1 pagemap, from the userspace perspective
2 ---------------------------------------
4 pagemap is a new (as of 2.6.25) set of interfaces in the kernel that allow
5 userspace programs to examine the page tables and related information by
6 reading files in /proc.
8 There are three components to pagemap:
10 * /proc/pid/pagemap. This file lets a userspace process find out which
11 physical frame each virtual page is mapped to. It contains one 64-bit
12 value for each virtual page, containing the following data (from
13 fs/proc/task_mmu.c, above pagemap_read):
15 * Bits 0-54 page frame number (PFN) if present
16 * Bits 0-4 swap type if swapped
17 * Bits 5-54 swap offset if swapped
18 * Bits 55-60 page shift (page size = 1<<page shift)
19 * Bit 61 reserved for future use
23 If the page is not present but in swap, then the PFN contains an
24 encoding of the swap file number and the page's offset into the
25 swap. Unmapped pages return a null PFN. This allows determining
26 precisely which pages are mapped (or in swap) and comparing mapped
27 pages between processes.
29 Efficient users of this interface will use /proc/pid/maps to
30 determine which areas of memory are actually mapped and llseek to
31 skip over unmapped regions.
33 * /proc/kpagecount. This file contains a 64-bit count of the number of
34 times each page is mapped, indexed by PFN.
36 * /proc/kpageflags. This file contains a 64-bit set of flags for each
39 The flags are (from fs/proc/page.c, above kpageflags_read):
62 Short descriptions to the page flags:
65 page is being locked for exclusive access, eg. by undergoing read/write IO
68 page is managed by the SLAB/SLOB/SLUB/SLQB kernel memory allocator
69 When compound page is used, SLUB/SLQB will only set this flag on the head
70 page; SLOB will not flag it at all.
73 a free memory block managed by the buddy system allocator
74 The buddy system organizes free memory in blocks of various orders.
75 An order N block has 2^N physically contiguous pages, with the BUDDY flag
76 set for and _only_ for the first page.
80 A compound page with order N consists of 2^N physically contiguous pages.
81 A compound page with order 2 takes the form of "HTTT", where H donates its
82 head page and T donates its tail page(s). The major consumers of compound
83 pages are hugeTLB pages (Documentation/vm/hugetlbpage.txt), the SLUB etc.
84 memory allocators and various device drivers. However in this interface,
85 only huge/giga pages are made visible to end users.
87 this is an integral part of a HugeTLB page
90 no page frame exists at the requested address
92 [IO related page flags]
93 1. ERROR IO error occurred
94 3. UPTODATE page has up-to-date data
95 ie. for file backed page: (in-memory data revision >= on-disk one)
96 4. DIRTY page has been written to, hence contains new data
97 ie. for file backed page: (in-memory data revision > on-disk one)
98 8. WRITEBACK page is being synced to disk
100 [LRU related page flags]
101 5. LRU page is in one of the LRU lists
102 6. ACTIVE page is in the active LRU list
103 18. UNEVICTABLE page is in the unevictable (non-)LRU list
104 It is somehow pinned and not a candidate for LRU page reclaims,
105 eg. ramfs pages, shmctl(SHM_LOCK) and mlock() memory segments
106 2. REFERENCED page has been referenced since last LRU list enqueue/requeue
107 9. RECLAIM page will be reclaimed soon after its pageout IO completed
108 11. MMAP a memory mapped page
109 12. ANON a memory mapped page that is not part of a file
110 13. SWAPCACHE page is mapped to swap space, ie. has an associated swap entry
111 14. SWAPBACKED page is backed by swap/RAM
113 The page-types tool in this directory can be used to query the above flags.
115 Using pagemap to do something useful:
117 The general procedure for using pagemap to find out about a process' memory
118 usage goes like this:
120 1. Read /proc/pid/maps to determine which parts of the memory space are
122 2. Select the maps you are interested in -- all of them, or a particular
123 library, or the stack or the heap, etc.
124 3. Open /proc/pid/pagemap and seek to the pages you would like to examine.
125 4. Read a u64 for each page from pagemap.
126 5. Open /proc/kpagecount and/or /proc/kpageflags. For each PFN you just
127 read, seek to that entry in the file, and read the data you want.
129 For example, to find the "unique set size" (USS), which is the amount of
130 memory that a process is using that is not shared with any other process,
131 you can go through every map in the process, find the PFNs, look those up
132 in kpagecount, and tally up the number of pages that are only referenced
137 Reading from any of the files will return -EINVAL if you are not starting
138 the read on an 8-byte boundary (e.g., if you seeked an odd number of bytes
139 into the file), or if the size of the read is not a multiple of 8 bytes.