요약·해설과 원문, 전문 번역을 서로 분리했습니다. API 이름, symbol, source path는 원문 표기를 사용합니다.
1. 요약·해설
원문의 핵심 논리와 kernel programming 관점의 보충 설명입니다. 아래의 전문 번역과는 별도로 작성했습니다.
2. 영어 원문 전체
번역 기준이 된 Linux v6.18.37 원문입니다. 줄 번호는 이 버전의 파일 좌표입니다.
원문 전체 펼치기
1
=============================
2
Examining Process Page Tables
3
=============================
5
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 and related information by
7
reading files in ``/proc``.
9
There are four components to pagemap:
11
* ``/proc/pid/pagemap``. This file lets a userspace process find out which
12
physical frame each virtual page is mapped to. It contains one 64-bit
13
value for each virtual page, containing the following data (from
14
``fs/proc/task_mmu.c``, above pagemap_read):
16
* Bits 0-54 page frame number (PFN) if present
17
* Bits 0-4 swap type if swapped
18
* Bits 5-54 swap offset if swapped
19
* Bit 55 pte is soft-dirty (see
20
Documentation/admin-guide/mm/soft-dirty.rst)
21
* Bit 56 page exclusively mapped (since 4.2)
22
* Bit 57 pte is uffd-wp write-protected (since 5.13) (see
23
Documentation/admin-guide/mm/userfaultfd.rst)
24
* Bit 58 pte is a guard region (since 6.15) (see madvise (2) man page)
25
* Bits 59-60 zero
26
* Bit 61 page is file-page or shared-anon (since 3.5)
27
* Bit 62 page swapped
28
* Bit 63 page present
30
Since Linux 4.0 only users with the CAP_SYS_ADMIN capability can get PFNs.
31
In 4.0 and 4.1 opens by unprivileged fail with -EPERM. Starting from
32
4.2 the PFN field is zeroed if the user does not have CAP_SYS_ADMIN.
33
Reason: information about PFNs helps in exploiting Rowhammer vulnerability.
35
If the page is not present but in swap, then the PFN contains an
36
encoding of the swap file number and the page's offset into the
37
swap. Unmapped pages return a null PFN. This allows determining
38
precisely which pages are mapped (or in swap) and comparing mapped
39
pages between processes.
41
Traditionally, bit 56 indicates that a page is mapped exactly once and bit
42
56 is clear when a page is mapped multiple times, even when mapped in the
43
same process multiple times. In some kernel configurations, the semantics
44
for pages part of a larger allocation (e.g., THP) can differ: bit 56 is set
45
if all pages part of the corresponding large allocation are *certainly*
46
mapped in the same process, even if the page is mapped multiple times in that
47
process. Bit 56 is clear when any page page of the larger allocation
48
is *maybe* mapped in a different process. In some cases, a large allocation
49
might be treated as "maybe mapped by multiple processes" even though this
50
is no longer the case.
52
Efficient users of this interface will use ``/proc/pid/maps`` to
53
determine which areas of memory are actually mapped and llseek to
54
skip over unmapped regions.
56
* ``/proc/kpagecount``. This file contains a 64-bit count of the number of
57
times each page is mapped, indexed by PFN. Some kernel configurations do
58
not track the precise number of times a page part of a larger allocation
59
(e.g., THP) is mapped. In these configurations, the average number of
60
mappings per page in this larger allocation is returned instead. However,
61
if any page of the large allocation is mapped, the returned value will
62
be at least 1.
64
The page-types tool in the tools/mm directory can be used to query the
65
number of times a page is mapped.
67
* ``/proc/kpageflags``. This file contains a 64-bit set of flags for each
68
page, indexed by PFN.
70
The flags are (from ``fs/proc/page.c``, above kpageflags_read):
72
0. LOCKED
73
1. ERROR
74
2. REFERENCED
75
3. UPTODATE
76
4. DIRTY
77
5. LRU
78
6. ACTIVE
79
7. SLAB
80
8. WRITEBACK
81
9. RECLAIM
82
10. BUDDY
83
11. MMAP
84
12. ANON
85
13. SWAPCACHE
86
14. SWAPBACKED
87
15. COMPOUND_HEAD
88
16. COMPOUND_TAIL
89
17. HUGE
90
18. UNEVICTABLE
91
19. HWPOISON
92
20. NOPAGE
93
21. KSM
94
22. THP
95
23. OFFLINE
96
24. ZERO_PAGE
97
25. IDLE
98
26. PGTABLE
100
* ``/proc/kpagecgroup``. This file contains a 64-bit inode number of the
101
memory cgroup each page is charged to, indexed by PFN. Only available when
102
CONFIG_MEMCG is set.
104
Short descriptions to the page flags
105
====================================
107
0 - LOCKED
108
The page is being locked for exclusive access, e.g. by undergoing read/write
109
IO.
110
7 - SLAB
111
The page is managed by the SLAB/SLUB kernel memory allocator.
112
When compound page is used, either will only set this flag on the head
113
page.
114
10 - BUDDY
115
A free memory block managed by the buddy system allocator.
116
The buddy system organizes free memory in blocks of various orders.
117
An order N block has 2^N physically contiguous pages, with the BUDDY flag
118
set for and _only_ for the first page.
119
15 - COMPOUND_HEAD
120
A compound page with order N consists of 2^N physically contiguous pages.
121
A compound page with order 2 takes the form of "HTTT", where H donates its
122
head page and T donates its tail page(s). The major consumers of compound
123
pages are hugeTLB pages (Documentation/admin-guide/mm/hugetlbpage.rst),
124
the SLUB etc. memory allocators and various device drivers.
125
However in this interface, only huge/giga pages are made visible
126
to end users.
127
16 - COMPOUND_TAIL
128
A compound page tail (see description above).
129
17 - HUGE
130
This is an integral part of a HugeTLB page.
131
19 - HWPOISON
132
Hardware detected memory corruption on this page: don't touch the data!
133
20 - NOPAGE
134
No page frame exists at the requested address.
135
21 - KSM
136
Identical memory pages dynamically shared between one or more processes.
137
22 - THP
138
Contiguous pages which construct THP of any size and mapped by any granularity.
139
23 - OFFLINE
140
The page is logically offline.
141
24 - ZERO_PAGE
142
Zero page for pfn_zero or huge_zero page.
143
25 - IDLE
144
The page has not been accessed since it was marked idle (see
145
Documentation/admin-guide/mm/idle_page_tracking.rst).
146
Note that this flag may be stale in case the page was accessed via
147
a PTE. To make sure the flag is up-to-date one has to read
148
``/sys/kernel/mm/page_idle/bitmap`` first.
149
26 - PGTABLE
150
The page is in use as a page table.
152
IO related page flags
153
---------------------
155
1 - ERROR
156
IO error occurred.
157
3 - UPTODATE
158
The page has up-to-date data.
159
ie. for file backed page: (in-memory data revision >= on-disk one)
160
4 - DIRTY
161
The page has been written to, hence contains new data.
162
i.e. for file backed page: (in-memory data revision > on-disk one)
163
8 - WRITEBACK
164
The page is being synced to disk.
166
LRU related page flags
167
----------------------
169
5 - LRU
170
The page is in one of the LRU lists.
171
6 - ACTIVE
172
The page is in the active LRU list.
173
18 - UNEVICTABLE
174
The page is in the unevictable (non-)LRU list It is somehow pinned and
175
not a candidate for LRU page reclaims, e.g. ramfs pages,
176
shmctl(SHM_LOCK) and mlock() memory segments.
177
2 - REFERENCED
178
The page has been referenced since last LRU list enqueue/requeue.
179
9 - RECLAIM
180
The page will be reclaimed soon after its pageout IO completed.
181
11 - MMAP
182
A memory mapped page.
183
12 - ANON
184
A memory mapped page that is not part of a file.
185
13 - SWAPCACHE
186
The page is mapped to swap space, i.e. has an associated swap entry.
187
14 - SWAPBACKED
188
The page is backed by swap/RAM.
190
The page-types tool in the tools/mm directory can be used to query the
191
above flags.
193
Exceptions for Shared Memory
194
============================
196
Page table entries for shared pages are cleared when the pages are zapped or
197
swapped out. This makes swapped out pages indistinguishable from never-allocated
198
ones.
200
In kernel space, the swap location can still be retrieved from the page cache.
201
However, values stored only on the normal PTE get lost irretrievably when the
202
page is swapped out (i.e. SOFT_DIRTY).
204
In user space, whether the page is present, swapped or none can be deduced with
205
the help of lseek and/or mincore system calls.
207
lseek() can differentiate between accessed pages (present or swapped out) and
208
holes (none/non-allocated) by specifying the SEEK_DATA flag on the file where
209
the pages are backed. For anonymous shared pages, the file can be found in
210
``/proc/pid/map_files/``.
212
mincore() can differentiate between pages in memory (present, including swap
213
cache) and out of memory (swapped out or none/non-allocated).
215
Other notes
216
===========
218
Reading from any of the files will return -EINVAL if you are not starting
219
the read on an 8-byte boundary (e.g., if you sought an odd number of bytes
220
into the file), or if the size of the read is not a multiple of 8 bytes.
222
Before Linux 3.11 pagemap bits 55-60 were used for "page-shift" (which is
223
always 12 at most architectures). Since Linux 3.11 their meaning changes
224
after first clear of soft-dirty bits. Since Linux 4.2 they are used for
225
flags unconditionally.
227
Pagemap Scan IOCTL
228
==================
230
The ``PAGEMAP_SCAN`` IOCTL on the pagemap file can be used to get or optionally
231
clear the info about page table entries. The following operations are supported
232
in this IOCTL:
234
- Scan the address range and get the memory ranges matching the provided criteria.
235
This is performed when the output buffer is specified.
236
- Write-protect the pages. The ``PM_SCAN_WP_MATCHING`` is used to write-protect
237
the pages of interest. The ``PM_SCAN_CHECK_WPASYNC`` aborts the operation if
238
non-Async Write Protected pages are found. The ``PM_SCAN_WP_MATCHING`` can be
239
used with or without ``PM_SCAN_CHECK_WPASYNC``.
240
- Both of those operations can be combined into one atomic operation where we can
241
get and write protect the pages as well.
243
Following flags about pages are currently supported:
245
- ``PAGE_IS_WPALLOWED`` - Page has async-write-protection enabled
246
- ``PAGE_IS_WRITTEN`` - Page has been written to from the time it was write protected
247
- ``PAGE_IS_FILE`` - Page is file backed
248
- ``PAGE_IS_PRESENT`` - Page is present in the memory
249
- ``PAGE_IS_SWAPPED`` - Page is in swapped
250
- ``PAGE_IS_PFNZERO`` - Page has zero PFN
251
- ``PAGE_IS_HUGE`` - Page is PMD-mapped THP or Hugetlb backed
252
- ``PAGE_IS_SOFT_DIRTY`` - Page is soft-dirty
253
- ``PAGE_IS_GUARD`` - Page is a part of a guard region
255
The ``struct pm_scan_arg`` is used as the argument of the IOCTL.
257
1. The size of the ``struct pm_scan_arg`` must be specified in the ``size``
258
field. This field will be helpful in recognizing the structure if extensions
259
are done later.
260
2. The flags can be specified in the ``flags`` field. The ``PM_SCAN_WP_MATCHING``
261
and ``PM_SCAN_CHECK_WPASYNC`` are the only added flags at this time. The get
262
operation is optionally performed depending upon if the output buffer is
263
provided or not.
264
3. The range is specified through ``start`` and ``end``.
265
4. The walk can abort before visiting the complete range such as the user buffer
266
can get full etc. The walk ending address is specified in``end_walk``.
267
5. The output buffer of ``struct page_region`` array and size is specified in
268
``vec`` and ``vec_len``.
269
6. The optional maximum requested pages are specified in the ``max_pages``.
270
7. The masks are specified in ``category_mask``, ``category_anyof_mask``,
271
``category_inverted`` and ``return_mask``.
273
Find pages which have been written and WP them as well::
275
struct pm_scan_arg arg = {
276
.size = sizeof(arg),
277
.flags = PM_SCAN_CHECK_WPASYNC | PM_SCAN_CHECK_WPASYNC,
278
..
279
.category_mask = PAGE_IS_WRITTEN,
280
.return_mask = PAGE_IS_WRITTEN,
281
};
283
Find pages which have been written, are file backed, not swapped and either
284
present or huge::
286
struct pm_scan_arg arg = {
287
.size = sizeof(arg),
288
.flags = 0,
289
..
290
.category_mask = PAGE_IS_WRITTEN | PAGE_IS_SWAPPED,
291
.category_inverted = PAGE_IS_SWAPPED,
292
.category_anyof_mask = PAGE_IS_PRESENT | PAGE_IS_HUGE,
293
.return_mask = PAGE_IS_WRITTEN | PAGE_IS_SWAPPED |
294
PAGE_IS_PRESENT | PAGE_IS_HUGE,
295
};
297
The ``PAGE_IS_WRITTEN`` flag can be considered as a better-performing alternative
298
of soft-dirty flag. It doesn't get affected by VMA merging of the kernel and hence
299
the user can find the true soft-dirty pages in case of normal pages. (There may
300
still be extra dirty pages reported for THP or Hugetlb pages.)
302
"PAGE_IS_WRITTEN" category is used with uffd write protect-enabled ranges to
303
implement memory dirty tracking in userspace:
305
1. The userfaultfd file descriptor is created with ``userfaultfd`` syscall.
306
2. The ``UFFD_FEATURE_WP_UNPOPULATED`` and ``UFFD_FEATURE_WP_ASYNC`` features
307
are set by ``UFFDIO_API`` IOCTL.
308
3. The memory range is registered with ``UFFDIO_REGISTER_MODE_WP`` mode
309
through ``UFFDIO_REGISTER`` IOCTL.
310
4. Then any part of the registered memory or the whole memory region must
311
be write protected using ``PAGEMAP_SCAN`` IOCTL with flag ``PM_SCAN_WP_MATCHING``
312
or the ``UFFDIO_WRITEPROTECT`` IOCTL can be used. Both of these perform the
313
same operation. The former is better in terms of performance.
314
5. Now the ``PAGEMAP_SCAN`` IOCTL can be used to either just find pages which
315
have been written to since they were last marked and/or optionally write protect
316
the pages as well.
3. 한국어 전문 번역
영어 원문의 문단 순서와 의미를 유지한 전체 번역입니다. 코드, 함수명, symbol과 URL은 원문 표기를 유지합니다.
Process page table 관찰
1-8pagemap은 Linux 2.6.25부터 제공되는 kernel interface 집합입니다. Userspace program이 `/proc` file을 읽어 page table과 관련 정보를 조사할 수 있게 합니다.
/proc/pid/pagemap
9-54| Interface | Record | 내용 |
|---|---|---|
| /proc/pid/pagemap | virtual page별 64-bit 값 | mapping된 PFN, swap encoding과 PTE 상태 |
| /proc/kpagecount | PFN별 64-bit count | 각 physical page가 mapping된 횟수 |
| /proc/kpageflags | PFN별 64-bit flag set | page 상태와 allocator·LRU 속성 |
| /proc/kpagecgroup | PFN별 64-bit inode number | page가 charge된 memory cgroup, CONFIG_MEMCG 필요 |
`/proc/pid/pagemap`에는 virtual page마다 하나의 64-bit 값이 있으며, 해당 virtual page가 mapping된 physical frame과 swap·PTE 상태를 다음 bit layout으로 표현합니다. Kernel 구현은 `fs/proc/task_mmu.c`의 `pagemap_read` 위쪽에서 확인할 수 있습니다.
| Bit | 의미 |
|---|---|
| 0-54 | present일 때 page frame number(PFN) |
| 0-4 | swapped일 때 swap type |
| 5-54 | swapped일 때 swap offset |
| 55 | PTE가 soft-dirty, Documentation/admin-guide/mm/soft-dirty.rst 참조 |
| 56 | page가 exclusively mapped, Linux 4.2부터 |
| 57 | PTE가 uffd-wp write-protected, Linux 5.13부터; userfaultfd.rst 참조 |
| 58 | PTE가 guard region, Linux 6.15부터; madvise(2) 참조 |
| 59-60 | 0 |
| 61 | file-page 또는 shared-anon page, Linux 3.5부터 |
| 62 | page swapped |
| 63 | page present |
Linux 4.0부터 PFN은 `CAP_SYS_ADMIN` capability가 있는 사용자만 얻을 수 있습니다. Linux 4.0과 4.1에서는 unprivileged open이 `-EPERM`으로 실패하고, Linux 4.2부터는 `CAP_SYS_ADMIN`이 없을 때 PFN field를 0으로 만듭니다. PFN 정보가 Rowhammer vulnerability 악용에 도움을 줄 수 있기 때문입니다.
Page가 present하지 않고 swap에 있으면 PFN field에는 swap file number와 해당 file 안의 page offset이 encoding됩니다. Unmapped page는 null PFN을 반환하므로 mapping되었거나 swap에 있는 page를 정확히 판별하고 process 사이의 mapped page를 비교할 수 있습니다.
전통적으로 bit 56은 page가 정확히 한 번 mapping됐을 때 set되고, 같은 process 안에서 여러 번 mapping된 경우까지 포함해 mapping이 여러 개면 clear됩니다. 다만 THP 같은 larger allocation의 page는 kernel configuration에 따라 의미가 다릅니다. Allocation의 모든 page가 확실히 같은 process에 mapping됐다면 그 process 안에서 중복 mapping되어도 set될 수 있고, 어느 page든 다른 process에 mapping됐을 가능성이 있으면 clear됩니다. 실제로는 더 이상 공유되지 않더라도 보수적으로 여러 process에 mapping된 것으로 취급될 수 있습니다.
효율적으로 사용하려면 `/proc/pid/maps`에서 실제 mapped area를 찾고 `llseek`으로 unmapped region을 건너뜁니다.
/proc/kpagecount
55-66`/proc/kpagecount`는 PFN으로 index된 page마다 64-bit mapping count를 제공합니다. 일부 kernel configuration은 THP 같은 larger allocation에 속한 개별 page의 정확한 mapping 횟수를 추적하지 않으며, 이때 allocation 전체의 page당 평균 mapping 수를 반환합니다. Large allocation의 어느 page든 mapping됐다면 반환 값은 최소 1입니다.
`tools/mm` directory의 `page-types` tool로 page mapping 횟수를 조회할 수 있습니다.
/proc/kpageflags와 kpagecgroup
67-103`/proc/kpageflags`는 PFN으로 index된 각 page의 64-bit flag set을 제공합니다. Flag 정의는 `fs/proc/page.c`의 `kpageflags_read` 위쪽에 있습니다.
| Bit | Flag |
|---|---|
| 0 | LOCKED |
| 1 | ERROR |
| 2 | REFERENCED |
| 3 | UPTODATE |
| 4 | DIRTY |
| 5 | LRU |
| 6 | ACTIVE |
| 7 | SLAB |
| 8 | WRITEBACK |
| 9 | RECLAIM |
| 10 | BUDDY |
| 11 | MMAP |
| 12 | ANON |
| 13 | SWAPCACHE |
| 14 | SWAPBACKED |
| 15 | COMPOUND_HEAD |
| 16 | COMPOUND_TAIL |
| 17 | HUGE |
| 18 | UNEVICTABLE |
| 19 | HWPOISON |
| 20 | NOPAGE |
| 21 | KSM |
| 22 | THP |
| 23 | OFFLINE |
| 24 | ZERO_PAGE |
| 25 | IDLE |
| 26 | PGTABLE |
`/proc/kpagecgroup`는 PFN으로 index된 각 page가 charge된 memory cgroup의 64-bit inode number를 제공합니다. `CONFIG_MEMCG`가 설정된 kernel에서만 사용할 수 있습니다.
Page flag 상세 설명
104-151| Flag | 설명 |
|---|---|
| LOCKED | read/write I/O 등 exclusive access를 위해 lock된 page |
| SLAB | SLAB/SLUB kernel memory allocator가 관리하며 compound page에서는 head에만 설정 |
| BUDDY | buddy allocator의 free block 첫 page에만 설정; order N은 2^N contiguous pages |
| COMPOUND_HEAD | order N compound allocation의 head. Order 2는 HTTT이며 H가 head, T가 tail |
| COMPOUND_TAIL | compound page의 tail |
| HUGE | HugeTLB page의 구성 page |
| HWPOISON | hardware가 memory corruption을 감지한 page |
| NOPAGE | 요청 address에 page frame이 없음 |
| KSM | 하나 이상의 process 사이에 동적으로 공유되는 동일 memory page |
| THP | 임의 크기의 THP를 구성하며 임의 granularity로 mapping된 contiguous page |
| OFFLINE | 논리적으로 offline된 page |
| ZERO_PAGE | pfn_zero의 zero page 또는 huge_zero page |
| IDLE | idle 표시 뒤 접근되지 않은 page. PTE 접근 시 stale할 수 있어 bitmap을 먼저 읽어야 함 |
| PGTABLE | page table로 사용 중인 page |
`BUDDY`의 order N block은 물리적으로 연속된 `2^N` page이며 첫 page에만 flag가 설정됩니다. Order 2 compound page를 `HTTT`로 표현하면 H는 head, T는 tail입니다. Compound page의 주요 사용자는 HugeTLB page, SLUB 같은 memory allocator와 여러 device driver이지만 이 interface에서는 huge/giga page만 userspace에 보입니다.
`IDLE` flag는 PTE를 통한 접근이 있었을 때 stale할 수 있습니다. 최신 상태를 보장하려면 `/sys/kernel/mm/page_idle/bitmap`을 먼저 읽습니다.
I/O와 LRU 관련 flag
152-192| I/O flag | 설명 |
|---|---|
| ERROR | I/O error 발생 |
| UPTODATE | 최신 data 보유; file-backed page라면 in-memory revision >= on-disk revision |
| DIRTY | write되어 새 data 보유; file-backed page라면 in-memory revision > on-disk revision |
| WRITEBACK | disk로 sync 중 |
| LRU flag | 설명 |
|---|---|
| LRU | LRU list 중 하나에 속함 |
| ACTIVE | active LRU list에 속함 |
| UNEVICTABLE | pin되어 reclaim 대상이 아닌 unevictable (non-)LRU page |
| REFERENCED | 마지막 LRU enqueue/requeue 뒤 reference됨 |
| RECLAIM | pageout I/O 완료 뒤 곧 reclaim될 page |
| MMAP | memory-mapped page |
| ANON | file 일부가 아닌 memory-mapped page |
| SWAPCACHE | swap entry가 연결되어 swap space에 mapping됨 |
| SWAPBACKED | swap/RAM으로 backing됨 |
`UNEVICTABLE`은 ramfs page, `shmctl(SHM_LOCK)`, `mlock()` memory segment처럼 pin되어 LRU reclaim 후보가 아닌 page입니다. `RECLAIM`은 pageout I/O가 끝난 뒤 곧 reclaim될 page를 나타냅니다.
`tools/mm/page-types`로 위 flag들을 조회할 수 있습니다.
정렬과 이전 bit 의미
215-226어느 pagemap 관련 file이든 read 시작 offset이 8-byte boundary가 아니거나 read size가 8의 배수가 아니면 `-EINVAL`을 반환합니다.
Linux 3.11 이전에는 pagemap bit 55-60이 대부분 architecture에서 항상 12인 `page-shift`에 쓰였습니다. Linux 3.11부터 첫 soft-dirty bit clear 뒤 의미가 바뀌었고, Linux 4.2부터는 조건 없이 flag로 사용합니다.
PAGEMAP_SCAN IOCTL
227-254pagemap file의 `PAGEMAP_SCAN` IOCTL은 page table entry 정보를 얻고 선택적으로 clear하는 데 사용합니다. Output buffer가 있으면 address range를 scan해 기준에 맞는 memory range를 반환합니다.
`PM_SCAN_WP_MATCHING`은 관심 page를 write-protect합니다. `PM_SCAN_CHECK_WPASYNC`는 async write-protected가 아닌 page를 만나면 operation을 중단합니다. 두 flag는 함께 또는 `PM_SCAN_WP_MATCHING`만 사용할 수 있으며, range 조회와 write protection을 하나의 atomic operation으로 결합할 수도 있습니다.
| Category | 의미 |
|---|---|
| PAGE_IS_WPALLOWED | async write protection이 enabled |
| PAGE_IS_WRITTEN | write-protect된 뒤 write됨 |
| PAGE_IS_FILE | file-backed page |
| PAGE_IS_PRESENT | memory에 present |
| PAGE_IS_SWAPPED | swap에 있음 |
| PAGE_IS_PFNZERO | PFN이 0 |
| PAGE_IS_HUGE | PMD-mapped THP 또는 Hugetlb-backed page |
| PAGE_IS_SOFT_DIRTY | soft-dirty page |
| PAGE_IS_GUARD | guard region의 일부 |
struct pm_scan_arg
255-272`struct pm_scan_arg`가 IOCTL argument로 사용됩니다.
| Field | 의미 |
|---|---|
| size | struct pm_scan_arg 크기; 향후 extension 식별에 사용 |
| flags | 현재 PM_SCAN_WP_MATCHING, PM_SCAN_CHECK_WPASYNC; output buffer 유무에 따라 get 선택 |
| start, end | scan address range |
| end_walk | user buffer full 등으로 range 전체 방문 전에 중단됐을 때 walk 종료 address |
| vec, vec_len | struct page_region output array와 크기 |
| max_pages | 선택적 최대 요청 page 수 |
| category_mask, category_anyof_mask, category_inverted, return_mask | match 조건과 반환 category mask |
Walk는 user buffer가 가득 차는 등의 이유로 전체 range를 방문하기 전에 끝날 수 있으며, 실제 종료 address는 `end_walk`에 기록됩니다. `category_mask`는 모두 일치해야 할 category, `category_anyof_mask`는 하나 이상 일치할 category, `category_inverted`는 반전 조건, `return_mask`는 userspace로 돌려줄 category를 정합니다.
PAGEMAP_SCAN 예제
273-300다음은 write된 page를 찾고 동시에 write-protect하는 원문 예제입니다.
struct pm_scan_arg arg = {
.size = sizeof(arg),
.flags = PM_SCAN_CHECK_WPASYNC | PM_SCAN_CHECK_WPASYNC,
..
.category_mask = PAGE_IS_WRITTEN,
.return_mask = PAGE_IS_WRITTEN,
};첫 예제의 `.flags = PM_SCAN_CHECK_WPASYNC | PM_SCAN_CHECK_WPASYNC` 중복은 Linux v6.18.37 원문을 그대로 보존한 것입니다.
다음은 write됐고 file-backed이며 swapped되지 않았고 present 또는 huge인 page를 찾는 예제입니다.
struct pm_scan_arg arg = {
.size = sizeof(arg),
.flags = 0,
..
.category_mask = PAGE_IS_WRITTEN | PAGE_IS_SWAPPED,
.category_inverted = PAGE_IS_SWAPPED,
.category_anyof_mask = PAGE_IS_PRESENT | PAGE_IS_HUGE,
.return_mask = PAGE_IS_WRITTEN | PAGE_IS_SWAPPED |
PAGE_IS_PRESENT | PAGE_IS_HUGE,
};`PAGE_IS_WRITTEN`은 soft-dirty flag보다 성능이 좋은 대안으로 볼 수 있습니다. Kernel의 VMA merging 영향을 받지 않아 normal page의 실제 soft-dirty page를 찾을 수 있습니다. 다만 THP 또는 Hugetlb page에서는 추가 dirty page가 보고될 수 있습니다.
Userspace dirty tracking workflow
301-316`PAGE_IS_WRITTEN` category는 userfaultfd write-protect-enabled range와 함께 userspace memory dirty tracking을 구현하는 데 사용합니다.
| 단계 | 작업 |
|---|---|
| 1 | userfaultfd syscall로 userfaultfd file descriptor 생성 |
| 2 | UFFDIO_API IOCTL로 UFFD_FEATURE_WP_UNPOPULATED와 UFFD_FEATURE_WP_ASYNC 설정 |
| 3 | UFFDIO_REGISTER IOCTL과 UFFDIO_REGISTER_MODE_WP로 memory range 등록 |
| 4 | PM_SCAN_WP_MATCHING PAGEMAP_SCAN 또는 UFFDIO_WRITEPROTECT로 range write-protect |
| 5 | PAGEMAP_SCAN으로 마지막 mark 뒤 write된 page를 찾고 선택적으로 다시 write-protect |
`PAGEMAP_SCAN`의 `PM_SCAN_WP_MATCHING`과 `UFFDIO_WRITEPROTECT`는 같은 write-protect operation을 수행하지만 전자가 성능 면에서 더 좋습니다. 이후 `PAGEMAP_SCAN`으로 마지막 표시 뒤 write된 page만 찾거나, 찾는 동시에 다시 write-protect할 수 있습니다.
운영 핵심
pagemap.rst:1-316pagemap 계열 interface는 virtual address와 PFN, mapping count, page flag, memory cgroup charge를 연결합니다. PFN 권한 제한과 THP·shared memory의 보수적 semantics를 고려해야 하며, 대량 dirty tracking에는 `PAGEMAP_SCAN`과 async userfaultfd write protection을 함께 사용합니다.