[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <1e5938cd-0e03-6cb9-4d5f-fee94fc1479e@suse.cz>
Date: Thu, 14 Jul 2016 17:22:25 +0200
From: Vlastimil Babka <vbabka@...e.cz>
To: Mel Gorman <mgorman@...hsingularity.net>,
Andrew Morton <akpm@...ux-foundation.org>,
Linux-MM <linux-mm@...ck.org>
Cc: Johannes Weiner <hannes@...xchg.org>,
Minchan Kim <minchan@...nel.org>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/4] mm, page_alloc: fix dirtyable highmem calculation
On 07/13/2016 12:00 PM, Mel Gorman wrote:
> From: Minchan Kim <minchan@...nel.org>
>
> Note from Mel: This may optionally be considered a fix to the mmotm patch
> mm-page_alloc-consider-dirtyable-memory-in-terms-of-nodes.patch
> but if so, please preserve credit for Minchan.
>
> When I tested vmscale in mmtest in 32bit, I found the benchmark was slow
> down 0.5 times.
>
> base node
> 1 global-1
> User 12.98 16.04
> System 147.61 166.42
> Elapsed 26.48 38.08
>
> With vmstat, I found IO wait avg is much increased compared to base.
>
> The reason was highmem_dirtyable_memory accumulates free pages and
> highmem_file_pages from HIGHMEM to MOVABLE zones which was wrong. With
> that, dirth_thresh in throtlle_vm_write is always 0 so that it calls
> congestion_wait frequently if writeback starts.
>
> With this patch, it is much recovered.
>
> base node fi
> 1 global-1 fix
> User 12.98 16.04 13.78
> System 147.61 166.42 143.92
> Elapsed 26.48 38.08 29.64
>
> Signed-off-by: Minchan Kim <minchan@...nel.org>
> Signed-off-by: Mel Gorman <mgorman@...hsingularity.net>
Acked-by: Vlastimil Babka <vbabka@...e.cz>
Just some nitpicks:
> ---
> mm/page-writeback.c | 16 ++++++++++------
> 1 file changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 0bca2376bd42..7b41d1290783 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -307,27 +307,31 @@ static unsigned long highmem_dirtyable_memory(unsigned long total)
> {
> #ifdef CONFIG_HIGHMEM
> int node;
> - unsigned long x = 0;
> + unsigned long x;
> int i;
> - unsigned long dirtyable = atomic_read(&highmem_file_pages);
> + unsigned long dirtyable = 0;
This wasn't necessary?
>
> for_each_node_state(node, N_HIGH_MEMORY) {
> for (i = ZONE_NORMAL + 1; i < MAX_NR_ZONES; i++) {
> struct zone *z;
> + unsigned long nr_pages;
>
> if (!is_highmem_idx(i))
> continue;
>
> z = &NODE_DATA(node)->node_zones[i];
> - dirtyable += zone_page_state(z, NR_FREE_PAGES);
> + if (!populated_zone(z))
> + continue;
>
> + nr_pages = zone_page_state(z, NR_FREE_PAGES);
> /* watch for underflows */
> - dirtyable -= min(dirtyable, high_wmark_pages(z));
> -
> - x += dirtyable;
> + nr_pages -= min(nr_pages, high_wmark_pages(z));
> + dirtyable += nr_pages;
> }
> }
>
> + x = dirtyable + atomic_read(&highmem_file_pages);
And then this addition wouldn't be necessary. BTW I think we could also
ditch the "x" variable and just use the "dirtyable" for the rest of the
function.
> +
> /*
> * Unreclaimable memory (kernel memory or anonymous memory
> * without swap) can bring down the dirtyable pages below
>
Powered by blists - more mailing lists