.c;beginline=6;endline=19;md5=bd426a634a43ec612e9fbf125dfcc949" |
9 | 9 | ||
10 | require rt-tests.inc | 10 | require rt-tests.inc |
11 | inherit ptest | ||
11 | 12 | ||
13 | SRC_URI += "file://run-ptest \ | ||
14 | file://rt_bmark.py \ | ||
15 | " | ||
12 | # Do not install hwlatdetect | 16 | # Do not install hwlatdetect |
13 | EXTRA_OEMAKE += "PYLIB=''" | 17 | EXTRA_OEMAKE += "PYLIB=''" |
14 | 18 | ||
@@ -17,4 +21,10 @@ do_install() { | |||
17 | INCLUDEDIR=${includedir} | 21 | INCLUDEDIR=${includedir} |
18 | } | 22 | } |
19 | 23 | ||
24 | do_install_ptest() { | ||
25 | cp ${WORKDIR}/rt_bmark.py ${D}${PTEST_PATH} | ||
26 | } | ||
27 | |||
28 | RDEPENDS_${PN}-ptest += " stress python python-subprocess python-multiprocessing python-datetime python-re python-lang" | ||
29 | |||
20 | FILES_${PN} += "${prefix}/src/backfire" | 30 | FILES_${PN} += "${prefix}/src/backfire" |