>> 1 .. _pagemap:
>> 2
1 ============================= 3 =============================
2 Examining Process Page Tables 4 Examining Process Page Tables
3 ============================= 5 =============================
4 6
5 pagemap is a new (as of 2.6.25) set of interfa 7 pagemap is a new (as of 2.6.25) set of interfaces in the kernel that allow
6 userspace programs to examine the page tables 8 userspace programs to examine the page tables and related information by
7 reading files in ``/proc``. 9 reading files in ``/proc``.
8 10
9 There are four components to pagemap: 11 There are four components to pagemap:
10 12
11 * ``/proc/pid/pagemap``. This file lets a us 13 * ``/proc/pid/pagemap``. This file lets a userspace process find out which
12 physical frame each virtual page is mapped 14 physical frame each virtual page is mapped to. It contains one 64-bit
13 value for each virtual page, containing the 15 value for each virtual page, containing the following data (from
14 ``fs/proc/task_mmu.c``, above pagemap_read) 16 ``fs/proc/task_mmu.c``, above pagemap_read):
15 17
16 * Bits 0-54 page frame number (PFN) if pr 18 * Bits 0-54 page frame number (PFN) if present
17 * Bits 0-4 swap type if swapped 19 * Bits 0-4 swap type if swapped
18 * Bits 5-54 swap offset if swapped 20 * Bits 5-54 swap offset if swapped
19 * Bit 55 pte is soft-dirty (see 21 * Bit 55 pte is soft-dirty (see
20 Documentation/admin-guide/mm/soft-dirty. !! 22 :ref:`Documentation/admin-guide/mm/soft-dirty.rst <soft_dirty>`)
21 * Bit 56 page exclusively mapped (sinc 23 * Bit 56 page exclusively mapped (since 4.2)
22 * Bit 57 pte is uffd-wp write-protecte !! 24 * Bits 57-60 zero
23 Documentation/admin-guide/mm/userfaultfd <<
24 * Bits 58-60 zero <<
25 * Bit 61 page is file-page or shared-a 25 * Bit 61 page is file-page or shared-anon (since 3.5)
26 * Bit 62 page swapped 26 * Bit 62 page swapped
27 * Bit 63 page present 27 * Bit 63 page present
28 28
29 Since Linux 4.0 only users with the CAP_SYS 29 Since Linux 4.0 only users with the CAP_SYS_ADMIN capability can get PFNs.
30 In 4.0 and 4.1 opens by unprivileged fail w 30 In 4.0 and 4.1 opens by unprivileged fail with -EPERM. Starting from
31 4.2 the PFN field is zeroed if the user doe 31 4.2 the PFN field is zeroed if the user does not have CAP_SYS_ADMIN.
32 Reason: information about PFNs helps in exp 32 Reason: information about PFNs helps in exploiting Rowhammer vulnerability.
33 33
34 If the page is not present but in swap, the 34 If the page is not present but in swap, then the PFN contains an
35 encoding of the swap file number and the pa 35 encoding of the swap file number and the page's offset into the
36 swap. Unmapped pages return a null PFN. Thi 36 swap. Unmapped pages return a null PFN. This allows determining
37 precisely which pages are mapped (or in swa 37 precisely which pages are mapped (or in swap) and comparing mapped
38 pages between processes. 38 pages between processes.
39 39
40 Efficient users of this interface will use 40 Efficient users of this interface will use ``/proc/pid/maps`` to
41 determine which areas of memory are actuall 41 determine which areas of memory are actually mapped and llseek to
42 skip over unmapped regions. 42 skip over unmapped regions.
43 43
44 * ``/proc/kpagecount``. This file contains a 44 * ``/proc/kpagecount``. This file contains a 64-bit count of the number of
45 times each page is mapped, indexed by PFN. 45 times each page is mapped, indexed by PFN.
46 46
47 The page-types tool in the tools/mm directory <<
48 number of times a page is mapped. <<
49 <<
50 * ``/proc/kpageflags``. This file contains a 47 * ``/proc/kpageflags``. This file contains a 64-bit set of flags for each
51 page, indexed by PFN. 48 page, indexed by PFN.
52 49
53 The flags are (from ``fs/proc/page.c``, abo 50 The flags are (from ``fs/proc/page.c``, above kpageflags_read):
54 51
55 0. LOCKED 52 0. LOCKED
56 1. ERROR 53 1. ERROR
57 2. REFERENCED 54 2. REFERENCED
58 3. UPTODATE 55 3. UPTODATE
59 4. DIRTY 56 4. DIRTY
60 5. LRU 57 5. LRU
61 6. ACTIVE 58 6. ACTIVE
62 7. SLAB 59 7. SLAB
63 8. WRITEBACK 60 8. WRITEBACK
64 9. RECLAIM 61 9. RECLAIM
65 10. BUDDY 62 10. BUDDY
66 11. MMAP 63 11. MMAP
67 12. ANON 64 12. ANON
68 13. SWAPCACHE 65 13. SWAPCACHE
69 14. SWAPBACKED 66 14. SWAPBACKED
70 15. COMPOUND_HEAD 67 15. COMPOUND_HEAD
71 16. COMPOUND_TAIL 68 16. COMPOUND_TAIL
72 17. HUGE 69 17. HUGE
73 18. UNEVICTABLE 70 18. UNEVICTABLE
74 19. HWPOISON 71 19. HWPOISON
75 20. NOPAGE 72 20. NOPAGE
76 21. KSM 73 21. KSM
77 22. THP 74 22. THP
78 23. OFFLINE !! 75 23. BALLOON
79 24. ZERO_PAGE 76 24. ZERO_PAGE
80 25. IDLE 77 25. IDLE
81 26. PGTABLE <<
82 78
83 * ``/proc/kpagecgroup``. This file contains 79 * ``/proc/kpagecgroup``. This file contains a 64-bit inode number of the
84 memory cgroup each page is charged to, inde 80 memory cgroup each page is charged to, indexed by PFN. Only available when
85 CONFIG_MEMCG is set. 81 CONFIG_MEMCG is set.
86 82
87 Short descriptions to the page flags 83 Short descriptions to the page flags
88 ==================================== 84 ====================================
89 85
90 0 - LOCKED 86 0 - LOCKED
91 The page is being locked for exclusive acce !! 87 page is being locked for exclusive access, e.g. by undergoing read/write IO
92 IO. <<
93 7 - SLAB 88 7 - SLAB
94 The page is managed by the SLAB/SLUB kernel !! 89 page is managed by the SLAB/SLOB/SLUB/SLQB kernel memory allocator
95 When compound page is used, either will onl !! 90 When compound page is used, SLUB/SLQB will only set this flag on the head
96 page. !! 91 page; SLOB will not flag it at all.
97 10 - BUDDY 92 10 - BUDDY
98 A free memory block managed by the buddy s !! 93 a free memory block managed by the buddy system allocator
99 The buddy system organizes free memory in 94 The buddy system organizes free memory in blocks of various orders.
100 An order N block has 2^N physically contig 95 An order N block has 2^N physically contiguous pages, with the BUDDY flag
101 set for and _only_ for the first page. 96 set for and _only_ for the first page.
102 15 - COMPOUND_HEAD 97 15 - COMPOUND_HEAD
103 A compound page with order N consists of 2 98 A compound page with order N consists of 2^N physically contiguous pages.
104 A compound page with order 2 takes the for 99 A compound page with order 2 takes the form of "HTTT", where H donates its
105 head page and T donates its tail page(s). 100 head page and T donates its tail page(s). The major consumers of compound
106 pages are hugeTLB pages (Documentation/adm !! 101 pages are hugeTLB pages
>> 102 (:ref:`Documentation/admin-guide/mm/hugetlbpage.rst <hugetlbpage>`),
107 the SLUB etc. memory allocators and vario 103 the SLUB etc. memory allocators and various device drivers.
108 However in this interface, only huge/giga 104 However in this interface, only huge/giga pages are made visible
109 to end users. 105 to end users.
110 16 - COMPOUND_TAIL 106 16 - COMPOUND_TAIL
111 A compound page tail (see description abov 107 A compound page tail (see description above).
112 17 - HUGE 108 17 - HUGE
113 This is an integral part of a HugeTLB page !! 109 this is an integral part of a HugeTLB page
114 19 - HWPOISON 110 19 - HWPOISON
115 Hardware detected memory corruption on thi !! 111 hardware detected memory corruption on this page: don't touch the data!
116 20 - NOPAGE 112 20 - NOPAGE
117 No page frame exists at the requested addr !! 113 no page frame exists at the requested address
118 21 - KSM 114 21 - KSM
119 Identical memory pages dynamically shared !! 115 identical memory pages dynamically shared between one or more processes
120 22 - THP 116 22 - THP
121 Contiguous pages which construct THP of an !! 117 contiguous pages which construct transparent hugepages
122 23 - OFFLINE !! 118 23 - BALLOON
123 The page is logically offline. !! 119 balloon compaction page
124 24 - ZERO_PAGE 120 24 - ZERO_PAGE
125 Zero page for pfn_zero or huge_zero page. !! 121 zero page for pfn_zero or huge_zero page
126 25 - IDLE 122 25 - IDLE
127 The page has not been accessed since it wa !! 123 page has not been accessed since it was marked idle (see
128 Documentation/admin-guide/mm/idle_page_tra !! 124 :ref:`Documentation/admin-guide/mm/idle_page_tracking.rst <idle_page_tracking>`).
129 Note that this flag may be stale in case t 125 Note that this flag may be stale in case the page was accessed via
130 a PTE. To make sure the flag is up-to-date 126 a PTE. To make sure the flag is up-to-date one has to read
131 ``/sys/kernel/mm/page_idle/bitmap`` first. 127 ``/sys/kernel/mm/page_idle/bitmap`` first.
132 26 - PGTABLE <<
133 The page is in use as a page table. <<
134 128
135 IO related page flags 129 IO related page flags
136 --------------------- 130 ---------------------
137 131
138 1 - ERROR 132 1 - ERROR
139 IO error occurred. !! 133 IO error occurred
140 3 - UPTODATE 134 3 - UPTODATE
141 The page has up-to-date data. !! 135 page has up-to-date data
142 ie. for file backed page: (in-memory data r 136 ie. for file backed page: (in-memory data revision >= on-disk one)
143 4 - DIRTY 137 4 - DIRTY
144 The page has been written to, hence contain !! 138 page has been written to, hence contains new data
145 i.e. for file backed page: (in-memory data 139 i.e. for file backed page: (in-memory data revision > on-disk one)
146 8 - WRITEBACK 140 8 - WRITEBACK
147 The page is being synced to disk. !! 141 page is being synced to disk
148 142
149 LRU related page flags 143 LRU related page flags
150 ---------------------- 144 ----------------------
151 145
152 5 - LRU 146 5 - LRU
153 The page is in one of the LRU lists. !! 147 page is in one of the LRU lists
154 6 - ACTIVE 148 6 - ACTIVE
155 The page is in the active LRU list. !! 149 page is in the active LRU list
156 18 - UNEVICTABLE 150 18 - UNEVICTABLE
157 The page is in the unevictable (non-)LRU li !! 151 page is in the unevictable (non-)LRU list It is somehow pinned and
158 not a candidate for LRU page reclaims, e.g. 152 not a candidate for LRU page reclaims, e.g. ramfs pages,
159 shmctl(SHM_LOCK) and mlock() memory segment !! 153 shmctl(SHM_LOCK) and mlock() memory segments
160 2 - REFERENCED 154 2 - REFERENCED
161 The page has been referenced since last LRU !! 155 page has been referenced since last LRU list enqueue/requeue
162 9 - RECLAIM 156 9 - RECLAIM
163 The page will be reclaimed soon after its p !! 157 page will be reclaimed soon after its pageout IO completed
164 11 - MMAP 158 11 - MMAP
165 A memory mapped page. !! 159 a memory mapped page
166 12 - ANON 160 12 - ANON
167 A memory mapped page that is not part of a !! 161 a memory mapped page that is not part of a file
168 13 - SWAPCACHE 162 13 - SWAPCACHE
169 The page is mapped to swap space, i.e. has !! 163 page is mapped to swap space, i.e. has an associated swap entry
170 14 - SWAPBACKED 164 14 - SWAPBACKED
171 The page is backed by swap/RAM. !! 165 page is backed by swap/RAM
172 166
173 The page-types tool in the tools/mm directory !! 167 The page-types tool in the tools/vm directory can be used to query the
174 above flags. 168 above flags.
175 169
176 Exceptions for Shared Memory !! 170 Using pagemap to do something useful
177 ============================ !! 171 ====================================
178 172
179 Page table entries for shared pages are cleare !! 173 The general procedure for using pagemap to find out about a process' memory
180 swapped out. This makes swapped out pages indi !! 174 usage goes like this:
181 ones. <<
182 <<
183 In kernel space, the swap location can still b <<
184 However, values stored only on the normal PTE <<
185 page is swapped out (i.e. SOFT_DIRTY). <<
186 <<
187 In user space, whether the page is present, sw <<
188 the help of lseek and/or mincore system calls. <<
189 <<
190 lseek() can differentiate between accessed pag <<
191 holes (none/non-allocated) by specifying the S <<
192 the pages are backed. For anonymous shared pag <<
193 ``/proc/pid/map_files/``. <<
194 175
195 mincore() can differentiate between pages in m !! 176 1. Read ``/proc/pid/maps`` to determine which parts of the memory space are
196 cache) and out of memory (swapped out or none/ !! 177 mapped to what.
>> 178 2. Select the maps you are interested in -- all of them, or a particular
>> 179 library, or the stack or the heap, etc.
>> 180 3. Open ``/proc/pid/pagemap`` and seek to the pages you would like to examine.
>> 181 4. Read a u64 for each page from pagemap.
>> 182 5. Open ``/proc/kpagecount`` and/or ``/proc/kpageflags``. For each PFN you
>> 183 just read, seek to that entry in the file, and read the data you want.
>> 184
>> 185 For example, to find the "unique set size" (USS), which is the amount of
>> 186 memory that a process is using that is not shared with any other process,
>> 187 you can go through every map in the process, find the PFNs, look those up
>> 188 in kpagecount, and tally up the number of pages that are only referenced
>> 189 once.
197 190
198 Other notes 191 Other notes
199 =========== 192 ===========
200 193
201 Reading from any of the files will return -EIN 194 Reading from any of the files will return -EINVAL if you are not starting
202 the read on an 8-byte boundary (e.g., if you s 195 the read on an 8-byte boundary (e.g., if you sought an odd number of bytes
203 into the file), or if the size of the read is 196 into the file), or if the size of the read is not a multiple of 8 bytes.
204 197
205 Before Linux 3.11 pagemap bits 55-60 were used 198 Before Linux 3.11 pagemap bits 55-60 were used for "page-shift" (which is
206 always 12 at most architectures). Since Linux 199 always 12 at most architectures). Since Linux 3.11 their meaning changes
207 after first clear of soft-dirty bits. Since Li 200 after first clear of soft-dirty bits. Since Linux 4.2 they are used for
208 flags unconditionally. 201 flags unconditionally.
209 <<
210 Pagemap Scan IOCTL <<
211 ================== <<
212 <<
213 The ``PAGEMAP_SCAN`` IOCTL on the pagemap file <<
214 clear the info about page table entries. The f <<
215 in this IOCTL: <<
216 <<
217 - Scan the address range and get the memory ra <<
218 This is performed when the output buffer is <<
219 - Write-protect the pages. The ``PM_SCAN_WP_MA <<
220 the pages of interest. The ``PM_SCAN_CHECK_W <<
221 non-Async Write Protected pages are found. T <<
222 used with or without ``PM_SCAN_CHECK_WPASYNC <<
223 - Both of those operations can be combined int <<
224 get and write protect the pages as well. <<
225 <<
226 Following flags about pages are currently supp <<
227 <<
228 - ``PAGE_IS_WPALLOWED`` - Page has async-write <<
229 - ``PAGE_IS_WRITTEN`` - Page has been written <<
230 - ``PAGE_IS_FILE`` - Page is file backed <<
231 - ``PAGE_IS_PRESENT`` - Page is present in the <<
232 - ``PAGE_IS_SWAPPED`` - Page is in swapped <<
233 - ``PAGE_IS_PFNZERO`` - Page has zero PFN <<
234 - ``PAGE_IS_HUGE`` - Page is PMD-mapped THP or <<
235 - ``PAGE_IS_SOFT_DIRTY`` - Page is soft-dirty <<
236 <<
237 The ``struct pm_scan_arg`` is used as the argu <<
238 <<
239 1. The size of the ``struct pm_scan_arg`` mus <<
240 field. This field will be helpful in recog <<
241 are done later. <<
242 2. The flags can be specified in the ``flags` <<
243 and ``PM_SCAN_CHECK_WPASYNC`` are the only <<
244 operation is optionally performed dependin <<
245 provided or not. <<
246 3. The range is specified through ``start`` a <<
247 4. The walk can abort before visiting the com <<
248 can get full etc. The walk ending address <<
249 5. The output buffer of ``struct page_region` <<
250 ``vec`` and ``vec_len``. <<
251 6. The optional maximum requested pages are s <<
252 7. The masks are specified in ``category_mask <<
253 ``category_inverted`` and ``return_mask``. <<
254 <<
255 Find pages which have been written and WP them <<
256 <<
257 struct pm_scan_arg arg = { <<
258 .size = sizeof(arg), <<
259 .flags = PM_SCAN_CHECK_WPASYNC | PM_SCAN_CH <<
260 .. <<
261 .category_mask = PAGE_IS_WRITTEN, <<
262 .return_mask = PAGE_IS_WRITTEN, <<
263 }; <<
264 <<
265 Find pages which have been written, are file b <<
266 present or huge:: <<
267 <<
268 struct pm_scan_arg arg = { <<
269 .size = sizeof(arg), <<
270 .flags = 0, <<
271 .. <<
272 .category_mask = PAGE_IS_WRITTEN | PAGE_IS_ <<
273 .category_inverted = PAGE_IS_SWAPPED, <<
274 .category_anyof_mask = PAGE_IS_PRESENT | PA <<
275 .return_mask = PAGE_IS_WRITTEN | PAGE_IS_SW <<
276 PAGE_IS_PRESENT | PAGE_IS_HU <<
277 }; <<
278 <<
279 The ``PAGE_IS_WRITTEN`` flag can be considered <<
280 of soft-dirty flag. It doesn't get affected by <<
281 the user can find the true soft-dirty pages in <<
282 still be extra dirty pages reported for THP or <<
283 <<
284 "PAGE_IS_WRITTEN" category is used with uffd w <<
285 implement memory dirty tracking in userspace: <<
286 <<
287 1. The userfaultfd file descriptor is created <<
288 2. The ``UFFD_FEATURE_WP_UNPOPULATED`` and `` <<
289 are set by ``UFFDIO_API`` IOCTL. <<
290 3. The memory range is registered with ``UFFD <<
291 through ``UFFDIO_REGISTER`` IOCTL. <<
292 4. Then any part of the registered memory or <<
293 be write protected using ``PAGEMAP_SCAN`` <<
294 or the ``UFFDIO_WRITEPROTECT`` IOCTL can b <<
295 same operation. The former is better in te <<
296 5. Now the ``PAGEMAP_SCAN`` IOCTL can be used <<
297 have been written to since they were last <<
298 the pages as well. <<
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.