1 ============================================== 1 ==================================================
2 page owner: Tracking about who allocated each 2 page owner: Tracking about who allocated each page
3 ============================================== 3 ==================================================
4 4
5 Introduction 5 Introduction
6 ============ 6 ============
7 7
8 page owner is for the tracking about who alloc 8 page owner is for the tracking about who allocated each page.
9 It can be used to debug memory leak or to find 9 It can be used to debug memory leak or to find a memory hogger.
10 When allocation happens, information about all 10 When allocation happens, information about allocation such as call stack
11 and order of pages is stored into certain stor 11 and order of pages is stored into certain storage for each page.
12 When we need to know about status of all pages 12 When we need to know about status of all pages, we can get and analyze
13 this information. 13 this information.
14 14
15 Although we already have tracepoint for tracin 15 Although we already have tracepoint for tracing page allocation/free,
16 using it for analyzing who allocate each page 16 using it for analyzing who allocate each page is rather complex. We need
17 to enlarge the trace buffer for preventing ove 17 to enlarge the trace buffer for preventing overlapping until userspace
18 program launched. And, launched program contin 18 program launched. And, launched program continually dump out the trace
19 buffer for later analysis and it would change 19 buffer for later analysis and it would change system behaviour with more
20 possibility rather than just keeping it in mem 20 possibility rather than just keeping it in memory, so bad for debugging.
21 21
22 page owner can also be used for various purpos 22 page owner can also be used for various purposes. For example, accurate
23 fragmentation statistics can be obtained throu 23 fragmentation statistics can be obtained through gfp flag information of
24 each page. It is already implemented and activ 24 each page. It is already implemented and activated if page owner is
25 enabled. Other usages are more than welcome. 25 enabled. Other usages are more than welcome.
26 26
27 It can also be used to show all the stacks and 27 It can also be used to show all the stacks and their current number of
28 allocated base pages, which gives us a quick o 28 allocated base pages, which gives us a quick overview of where the memory
29 is going without the need to screen through al 29 is going without the need to screen through all the pages and match the
30 allocation and free operation. 30 allocation and free operation.
31 31
32 page owner is disabled by default. So, if you' 32 page owner is disabled by default. So, if you'd like to use it, you need
33 to add "page_owner=on" to your boot cmdline. I 33 to add "page_owner=on" to your boot cmdline. If the kernel is built
34 with page owner and page owner is disabled in 34 with page owner and page owner is disabled in runtime due to not enabling
35 boot option, runtime overhead is marginal. If 35 boot option, runtime overhead is marginal. If disabled in runtime, it
36 doesn't require memory to store owner informat 36 doesn't require memory to store owner information, so there is no runtime
37 memory overhead. And, page owner inserts just 37 memory overhead. And, page owner inserts just two unlikely branches into
38 the page allocator hotpath and if not enabled, 38 the page allocator hotpath and if not enabled, then allocation is done
39 like as the kernel without page owner. These t 39 like as the kernel without page owner. These two unlikely branches should
40 not affect to allocation performance, especial 40 not affect to allocation performance, especially if the static keys jump
41 label patching functionality is available. Fol 41 label patching functionality is available. Following is the kernel's code
42 size change due to this facility. 42 size change due to this facility.
43 43
44 Although enabling page owner increases kernel 44 Although enabling page owner increases kernel size by several kilobytes,
45 most of this code is outside page allocator an 45 most of this code is outside page allocator and its hot path. Building
46 the kernel with page owner and turning it on i 46 the kernel with page owner and turning it on if needed would be great
47 option to debug kernel memory problem. 47 option to debug kernel memory problem.
48 48
49 There is one notice that is caused by implemen 49 There is one notice that is caused by implementation detail. page owner
50 stores information into the memory from struct 50 stores information into the memory from struct page extension. This memory
51 is initialized some time later than that page 51 is initialized some time later than that page allocator starts in sparse
52 memory system, so, until initialization, many 52 memory system, so, until initialization, many pages can be allocated and
53 they would have no owner information. To fix i 53 they would have no owner information. To fix it up, these early allocated
54 pages are investigated and marked as allocated 54 pages are investigated and marked as allocated in initialization phase.
55 Although it doesn't mean that they have the ri 55 Although it doesn't mean that they have the right owner information,
56 at least, we can tell whether the page is allo 56 at least, we can tell whether the page is allocated or not,
57 more accurately. On 2GB memory x86-64 VM box, 57 more accurately. On 2GB memory x86-64 VM box, 13343 early allocated pages
58 are caught and marked, although they are mostl 58 are caught and marked, although they are mostly allocated from struct
59 page extension feature. Anyway, after that, no 59 page extension feature. Anyway, after that, no page is left in
60 un-tracking state. 60 un-tracking state.
61 61
62 Usage 62 Usage
63 ===== 63 =====
64 64
65 1) Build user-space helper:: 65 1) Build user-space helper::
66 66
67 cd tools/mm 67 cd tools/mm
68 make page_owner_sort 68 make page_owner_sort
69 69
70 2) Enable page owner: add "page_owner=on" to b 70 2) Enable page owner: add "page_owner=on" to boot cmdline.
71 71
72 3) Do the job that you want to debug. 72 3) Do the job that you want to debug.
73 73
74 4) Analyze information from page owner:: 74 4) Analyze information from page owner::
75 75
76 cat /sys/kernel/debug/page_owner_stack 76 cat /sys/kernel/debug/page_owner_stacks/show_stacks > stacks.txt
77 cat stacks.txt 77 cat stacks.txt
78 post_alloc_hook+0x177/0x1a0 78 post_alloc_hook+0x177/0x1a0
79 get_page_from_freelist+0xd01/0xd80 79 get_page_from_freelist+0xd01/0xd80
80 __alloc_pages+0x39e/0x7e0 80 __alloc_pages+0x39e/0x7e0
81 allocate_slab+0xbc/0x3f0 81 allocate_slab+0xbc/0x3f0
82 ___slab_alloc+0x528/0x8a0 82 ___slab_alloc+0x528/0x8a0
83 kmem_cache_alloc+0x224/0x3b0 83 kmem_cache_alloc+0x224/0x3b0
84 sk_prot_alloc+0x58/0x1a0 84 sk_prot_alloc+0x58/0x1a0
85 sk_alloc+0x32/0x4f0 85 sk_alloc+0x32/0x4f0
86 inet_create+0x427/0xb50 86 inet_create+0x427/0xb50
87 __sock_create+0x2e4/0x650 87 __sock_create+0x2e4/0x650
88 inet_ctl_sock_create+0x30/0x180 88 inet_ctl_sock_create+0x30/0x180
89 igmp_net_init+0xc1/0x130 89 igmp_net_init+0xc1/0x130
90 ops_init+0x167/0x410 90 ops_init+0x167/0x410
91 setup_net+0x304/0xa60 91 setup_net+0x304/0xa60
92 copy_net_ns+0x29b/0x4a0 92 copy_net_ns+0x29b/0x4a0
93 create_new_namespaces+0x4a1/0x820 93 create_new_namespaces+0x4a1/0x820
94 nr_base_pages: 16 94 nr_base_pages: 16
95 ... 95 ...
96 ... 96 ...
97 echo 7000 > /sys/kernel/debug/page_own 97 echo 7000 > /sys/kernel/debug/page_owner_stacks/count_threshold
98 cat /sys/kernel/debug/page_owner_stack 98 cat /sys/kernel/debug/page_owner_stacks/show_stacks> stacks_7000.txt
99 cat stacks_7000.txt 99 cat stacks_7000.txt
100 post_alloc_hook+0x177/0x1a0 100 post_alloc_hook+0x177/0x1a0
101 get_page_from_freelist+0xd01/0xd80 101 get_page_from_freelist+0xd01/0xd80
102 __alloc_pages+0x39e/0x7e0 102 __alloc_pages+0x39e/0x7e0
103 alloc_pages_mpol+0x22e/0x490 103 alloc_pages_mpol+0x22e/0x490
104 folio_alloc+0xd5/0x110 104 folio_alloc+0xd5/0x110
105 filemap_alloc_folio+0x78/0x230 105 filemap_alloc_folio+0x78/0x230
106 page_cache_ra_order+0x287/0x6f0 106 page_cache_ra_order+0x287/0x6f0
107 filemap_get_pages+0x517/0x1160 107 filemap_get_pages+0x517/0x1160
108 filemap_read+0x304/0x9f0 108 filemap_read+0x304/0x9f0
109 xfs_file_buffered_read+0xe6/0x1d0 [xf 109 xfs_file_buffered_read+0xe6/0x1d0 [xfs]
110 xfs_file_read_iter+0x1f0/0x380 [xfs] 110 xfs_file_read_iter+0x1f0/0x380 [xfs]
111 __kernel_read+0x3b9/0x730 111 __kernel_read+0x3b9/0x730
112 kernel_read_file+0x309/0x4d0 112 kernel_read_file+0x309/0x4d0
113 __do_sys_finit_module+0x381/0x730 113 __do_sys_finit_module+0x381/0x730
114 do_syscall_64+0x8d/0x150 114 do_syscall_64+0x8d/0x150
115 entry_SYSCALL_64_after_hwframe+0x62/0 115 entry_SYSCALL_64_after_hwframe+0x62/0x6a
116 nr_base_pages: 20824 116 nr_base_pages: 20824
117 ... 117 ...
118 118
119 cat /sys/kernel/debug/page_owner > pag 119 cat /sys/kernel/debug/page_owner > page_owner_full.txt
120 ./page_owner_sort page_owner_full.txt 120 ./page_owner_sort page_owner_full.txt sorted_page_owner.txt
121 121
122 The general output of ``page_owner_full.txt 122 The general output of ``page_owner_full.txt`` is as follows::
123 123
124 Page allocated via order XXX, ... 124 Page allocated via order XXX, ...
125 PFN XXX ... 125 PFN XXX ...
126 // Detailed stack 126 // Detailed stack
127 127
128 Page allocated via order XXX, ... 128 Page allocated via order XXX, ...
129 PFN XXX ... 129 PFN XXX ...
130 // Detailed stack 130 // Detailed stack
131 By default, it will do full pfn dump, to s 131 By default, it will do full pfn dump, to start with a given pfn,
132 page_owner supports fseek. 132 page_owner supports fseek.
133 133
134 FILE *fp = fopen("/sys/kernel/debug/page_o 134 FILE *fp = fopen("/sys/kernel/debug/page_owner", "r");
135 fseek(fp, pfn_start, SEEK_SET); 135 fseek(fp, pfn_start, SEEK_SET);
136 136
137 The ``page_owner_sort`` tool ignores ``PFN` 137 The ``page_owner_sort`` tool ignores ``PFN`` rows, puts the remaining rows
138 in buf, uses regexp to extract the page ord 138 in buf, uses regexp to extract the page order value, counts the times
139 and pages of buf, and finally sorts them ac 139 and pages of buf, and finally sorts them according to the parameter(s).
140 140
141 See the result about who allocated each pag 141 See the result about who allocated each page
142 in the ``sorted_page_owner.txt``. General o 142 in the ``sorted_page_owner.txt``. General output::
143 143
144 XXX times, XXX pages: 144 XXX times, XXX pages:
145 Page allocated via order XXX, ... 145 Page allocated via order XXX, ...
146 // Detailed stack 146 // Detailed stack
147 147
148 By default, ``page_owner_sort`` is sorted a 148 By default, ``page_owner_sort`` is sorted according to the times of buf.
149 If you want to sort by the page nums of buf 149 If you want to sort by the page nums of buf, use the ``-m`` parameter.
150 The detailed parameters are: 150 The detailed parameters are:
151 151
152 fundamental function:: 152 fundamental function::
153 153
154 Sort: 154 Sort:
155 -a Sort by memory 155 -a Sort by memory allocation time.
156 -m Sort by total 156 -m Sort by total memory.
157 -p Sort by pid. 157 -p Sort by pid.
158 -P Sort by tgid. 158 -P Sort by tgid.
159 -n Sort by task c 159 -n Sort by task command name.
160 -r Sort by memory 160 -r Sort by memory release time.
161 -s Sort by stack 161 -s Sort by stack trace.
162 -t Sort by times 162 -t Sort by times (default).
163 --sort <order> Specify sortin 163 --sort <order> Specify sorting order. Sorting syntax is [+|-]key[,[+|-]key[,...]].
164 Choose a key f 164 Choose a key from the **STANDARD FORMAT SPECIFIERS** section. The "+" is
165 optional since 165 optional since default direction is increasing numerical or lexicographic
166 order. Mixed u 166 order. Mixed use of abbreviated and complete-form of keys is allowed.
167 167
168 Examples: 168 Examples:
169 ./page_owner_s 169 ./page_owner_sort <input> <output> --sort=n,+pid,-tgid
170 ./page_owner_s 170 ./page_owner_sort <input> <output> --sort=at
171 171
172 additional function:: 172 additional function::
173 173
174 Cull: 174 Cull:
175 --cull <rules> 175 --cull <rules>
176 Specify cullin 176 Specify culling rules.Culling syntax is key[,key[,...]].Choose a
177 multi-letter k 177 multi-letter key from the **STANDARD FORMAT SPECIFIERS** section.
178 178
179 <rules> is a single argument i 179 <rules> is a single argument in the form of a comma-separated list,
180 which offers a way to specify 180 which offers a way to specify individual culling rules. The recognized
181 keywords are described in the 181 keywords are described in the **STANDARD FORMAT SPECIFIERS** section below.
182 <rules> can be specified by th 182 <rules> can be specified by the sequence of keys k1,k2, ..., as described in
183 the STANDARD SORT KEYS section 183 the STANDARD SORT KEYS section below. Mixed use of abbreviated and
184 complete-form of keys is allow 184 complete-form of keys is allowed.
185 185
186 Examples: 186 Examples:
187 ./page_owner_s 187 ./page_owner_sort <input> <output> --cull=stacktrace
188 ./page_owner_s 188 ./page_owner_sort <input> <output> --cull=st,pid,name
189 ./page_owner_s 189 ./page_owner_sort <input> <output> --cull=n,f
190 190
191 Filter: 191 Filter:
192 -f Filter out the 192 -f Filter out the information of blocks whose memory has been released.
193 193
194 Select: 194 Select:
195 --pid <pidlist> Select 195 --pid <pidlist> Select by pid. This selects the blocks whose process ID
196 number 196 numbers appear in <pidlist>.
197 --tgid <tgidlist> Select 197 --tgid <tgidlist> Select by tgid. This selects the blocks whose thread
198 group 198 group ID numbers appear in <tgidlist>.
199 --name <cmdlist> Select 199 --name <cmdlist> Select by task command name. This selects the blocks whose
200 task c 200 task command name appear in <cmdlist>.
201 201
202 <pidlist>, <tgidlist>, <cmdlis 202 <pidlist>, <tgidlist>, <cmdlist> are single arguments in the form of a comma-separated list,
203 which offers a way to specify 203 which offers a way to specify individual selecting rules.
204 204
205 205
206 Examples: 206 Examples:
207 ./page_owner_s 207 ./page_owner_sort <input> <output> --pid=1
208 ./page_owner_s 208 ./page_owner_sort <input> <output> --tgid=1,2,3
209 ./page_owner_s 209 ./page_owner_sort <input> <output> --name name1,name2
210 210
211 STANDARD FORMAT SPECIFIERS 211 STANDARD FORMAT SPECIFIERS
212 ========================== 212 ==========================
213 :: 213 ::
214 214
215 For --sort option: 215 For --sort option:
216 216
217 KEY LONG DESCRI 217 KEY LONG DESCRIPTION
218 p pid proces 218 p pid process ID
219 tg tgid thread 219 tg tgid thread group ID
220 n name task c 220 n name task command name
221 st stacktrace stack 221 st stacktrace stack trace of the page allocation
222 T txt full t 222 T txt full text of block
223 ft free_ts timest 223 ft free_ts timestamp of the page when it was released
224 at alloc_ts timest 224 at alloc_ts timestamp of the page when it was allocated
225 ator allocator memory 225 ator allocator memory allocator for pages
226 226
227 For --cull option: 227 For --cull option:
228 228
229 KEY LONG DESCRI 229 KEY LONG DESCRIPTION
230 p pid proces 230 p pid process ID
231 tg tgid thread 231 tg tgid thread group ID
232 n name task c 232 n name task command name
233 f free whethe 233 f free whether the page has been released or not
234 st stacktrace stack 234 st stacktrace stack trace of the page allocation
235 ator allocator memory 235 ator allocator memory allocator for pages
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.