Subject: + memcg-add-per-cgroup-writeback-pages-accounting.patch added to -mm tree
To: handai.szj@xxxxxxxxxx,fengguang.wu@xxxxxxxxx,gthelen@xxxxxxxxxx,hannes@xxxxxxxxxxx,kamezawa.hiroyu@xxxxxxxxxxxxxx,mhocko@xxxxxxx
From: akpm@xxxxxxxxxxxxxxxxxxxx
Date: Thu, 22 Aug 2013 15:43:33 -0700
The patch titled
Subject: memcg: add per cgroup writeback pages accounting
has been added to the -mm tree. Its filename is
memcg-add-per-cgroup-writeback-pages-accounting.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/memcg-add-per-cgroup-writeback-pages-accounting.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/memcg-add-per-cgroup-writeback-pages-accounting.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Sha Zhengju <handai.szj@xxxxxxxxxx>
Subject: memcg: add per cgroup writeback pages accounting
Add memcg routines to count writeback pages, later dirty pages will also
be accounted.
After Kame's commit 89c06bd5 ("memcg: use new logic for page stat
accounting"), we can use 'struct page' flag to test page state instead of
per page_cgroup flag. But memcg has a feature to move a page from a
cgroup to another one and may have race between "move" and "page stat
accounting". So in order to avoid the race we have designed a new lock:
mem_cgroup_begin_update_page_stat()
modify page information -->(a)
mem_cgroup_update_page_stat() -->(b)
mem_cgroup_end_update_page_stat()
It requires both (a) and (b)(writeback pages accounting) to be pretected
in mem_cgroup_{begin/end}_update_page_stat(). It's full no-op for
!CONFIG_MEMCG, almost no-op if memcg is disabled (but compiled in), rcu
read lock in the most cases (no task is moving), and spin_lock_irqsave on
top in the slow path.
There're two writeback interfaces to modify: test_{clear/set}_page_writeback().
And the lock order is:
--> memcg->move_lock
--> mapping->tree_lock
Signed-off-by: Sha Zhengju <handai.szj@xxxxxxxxxx>
Acked-by: Michal Hocko <mhocko@xxxxxxx>
Reviewed-by: Greg Thelen <gthelen@xxxxxxxxxx>
Cc: Fengguang Wu <fengguang.wu@xxxxxxxxx>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/memcontrol.h | 1 +
mm/memcontrol.c | 30 +++++++++++++++++++++++-------
mm/page-writeback.c | 15 +++++++++++++++
3 files changed, 39 insertions(+), 7 deletions(-)
diff -puN include/linux/memcontrol.h~memcg-add-per-cgroup-writeback-pages-accounting include/linux/memcontrol.h
--- a/include/linux/memcontrol.h~memcg-add-per-cgroup-writeback-pages-accounting
+++ a/include/linux/memcontrol.h
@@ -42,6 +42,7 @@ enum mem_cgroup_stat_index {
MEM_CGROUP_STAT_RSS, /* # of pages charged as anon rss */
MEM_CGROUP_STAT_RSS_HUGE, /* # of pages charged as anon huge */
MEM_CGROUP_STAT_FILE_MAPPED, /* # of pages charged as file rss */
+ MEM_CGROUP_STAT_WRITEBACK, /* # of pages under writeback */
MEM_CGROUP_STAT_SWAP, /* # of pages, swapped out */
MEM_CGROUP_STAT_NSTATS,
};
diff -puN mm/memcontrol.c~memcg-add-per-cgroup-writeback-pages-accounting mm/memcontrol.c
--- a/mm/memcontrol.c~memcg-add-per-cgroup-writeback-pages-accounting
+++ a/mm/memcontrol.c
@@ -89,6 +89,7 @@ static const char * const mem_cgroup_sta
"rss",
"rss_huge",
"mapped_file",
+ "writeback",
"swap",
};
@@ -3654,6 +3655,20 @@ void mem_cgroup_split_huge_fixup(struct
}
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+static inline
+void mem_cgroup_move_account_page_stat(struct mem_cgroup *from,
+ struct mem_cgroup *to,
+ unsigned int nr_pages,
+ enum mem_cgroup_stat_index idx)
+{
+ /* Update stat data for mem_cgroup */
+ preempt_disable();
+ WARN_ON_ONCE(from->stat->count[idx] < nr_pages);
+ __this_cpu_add(from->stat->count[idx], -nr_pages);
+ __this_cpu_add(to->stat->count[idx], nr_pages);
+ preempt_enable();
+}
+
/**
* mem_cgroup_move_account - move account of the page
* @page: the page
@@ -3699,13 +3714,14 @@ static int mem_cgroup_move_account(struc
move_lock_mem_cgroup(from, &flags);
- if (!anon && page_mapped(page)) {
- /* Update mapped_file data for mem_cgroup */
- preempt_disable();
- __this_cpu_dec(from->stat->count[MEM_CGROUP_STAT_FILE_MAPPED]);
- __this_cpu_inc(to->stat->count[MEM_CGROUP_STAT_FILE_MAPPED]);
- preempt_enable();
- }
+ if (!anon && page_mapped(page))
+ mem_cgroup_move_account_page_stat(from, to, nr_pages,
+ MEM_CGROUP_STAT_FILE_MAPPED);
+
+ if (PageWriteback(page))
+ mem_cgroup_move_account_page_stat(from, to, nr_pages,
+ MEM_CGROUP_STAT_WRITEBACK);
+
mem_cgroup_charge_statistics(from, page, anon, -nr_pages);
/* caller should have done css_get */
diff -puN mm/page-writeback.c~memcg-add-per-cgroup-writeback-pages-accounting mm/page-writeback.c
--- a/mm/page-writeback.c~memcg-add-per-cgroup-writeback-pages-accounting
+++ a/mm/page-writeback.c
@@ -2143,11 +2143,17 @@ EXPORT_SYMBOL(account_page_dirtied);
/*
* Helper function for set_page_writeback family.
+ *
+ * The caller must hold mem_cgroup_begin/end_update_page_stat() lock
+ * while calling this function.
+ * See test_set_page_writeback for example.
+ *
* NOTE: Unlike account_page_dirtied this does not rely on being atomic
* wrt interrupts.
*/
void account_page_writeback(struct page *page)
{
+ mem_cgroup_inc_page_stat(page, MEM_CGROUP_STAT_WRITEBACK);
inc_zone_page_state(page, NR_WRITEBACK);
}
EXPORT_SYMBOL(account_page_writeback);
@@ -2364,7 +2370,10 @@ int test_clear_page_writeback(struct pag
{
struct address_space *mapping = page_mapping(page);
int ret;
+ bool locked;
+ unsigned long memcg_flags;
+ mem_cgroup_begin_update_page_stat(page, &locked, &memcg_flags);
if (mapping) {
struct backing_dev_info *bdi = mapping->backing_dev_info;
unsigned long flags;
@@ -2385,9 +2394,11 @@ int test_clear_page_writeback(struct pag
ret = TestClearPageWriteback(page);
}
if (ret) {
+ mem_cgroup_dec_page_stat(page, MEM_CGROUP_STAT_WRITEBACK);
dec_zone_page_state(page, NR_WRITEBACK);
inc_zone_page_state(page, NR_WRITTEN);
}
+ mem_cgroup_end_update_page_stat(page, &locked, &memcg_flags);
return ret;
}
@@ -2395,7 +2406,10 @@ int test_set_page_writeback(struct page
{
struct address_space *mapping = page_mapping(page);
int ret;
+ bool locked;
+ unsigned long memcg_flags;
+ mem_cgroup_begin_update_page_stat(page, &locked, &memcg_flags);
if (mapping) {
struct backing_dev_info *bdi = mapping->backing_dev_info;
unsigned long flags;
@@ -2422,6 +2436,7 @@ int test_set_page_writeback(struct page
}
if (!ret)
account_page_writeback(page);
+ mem_cgroup_end_update_page_stat(page, &locked, &memcg_flags);
return ret;
}
_
Patches currently in -mm which might be from handai.szj@xxxxxxxxxx are
memcg-correct-resource_max-to-ullong_max.patch
memcg-rename-resource_max-to-res_counter_max.patch
memcg-avoid-overflow-caused-by-page_align.patch
memcg-reduce-function-dereference.patch
memcg-remove-memcg_nr_file_mapped.patch
memcg-check-for-proper-lock-held-in-mem_cgroup_update_page_stat.patch
memcg-add-per-cgroup-writeback-pages-accounting.patch
memcg-document-cgroup-dirty-writeback-memory-statistics.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
