On Fri, 28 Jan 2011 08:52:15 +0100
Johannes Weiner <hannes@xxxxxxxxxxx> wrote:
> > @@ -1853,7 +1869,14 @@ static int __mem_cgroup_do_charge(struct
> > * Check the limit again to see if the reclaim reduced the
> > * current usage of the cgroup before giving up
> > */
> > - if (ret || mem_cgroup_check_under_limit(mem_over_limit))
> > + if (mem_cgroup_check_margin(mem_over_limit) >= csize)
> > + return CHARGE_RETRY;
> > +
> > + /*
> > + * If the charge size is a PAGE_SIZE, it's not hopeless while
> > + * we can reclaim a page.
> > + */
> > + if (csize == PAGE_SIZE && ret)
> > return CHARGE_RETRY;
>
> That makes sense.
>
here.
It's okay to add your Signed-off because this one is very similar to yours.
==
From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Current memory cgroup's code tends to assume page_size == PAGE_SIZE
and arrangement for THP is not enough yet.
This is one of fixes for supporing THP. This adds
mem_cgroup_get_margin() and checks whether there are required amount of
free resource after memory reclaim. By this, THP page allocation
can know whether it really succeeded or not and avoid infinite-loop
and hangup.
Total fixes for do_charge()/reclaim memory will follow this patch.
Changelog:
- rename check_margin -> get_margin.
- style fix.
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
---
include/linux/res_counter.h | 11 +++++++++++
mm/memcontrol.c | 22 +++++++++++++++++++++-
2 files changed, 32 insertions(+), 1 deletion(-)
Index: mmotm-0125/include/linux/res_counter.h
===================================================================
--- mmotm-0125.orig/include/linux/res_counter.h
+++ mmotm-0125/include/linux/res_counter.h
@@ -182,6 +182,17 @@ static inline bool res_counter_check_und
return ret;
}
+static inline s64 res_counter_get_margin(struct res_counter *cnt)
+{
+ s64 ret;
+ unsigned long flags;
+
+ spin_lock_irqsave(&cnt->lock, flags);
+ ret = cnt->limit - cnt->usage;
+ spin_unlock_irqrestore(&cnt->lock, flags);
+ return ret;
+}
+
static inline bool res_counter_check_under_soft_limit(struct res_counter *cnt)
{
bool ret;
Index: mmotm-0125/mm/memcontrol.c
===================================================================
--- mmotm-0125.orig/mm/memcontrol.c
+++ mmotm-0125/mm/memcontrol.c
@@ -1111,6 +1111,19 @@ static bool mem_cgroup_check_under_limit
return false;
}
+static s64 mem_cgroup_get_margin(struct mem_cgroup *mem)
+{
+ s64 mem_margin = res_counter_get_margin(&mem->res);
+ s64 memsw_margin;
+
+ if (do_swap_account)
+ memsw_margin = res_counter_get_margin(&mem->memsw);
+ else
+ memsw_margin = RESOURCE_MAX;
+
+ return min(mem_margin, memsw_margin);
+}
+
static unsigned int get_swappiness(struct mem_cgroup *memcg)
{
struct cgroup *cgrp = memcg->css.cgroup;
@@ -1853,7 +1866,14 @@ static int __mem_cgroup_do_charge(struct
* Check the limit again to see if the reclaim reduced the
* current usage of the cgroup before giving up
*/
- if (ret || mem_cgroup_check_under_limit(mem_over_limit))
+ if (mem_cgroup_get_margin(mem_over_limit) >= csize)
+ return CHARGE_RETRY;
+
+ /*
+ * If the charge size is a PAGE_SIZE, it's not hopeless while
+ * we can reclaim a page.
+ */
+ if (csize == PAGE_SIZE && ret)
return CHARGE_RETRY;
/*
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>