1 page owner: Tracking about who allocated each page
2 -----------------------------------------------------------
6 page owner is for the tracking about who allocated each page.
7 It can be used to debug memory leak or to find a memory hogger.
8 When allocation happens, information about allocation such as call stack
9 and order of pages is stored into certain storage for each page.
10 When we need to know about status of all pages, we can get and analyze
13 Although we already have tracepoint for tracing page allocation/free,
14 using it for analyzing who allocate each page is rather complex. We need
15 to enlarge the trace buffer for preventing overlapping until userspace
16 program launched. And, launched program continually dump out the trace
17 buffer for later analysis and it would change system behviour with more
18 possibility rather than just keeping it in memory, so bad for debugging.
20 page owner can also be used for various purposes. For example, accurate
21 fragmentation statistics can be obtained through gfp flag information of
22 each page. It is already implemented and activated if page owner is
23 enabled. Other usages are more than welcome.
25 page owner is disabled in default. So, if you'd like to use it, you need
26 to add "page_owner=on" into your boot cmdline. If the kernel is built
27 with page owner and page owner is disabled in runtime due to no enabling
28 boot option, runtime overhead is marginal. If disabled in runtime, it
29 doesn't require memory to store owner information, so there is no runtime
30 memory overhead. And, page owner inserts just two unlikely branches into
31 the page allocator hotpath and if not enabled, then allocation is done
32 like as the kernel without page owner. These two unlikely branches should
33 not affect to allocation performance, especially if the static keys jump
34 label patching functionality is available. Following is the kernel's code
35 size change due to this facility.
38 text data bss dec hex filename
39 40662 1493 644 42799 a72f mm/page_alloc.o
42 text data bss dec hex filename
43 40892 1493 644 43029 a815 mm/page_alloc.o
44 1427 24 8 1459 5b3 mm/page_ext.o
45 2722 50 0 2772 ad4 mm/page_owner.o
47 Although, roughly, 4 KB code is added in total, page_alloc.o increase by
48 230 bytes and only half of it is in hotpath. Building the kernel with
49 page owner and turning it on if needed would be great option to debug
50 kernel memory problem.
52 There is one notice that is caused by implementation detail. page owner
53 stores information into the memory from struct page extension. This memory
54 is initialized some time later than that page allocator starts in sparse
55 memory system, so, until initialization, many pages can be allocated and
56 they would have no owner information. To fix it up, these early allocated
57 pages are investigated and marked as allocated in initialization phase.
58 Although it doesn't mean that they have the right owner information,
59 at least, we can tell whether the page is allocated or not,
60 more accurately. On 2GB memory x86-64 VM box, 13343 early allocated pages
61 are catched and marked, although they are mostly allocated from struct
62 page extension feature. Anyway, after that, no page is left in
67 1) Build user-space helper
72 Add "page_owner=on" to boot cmdline.
74 3) Do the job what you want to debug
76 4) Analyze information from page owner
77 cat /sys/kernel/debug/page_owner > page_owner_full.txt
78 grep -v ^PFN page_owner_full.txt > page_owner.txt
79 ./page_owner_sort page_owner.txt sorted_page_owner.txt
81 See the result about who allocated each page
82 in the sorted_page_owner.txt.