@@ -133,18 +133,9 @@ Under below explanation, we assume CONFIG_MEM_RES_CTRL_SWAP=y.
8. LRU
======
- Each memcg has its own private LRU. Now, its handling is under global
- VM's control (means that it's handled under global pgdat->lru_lock).
- Almost all routines around memcg's LRU is called by global LRU's
- list management functions under pgdat->lru_lock.
-
- A special function is mem_cgroup_isolate_pages(). This scans
- memcg's private LRU and call __isolate_lru_page() to extract a page
- from LRU.
-
- (By __isolate_lru_page(), the page is removed from both of global and
- private LRU.)
-
+ Each memcg has its own vector of LRUs (inactive anon, active anon,
+ inactive file, active file, unevictable) of pages from each node,
+ each LRU handled under a single lru_lock for that memcg and node.
9. Typical Tests.
=================
@@ -285,20 +285,17 @@ When oom event notifier is registered, event will be delivered.
2.6 Locking
-----------
- lock_page_cgroup()/unlock_page_cgroup() should not be called under
- the i_pages lock.
+Lock order is as follows:
- Other lock order is following:
+ Page lock (PG_locked bit of page->flags)
+ mm->page_table_lock or split pte_lock
+ lock_page_memcg (memcg->move_lock)
+ mapping->i_pages lock
+ lruvec->lru_lock.
- PG_locked.
- mm->page_table_lock
- pgdat->lru_lock
- lock_page_cgroup.
-
- In many cases, just lock_page_cgroup() is called.
-
- per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by
- pgdat->lru_lock, it has no lock of its own.
+Per-node-per-memcgroup LRU (cgroup's private LRU) is guarded by
+lruvec->lru_lock; PG_lru bit of page->flags is cleared before
+isolating a page from its LRU under lruvec->lru_lock.
2.7 Kernel Memory Extension (CONFIG_MEMCG_KMEM)
-----------------------------------------------
@@ -69,7 +69,7 @@ When pages are freed in batch, the also mm_page_free_batched is triggered.
Broadly speaking, pages are taken off the LRU lock in bulk and
freed in batch with a page list. Significant amounts of activity here could
indicate that the system is under memory pressure and can also indicate
-contention on the zone->lru_lock.
+contention on the lruvec->lru_lock.
4. Per-CPU Allocator Activity
=============================
@@ -33,7 +33,7 @@ reclaim in Linux. The problems have been observed at customer sites on large
memory x86_64 systems.
To illustrate this with an example, a non-NUMA x86_64 platform with 128GB of
-main memory will have over 32 million 4k pages in a single zone. When a large
+main memory will have over 32 million 4k pages in a single node. When a large
fraction of these pages are not evictable for any reason [see below], vmscan
will spend a lot of time scanning the LRU lists looking for the small fraction
of pages that are evictable. This can result in a situation where all CPUs are
@@ -55,7 +55,7 @@ unevictable, either by definition or by circumstance, in the future.
The Unevictable Page List
-------------------------
-The Unevictable LRU infrastructure consists of an additional, per-zone, LRU list
+The Unevictable LRU infrastructure consists of an additional, per-node, LRU list
called the "unevictable" list and an associated page flag, PG_unevictable, to
indicate that the page is being managed on the unevictable list.
@@ -84,15 +84,9 @@ The unevictable list does not differentiate between file-backed and anonymous,
swap-backed pages. This differentiation is only important while the pages are,
in fact, evictable.
-The unevictable list benefits from the "arrayification" of the per-zone LRU
+The unevictable list benefits from the "arrayification" of the per-node LRU
lists and statistics originally proposed and posted by Christoph Lameter.
-The unevictable list does not use the LRU pagevec mechanism. Rather,
-unevictable pages are placed directly on the page's zone's unevictable list
-under the zone lru_lock. This allows us to prevent the stranding of pages on
-the unevictable list when one task has the page isolated from the LRU and other
-tasks are changing the "evictability" state of the page.
-
Memory Control Group Interaction
--------------------------------
@@ -101,8 +95,8 @@ The unevictable LRU facility interacts with the memory control group [aka
memory controller; see Documentation/admin-guide/cgroup-v1/memory.rst] by extending the
lru_list enum.
-The memory controller data structure automatically gets a per-zone unevictable
-list as a result of the "arrayification" of the per-zone LRU lists (one per
+The memory controller data structure automatically gets a per-node unevictable
+list as a result of the "arrayification" of the per-node LRU lists (one per
lru_list enum element). The memory controller tracks the movement of pages to
and from the unevictable list.
@@ -196,7 +190,7 @@ for the sake of expediency, to leave a unevictable page on one of the regular
active/inactive LRU lists for vmscan to deal with. vmscan checks for such
pages in all of the shrink_{active|inactive|page}_list() functions and will
"cull" such pages that it encounters: that is, it diverts those pages to the
-unevictable list for the zone being scanned.
+unevictable list for the node being scanned.
There may be situations where a page is mapped into a VM_LOCKED VMA, but the
page is not marked as PG_mlocked. Such pages will make it all the way to
@@ -328,7 +322,7 @@ If the page was NOT already mlocked, mlock_vma_page() attempts to isolate the
page from the LRU, as it is likely on the appropriate active or inactive list
at that time. If the isolate_lru_page() succeeds, mlock_vma_page() will put
back the page - by calling putback_lru_page() - which will notice that the page
-is now mlocked and divert the page to the zone's unevictable list. If
+is now mlocked and divert the page to the node's unevictable list. If
mlock_vma_page() is unable to isolate the page from the LRU, vmscan will handle
it later if and when it attempts to reclaim the page.
@@ -603,7 +597,7 @@ Some examples of these unevictable pages on the LRU lists are:
unevictable list in mlock_vma_page().
shrink_inactive_list() also diverts any unevictable pages that it finds on the
-inactive lists to the appropriate zone's unevictable list.
+inactive lists to the appropriate node's unevictable list.
shrink_inactive_list() should only see SHM_LOCK'd pages that became SHM_LOCK'd
after shrink_active_list() had moved them to the inactive list, or pages mapped
@@ -78,7 +78,7 @@ struct page {
struct { /* Page cache and anonymous pages */
/**
* @lru: Pageout list, eg. active_list protected by
- * pgdat->lru_lock. Sometimes used as a generic list
+ * lruvec->lru_lock. Sometimes used as a generic list
* by the page owner.
*/
struct list_head lru;
@@ -113,8 +113,7 @@ static inline bool free_area_empty(struct free_area *area, int migratetype)
struct pglist_data;
/*
- * zone->lock and the zone lru_lock are two of the hottest locks in the kernel.
- * So add a wild amount of padding here to ensure that they fall into separate
+ * Add a wild amount of padding here to ensure datas fall into separate
* cachelines. There are very few zone structures in the machine, so space
* consumption is not a concern here.
*/
@@ -102,8 +102,8 @@
* ->swap_lock (try_to_unmap_one)
* ->private_lock (try_to_unmap_one)
* ->i_pages lock (try_to_unmap_one)
- * ->pgdat->lru_lock (follow_page->mark_page_accessed)
- * ->pgdat->lru_lock (check_pte_range->isolate_lru_page)
+ * ->lruvec->lru_lock (follow_page->mark_page_accessed)
+ * ->lruvec->lru_lock (check_pte_range->isolate_lru_page)
* ->private_lock (page_remove_rmap->set_page_dirty)
* ->i_pages lock (page_remove_rmap->set_page_dirty)
* bdi.wb->list_lock (page_remove_rmap->set_page_dirty)
@@ -28,12 +28,12 @@
* hugetlb_fault_mutex (hugetlbfs specific page fault mutex)
* anon_vma->rwsem
* mm->page_table_lock or pte_lock
- * pgdat->lru_lock (in mark_page_accessed, isolate_lru_page)
* swap_lock (in swap_duplicate, swap_info_get)
* mmlist_lock (in mmput, drain_mmlist and others)
* mapping->private_lock (in __set_page_dirty_buffers)
- * mem_cgroup_{begin,end}_page_stat (memcg->move_lock)
+ * lock_page_memcg move_lock (in __set_page_dirty_buffers)
* i_pages lock (widely used)
+ * lruvec->lru_lock (in lock_page_lruvec_irq)
* inode->i_lock (in set_page_dirty's __mark_inode_dirty)
* bdi.wb->list_lock (in set_page_dirty's __mark_inode_dirty)
* sb_lock (within inode_lock in fs/fs-writeback.c)
@@ -1614,14 +1614,16 @@ static __always_inline void update_lru_sizes(struct lruvec *lruvec,
}
/**
- * pgdat->lru_lock is heavily contended. Some of the functions that
+ * Isolating page from the lruvec to fill in @dst list by nr_to_scan times.
+ *
+ * lruvec->lru_lock is heavily contended. Some of the functions that
* shrink the lists perform better by taking out a batch of pages
* and working on them outside the LRU lock.
*
* For pagecache intensive workloads, this function is the hottest
* spot in the kernel (apart from copy_*_user functions).
*
- * Appropriate locks must be held before calling this function.
+ * Lru_lock must be held before calling this function.
*
* @nr_to_scan: The number of eligible pages to look through on the list.
* @lruvec: The LRU vector to pull pages from.
@@ -1815,25 +1817,11 @@ static int too_many_isolated(struct pglist_data *pgdat, int file,
}
/*
- * This moves pages from @list to corresponding LRU list.
- *
- * We move them the other way if the page is referenced by one or more
- * processes, from rmap.
- *
- * If the pages are mostly unmapped, the processing is fast and it is
- * appropriate to hold zone_lru_lock across the whole operation. But if
- * the pages are mapped, the processing is slow (page_referenced()) so we
- * should drop zone_lru_lock around each page. It's impossible to balance
- * this, so instead we remove the pages from the LRU while processing them.
- * It is safe to rely on PG_active against the non-LRU pages in here because
- * nobody will play with that bit on a non-LRU page.
- *
- * The downside is that we have to touch page->_refcount against each page.
- * But we had to alter page->flags anyway.
+ * move_pages_to_lru() moves pages from private @list to appropriate LRU list.
+ * On return, @list is reused as a list of pages to be freed by the caller.
*
* Returns the number of pages moved to the given lruvec.
*/
-
static unsigned noinline_for_stack move_pages_to_lru(struct lruvec *lruvec,
struct list_head *list)
{
@@ -2012,6 +2000,23 @@ static int current_may_throttle(void)
return nr_reclaimed;
}
+/*
+ * shrink_active_list() moves pages from the active LRU to the inactive LRU.
+ *
+ * We move them the other way if the page is referenced by one or more
+ * processes.
+ *
+ * If the pages are mostly unmapped, the processing is fast and it is
+ * appropriate to hold lru_lock across the whole operation. But if
+ * the pages are mapped, the processing is slow (page_referenced()), so
+ * we should drop lru_lock around each page. It's impossible to balance
+ * this, so instead we remove the pages from the LRU while processing them.
+ * It is safe to rely on PG_active against the non-LRU pages in here because
+ * nobody will play with that bit on a non-LRU page.
+ *
+ * The downside is that we have to touch page->_refcount against each page.
+ * But we had to alter page->flags anyway.
+ */
static void shrink_active_list(unsigned long nr_to_scan,
struct lruvec *lruvec,
struct scan_control *sc,