linux - 进程smaps中 "Referenced"的含义

Question

我试图理解进程的 smaps 文件中“引用”的含义。根据内核文档:

"Referenced" indicates the amount of memory currently marked as referenced or accessed."

我理解“访问”的意思，但“引用”是什么意思？如果都是访问内存，那么我期望是“RSS”+“Swap”，但是在下面的例子中，不是这样的。

例子:

7f13e7577000-7f13e75c1000 rw-p 00000000 00:00 0                          [heap]
Size:                296 kB
Rss:                 156 kB
Pss:                 156 kB
Shared_Clean:          0 kB
Shared_Dirty:          0 kB
Private_Clean:         4 kB
Private_Dirty:       152 kB
Referenced:          128 kB
Anonymous:           156 kB
AnonHugePages:         0 kB
Swap:                 56 kB
KernelPageSize:        4 kB
MMUPageSize:           4 kB
Locked:                0 kB

Answer 1

来自Linux kernel source code定义内存页的referenced标志位，

PG_referenced, PG_reclaim are used to reclaim pages for anonymous, file-backed pagecache.

referenced 标志位用于 page_check_references()哪里被进一步记录为，

/*
* All mapped pages start out with page table
* references from the instantiating fault, so we need
* to look twice if a mapped file page is used more
* than once.
*
* Mark it and spare it for another trip around the
* inactive list. Another page table reference will
* lead to its activation.
*
* Note: the mark is set for activated pages as well
* so that recently deactivated but used pages are
* quickly recovered.
*/

还有 mark_page_accessed()是从 Linux 内核中的不同位置调用的函数记录了页面的 referenced 标志位在其生命周期的初始部分的更新:
(即从新分配页面时开始)

* Mark a page as having seen activity.
*
* inactive,unreferenced  ->  inactive,referenced
* inactive,referenced    ->  active,unreferenced
* active,unreferenced    ->  active,referenced
*
* When a newly allocated page is not yet visible, so safe for non-atomic ops,
* __SetPageReferenced(page) may be substituted for mark_page_accessed(page).
*/

i.e. referenced = active page-cache that is NOT going to be reclaimed anytime soon.