1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
|
/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
#include "mozilla/SystemMemoryReporter.h"
#include "mozilla/Attributes.h"
#include "mozilla/LinuxUtils.h"
#include "mozilla/PodOperations.h"
#include "mozilla/Preferences.h"
#include "mozilla/TaggedAnonymousMemory.h"
#include "mozilla/Unused.h"
#include "nsDataHashtable.h"
#include "nsIMemoryReporter.h"
#include "nsPrintfCString.h"
#include "nsString.h"
#include "nsTHashtable.h"
#include "nsHashKeys.h"
#include <dirent.h>
#include <inttypes.h>
#include <stdio.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>
#include <errno.h>
// This file implements a Linux-specific, system-wide memory reporter. It
// gathers all the useful memory measurements obtainable from the OS in a
// single place, giving a high-level view of memory consumption for the entire
// machine/device.
//
// Other memory reporters measure part of a single process's memory consumption.
// This reporter is different in that it measures memory consumption of many
// processes, and they end up in a single reports tree. This is a slight abuse
// of the memory reporting infrastructure, and therefore the results are given
// their own "process" called "System", which means they show up in about:memory
// in their own section, distinct from the per-process sections.
namespace mozilla {
namespace SystemMemoryReporter {
#if !defined(XP_LINUX)
#error "This won't work if we're not on Linux."
#endif
/**
* RAII helper that will close an open DIR handle.
*/
struct MOZ_STACK_CLASS AutoDir
{
explicit AutoDir(DIR* aDir) : mDir(aDir) {}
~AutoDir() { if (mDir) closedir(mDir); };
DIR* mDir;
};
/**
* RAII helper that will close an open FILE handle.
*/
struct MOZ_STACK_CLASS AutoFile
{
explicit AutoFile(FILE* aFile) : mFile(aFile) {}
~AutoFile() { if (mFile) fclose(mFile); }
FILE* mFile;
};
static bool
EndsWithLiteral(const nsCString& aHaystack, const char* aNeedle)
{
int32_t idx = aHaystack.RFind(aNeedle);
return idx != -1 && idx + strlen(aNeedle) == aHaystack.Length();
}
static void
GetDirname(const nsCString& aPath, nsACString& aOut)
{
int32_t idx = aPath.RFind("/");
if (idx == -1) {
aOut.Truncate();
} else {
aOut.Assign(Substring(aPath, 0, idx));
}
}
static void
GetBasename(const nsCString& aPath, nsACString& aOut)
{
nsCString out;
int32_t idx = aPath.RFind("/");
if (idx == -1) {
out.Assign(aPath);
} else {
out.Assign(Substring(aPath, idx + 1));
}
// On Android, some entries in /dev/ashmem end with "(deleted)" (e.g.
// "/dev/ashmem/libxul.so(deleted)"). We don't care about this modifier, so
// cut it off when getting the entry's basename.
if (EndsWithLiteral(out, "(deleted)")) {
out.Assign(Substring(out, 0, out.RFind("(deleted)")));
}
out.StripChars(" ");
aOut.Assign(out);
}
static bool
IsNumeric(const char* aStr)
{
MOZ_ASSERT(*aStr); // shouldn't see empty strings
while (*aStr) {
if (!isdigit(*aStr)) {
return false;
}
++aStr;
}
return true;
}
static bool
IsAnonymous(const nsACString& aName)
{
// Recent kernels have multiple [stack:nnnn] entries, where |nnnn| is a
// thread ID. However, the entire virtual memory area containing a thread's
// stack pointer is considered the stack for that thread, even if it was
// merged with an adjacent area containing non-stack data. So we treat them
// as regular anonymous memory. However, see below about tagged anonymous
// memory.
return aName.IsEmpty() ||
StringBeginsWith(aName, NS_LITERAL_CSTRING("[stack:"));
}
class SystemReporter final : public nsIMemoryReporter
{
~SystemReporter() {}
public:
NS_DECL_THREADSAFE_ISUPPORTS
#define REPORT(_path, _units, _amount, _desc) \
do { \
size_t __amount = _amount; /* evaluate _amount only once */ \
if (__amount > 0) { \
aHandleReport->Callback(NS_LITERAL_CSTRING("System"), _path, \
KIND_OTHER, _units, __amount, _desc, aData); \
} \
} while (0)
NS_IMETHOD CollectReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData, bool aAnonymize) override
{
// There is lots of privacy-sensitive data in /proc. Just skip this
// reporter entirely when anonymization is required.
if (aAnonymize) {
return NS_OK;
}
if (!Preferences::GetBool("memory.system_memory_reporter")) {
return NS_OK;
}
// Read relevant fields from /proc/meminfo.
int64_t memTotal = 0, memFree = 0;
nsresult rv1 = ReadMemInfo(&memTotal, &memFree);
// Collect per-process reports from /proc/<pid>/smaps.
int64_t totalPss = 0;
nsresult rv2 = CollectProcessReports(aHandleReport, aData, &totalPss);
// Report the non-process numbers.
if (NS_SUCCEEDED(rv1) && NS_SUCCEEDED(rv2)) {
int64_t other = memTotal - memFree - totalPss;
REPORT(NS_LITERAL_CSTRING("mem/other"), UNITS_BYTES, other,
NS_LITERAL_CSTRING(
"Memory which is neither owned by any user-space process nor free. Note that "
"this includes memory holding cached files from the disk which can be "
"reclaimed by the OS at any time."));
REPORT(NS_LITERAL_CSTRING("mem/free"), UNITS_BYTES, memFree,
NS_LITERAL_CSTRING(
"Memory which is free and not being used for any purpose."));
}
// Report reserved memory not included in memTotal.
CollectPmemReports(aHandleReport, aData);
// Report zram usage statistics.
CollectZramReports(aHandleReport, aData);
// Report kgsl graphics memory usage.
CollectKgslReports(aHandleReport, aData);
// Report ION memory usage.
CollectIonReports(aHandleReport, aData);
return NS_OK;
}
private:
// These are the cross-cutting measurements across all processes.
class ProcessSizes
{
public:
void Add(const nsACString& aKey, size_t aSize)
{
mTagged.Put(aKey, mTagged.Get(aKey) + aSize);
}
void Report(nsIHandleReportCallback* aHandleReport, nsISupports* aData)
{
for (auto iter = mTagged.Iter(); !iter.Done(); iter.Next()) {
nsCStringHashKey::KeyType key = iter.Key();
size_t amount = iter.UserData();
nsAutoCString path("processes/");
path.Append(key);
nsAutoCString desc("This is the sum of all processes' '");
desc.Append(key);
desc.AppendLiteral("' numbers.");
REPORT(path, UNITS_BYTES, amount, desc);
}
}
private:
nsDataHashtable<nsCStringHashKey, size_t> mTagged;
};
nsresult ReadMemInfo(int64_t* aMemTotal, int64_t* aMemFree)
{
FILE* f = fopen("/proc/meminfo", "r");
if (!f) {
return NS_ERROR_FAILURE;
}
int n1 = fscanf(f, "MemTotal: %" SCNd64 " kB\n", aMemTotal);
int n2 = fscanf(f, "MemFree: %" SCNd64 " kB\n", aMemFree);
fclose(f);
if (n1 != 1 || n2 != 1) {
return NS_ERROR_FAILURE;
}
// Convert from KB to B.
*aMemTotal *= 1024;
*aMemFree *= 1024;
return NS_OK;
}
nsresult CollectProcessReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData,
int64_t* aTotalPss)
{
*aTotalPss = 0;
ProcessSizes processSizes;
DIR* d = opendir("/proc");
if (NS_WARN_IF(!d)) {
return NS_ERROR_FAILURE;
}
struct dirent* ent;
while ((ent = readdir(d))) {
struct stat statbuf;
const char* pidStr = ent->d_name;
// Don't check the return value of stat() -- it can return -1 for these
// directories even when it has succeeded, apparently.
stat(pidStr, &statbuf);
if (S_ISDIR(statbuf.st_mode) && IsNumeric(pidStr)) {
nsCString processName("process(");
// Get the command name from cmdline. If that fails, the pid is still
// shown.
nsPrintfCString cmdlinePath("/proc/%s/cmdline", pidStr);
FILE* f = fopen(cmdlinePath.get(), "r");
if (f) {
static const size_t len = 256;
char buf[len];
if (fgets(buf, len, f)) {
processName.Append(buf);
// A hack: replace forward slashes with '\\' so they aren't treated
// as path separators. Consumers of this reporter (such as
// about:memory) have to undo this change.
processName.ReplaceChar('/', '\\');
processName.AppendLiteral(", ");
}
fclose(f);
}
processName.AppendLiteral("pid=");
processName.Append(pidStr);
processName.Append(')');
// Read the PSS values from the smaps file.
nsPrintfCString smapsPath("/proc/%s/smaps", pidStr);
f = fopen(smapsPath.get(), "r");
if (!f) {
// Processes can terminate between the readdir() call above and now,
// so just skip if we can't open the file.
continue;
}
ParseMappings(f, processName, aHandleReport, aData, &processSizes,
aTotalPss);
fclose(f);
// Report the open file descriptors for this process.
nsPrintfCString procFdPath("/proc/%s/fd", pidStr);
CollectOpenFileReports(aHandleReport, aData, procFdPath, processName);
}
}
closedir(d);
// Report the "processes/" tree.
processSizes.Report(aHandleReport, aData);
return NS_OK;
}
void ParseMappings(FILE* aFile,
const nsACString& aProcessName,
nsIHandleReportCallback* aHandleReport,
nsISupports* aData,
ProcessSizes* aProcessSizes,
int64_t* aTotalPss)
{
// The first line of an entry in /proc/<pid>/smaps looks just like an entry
// in /proc/<pid>/maps:
//
// address perms offset dev inode pathname
// 02366000-025d8000 rw-p 00000000 00:00 0 [heap]
//
// Each of the following lines contains a key and a value, separated
// by ": ", where the key does not contain either of those characters.
// Assuming more than this about the structure of those lines has
// failed to be future-proof in the past, so we avoid doing so.
//
// This makes it difficult to detect the start of a new entry
// until it's been removed from the stdio buffer, so we just loop
// over all lines in the file in this routine.
const int argCount = 8;
unsigned long long addrStart, addrEnd;
char perms[5];
unsigned long long offset;
// The 2.6 and 3.0 kernels allocate 12 bits for the major device number and
// 20 bits for the minor device number. Future kernels might allocate more.
// 64 bits ought to be enough for anybody.
char devMajor[17];
char devMinor[17];
unsigned int inode;
char line[1025];
// This variable holds the path of the current entry, or is void
// if we're scanning for the start of a new entry.
nsAutoCString currentPath;
int pathOffset;
currentPath.SetIsVoid(true);
while (fgets(line, sizeof(line), aFile)) {
if (currentPath.IsVoid()) {
int n = sscanf(line,
"%llx-%llx %4s %llx "
"%16[0-9a-fA-F]:%16[0-9a-fA-F] %u %n",
&addrStart, &addrEnd, perms, &offset, devMajor,
devMinor, &inode, &pathOffset);
if (n >= argCount - 1) {
currentPath.Assign(line + pathOffset);
currentPath.StripChars("\n");
}
continue;
}
// Now that we have a name and other metadata, scan for the PSS.
size_t pss_kb;
int n = sscanf(line, "Pss: %zu", &pss_kb);
if (n < 1) {
continue;
}
size_t pss = pss_kb * 1024;
if (pss > 0) {
nsAutoCString name, description, tag;
GetReporterNameAndDescription(currentPath.get(), perms, name, description, tag);
nsAutoCString processMemPath("mem/processes/");
processMemPath.Append(aProcessName);
processMemPath.Append('/');
processMemPath.Append(name);
REPORT(processMemPath, UNITS_BYTES, pss, description);
// Increment the appropriate aProcessSizes values, and the total.
aProcessSizes->Add(tag, pss);
*aTotalPss += pss;
}
// Now that we've seen the PSS, we're done with this entry.
currentPath.SetIsVoid(true);
}
}
void GetReporterNameAndDescription(const char* aPath,
const char* aPerms,
nsACString& aName,
nsACString& aDesc,
nsACString& aTag)
{
aName.Truncate();
aDesc.Truncate();
aTag.Truncate();
// If aPath points to a file, we have its absolute path; it might
// also be a bracketed pseudo-name (see below). In either case
// there is also some whitespace to trim.
nsAutoCString absPath;
absPath.Append(aPath);
absPath.StripChars(" ");
if (absPath.EqualsLiteral("[heap]")) {
aName.AppendLiteral("anonymous/brk-heap");
aDesc.AppendLiteral(
"Memory in anonymous mappings within the boundaries defined by "
"brk() / sbrk(). This is likely to be just a portion of the "
"application's heap; the remainder lives in other anonymous mappings. "
"This corresponds to '[heap]' in /proc/<pid>/smaps.");
aTag = aName;
} else if (absPath.EqualsLiteral("[stack]")) {
aName.AppendLiteral("stack/main-thread");
aDesc.AppendPrintf(
"The stack size of the process's main thread. This corresponds to "
"'[stack]' in /proc/<pid>/smaps.");
aTag = aName;
} else if (MozTaggedMemoryIsSupported() &&
StringBeginsWith(absPath, NS_LITERAL_CSTRING("[stack:"))) {
// If tagged memory is supported, we can be reasonably sure that
// the virtual memory area containing the stack hasn't been
// merged with unrelated heap memory. (This prevents the
// "[stack:" entries from reaching the IsAnonymous case below.)
pid_t tid = atoi(absPath.get() + 7);
nsAutoCString threadName, escapedThreadName;
LinuxUtils::GetThreadName(tid, threadName);
if (threadName.IsEmpty()) {
threadName.AssignLiteral("<unknown>");
}
escapedThreadName.Assign(threadName);
escapedThreadName.StripChars("()");
escapedThreadName.ReplaceChar('/', '\\');
aName.AppendLiteral("stack/non-main-thread");
aName.AppendLiteral("/name(");
aName.Append(escapedThreadName);
aName.Append(')');
aTag = aName;
aName.AppendPrintf("/thread(%d)", tid);
aDesc.AppendPrintf("The stack size of a non-main thread named '%s' with "
"thread ID %d. This corresponds to '[stack:%d]' "
"in /proc/%d/smaps.", threadName.get(), tid, tid);
} else if (absPath.EqualsLiteral("[vdso]")) {
aName.AppendLiteral("vdso");
aDesc.AppendLiteral(
"The virtual dynamically-linked shared object, also known as the "
"'vsyscall page'. This is a memory region mapped by the operating "
"system for the purpose of allowing processes to perform some "
"privileged actions without the overhead of a syscall.");
aTag = aName;
} else if (StringBeginsWith(absPath, NS_LITERAL_CSTRING("[anon:")) &&
EndsWithLiteral(absPath, "]")) {
// It's tagged memory; see also "mfbt/TaggedAnonymousMemory.h".
nsAutoCString tag(Substring(absPath, 6, absPath.Length() - 7));
aName.AppendLiteral("anonymous/");
aName.Append(tag);
aTag = aName;
aDesc.AppendLiteral("Memory in anonymous mappings tagged with '");
aDesc.Append(tag);
aDesc.Append('\'');
} else if (!IsAnonymous(absPath)) {
// We now know it's an actual file. Truncate this to its
// basename, and put the absolute path in the description.
nsAutoCString basename, dirname;
GetBasename(absPath, basename);
GetDirname(absPath, dirname);
// Hack: A file is a shared library if the basename contains ".so" and
// its dirname contains "/lib", or if the basename ends with ".so".
if (EndsWithLiteral(basename, ".so") ||
(basename.Find(".so") != -1 && dirname.Find("/lib") != -1)) {
aName.AppendLiteral("shared-libraries/");
aTag = aName;
if (strncmp(aPerms, "r-x", 3) == 0) {
aTag.AppendLiteral("read-executable");
} else if (strncmp(aPerms, "rw-", 3) == 0) {
aTag.AppendLiteral("read-write");
} else if (strncmp(aPerms, "r--", 3) == 0) {
aTag.AppendLiteral("read-only");
} else {
aTag.AppendLiteral("other");
}
} else {
aName.AppendLiteral("other-files");
if (EndsWithLiteral(basename, ".xpi")) {
aName.AppendLiteral("/extensions");
} else if (dirname.Find("/fontconfig") != -1) {
aName.AppendLiteral("/fontconfig");
} else {
aName.AppendLiteral("/misc");
}
aTag = aName;
aName.Append('/');
}
aName.Append(basename);
aDesc.Append(absPath);
} else {
if (MozTaggedMemoryIsSupported()) {
aName.AppendLiteral("anonymous/untagged");
aDesc.AppendLiteral("Memory in untagged anonymous mappings.");
aTag = aName;
} else {
aName.AppendLiteral("anonymous/outside-brk");
aDesc.AppendLiteral("Memory in anonymous mappings outside the "
"boundaries defined by brk() / sbrk().");
aTag = aName;
}
}
aName.AppendLiteral("/[");
aName.Append(aPerms);
aName.Append(']');
// Append the permissions. This is useful for non-verbose mode in
// about:memory when the filename is long and goes of the right side of the
// window.
aDesc.AppendLiteral(" [");
aDesc.Append(aPerms);
aDesc.Append(']');
}
void CollectPmemReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData)
{
// The pmem subsystem allocates physically contiguous memory for
// interfacing with hardware. In order to ensure availability,
// this memory is reserved during boot, and allocations are made
// within these regions at runtime.
//
// There are typically several of these pools allocated at boot.
// The /sys/kernel/pmem_regions directory contains a subdirectory
// for each one. Within each subdirectory, the files we care
// about are "size" (the total amount of physical memory) and
// "mapped_regions" (a list of the current allocations within that
// area).
DIR* d = opendir("/sys/kernel/pmem_regions");
if (!d) {
return;
}
struct dirent* ent;
while ((ent = readdir(d))) {
const char* name = ent->d_name;
uint64_t size;
int scanned;
// Skip "." and ".." (and any other dotfiles).
if (name[0] == '.') {
continue;
}
// Read the total size. The file gives the size in decimal and
// hex, in the form "13631488(0xd00000)"; we parse the former.
nsPrintfCString sizePath("/sys/kernel/pmem_regions/%s/size", name);
FILE* sizeFile = fopen(sizePath.get(), "r");
if (NS_WARN_IF(!sizeFile)) {
continue;
}
scanned = fscanf(sizeFile, "%" SCNu64, &size);
fclose(sizeFile);
if (NS_WARN_IF(scanned != 1)) {
continue;
}
// Read mapped regions; format described below.
uint64_t freeSize = size;
nsPrintfCString regionsPath("/sys/kernel/pmem_regions/%s/mapped_regions",
name);
FILE* regionsFile = fopen(regionsPath.get(), "r");
if (regionsFile) {
static const size_t bufLen = 4096;
char buf[bufLen];
while (fgets(buf, bufLen, regionsFile)) {
int pid;
// Skip header line.
if (strncmp(buf, "pid #", 5) == 0) {
continue;
}
// Line format: "pid N:" + zero or more "(Start,Len) ".
// N is decimal; Start and Len are in hex.
scanned = sscanf(buf, "pid %d", &pid);
if (NS_WARN_IF(scanned != 1)) {
continue;
}
for (const char* nextParen = strchr(buf, '(');
nextParen != nullptr;
nextParen = strchr(nextParen + 1, '(')) {
uint64_t mapStart, mapLen;
scanned = sscanf(nextParen + 1, "%" SCNx64 ",%" SCNx64,
&mapStart, &mapLen);
if (NS_WARN_IF(scanned != 2)) {
break;
}
nsPrintfCString path("mem/pmem/used/%s/segment(pid=%d, "
"offset=0x%" PRIx64 ")", name, pid, mapStart);
nsPrintfCString desc("Physical memory reserved for the \"%s\" pool "
"and allocated to a buffer.", name);
REPORT(path, UNITS_BYTES, mapLen, desc);
freeSize -= mapLen;
}
}
fclose(regionsFile);
}
nsPrintfCString path("mem/pmem/free/%s", name);
nsPrintfCString desc("Physical memory reserved for the \"%s\" pool and "
"unavailable to the rest of the system, but not "
"currently allocated.", name);
REPORT(path, UNITS_BYTES, freeSize, desc);
}
closedir(d);
}
void
CollectIonReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData)
{
// ION is a replacement for PMEM (and other similar allocators).
//
// More details from http://lwn.net/Articles/480055/
// "Like its PMEM-like predecessors, ION manages one or more memory pools,
// some of which are set aside at boot time to combat fragmentation or to
// serve special hardware needs. GPUs, display controllers, and cameras
// are some of the hardware blocks that may have special memory
// requirements."
//
// The file format starts as follows:
// client pid size
// ----------------------------------------------------
// adsprpc-smd 1 4096
// fd900000.qcom,mdss_mdp 1 1658880
// ----------------------------------------------------
// orphaned allocations (info is from last known client):
// Homescreen 24100 294912 0 1
// b2g 23987 1658880 0 1
// mdss_fb0 401 1658880 0 1
// b2g 23987 4096 0 1
// Built-in Keyboa 24205 61440 0 1
// ----------------------------------------------------
// <other stuff>
//
// For our purposes we only care about the first portion of the file noted
// above which contains memory alloations (both sections). The term
// "orphaned" is misleading, it appears that every allocation not by the
// first process is considered orphaned on FxOS devices.
// The first three fields of each entry interest us:
// 1) client - Essentially the process name. We limit client names to 63
// characters, in theory they should never be greater than 15
// due to thread name length limitations.
// 2) pid - The ID of the allocating process, read as a uint32_t.
// 3) size - The size of the allocation in bytes, read as as a uint64_t.
const char* const kFormatString = "%63s %" SCNu32 " %" SCNu64;
const int kNumFields = 3;
const size_t kStringSize = 64;
const char* const kIonIommuPath = "/sys/kernel/debug/ion/iommu";
FILE* iommu = fopen(kIonIommuPath, "r");
if (!iommu) {
return;
}
AutoFile iommuGuard(iommu);
const size_t kBufferLen = 256;
char buffer[kBufferLen];
char client[kStringSize];
uint32_t pid;
uint64_t size;
// Ignore the header line.
Unused << fgets(buffer, kBufferLen, iommu);
// Ignore the separator line.
Unused << fgets(buffer, kBufferLen, iommu);
const char* const kSep = "----";
const size_t kSepLen = 4;
// Read non-orphaned entries.
while (fgets(buffer, kBufferLen, iommu) &&
strncmp(kSep, buffer, kSepLen) != 0) {
if (sscanf(buffer, kFormatString, client, &pid, &size) == kNumFields) {
nsPrintfCString entryPath("ion-memory/%s (pid=%d)", client, pid);
REPORT(entryPath, UNITS_BYTES, size,
NS_LITERAL_CSTRING("An ION kernel memory allocation."));
}
}
// Ignore the orphaned header.
Unused << fgets(buffer, kBufferLen, iommu);
// Read orphaned entries.
while (fgets(buffer, kBufferLen, iommu) &&
strncmp(kSep, buffer, kSepLen) != 0) {
if (sscanf(buffer, kFormatString, client, &pid, &size) == kNumFields) {
nsPrintfCString entryPath("ion-memory/%s (pid=%d)", client, pid);
REPORT(entryPath, UNITS_BYTES, size,
NS_LITERAL_CSTRING("An ION kernel memory allocation."));
}
}
// Ignore the rest of the file.
}
uint64_t
ReadSizeFromFile(const char* aFilename)
{
FILE* sizeFile = fopen(aFilename, "r");
if (NS_WARN_IF(!sizeFile)) {
return 0;
}
uint64_t size = 0;
Unused << fscanf(sizeFile, "%" SCNu64, &size);
fclose(sizeFile);
return size;
}
void
CollectZramReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData)
{
// zram usage stats files can be found under:
// /sys/block/zram<id>
// |--> disksize - Maximum amount of uncompressed data that can be
// stored on the disk (bytes)
// |--> orig_data_size - Uncompressed size of data in the disk (bytes)
// |--> compr_data_size - Compressed size of the data in the disk (bytes)
// |--> num_reads - Number of attempted reads to the disk (count)
// |--> num_writes - Number of attempted writes to the disk (count)
//
// Each file contains a single integer value in decimal form.
DIR* d = opendir("/sys/block");
if (!d) {
return;
}
struct dirent* ent;
while ((ent = readdir(d))) {
const char* name = ent->d_name;
// Skip non-zram entries.
if (strncmp("zram", name, 4) != 0) {
continue;
}
// Report disk size statistics.
nsPrintfCString diskSizeFile("/sys/block/%s/disksize", name);
nsPrintfCString origSizeFile("/sys/block/%s/orig_data_size", name);
uint64_t diskSize = ReadSizeFromFile(diskSizeFile.get());
uint64_t origSize = ReadSizeFromFile(origSizeFile.get());
uint64_t unusedSize = diskSize - origSize;
nsPrintfCString diskUsedPath("zram-disksize/%s/used", name);
nsPrintfCString diskUsedDesc(
"The uncompressed size of data stored in \"%s.\" "
"This excludes zero-filled pages since "
"no memory is allocated for them.", name);
REPORT(diskUsedPath, UNITS_BYTES, origSize, diskUsedDesc);
nsPrintfCString diskUnusedPath("zram-disksize/%s/unused", name);
nsPrintfCString diskUnusedDesc(
"The amount of uncompressed data that can still be "
"be stored in \"%s\"", name);
REPORT(diskUnusedPath, UNITS_BYTES, unusedSize, diskUnusedDesc);
// Report disk accesses.
nsPrintfCString readsFile("/sys/block/%s/num_reads", name);
nsPrintfCString writesFile("/sys/block/%s/num_writes", name);
uint64_t reads = ReadSizeFromFile(readsFile.get());
uint64_t writes = ReadSizeFromFile(writesFile.get());
nsPrintfCString readsDesc(
"The number of reads (failed or successful) done on "
"\"%s\"", name);
nsPrintfCString readsPath("zram-accesses/%s/reads", name);
REPORT(readsPath, UNITS_COUNT_CUMULATIVE, reads, readsDesc);
nsPrintfCString writesDesc(
"The number of writes (failed or successful) done "
"on \"%s\"", name);
nsPrintfCString writesPath("zram-accesses/%s/writes", name);
REPORT(writesPath, UNITS_COUNT_CUMULATIVE, writes, writesDesc);
// Report compressed data size.
nsPrintfCString comprSizeFile("/sys/block/%s/compr_data_size", name);
uint64_t comprSize = ReadSizeFromFile(comprSizeFile.get());
nsPrintfCString comprSizeDesc(
"The compressed size of data stored in \"%s\"",
name);
nsPrintfCString comprSizePath("zram-compr-data-size/%s", name);
REPORT(comprSizePath, UNITS_BYTES, comprSize, comprSizeDesc);
}
closedir(d);
}
void
CollectOpenFileReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData,
const nsACString& aProcPath,
const nsACString& aProcessName)
{
// All file descriptors opened by a process are listed under
// /proc/<pid>/fd/<numerical_fd>. Each entry is a symlink that points to the
// path that was opened. This can be an actual file, a socket, a pipe, an
// anon_inode, or possibly an uncategorized device.
const char kFilePrefix[] = "/";
const char kSocketPrefix[] = "socket:";
const char kPipePrefix[] = "pipe:";
const char kAnonInodePrefix[] = "anon_inode:";
const nsCString procPath(aProcPath);
DIR* d = opendir(procPath.get());
if (!d) {
return;
}
char linkPath[PATH_MAX + 1];
struct dirent* ent;
while ((ent = readdir(d))) {
const char* fd = ent->d_name;
// Skip "." and ".." (and any other dotfiles).
if (fd[0] == '.') {
continue;
}
nsPrintfCString fullPath("%s/%s", procPath.get(), fd);
ssize_t linkPathSize = readlink(fullPath.get(), linkPath, PATH_MAX);
if (linkPathSize > 0) {
linkPath[linkPathSize] = '\0';
#define CHECK_PREFIX(prefix) \
(strncmp(linkPath, prefix, sizeof(prefix) - 1) == 0)
const char* category = nullptr;
const char* descriptionPrefix = nullptr;
if (CHECK_PREFIX(kFilePrefix)) {
category = "files"; // No trailing slash, the file path will have one
descriptionPrefix = "An open";
} else if (CHECK_PREFIX(kSocketPrefix)) {
category = "sockets/";
descriptionPrefix = "A socket";
} else if (CHECK_PREFIX(kPipePrefix)) {
category = "pipes/";
descriptionPrefix = "A pipe";
} else if (CHECK_PREFIX(kAnonInodePrefix)) {
category = "anon_inodes/";
descriptionPrefix = "An anon_inode";
} else {
category = "";
descriptionPrefix = "An uncategorized";
}
#undef CHECK_PREFIX
const nsCString processName(aProcessName);
nsPrintfCString entryPath("open-fds/%s/%s%s/%s",
processName.get(), category, linkPath, fd);
nsPrintfCString entryDescription("%s file descriptor opened by the process",
descriptionPrefix);
REPORT(entryPath, UNITS_COUNT, 1, entryDescription);
}
}
closedir(d);
}
void
CollectKgslReports(nsIHandleReportCallback* aHandleReport,
nsISupports* aData)
{
// Each process that uses kgsl memory will have an entry under
// /sys/kernel/debug/kgsl/proc/<pid>/mem. This file format includes a
// header and then entries with types as follows:
// gpuaddr useraddr size id flags type usage sglen
// hexaddr hexaddr int int str str str int
// We care primarily about the usage and size.
// For simplicity numbers will be uint64_t, strings 63 chars max.
const char* const kScanFormat =
"%" SCNx64 " %" SCNx64 " %" SCNu64 " %" SCNu64
" %63s %63s %63s %" SCNu64;
const int kNumFields = 8;
const size_t kStringSize = 64;
DIR* d = opendir("/sys/kernel/debug/kgsl/proc/");
if (!d) {
return;
}
AutoDir dirGuard(d);
struct dirent* ent;
while ((ent = readdir(d))) {
const char* pid = ent->d_name;
// Skip "." and ".." (and any other dotfiles).
if (pid[0] == '.') {
continue;
}
nsPrintfCString memPath("/sys/kernel/debug/kgsl/proc/%s/mem", pid);
FILE* memFile = fopen(memPath.get(), "r");
if (NS_WARN_IF(!memFile)) {
continue;
}
AutoFile fileGuard(memFile);
// Attempt to map the pid to a more useful name.
nsAutoCString procName;
LinuxUtils::GetThreadName(atoi(pid), procName);
if (procName.IsEmpty()) {
procName.Append("pid=");
procName.Append(pid);
} else {
procName.Append(" (pid=");
procName.Append(pid);
procName.Append(")");
}
uint64_t gpuaddr, useraddr, size, id, sglen;
char flags[kStringSize];
char type[kStringSize];
char usage[kStringSize];
// Bypass the header line.
char buff[1024];
Unused << fgets(buff, 1024, memFile);
while (fscanf(memFile, kScanFormat, &gpuaddr, &useraddr, &size, &id,
flags, type, usage, &sglen) == kNumFields) {
nsPrintfCString entryPath("kgsl-memory/%s/%s", procName.get(), usage);
REPORT(entryPath, UNITS_BYTES, size,
NS_LITERAL_CSTRING("A kgsl graphics memory allocation."));
}
}
}
};
NS_IMPL_ISUPPORTS(SystemReporter, nsIMemoryReporter)
void
Init()
{
RegisterStrongMemoryReporter(new SystemReporter());
}
} // namespace SystemMemoryReporter
} // namespace mozilla
|