From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
To: Johannes Weiner <hannes@cmpxchg.org>
Cc: "linux-mm@kvack.org" <linux-mm@kvack.org>,
"balbir@linux.vnet.ibm.com" <balbir@linux.vnet.ibm.com>,
"nishimura@mxp.nes.nec.co.jp" <nishimura@mxp.nes.nec.co.jp>,
"hugh.dickins@tiscali.co.uk" <hugh.dickins@tiscali.co.uk>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [RFC][PATCH 3/5] count cache-only swaps
Date: Wed, 27 May 2009 08:49:50 +0900 [thread overview]
Message-ID: <20090527084950.db645ae1.kamezawa.hiroyu@jp.fujitsu.com> (raw)
In-Reply-To: <20090526173736.GA2843@cmpxchg.org>
On Tue, 26 May 2009 19:37:36 +0200
Johannes Weiner <hannes@cmpxchg.org> wrote:
> On Tue, May 26, 2009 at 12:16:38PM +0900, KAMEZAWA Hiroyuki wrote:
> > From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> >
> > This patch adds a counter for unused swap caches.
> > Maybe useful to see "we're really under shortage of swap".
> >
> > The value can be seen as kernel message at Sysrq-m etc.
> >
> > Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> > ---
> > include/linux/swap.h | 3 +++
> > mm/swap_state.c | 2 ++
> > mm/swapfile.c | 23 ++++++++++++++++++++---
> > 3 files changed, 25 insertions(+), 3 deletions(-)
> >
> > Index: new-trial-swapcount/include/linux/swap.h
> > ===================================================================
> > --- new-trial-swapcount.orig/include/linux/swap.h
> > +++ new-trial-swapcount/include/linux/swap.h
> > @@ -155,6 +155,7 @@ struct swap_info_struct {
> > unsigned int max;
> > unsigned int inuse_pages;
> > unsigned int old_block_size;
> > + unsigned int cache_only;
> > };
> >
> > struct swap_list_t {
> > @@ -298,6 +299,7 @@ extern struct page *swapin_readahead(swp
> > /* linux/mm/swapfile.c */
> > extern long nr_swap_pages;
> > extern long total_swap_pages;
> > +extern long nr_cache_only_swaps;
> > extern void si_swapinfo(struct sysinfo *);
> > extern swp_entry_t get_swap_page(void);
> > extern swp_entry_t get_swap_page_of_type(int);
> > @@ -358,6 +360,7 @@ static inline void mem_cgroup_uncharge_s
> > #define nr_swap_pages 0L
> > #define total_swap_pages 0L
> > #define total_swapcache_pages 0UL
> > +#define nr_cache_only_swaps 0UL
> >
> > #define si_swapinfo(val) \
> > do { (val)->freeswap = (val)->totalswap = 0; } while (0)
> > Index: new-trial-swapcount/mm/swapfile.c
> > ===================================================================
> > --- new-trial-swapcount.orig/mm/swapfile.c
> > +++ new-trial-swapcount/mm/swapfile.c
> > @@ -39,6 +39,7 @@ static DEFINE_SPINLOCK(swap_lock);
> > static unsigned int nr_swapfiles;
> > long nr_swap_pages;
> > long total_swap_pages;
> > +long nr_cache_only_swaps;
> > static int swap_overflow;
> > static int least_priority;
> >
> > @@ -306,9 +307,11 @@ checks:
> > si->lowest_bit = si->max;
> > si->highest_bit = 0;
> > }
> > - if (cache) /* at usual swap-out via vmscan.c */
> > + if (cache) {/* at usual swap-out via vmscan.c */
> > si->swap_map[offset] = make_swap_count(0, 1);
> > - else /* at suspend */
> > + si->cache_only++;
> > + nr_cache_only_swaps++;
> > + } else /* at suspend */
> > si->swap_map[offset] = make_swap_count(1, 0);
> > si->cluster_next = offset + 1;
> > si->flags -= SWP_SCANNING;
> > @@ -513,7 +516,10 @@ static int swap_entry_free(struct swap_i
> > } else { /* dropping swap cache flag */
> > VM_BUG_ON(!has_cache);
> > p->swap_map[offset] = make_swap_count(count, 0);
> > -
> > + if (!count) {
> > + p->cache_only--;
> > + nr_cache_only_swaps--;
> > + }
> > }
> > /* return code. */
> > count = p->swap_map[offset];
> > @@ -529,6 +535,11 @@ static int swap_entry_free(struct swap_i
> > p->inuse_pages--;
> > mem_cgroup_uncharge_swap(ent);
> > }
> > + if (swap_has_cache(count) && !swap_count(count)) {
> > + nr_cache_only_swaps++;
> > + p->cache_only++;
> > + }
> > +
> > return count;
> > }
> >
> > @@ -1128,6 +1139,8 @@ static int try_to_unuse(unsigned int typ
> > if (swap_count(*swap_map) == SWAP_MAP_MAX) {
> > spin_lock(&swap_lock);
> > *swap_map = make_swap_count(0, 1);
> > + si->cache_only++;
> > + nr_cache_only_swaps++;
> > spin_unlock(&swap_lock);
> > reset_overflow = 1;
> > }
> > @@ -2033,6 +2046,10 @@ static int __swap_duplicate(swp_entry_t
> > if (count < SWAP_MAP_MAX - 1) {
> > p->swap_map[offset] = make_swap_count(count + 1,
> > has_cache);
> > + if (has_cache && !count) {
> > + p->cache_only--;
> > + nr_cache_only_swaps--;
> > + }
> > result = 1;
> > } else if (count <= SWAP_MAP_MAX) {
> > if (swap_overflow++ < 5)
> > Index: new-trial-swapcount/mm/swap_state.c
> > ===================================================================
> > --- new-trial-swapcount.orig/mm/swap_state.c
> > +++ new-trial-swapcount/mm/swap_state.c
> > @@ -63,6 +63,8 @@ void show_swap_cache_info(void)
> > swap_cache_info.find_success, swap_cache_info.find_total);
> > printk("Free swap = %ldkB\n", nr_swap_pages << (PAGE_SHIFT - 10));
> > printk("Total swap = %lukB\n", total_swap_pages << (PAGE_SHIFT - 10));
> > + printk("Cache only swap = %lukB\n",
> > + nr_cache_only_swaps << (PAGE_SHIFT - 10));
> > }
>
> This is shown rather seldomly (sysrq and oom), for that purpose two
> counters are overkill. Maybe remove the global one and sum up the
> per-swapdevice counters on demand?
>
One for scanning (as 5/5), One for checking we should scan or not.
But yes, I feel 2 counters may be overkilling....
Maybe I'll remove global counter in the next version.
Thanks,
-Kame
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2009-05-26 23:50 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-05-26 3:12 [RFC][PATCH] memcg: fix swap account (26/May)[0/5] KAMEZAWA Hiroyuki
2009-05-26 3:14 ` [RFC][PATCH 1/5] change swap cache interfaces KAMEZAWA Hiroyuki
2009-05-26 3:15 ` [RFC][PATCH 2/5] add SWAP_HAS_CACHE flag to swap_map KAMEZAWA Hiroyuki
2009-05-27 4:02 ` Daisuke Nishimura
2009-05-27 4:36 ` KAMEZAWA Hiroyuki
2009-05-27 5:00 ` Daisuke Nishimura
2009-05-28 0:41 ` Daisuke Nishimura
2009-05-28 1:05 ` KAMEZAWA Hiroyuki
2009-05-28 1:40 ` Daisuke Nishimura
2009-05-28 1:44 ` KAMEZAWA Hiroyuki
2009-05-26 3:16 ` [RFC][PATCH 3/5] count cache-only swaps KAMEZAWA Hiroyuki
2009-05-26 17:37 ` Johannes Weiner
2009-05-26 23:49 ` KAMEZAWA Hiroyuki [this message]
2009-05-26 3:17 ` [RFC][PATCH 4/5] memcg: fix swap account KAMEZAWA Hiroyuki
2009-05-26 3:18 ` [RFC][PATCH 5/5] (experimental) chase and free cache only swap KAMEZAWA Hiroyuki
2009-05-26 18:14 ` Johannes Weiner
2009-05-27 0:08 ` KAMEZAWA Hiroyuki
2009-05-27 1:26 ` Johannes Weiner
2009-05-27 1:31 ` KAMEZAWA Hiroyuki
2009-05-27 2:06 ` Johannes Weiner
2009-05-27 5:14 ` KAMEZAWA Hiroyuki
2009-05-27 6:30 ` Daisuke Nishimura
2009-05-27 6:50 ` KAMEZAWA Hiroyuki
2009-05-27 6:43 ` [RFC][PATCH] memcg: fix swap account (26/May)[0/5] KAMEZAWA Hiroyuki
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20090527084950.db645ae1.kamezawa.hiroyu@jp.fujitsu.com \
--to=kamezawa.hiroyu@jp.fujitsu.com \
--cc=balbir@linux.vnet.ibm.com \
--cc=hannes@cmpxchg.org \
--cc=hugh.dickins@tiscali.co.uk \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=nishimura@mxp.nes.nec.co.jp \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox