blob: 9c2a61554eb942ec6ece3bb43268dcd5db8ff60e [file] [log] [blame]
Austin Schuh745610d2015-09-06 18:19:50 -07001// -*- Mode: C++; c-basic-offset: 2; indent-tabs-mode: nil -*-
2// Copyright (c) 2006, Google Inc.
3// All rights reserved.
Brian Silverman20350ac2021-11-17 18:19:55 -08004//
Austin Schuh745610d2015-09-06 18:19:50 -07005// Redistribution and use in source and binary forms, with or without
6// modification, are permitted provided that the following conditions are
7// met:
Brian Silverman20350ac2021-11-17 18:19:55 -08008//
Austin Schuh745610d2015-09-06 18:19:50 -07009// * Redistributions of source code must retain the above copyright
10// notice, this list of conditions and the following disclaimer.
11// * Redistributions in binary form must reproduce the above
12// copyright notice, this list of conditions and the following disclaimer
13// in the documentation and/or other materials provided with the
14// distribution.
15// * Neither the name of Google Inc. nor the names of its
16// contributors may be used to endorse or promote products derived from
17// this software without specific prior written permission.
Brian Silverman20350ac2021-11-17 18:19:55 -080018//
Austin Schuh745610d2015-09-06 18:19:50 -070019// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30
31#include <config.h>
32#if (defined(_WIN32) || defined(__MINGW32__)) && !defined(__CYGWIN__) && !defined(__CYGWIN32)
33# define PLATFORM_WINDOWS 1
34#endif
35
36#include <ctype.h> // for isspace()
37#include <stdlib.h> // for getenv()
38#include <stdio.h> // for snprintf(), sscanf()
39#include <string.h> // for memmove(), memchr(), etc.
40#include <fcntl.h> // for open()
41#include <errno.h> // for errno
42#ifdef HAVE_UNISTD_H
43#include <unistd.h> // for read()
44#endif
45#if defined __MACH__ // Mac OS X, almost certainly
46#include <mach-o/dyld.h> // for iterating over dll's in ProcMapsIter
47#include <mach-o/loader.h> // for iterating over dll's in ProcMapsIter
48#include <sys/types.h>
49#include <sys/sysctl.h> // how we figure out numcpu's on OS X
50#elif defined __FreeBSD__
51#include <sys/sysctl.h>
52#elif defined __sun__ // Solaris
53#include <procfs.h> // for, e.g., prmap_t
54#elif defined(PLATFORM_WINDOWS)
55#include <process.h> // for getpid() (actually, _getpid())
56#include <shlwapi.h> // for SHGetValueA()
57#include <tlhelp32.h> // for Module32First()
58#endif
59#include "base/sysinfo.h"
60#include "base/commandlineflags.h"
61#include "base/dynamic_annotations.h" // for RunningOnValgrind
62#include "base/logging.h"
Austin Schuh745610d2015-09-06 18:19:50 -070063
64#ifdef PLATFORM_WINDOWS
65#ifdef MODULEENTRY32
66// In a change from the usual W-A pattern, there is no A variant of
67// MODULEENTRY32. Tlhelp32.h #defines the W variant, but not the A.
68// In unicode mode, tlhelp32.h #defines MODULEENTRY32 to be
69// MODULEENTRY32W. These #undefs are the only way I see to get back
70// access to the original, ascii struct (and related functions).
71#undef MODULEENTRY32
72#undef Module32First
73#undef Module32Next
74#undef PMODULEENTRY32
75#undef LPMODULEENTRY32
76#endif /* MODULEENTRY32 */
77// MinGW doesn't seem to define this, perhaps some windowsen don't either.
78#ifndef TH32CS_SNAPMODULE32
79#define TH32CS_SNAPMODULE32 0
80#endif /* TH32CS_SNAPMODULE32 */
81#endif /* PLATFORM_WINDOWS */
82
83// Re-run fn until it doesn't cause EINTR.
84#define NO_INTR(fn) do {} while ((fn) < 0 && errno == EINTR)
85
86// open/read/close can set errno, which may be illegal at this
87// time, so prefer making the syscalls directly if we can.
88#ifdef HAVE_SYS_SYSCALL_H
89# include <sys/syscall.h>
90#endif
91#ifdef SYS_open // solaris 11, at least sometimes, only defines SYS_openat
92# define safeopen(filename, mode) syscall(SYS_open, filename, mode)
93#else
94# define safeopen(filename, mode) open(filename, mode)
95#endif
96#ifdef SYS_read
97# define saferead(fd, buffer, size) syscall(SYS_read, fd, buffer, size)
98#else
99# define saferead(fd, buffer, size) read(fd, buffer, size)
100#endif
101#ifdef SYS_close
102# define safeclose(fd) syscall(SYS_close, fd)
103#else
104# define safeclose(fd) close(fd)
105#endif
106
107// ----------------------------------------------------------------------
108// GetenvBeforeMain()
109// GetUniquePathFromEnv()
110// Some non-trivial getenv-related functions.
111// ----------------------------------------------------------------------
112
Brian Silverman20350ac2021-11-17 18:19:55 -0800113// we reimplement memcmp and friends to avoid depending on any glibc
114// calls too early in the process lifetime. This allows us to use
115// GetenvBeforeMain from inside ifunc handler
116static int slow_memcmp(const void *_a, const void *_b, size_t n) {
117 const uint8_t *a = reinterpret_cast<const uint8_t *>(_a);
118 const uint8_t *b = reinterpret_cast<const uint8_t *>(_b);
119 while (n-- != 0) {
120 uint8_t ac = *a++;
121 uint8_t bc = *b++;
122 if (ac != bc) {
123 if (ac < bc) {
124 return -1;
125 }
126 return 1;
127 }
128 }
129 return 0;
130}
131
132static const char *slow_memchr(const char *s, int c, size_t n) {
133 uint8_t ch = static_cast<uint8_t>(c);
134 while (n--) {
135 if (*s++ == ch) {
136 return s - 1;
137 }
138 }
139 return 0;
140}
141
142static size_t slow_strlen(const char *s) {
143 const char *s2 = slow_memchr(s, '\0', static_cast<size_t>(-1));
144 return s2 - s;
145}
146
Austin Schuh745610d2015-09-06 18:19:50 -0700147// It's not safe to call getenv() in the malloc hooks, because they
148// might be called extremely early, before libc is done setting up
149// correctly. In particular, the thread library may not be done
150// setting up errno. So instead, we use the built-in __environ array
151// if it exists, and otherwise read /proc/self/environ directly, using
152// system calls to read the file, and thus avoid setting errno.
153// /proc/self/environ has a limit of how much data it exports (around
154// 8K), so it's not an ideal solution.
155const char* GetenvBeforeMain(const char* name) {
Brian Silverman20350ac2021-11-17 18:19:55 -0800156 const int namelen = slow_strlen(name);
Austin Schuh745610d2015-09-06 18:19:50 -0700157#if defined(HAVE___ENVIRON) // if we have it, it's declared in unistd.h
158 if (__environ) { // can exist but be NULL, if statically linked
Austin Schuh745610d2015-09-06 18:19:50 -0700159 for (char** p = __environ; *p; p++) {
Brian Silverman20350ac2021-11-17 18:19:55 -0800160 if (!slow_memcmp(*p, name, namelen) && (*p)[namelen] == '=')
161 return *p + namelen+1;
Austin Schuh745610d2015-09-06 18:19:50 -0700162 }
163 return NULL;
164 }
165#endif
166#if defined(PLATFORM_WINDOWS)
167 // TODO(mbelshe) - repeated calls to this function will overwrite the
168 // contents of the static buffer.
169 static char envvar_buf[1024]; // enough to hold any envvar we care about
170 if (!GetEnvironmentVariableA(name, envvar_buf, sizeof(envvar_buf)-1))
171 return NULL;
172 return envvar_buf;
173#endif
174 // static is ok because this function should only be called before
175 // main(), when we're single-threaded.
176 static char envbuf[16<<10];
177 if (*envbuf == '\0') { // haven't read the environ yet
178 int fd = safeopen("/proc/self/environ", O_RDONLY);
179 // The -2 below guarantees the last two bytes of the buffer will be \0\0
180 if (fd == -1 || // unable to open the file, fall back onto libc
181 saferead(fd, envbuf, sizeof(envbuf) - 2) < 0) { // error reading file
182 RAW_VLOG(1, "Unable to open /proc/self/environ, falling back "
183 "on getenv(\"%s\"), which may not work", name);
184 if (fd != -1) safeclose(fd);
185 return getenv(name);
186 }
187 safeclose(fd);
188 }
Austin Schuh745610d2015-09-06 18:19:50 -0700189 const char* p = envbuf;
190 while (*p != '\0') { // will happen at the \0\0 that terminates the buffer
191 // proc file has the format NAME=value\0NAME=value\0NAME=value\0...
Brian Silvermana659aaa2022-01-02 00:19:11 -0800192 const char* endp = (const char*)slow_memchr(p, '\0',
Brian Silverman20350ac2021-11-17 18:19:55 -0800193 sizeof(envbuf) - (p - envbuf));
Austin Schuh745610d2015-09-06 18:19:50 -0700194 if (endp == NULL) // this entry isn't NUL terminated
195 return NULL;
Brian Silverman20350ac2021-11-17 18:19:55 -0800196 else if (!slow_memcmp(p, name, namelen) && p[namelen] == '=') // it's a match
Austin Schuh745610d2015-09-06 18:19:50 -0700197 return p + namelen+1; // point after =
198 p = endp + 1;
199 }
200 return NULL; // env var never found
201}
202
203extern "C" {
204 const char* TCMallocGetenvSafe(const char* name) {
205 return GetenvBeforeMain(name);
206 }
207}
208
209// This takes as an argument an environment-variable name (like
210// CPUPROFILE) whose value is supposed to be a file-path, and sets
211// path to that path, and returns true. If the env var doesn't exist,
212// or is the empty string, leave path unchanged and returns false.
213// The reason this is non-trivial is that this function handles munged
214// pathnames. Here's why:
215//
216// If we're a child process of the 'main' process, we can't just use
217// getenv("CPUPROFILE") -- the parent process will be using that path.
218// Instead we append our pid to the pathname. How do we tell if we're a
219// child process? Ideally we'd set an environment variable that all
220// our children would inherit. But -- and this is seemingly a bug in
221// gcc -- if you do a setenv() in a shared libarary in a global
222// constructor, the environment setting is lost by the time main() is
223// called. The only safe thing we can do in such a situation is to
224// modify the existing envvar. So we do a hack: in the parent, we set
225// the high bit of the 1st char of CPUPROFILE. In the child, we
226// notice the high bit is set and append the pid(). This works
227// assuming cpuprofile filenames don't normally have the high bit set
228// in their first character! If that assumption is violated, we'll
229// still get a profile, but one with an unexpected name.
230// TODO(csilvers): set an envvar instead when we can do it reliably.
231bool GetUniquePathFromEnv(const char* env_name, char* path) {
232 char* envval = getenv(env_name);
233 if (envval == NULL || *envval == '\0')
234 return false;
235 if (envval[0] & 128) { // high bit is set
236 snprintf(path, PATH_MAX, "%c%s_%u", // add pid and clear high bit
237 envval[0] & 127, envval+1, (unsigned int)(getpid()));
238 } else {
239 snprintf(path, PATH_MAX, "%s", envval);
240 envval[0] |= 128; // set high bit for kids to see
241 }
242 return true;
243}
244
Brian Silverman20350ac2021-11-17 18:19:55 -0800245int GetSystemCPUsCount()
246{
247#if defined(PLATFORM_WINDOWS)
Austin Schuh745610d2015-09-06 18:19:50 -0700248 // Get the number of processors.
249 SYSTEM_INFO info;
250 GetSystemInfo(&info);
Brian Silverman20350ac2021-11-17 18:19:55 -0800251 return info.dwNumberOfProcessors;
Austin Schuh745610d2015-09-06 18:19:50 -0700252#else
Brian Silverman20350ac2021-11-17 18:19:55 -0800253 long rv = sysconf(_SC_NPROCESSORS_ONLN);
254 if (rv < 0) {
255 return 1;
256 }
257 return static_cast<int>(rv);
Austin Schuh745610d2015-09-06 18:19:50 -0700258#endif
259}
260
Austin Schuh745610d2015-09-06 18:19:50 -0700261// ----------------------------------------------------------------------
262
263#if defined __linux__ || defined __FreeBSD__ || defined __sun__ || defined __CYGWIN__ || defined __CYGWIN32__
264static void ConstructFilename(const char* spec, pid_t pid,
265 char* buf, int buf_size) {
266 CHECK_LT(snprintf(buf, buf_size,
267 spec,
268 static_cast<int>(pid ? pid : getpid())), buf_size);
269}
270#endif
271
272// A templatized helper function instantiated for Mach (OS X) only.
273// It can handle finding info for both 32 bits and 64 bits.
274// Returns true if it successfully handled the hdr, false else.
275#ifdef __MACH__ // Mac OS X, almost certainly
276template<uint32_t kMagic, uint32_t kLCSegment,
277 typename MachHeader, typename SegmentCommand>
278static bool NextExtMachHelper(const mach_header* hdr,
279 int current_image, int current_load_cmd,
280 uint64 *start, uint64 *end, char **flags,
281 uint64 *offset, int64 *inode, char **filename,
282 uint64 *file_mapping, uint64 *file_pages,
283 uint64 *anon_mapping, uint64 *anon_pages,
284 dev_t *dev) {
285 static char kDefaultPerms[5] = "r-xp";
286 if (hdr->magic != kMagic)
287 return false;
288 const char* lc = (const char *)hdr + sizeof(MachHeader);
289 // TODO(csilvers): make this not-quadradic (increment and hold state)
290 for (int j = 0; j < current_load_cmd; j++) // advance to *our* load_cmd
291 lc += ((const load_command *)lc)->cmdsize;
292 if (((const load_command *)lc)->cmd == kLCSegment) {
293 const intptr_t dlloff = _dyld_get_image_vmaddr_slide(current_image);
294 const SegmentCommand* sc = (const SegmentCommand *)lc;
295 if (start) *start = sc->vmaddr + dlloff;
296 if (end) *end = sc->vmaddr + sc->vmsize + dlloff;
297 if (flags) *flags = kDefaultPerms; // can we do better?
298 if (offset) *offset = sc->fileoff;
299 if (inode) *inode = 0;
300 if (filename)
301 *filename = const_cast<char*>(_dyld_get_image_name(current_image));
302 if (file_mapping) *file_mapping = 0;
303 if (file_pages) *file_pages = 0; // could we use sc->filesize?
304 if (anon_mapping) *anon_mapping = 0;
305 if (anon_pages) *anon_pages = 0;
306 if (dev) *dev = 0;
307 return true;
308 }
309
310 return false;
311}
312#endif
313
314// Finds |c| in |text|, and assign '\0' at the found position.
315// The original character at the modified position should be |c|.
316// A pointer to the modified position is stored in |endptr|.
317// |endptr| should not be NULL.
318static bool ExtractUntilChar(char *text, int c, char **endptr) {
319 CHECK_NE(text, NULL);
320 CHECK_NE(endptr, NULL);
321 char *found;
322 found = strchr(text, c);
323 if (found == NULL) {
324 *endptr = NULL;
325 return false;
326 }
327
328 *endptr = found;
329 *found = '\0';
330 return true;
331}
332
333// Increments |*text_pointer| while it points a whitespace character.
334// It is to follow sscanf's whilespace handling.
335static void SkipWhileWhitespace(char **text_pointer, int c) {
336 if (isspace(c)) {
337 while (isspace(**text_pointer) && isspace(*((*text_pointer) + 1))) {
338 ++(*text_pointer);
339 }
340 }
341}
342
343template<class T>
344static T StringToInteger(char *text, char **endptr, int base) {
345 assert(false);
346 return T();
347}
348
349template<>
350int StringToInteger<int>(char *text, char **endptr, int base) {
351 return strtol(text, endptr, base);
352}
353
354template<>
355int64 StringToInteger<int64>(char *text, char **endptr, int base) {
356 return strtoll(text, endptr, base);
357}
358
359template<>
360uint64 StringToInteger<uint64>(char *text, char **endptr, int base) {
361 return strtoull(text, endptr, base);
362}
363
364template<typename T>
365static T StringToIntegerUntilChar(
366 char *text, int base, int c, char **endptr_result) {
367 CHECK_NE(endptr_result, NULL);
368 *endptr_result = NULL;
369
370 char *endptr_extract;
371 if (!ExtractUntilChar(text, c, &endptr_extract))
372 return 0;
373
374 T result;
375 char *endptr_strto;
376 result = StringToInteger<T>(text, &endptr_strto, base);
377 *endptr_extract = c;
378
379 if (endptr_extract != endptr_strto)
380 return 0;
381
382 *endptr_result = endptr_extract;
383 SkipWhileWhitespace(endptr_result, c);
384
385 return result;
386}
387
388static char *CopyStringUntilChar(
389 char *text, unsigned out_len, int c, char *out) {
390 char *endptr;
391 if (!ExtractUntilChar(text, c, &endptr))
392 return NULL;
393
394 strncpy(out, text, out_len);
395 out[out_len-1] = '\0';
396 *endptr = c;
397
398 SkipWhileWhitespace(&endptr, c);
399 return endptr;
400}
401
402template<typename T>
403static bool StringToIntegerUntilCharWithCheck(
404 T *outptr, char *text, int base, int c, char **endptr) {
405 *outptr = StringToIntegerUntilChar<T>(*endptr, base, c, endptr);
406 if (*endptr == NULL || **endptr == '\0') return false;
407 ++(*endptr);
408 return true;
409}
410
411static bool ParseProcMapsLine(char *text, uint64 *start, uint64 *end,
412 char *flags, uint64 *offset,
413 int *major, int *minor, int64 *inode,
414 unsigned *filename_offset) {
415#if defined(__linux__)
416 /*
417 * It's similar to:
418 * sscanf(text, "%"SCNx64"-%"SCNx64" %4s %"SCNx64" %x:%x %"SCNd64" %n",
419 * start, end, flags, offset, major, minor, inode, filename_offset)
420 */
421 char *endptr = text;
422 if (endptr == NULL || *endptr == '\0') return false;
423
424 if (!StringToIntegerUntilCharWithCheck(start, endptr, 16, '-', &endptr))
425 return false;
426
427 if (!StringToIntegerUntilCharWithCheck(end, endptr, 16, ' ', &endptr))
428 return false;
429
430 endptr = CopyStringUntilChar(endptr, 5, ' ', flags);
431 if (endptr == NULL || *endptr == '\0') return false;
432 ++endptr;
433
434 if (!StringToIntegerUntilCharWithCheck(offset, endptr, 16, ' ', &endptr))
435 return false;
436
437 if (!StringToIntegerUntilCharWithCheck(major, endptr, 16, ':', &endptr))
438 return false;
439
440 if (!StringToIntegerUntilCharWithCheck(minor, endptr, 16, ' ', &endptr))
441 return false;
442
443 if (!StringToIntegerUntilCharWithCheck(inode, endptr, 10, ' ', &endptr))
444 return false;
445
446 *filename_offset = (endptr - text);
447 return true;
448#else
449 return false;
450#endif
451}
452
453ProcMapsIterator::ProcMapsIterator(pid_t pid) {
454 Init(pid, NULL, false);
455}
456
457ProcMapsIterator::ProcMapsIterator(pid_t pid, Buffer *buffer) {
458 Init(pid, buffer, false);
459}
460
461ProcMapsIterator::ProcMapsIterator(pid_t pid, Buffer *buffer,
462 bool use_maps_backing) {
463 Init(pid, buffer, use_maps_backing);
464}
465
466void ProcMapsIterator::Init(pid_t pid, Buffer *buffer,
467 bool use_maps_backing) {
468 pid_ = pid;
469 using_maps_backing_ = use_maps_backing;
470 dynamic_buffer_ = NULL;
471 if (!buffer) {
472 // If the user didn't pass in any buffer storage, allocate it
473 // now. This is the normal case; the signal handler passes in a
474 // static buffer.
475 buffer = dynamic_buffer_ = new Buffer;
476 } else {
477 dynamic_buffer_ = NULL;
478 }
479
480 ibuf_ = buffer->buf_;
481
482 stext_ = etext_ = nextline_ = ibuf_;
483 ebuf_ = ibuf_ + Buffer::kBufSize - 1;
484 nextline_ = ibuf_;
485
486#if defined(__linux__) || defined(__CYGWIN__) || defined(__CYGWIN32__)
487 if (use_maps_backing) { // don't bother with clever "self" stuff in this case
488 ConstructFilename("/proc/%d/maps_backing", pid, ibuf_, Buffer::kBufSize);
489 } else if (pid == 0) {
490 // We have to kludge a bit to deal with the args ConstructFilename
491 // expects. The 1 is never used -- it's only impt. that it's not 0.
492 ConstructFilename("/proc/self/maps", 1, ibuf_, Buffer::kBufSize);
493 } else {
494 ConstructFilename("/proc/%d/maps", pid, ibuf_, Buffer::kBufSize);
495 }
496 // No error logging since this can be called from the crash dump
497 // handler at awkward moments. Users should call Valid() before
498 // using.
499 NO_INTR(fd_ = open(ibuf_, O_RDONLY));
500#elif defined(__FreeBSD__)
501 // We don't support maps_backing on freebsd
502 if (pid == 0) {
503 ConstructFilename("/proc/curproc/map", 1, ibuf_, Buffer::kBufSize);
504 } else {
505 ConstructFilename("/proc/%d/map", pid, ibuf_, Buffer::kBufSize);
506 }
507 NO_INTR(fd_ = open(ibuf_, O_RDONLY));
508#elif defined(__sun__)
509 if (pid == 0) {
510 ConstructFilename("/proc/self/map", 1, ibuf_, Buffer::kBufSize);
511 } else {
512 ConstructFilename("/proc/%d/map", pid, ibuf_, Buffer::kBufSize);
513 }
514 NO_INTR(fd_ = open(ibuf_, O_RDONLY));
515#elif defined(__MACH__)
516 current_image_ = _dyld_image_count(); // count down from the top
517 current_load_cmd_ = -1;
518#elif defined(PLATFORM_WINDOWS)
519 snapshot_ = CreateToolhelp32Snapshot(TH32CS_SNAPMODULE |
520 TH32CS_SNAPMODULE32,
521 GetCurrentProcessId());
522 memset(&module_, 0, sizeof(module_));
523#else
524 fd_ = -1; // so Valid() is always false
525#endif
526
527}
528
529ProcMapsIterator::~ProcMapsIterator() {
530#if defined(PLATFORM_WINDOWS)
531 if (snapshot_ != INVALID_HANDLE_VALUE) CloseHandle(snapshot_);
532#elif defined(__MACH__)
533 // no cleanup necessary!
534#else
535 if (fd_ >= 0) NO_INTR(close(fd_));
536#endif
537 delete dynamic_buffer_;
538}
539
540bool ProcMapsIterator::Valid() const {
541#if defined(PLATFORM_WINDOWS)
542 return snapshot_ != INVALID_HANDLE_VALUE;
543#elif defined(__MACH__)
544 return 1;
545#else
546 return fd_ != -1;
547#endif
548}
549
550bool ProcMapsIterator::Next(uint64 *start, uint64 *end, char **flags,
551 uint64 *offset, int64 *inode, char **filename) {
552 return NextExt(start, end, flags, offset, inode, filename, NULL, NULL,
553 NULL, NULL, NULL);
554}
555
556// This has too many arguments. It should really be building
557// a map object and returning it. The problem is that this is called
558// when the memory allocator state is undefined, hence the arguments.
559bool ProcMapsIterator::NextExt(uint64 *start, uint64 *end, char **flags,
560 uint64 *offset, int64 *inode, char **filename,
561 uint64 *file_mapping, uint64 *file_pages,
562 uint64 *anon_mapping, uint64 *anon_pages,
563 dev_t *dev) {
564
565#if defined(__linux__) || defined(__FreeBSD__) || defined(__CYGWIN__) || defined(__CYGWIN32__)
566 do {
567 // Advance to the start of the next line
568 stext_ = nextline_;
569
570 // See if we have a complete line in the buffer already
571 nextline_ = static_cast<char *>(memchr (stext_, '\n', etext_ - stext_));
572 if (!nextline_) {
573 // Shift/fill the buffer so we do have a line
574 int count = etext_ - stext_;
575
576 // Move the current text to the start of the buffer
577 memmove(ibuf_, stext_, count);
578 stext_ = ibuf_;
579 etext_ = ibuf_ + count;
580
581 int nread = 0; // fill up buffer with text
582 while (etext_ < ebuf_) {
583 NO_INTR(nread = read(fd_, etext_, ebuf_ - etext_));
584 if (nread > 0)
585 etext_ += nread;
586 else
587 break;
588 }
589
590 // Zero out remaining characters in buffer at EOF to avoid returning
591 // garbage from subsequent calls.
592 if (etext_ != ebuf_ && nread == 0) {
593 memset(etext_, 0, ebuf_ - etext_);
594 }
595 *etext_ = '\n'; // sentinel; safe because ibuf extends 1 char beyond ebuf
596 nextline_ = static_cast<char *>(memchr (stext_, '\n', etext_ + 1 - stext_));
597 }
598 *nextline_ = 0; // turn newline into nul
599 nextline_ += ((nextline_ < etext_)? 1 : 0); // skip nul if not end of text
600 // stext_ now points at a nul-terminated line
601 uint64 tmpstart, tmpend, tmpoffset;
602 int64 tmpinode;
603 int major, minor;
604 unsigned filename_offset = 0;
605#if defined(__linux__)
606 // for now, assume all linuxes have the same format
607 if (!ParseProcMapsLine(
608 stext_,
609 start ? start : &tmpstart,
610 end ? end : &tmpend,
611 flags_,
612 offset ? offset : &tmpoffset,
613 &major, &minor,
614 inode ? inode : &tmpinode, &filename_offset)) continue;
615#elif defined(__CYGWIN__) || defined(__CYGWIN32__)
616 // cygwin is like linux, except the third field is the "entry point"
617 // rather than the offset (see format_process_maps at
618 // http://cygwin.com/cgi-bin/cvsweb.cgi/src/winsup/cygwin/fhandler_process.cc?rev=1.89&content-type=text/x-cvsweb-markup&cvsroot=src
619 // Offset is always be 0 on cygwin: cygwin implements an mmap
620 // by loading the whole file and then calling NtMapViewOfSection.
621 // Cygwin also seems to set its flags kinda randomly; use windows default.
622 char tmpflags[5];
623 if (offset)
624 *offset = 0;
625 strcpy(flags_, "r-xp");
626 if (sscanf(stext_, "%llx-%llx %4s %llx %x:%x %lld %n",
627 start ? start : &tmpstart,
628 end ? end : &tmpend,
629 tmpflags,
630 &tmpoffset,
631 &major, &minor,
632 inode ? inode : &tmpinode, &filename_offset) != 7) continue;
633#elif defined(__FreeBSD__)
634 // For the format, see http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/fs/procfs/procfs_map.c?rev=1.31&content-type=text/x-cvsweb-markup
635 tmpstart = tmpend = tmpoffset = 0;
636 tmpinode = 0;
637 major = minor = 0; // can't get this info in freebsd
638 if (inode)
639 *inode = 0; // nor this
640 if (offset)
641 *offset = 0; // seems like this should be in there, but maybe not
642 // start end resident privateresident obj(?) prot refcnt shadowcnt
643 // flags copy_on_write needs_copy type filename:
644 // 0x8048000 0x804a000 2 0 0xc104ce70 r-x 1 0 0x0 COW NC vnode /bin/cat
645 if (sscanf(stext_, "0x%" SCNx64 " 0x%" SCNx64 " %*d %*d %*p %3s %*d %*d 0x%*x %*s %*s %*s %n",
646 start ? start : &tmpstart,
647 end ? end : &tmpend,
648 flags_,
649 &filename_offset) != 3) continue;
650#endif
651
652 // Depending on the Linux kernel being used, there may or may not be a space
653 // after the inode if there is no filename. sscanf will in such situations
654 // nondeterministically either fill in filename_offset or not (the results
655 // differ on multiple calls in the same run even with identical arguments).
656 // We don't want to wander off somewhere beyond the end of the string.
657 size_t stext_length = strlen(stext_);
658 if (filename_offset == 0 || filename_offset > stext_length)
659 filename_offset = stext_length;
660
661 // We found an entry
662 if (flags) *flags = flags_;
663 if (filename) *filename = stext_ + filename_offset;
664 if (dev) *dev = minor | (major << 8);
665
666 if (using_maps_backing_) {
667 // Extract and parse physical page backing info.
668 char *backing_ptr = stext_ + filename_offset +
669 strlen(stext_+filename_offset);
670
671 // find the second '('
672 int paren_count = 0;
673 while (--backing_ptr > stext_) {
674 if (*backing_ptr == '(') {
675 ++paren_count;
676 if (paren_count >= 2) {
677 uint64 tmp_file_mapping;
678 uint64 tmp_file_pages;
679 uint64 tmp_anon_mapping;
680 uint64 tmp_anon_pages;
681
682 sscanf(backing_ptr+1, "F %" SCNx64 " %" SCNd64 ") (A %" SCNx64 " %" SCNd64 ")",
683 file_mapping ? file_mapping : &tmp_file_mapping,
684 file_pages ? file_pages : &tmp_file_pages,
685 anon_mapping ? anon_mapping : &tmp_anon_mapping,
686 anon_pages ? anon_pages : &tmp_anon_pages);
687 // null terminate the file name (there is a space
688 // before the first (.
689 backing_ptr[-1] = 0;
690 break;
691 }
692 }
693 }
694 }
695
696 return true;
697 } while (etext_ > ibuf_);
698#elif defined(__sun__)
699 // This is based on MA_READ == 4, MA_WRITE == 2, MA_EXEC == 1
700 static char kPerms[8][4] = { "---", "--x", "-w-", "-wx",
701 "r--", "r-x", "rw-", "rwx" };
702 COMPILE_ASSERT(MA_READ == 4, solaris_ma_read_must_equal_4);
703 COMPILE_ASSERT(MA_WRITE == 2, solaris_ma_write_must_equal_2);
704 COMPILE_ASSERT(MA_EXEC == 1, solaris_ma_exec_must_equal_1);
705 Buffer object_path;
706 int nread = 0; // fill up buffer with text
707 NO_INTR(nread = read(fd_, ibuf_, sizeof(prmap_t)));
708 if (nread == sizeof(prmap_t)) {
709 long inode_from_mapname = 0;
710 prmap_t* mapinfo = reinterpret_cast<prmap_t*>(ibuf_);
711 // Best-effort attempt to get the inode from the filename. I think the
712 // two middle ints are major and minor device numbers, but I'm not sure.
713 sscanf(mapinfo->pr_mapname, "ufs.%*d.%*d.%ld", &inode_from_mapname);
714
715 if (pid_ == 0) {
716 CHECK_LT(snprintf(object_path.buf_, Buffer::kBufSize,
717 "/proc/self/path/%s", mapinfo->pr_mapname),
718 Buffer::kBufSize);
719 } else {
720 CHECK_LT(snprintf(object_path.buf_, Buffer::kBufSize,
721 "/proc/%d/path/%s",
722 static_cast<int>(pid_), mapinfo->pr_mapname),
723 Buffer::kBufSize);
724 }
725 ssize_t len = readlink(object_path.buf_, current_filename_, PATH_MAX);
726 CHECK_LT(len, PATH_MAX);
727 if (len < 0)
728 len = 0;
729 current_filename_[len] = '\0';
730
731 if (start) *start = mapinfo->pr_vaddr;
732 if (end) *end = mapinfo->pr_vaddr + mapinfo->pr_size;
733 if (flags) *flags = kPerms[mapinfo->pr_mflags & 7];
734 if (offset) *offset = mapinfo->pr_offset;
735 if (inode) *inode = inode_from_mapname;
736 if (filename) *filename = current_filename_;
737 if (file_mapping) *file_mapping = 0;
738 if (file_pages) *file_pages = 0;
739 if (anon_mapping) *anon_mapping = 0;
740 if (anon_pages) *anon_pages = 0;
741 if (dev) *dev = 0;
742 return true;
743 }
744#elif defined(__MACH__)
745 // We return a separate entry for each segment in the DLL. (TODO(csilvers):
746 // can we do better?) A DLL ("image") has load-commands, some of which
747 // talk about segment boundaries.
748 // cf image_for_address from http://svn.digium.com/view/asterisk/team/oej/minivoicemail/dlfcn.c?revision=53912
749 for (; current_image_ >= 0; current_image_--) {
750 const mach_header* hdr = _dyld_get_image_header(current_image_);
751 if (!hdr) continue;
752 if (current_load_cmd_ < 0) // set up for this image
753 current_load_cmd_ = hdr->ncmds; // again, go from the top down
754
755 // We start with the next load command (we've already looked at this one).
756 for (current_load_cmd_--; current_load_cmd_ >= 0; current_load_cmd_--) {
757#ifdef MH_MAGIC_64
758 if (NextExtMachHelper<MH_MAGIC_64, LC_SEGMENT_64,
759 struct mach_header_64, struct segment_command_64>(
760 hdr, current_image_, current_load_cmd_,
761 start, end, flags, offset, inode, filename,
762 file_mapping, file_pages, anon_mapping,
763 anon_pages, dev)) {
764 return true;
765 }
766#endif
767 if (NextExtMachHelper<MH_MAGIC, LC_SEGMENT,
768 struct mach_header, struct segment_command>(
769 hdr, current_image_, current_load_cmd_,
770 start, end, flags, offset, inode, filename,
771 file_mapping, file_pages, anon_mapping,
772 anon_pages, dev)) {
773 return true;
774 }
775 }
776 // If we get here, no more load_cmd's in this image talk about
777 // segments. Go on to the next image.
778 }
779#elif defined(PLATFORM_WINDOWS)
780 static char kDefaultPerms[5] = "r-xp";
781 BOOL ok;
782 if (module_.dwSize == 0) { // only possible before first call
783 module_.dwSize = sizeof(module_);
784 ok = Module32First(snapshot_, &module_);
785 } else {
786 ok = Module32Next(snapshot_, &module_);
787 }
788 if (ok) {
789 uint64 base_addr = reinterpret_cast<DWORD_PTR>(module_.modBaseAddr);
790 if (start) *start = base_addr;
791 if (end) *end = base_addr + module_.modBaseSize;
792 if (flags) *flags = kDefaultPerms;
793 if (offset) *offset = 0;
794 if (inode) *inode = 0;
795 if (filename) *filename = module_.szExePath;
796 if (file_mapping) *file_mapping = 0;
797 if (file_pages) *file_pages = 0;
798 if (anon_mapping) *anon_mapping = 0;
799 if (anon_pages) *anon_pages = 0;
800 if (dev) *dev = 0;
801 return true;
802 }
803#endif
804
805 // We didn't find anything
806 return false;
807}
808
809int ProcMapsIterator::FormatLine(char* buffer, int bufsize,
810 uint64 start, uint64 end, const char *flags,
811 uint64 offset, int64 inode,
812 const char *filename, dev_t dev) {
813 // We assume 'flags' looks like 'rwxp' or 'rwx'.
814 char r = (flags && flags[0] == 'r') ? 'r' : '-';
815 char w = (flags && flags[0] && flags[1] == 'w') ? 'w' : '-';
816 char x = (flags && flags[0] && flags[1] && flags[2] == 'x') ? 'x' : '-';
817 // p always seems set on linux, so we set the default to 'p', not '-'
818 char p = (flags && flags[0] && flags[1] && flags[2] && flags[3] != 'p')
819 ? '-' : 'p';
820
821 const int rc = snprintf(buffer, bufsize,
822 "%08" PRIx64 "-%08" PRIx64 " %c%c%c%c %08" PRIx64 " %02x:%02x %-11" PRId64 " %s\n",
823 start, end, r,w,x,p, offset,
824 static_cast<int>(dev/256), static_cast<int>(dev%256),
825 inode, filename);
826 return (rc < 0 || rc >= bufsize) ? 0 : rc;
827}
828
829namespace tcmalloc {
830
831// Helper to add the list of mapped shared libraries to a profile.
832// Fill formatted "/proc/self/maps" contents into buffer 'buf' of size 'size'
833// and return the actual size occupied in 'buf'. We fill wrote_all to true
834// if we successfully wrote all proc lines to buf, false else.
835// We do not provision for 0-terminating 'buf'.
836int FillProcSelfMaps(char buf[], int size, bool* wrote_all) {
837 ProcMapsIterator::Buffer iterbuf;
838 ProcMapsIterator it(0, &iterbuf); // 0 means "current pid"
839
840 uint64 start, end, offset;
841 int64 inode;
842 char *flags, *filename;
843 int bytes_written = 0;
844 *wrote_all = true;
845 while (it.Next(&start, &end, &flags, &offset, &inode, &filename)) {
846 const int line_length = it.FormatLine(buf + bytes_written,
847 size - bytes_written,
848 start, end, flags, offset,
849 inode, filename, 0);
850 if (line_length == 0)
851 *wrote_all = false; // failed to write this line out
852 else
853 bytes_written += line_length;
854
855 }
856 return bytes_written;
857}
858
859// Dump the same data as FillProcSelfMaps reads to fd.
860// It seems easier to repeat parts of FillProcSelfMaps here than to
861// reuse it via a call.
862void DumpProcSelfMaps(RawFD fd) {
863 ProcMapsIterator::Buffer iterbuf;
864 ProcMapsIterator it(0, &iterbuf); // 0 means "current pid"
865
866 uint64 start, end, offset;
867 int64 inode;
868 char *flags, *filename;
869 ProcMapsIterator::Buffer linebuf;
870 while (it.Next(&start, &end, &flags, &offset, &inode, &filename)) {
871 int written = it.FormatLine(linebuf.buf_, sizeof(linebuf.buf_),
872 start, end, flags, offset, inode, filename,
873 0);
874 RawWrite(fd, linebuf.buf_, written);
875 }
876}
877
878} // namespace tcmalloc