1 ==============================================
2 page owner: Tracking about who allocated each
3 ==============================================
4
5 Introduction
6 ============
7
8 page owner is for the tracking about who alloc
9 It can be used to debug memory leak or to find
10 When allocation happens, information about all
11 and order of pages is stored into certain stor
12 When we need to know about status of all pages
13 this information.
14
15 Although we already have tracepoint for tracin
16 using it for analyzing who allocate each page
17 to enlarge the trace buffer for preventing ove
18 program launched. And, launched program contin
19 buffer for later analysis and it would change
20 possibility rather than just keeping it in mem
21
22 page owner can also be used for various purpos
23 fragmentation statistics can be obtained throu
24 each page. It is already implemented and activ
25 enabled. Other usages are more than welcome.
26
27 It can also be used to show all the stacks and
28 allocated base pages, which gives us a quick o
29 is going without the need to screen through al
30 allocation and free operation.
31
32 page owner is disabled by default. So, if you'
33 to add "page_owner=on" to your boot cmdline. I
34 with page owner and page owner is disabled in
35 boot option, runtime overhead is marginal. If
36 doesn't require memory to store owner informat
37 memory overhead. And, page owner inserts just
38 the page allocator hotpath and if not enabled,
39 like as the kernel without page owner. These t
40 not affect to allocation performance, especial
41 label patching functionality is available. Fol
42 size change due to this facility.
43
44 Although enabling page owner increases kernel
45 most of this code is outside page allocator an
46 the kernel with page owner and turning it on i
47 option to debug kernel memory problem.
48
49 There is one notice that is caused by implemen
50 stores information into the memory from struct
51 is initialized some time later than that page
52 memory system, so, until initialization, many
53 they would have no owner information. To fix i
54 pages are investigated and marked as allocated
55 Although it doesn't mean that they have the ri
56 at least, we can tell whether the page is allo
57 more accurately. On 2GB memory x86-64 VM box,
58 are caught and marked, although they are mostl
59 page extension feature. Anyway, after that, no
60 un-tracking state.
61
62 Usage
63 =====
64
65 1) Build user-space helper::
66
67 cd tools/mm
68 make page_owner_sort
69
70 2) Enable page owner: add "page_owner=on" to b
71
72 3) Do the job that you want to debug.
73
74 4) Analyze information from page owner::
75
76 cat /sys/kernel/debug/page_owner_stack
77 cat stacks.txt
78 post_alloc_hook+0x177/0x1a0
79 get_page_from_freelist+0xd01/0xd80
80 __alloc_pages+0x39e/0x7e0
81 allocate_slab+0xbc/0x3f0
82 ___slab_alloc+0x528/0x8a0
83 kmem_cache_alloc+0x224/0x3b0
84 sk_prot_alloc+0x58/0x1a0
85 sk_alloc+0x32/0x4f0
86 inet_create+0x427/0xb50
87 __sock_create+0x2e4/0x650
88 inet_ctl_sock_create+0x30/0x180
89 igmp_net_init+0xc1/0x130
90 ops_init+0x167/0x410
91 setup_net+0x304/0xa60
92 copy_net_ns+0x29b/0x4a0
93 create_new_namespaces+0x4a1/0x820
94 nr_base_pages: 16
95 ...
96 ...
97 echo 7000 > /sys/kernel/debug/page_own
98 cat /sys/kernel/debug/page_owner_stack
99 cat stacks_7000.txt
100 post_alloc_hook+0x177/0x1a0
101 get_page_from_freelist+0xd01/0xd80
102 __alloc_pages+0x39e/0x7e0
103 alloc_pages_mpol+0x22e/0x490
104 folio_alloc+0xd5/0x110
105 filemap_alloc_folio+0x78/0x230
106 page_cache_ra_order+0x287/0x6f0
107 filemap_get_pages+0x517/0x1160
108 filemap_read+0x304/0x9f0
109 xfs_file_buffered_read+0xe6/0x1d0 [xf
110 xfs_file_read_iter+0x1f0/0x380 [xfs]
111 __kernel_read+0x3b9/0x730
112 kernel_read_file+0x309/0x4d0
113 __do_sys_finit_module+0x381/0x730
114 do_syscall_64+0x8d/0x150
115 entry_SYSCALL_64_after_hwframe+0x62/0
116 nr_base_pages: 20824
117 ...
118
119 cat /sys/kernel/debug/page_owner > pag
120 ./page_owner_sort page_owner_full.txt
121
122 The general output of ``page_owner_full.txt
123
124 Page allocated via order XXX, ...
125 PFN XXX ...
126 // Detailed stack
127
128 Page allocated via order XXX, ...
129 PFN XXX ...
130 // Detailed stack
131 By default, it will do full pfn dump, to s
132 page_owner supports fseek.
133
134 FILE *fp = fopen("/sys/kernel/debug/page_o
135 fseek(fp, pfn_start, SEEK_SET);
136
137 The ``page_owner_sort`` tool ignores ``PFN`
138 in buf, uses regexp to extract the page ord
139 and pages of buf, and finally sorts them ac
140
141 See the result about who allocated each pag
142 in the ``sorted_page_owner.txt``. General o
143
144 XXX times, XXX pages:
145 Page allocated via order XXX, ...
146 // Detailed stack
147
148 By default, ``page_owner_sort`` is sorted a
149 If you want to sort by the page nums of buf
150 The detailed parameters are:
151
152 fundamental function::
153
154 Sort:
155 -a Sort by memory
156 -m Sort by total
157 -p Sort by pid.
158 -P Sort by tgid.
159 -n Sort by task c
160 -r Sort by memory
161 -s Sort by stack
162 -t Sort by times
163 --sort <order> Specify sortin
164 Choose a key f
165 optional since
166 order. Mixed u
167
168 Examples:
169 ./page_owner_s
170 ./page_owner_s
171
172 additional function::
173
174 Cull:
175 --cull <rules>
176 Specify cullin
177 multi-letter k
178
179 <rules> is a single argument i
180 which offers a way to specify
181 keywords are described in the
182 <rules> can be specified by th
183 the STANDARD SORT KEYS section
184 complete-form of keys is allow
185
186 Examples:
187 ./page_owner_s
188 ./page_owner_s
189 ./page_owner_s
190
191 Filter:
192 -f Filter out the
193
194 Select:
195 --pid <pidlist> Select
196 number
197 --tgid <tgidlist> Select
198 group
199 --name <cmdlist> Select
200 task c
201
202 <pidlist>, <tgidlist>, <cmdlis
203 which offers a way to specify
204
205
206 Examples:
207 ./page_owner_s
208 ./page_owner_s
209 ./page_owner_s
210
211 STANDARD FORMAT SPECIFIERS
212 ==========================
213 ::
214
215 For --sort option:
216
217 KEY LONG DESCRI
218 p pid proces
219 tg tgid thread
220 n name task c
221 st stacktrace stack
222 T txt full t
223 ft free_ts timest
224 at alloc_ts timest
225 ator allocator memory
226
227 For --cull option:
228
229 KEY LONG DESCRI
230 p pid proces
231 tg tgid thread
232 n name task c
233 f free whethe
234 st stacktrace stack
235 ator allocator memory
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.