1#!/usr/bin/env python
2
3# Copyright (C) 2015 The Android Open Source Project
4#
5# Licensed under the Apache License, Version 2.0 (the 'License');
6# you may not use this file except in compliance with the License.
7# You may obtain a copy of the License at
8#
9#      http://www.apache.org/licenses/LICENSE-2.0
10#
11# Unless required by applicable law or agreed to in writing, software
12# distributed under the License is distributed on an 'AS IS' BASIS,
13# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14# See the License for the specific language governing permissions and
15# limitations under the License.
16
17"""
18Generates storage benchmark from captured strace output.
19
20Currently assumes that all mmap'ed regions are resource accesses, and emulates as pread().
21
22Usage:
23$ adb shell strace -p `pid zygote` -o /data/local/tmp/trace -f -ff -y -ttt -e trace=file,desc,munmap
24$ adb pull /data/local/tmp/trace*
25$ python benchgen.py trace.*
26
27"""
28
29import re, sys, collections, traceback, argparse
30
31from operator import itemgetter
32from collections import defaultdict
33
34class Event:
35    def __init__(self, thread, time, call, args, ret):
36        self.thread = thread
37        self.time = time
38        self.call = call
39        self.args = args
40        self.ret = ret
41
42    def __repr__(self):
43        return "%s(%s)=%s" % (self.call, repr(self.args), self.ret)
44
45
46class File:
47    def __init__(self, name, ident):
48        self.name = name
49        self.ident = ident
50        self.size = 0
51
52    def __repr__(self):
53        return self.name
54
55
56events = []
57files = {}
58
59def find_file(name):
60    name = name.strip('<>"')
61    if name not in files:
62        files[name] = File(name, len(files))
63    return files[name]
64
65def extract_file(e, arg):
66    if "<" in arg:
67        fd, path = arg.split("<")
68        path = path.strip(">")
69        handle = "t%sf%s" % (e.thread, fd)
70        return (fd, find_file(path), handle)
71    else:
72        return (None, None, None)
73
74def parse_args(s):
75    args = []
76    arg = ""
77    esc = False
78    quot = False
79    for c in s:
80        if esc:
81            esc = False
82            arg += c
83            continue
84
85        if c == '"':
86            if quot:
87                quot = False
88                continue
89            else:
90                quot = True
91                continue
92
93        if c == '\\':
94            esc = True
95            continue
96
97        if c == ',' and not quot:
98            args.append(arg.strip())
99            arg = ""
100        else:
101            arg += c
102
103    args.append(arg.strip())
104    return args
105
106
107bufsize = 1048576
108interesting = ["mmap2","read","write","pread64","pwrite64","fsync","fdatasync","openat","close","lseek","_llseek"]
109
110re_event = re.compile(r"^([\d\.]+) (.+?)\((.+?)\) = (.+?)$")
111re_arg = re.compile(r'''((?:[^,"']|"[^"]*"|'[^']*')+)''')
112for fn in sys.argv[1:]:
113    with open(fn) as f:
114        thread = int(fn.split(".")[-1])
115        for line in f:
116            line = re_event.match(line)
117            if not line: continue
118
119            time, call, args, ret = line.groups()
120            if call not in interesting: continue
121            if "/data/" not in args: continue
122
123            time = float(time)
124            args = parse_args(args)
125            events.append(Event(thread, time, call, args, ret))
126
127
128with open("BenchmarkGen.h", 'w') as bench:
129    print >>bench, """/*
130 * Copyright (C) 2015 The Android Open Source Project
131 *
132 * Licensed under the Apache License, Version 2.0 (the "License");
133 * you may not use this file except in compliance with the License.
134 * You may obtain a copy of the License at
135 *
136 *      http://www.apache.org/licenses/LICENSE-2.0
137 *
138 * Unless required by applicable law or agreed to in writing, software
139 * distributed under the License is distributed on an "AS IS" BASIS,
140 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
141 * See the License for the specific language governing permissions and
142 * limitations under the License.
143 */
144
145
146/******************************************************************
147 * THIS CODE WAS GENERATED BY benchgen.py, DO NOT MODIFY DIRECTLY *
148 ******************************************************************/
149
150
151#include <android-base/logging.h>
152
153#include <stdlib.h>
154#include <sys/types.h>
155#include <sys/stat.h>
156#include <sys/sendfile.h>
157#include <fcntl.h>
158
159#include <algorithm>
160#include <functional>
161#include <string>
162
163#include <Utils.h>
164
165namespace android {
166namespace vold {
167
168static status_t BenchmarkRun(std::function<bool(int)> checkpoint) {
169
170"""
171
172    print >>bench, "char* buf = (char*) malloc(%d);" % (bufsize)
173
174    nread = 0
175    nwrite = 0
176    nsync = 0
177    events = sorted(events, key=lambda e: e.time)
178    active = set()
179    defined = set()
180    i = 0
181    total = len(events)
182    for e in events:
183        i += 1
184        if i % 256 == 0:
185            print >>bench, "if (!checkpoint(%d)) return -1;" % (50 + ((i * 50) / total))
186
187        if e.call == "openat":
188            fd, f, handle = extract_file(e, e.ret)
189            if f:
190                active.add(handle)
191                if handle not in defined:
192                    print >>bench, "int",
193                    defined.add(handle)
194                create_mode = ''
195                if 'O_CREAT' in e.args[2]:
196                    assert len(e.args) > 3, 'File creation lacks a mode?'
197                    create_mode = ', ' + e.args[3]
198                print >>bench, '%s = TEMP_FAILURE_RETRY(open("file%s", %s%s));' \
199                        % (handle, f.ident, e.args[2], create_mode)
200
201        elif e.call == "close":
202            fd, f, handle = extract_file(e, e.args[0])
203            if handle in active:
204                active.remove(handle)
205                print >>bench, 'close(%s);' % (handle)
206
207        elif e.call == "lseek":
208            fd, f, handle = extract_file(e, e.args[0])
209            if handle in active:
210                print >>bench, 'TEMP_FAILURE_RETRY(lseek(%s, %s, %s));' % (handle, e.args[1], e.args[2])
211
212        elif e.call == "_llseek":
213            fd, f, handle = extract_file(e, e.args[0])
214            if handle in active:
215                print >>bench, 'TEMP_FAILURE_RETRY(lseek(%s, %s, %s));' % (handle, e.args[1], e.args[3])
216
217        elif e.call == "read":
218            fd, f, handle = extract_file(e, e.args[0])
219            if handle in active:
220                # TODO: track actual file size instead of guessing
221                count = min(int(e.args[2]), bufsize)
222                f.size += count
223                print >>bench, 'TEMP_FAILURE_RETRY(read(%s, buf, %d));' % (handle, count)
224                nread += 1
225
226        elif e.call == "write":
227            fd, f, handle = extract_file(e, e.args[0])
228            if handle in active:
229                # TODO: track actual file size instead of guessing
230                count = min(int(e.args[2]), bufsize)
231                f.size += count
232                print >>bench, 'TEMP_FAILURE_RETRY(write(%s, buf, %d));' % (handle, count)
233                nwrite += 1
234
235        elif e.call == "pread64":
236            fd, f, handle = extract_file(e, e.args[0])
237            if handle in active:
238                f.size = max(f.size, int(e.args[2]) + int(e.args[3]))
239                count = min(int(e.args[2]), bufsize)
240                print >>bench, 'TEMP_FAILURE_RETRY(pread(%s, buf, %d, %s));' % (handle, count, e.args[3])
241                nread += 1
242
243        elif e.call == "pwrite64":
244            fd, f, handle = extract_file(e, e.args[0])
245            if handle in active:
246                f.size = max(f.size, int(e.args[2]) + int(e.args[3]))
247                count = min(int(e.args[2]), bufsize)
248                print >>bench, 'TEMP_FAILURE_RETRY(pwrite(%s, buf, %d, %s));' % (handle, count, e.args[3])
249                nwrite += 1
250
251        elif e.call == "fsync":
252            fd, f, handle = extract_file(e, e.args[0])
253            if handle in active:
254                print >>bench, 'TEMP_FAILURE_RETRY(fsync(%s));' % (handle)
255                nsync += 1
256
257        elif e.call == "fdatasync":
258            fd, f, handle = extract_file(e, e.args[0])
259            if handle in active:
260                print >>bench, 'TEMP_FAILURE_RETRY(fdatasync(%s));' % (handle)
261                nsync += 1
262
263        elif e.call == "mmap2":
264            fd, f, handle = extract_file(e, e.args[4])
265            if handle in active:
266                count = min(int(e.args[1]), bufsize)
267                offset = int(e.args[5], 0)
268                f.size = max(f.size, count + offset)
269                print >>bench, 'TEMP_FAILURE_RETRY(pread(%s, buf, %s, %s)); // mmap2' % (handle, count, offset)
270                nread += 1
271
272    for handle in active:
273        print >>bench, 'close(%s);' % (handle)
274
275    print >>bench, """
276free(buf);
277return 0;
278}
279
280static status_t CreateFile(const char* name, int len) {
281    int chunk = std::min(len, 65536);
282    int out = -1;
283    std::string buf;
284
285    if (android::vold::ReadRandomBytes(chunk, buf) != OK) {
286        LOG(ERROR) << "Failed to read random data";
287        return -EIO;
288    }
289    if ((out = TEMP_FAILURE_RETRY(open(name, O_WRONLY|O_CREAT|O_TRUNC, 0644))) < 0) {
290        PLOG(ERROR) << "Failed to open " << name;
291        return -errno;
292    }
293
294    while (len > 0) {
295        int n = write(out, buf.c_str(), std::min(len, chunk));
296        if (n < 0) {
297            PLOG(ERROR) << "Failed to write";
298            close(out);
299            return -errno;
300        }
301        len -= n;
302    }
303
304    close(out);
305    return OK;
306}
307
308static status_t BenchmarkCreate(std::function<bool(int)> checkpoint) {
309status_t res = 0;
310res |= CreateFile("stub", 0);
311"""
312    i = 0
313    total = len(files.values())
314    for f in files.values():
315        i += 1
316        if i % 12 == 0:
317            print >>bench, "if (!checkpoint(%d)) return -1;" % ((i * 50) / total)
318
319        print >>bench, 'res |= CreateFile("file%s", %d);' % (f.ident, f.size)
320
321    print >>bench, """
322return res;
323}
324
325static status_t BenchmarkDestroy() {
326status_t res = 0;
327res |= unlink("stub");
328"""
329    for f in files.values():
330        print >>bench, 'res |= unlink("file%s");' % (f.ident)
331
332    print >>bench, """
333return res;
334}
335
336static std::string BenchmarkIdent() {"""
337    print >>bench, """return "r%d:w%d:s%d";""" % (nread, nwrite, nsync)
338    print >>bench, """}
339
340}  // namespace vold
341}  // namespace android
342"""
343
344
345size = sum([ f.size for f in files.values() ])
346print "Found", len(files), "data files accessed, total size", (size/1024), "kB"
347
348types = defaultdict(int)
349for e in events:
350    types[e.call] += 1
351
352print "Found syscalls:"
353for t, n in types.iteritems():
354    print str(n).rjust(8), t
355
356print
357