All of lore.kernel.org
 help / color / mirror / Atom feed
From: Hillf Danton <hdanton@sina.com>
To: Mel Gorman <mgorman@techsingularity.net>
Cc: Linux-MM <linux-mm@kvack.org>, NeilBrown <neilb@suse.de>,
	Dave Chinner <david@fromorbit.com>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 1/5] mm/vmscan: Throttle reclaim until some writeback completes if congested
Date: Wed, 22 Sep 2021 20:16:20 +0800	[thread overview]
Message-ID: <20210922121620.4716-1-hdanton@sina.com> (raw)
In-Reply-To: <20210920085436.20939-2-mgorman@techsingularity.net>

On Mon, 20 Sep 2021 09:54:32 +0100 Mel Gorman wrote:
> +static void
> +reclaim_throttle(pg_data_t *pgdat, enum vmscan_throttle_state reason,
> +							long timeout)
> +{
> +	wait_queue_head_t *wqh =3D &pgdat->reclaim_wait;
> +	unsigned long start =3D jiffies;
> +	long ret;
> +	DEFINE_WAIT(wait);
> +
> +	atomic_inc(&pgdat->nr_reclaim_throttled);
> +	WRITE_ONCE(pgdat->nr_reclaim_start,
> +		 node_page_state(pgdat, NR_THROTTLED_WRITTEN));

Missing wakeup could happen if the current sleeper overwrites 
pgdat->nr_reclaim_start set by the existing sleeper.

	if (1 == atomic_inc_and_return(&pgdat->nr_reclaim_throttled))
		WRITE_ONCE(pgdat->nr_reclaim_start,
				node_page_state(pgdat, NR_THROTTLED_WRITTEN));

> +
> +	prepare_to_wait(wqh, &wait, TASK_INTERRUPTIBLE);
> +	ret =3D schedule_timeout(timeout);
> +	finish_wait(wqh, &wait);
> +	atomic_dec(&pgdat->nr_reclaim_throttled);
> +
> +	trace_mm_vmscan_throttled(pgdat->node_id, jiffies_to_usecs(timeout),
> +				jiffies_to_usecs(jiffies - start),
> +				reason);
> +}
> +
> +/*
> + * Account for pages written if tasks are throttled waiting on dirty
> + * pages to clean. If enough pages have been cleaned since throttling
> + * started then wakeup the throttled tasks.
> + */
> +void __acct_reclaim_writeback(pg_data_t *pgdat, struct page *page)
> +{
> +	unsigned long nr_written;
> +	int nr_throttled =3D atomic_read(&pgdat->nr_reclaim_throttled);
> +
> +	__inc_node_page_state(page, NR_THROTTLED_WRITTEN);
> +	nr_written =3D node_page_state(pgdat, NR_THROTTLED_WRITTEN) -
> +		READ_ONCE(pgdat->nr_reclaim_start);
> +
> +	if (nr_written > SWAP_CLUSTER_MAX * nr_throttled)
> +		wake_up_interruptible_all(&pgdat->reclaim_wait);
> +}


  parent reply	other threads:[~2021-09-22 12:16 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-20  8:54 [RFC PATCH 0/5] Remove dependency on congestion_wait in mm/ Mel Gorman
2021-09-20  8:54 ` [PATCH 1/5] mm/vmscan: Throttle reclaim until some writeback completes if congested Mel Gorman
2021-09-20 23:19   ` NeilBrown
2021-09-21 11:12     ` Mel Gorman
2021-09-21 21:27       ` NeilBrown
2021-09-21  0:13   ` NeilBrown
2021-09-21 10:58     ` Mel Gorman
2021-09-21 21:40       ` NeilBrown
2021-09-22  6:04       ` Dave Chinner
2021-09-22  8:03         ` Mel Gorman
2021-09-22 12:16   ` Hillf Danton [this message]
2021-09-22 14:13     ` Mel Gorman
2021-09-20  8:54 ` [PATCH 2/5] mm/vmscan: Throttle reclaim and compaction when too may pages are isolated Mel Gorman
2021-09-20 23:27   ` NeilBrown
2021-09-21 11:03     ` Mel Gorman
2021-09-21 18:45   ` Yang Shi
2021-09-21 18:45     ` Yang Shi
2021-09-22  8:11     ` Mel Gorman
2021-09-20  8:54 ` [PATCH 3/5] mm/vmscan: Throttle reclaim when no progress is being made Mel Gorman
2021-09-20 23:31   ` NeilBrown
2021-09-21 11:16     ` Mel Gorman
2021-09-21 21:46       ` NeilBrown
2021-09-22  9:21         ` Mel Gorman
2021-09-20  8:54 ` [PATCH 4/5] mm/writeback: Throttle based on page writeback instead of congestion Mel Gorman
2021-09-20  8:54 ` [PATCH 5/5] mm/page_alloc: Remove the throttling logic from the page allocator Mel Gorman
2021-09-20 11:42 ` [RFC PATCH 0/5] Remove dependency on congestion_wait in mm/ Matthew Wilcox
2021-09-20 12:50   ` Mel Gorman
2021-09-20 14:11     ` David Sterba
2021-09-21 11:18       ` Mel Gorman
2021-09-20 19:51   ` Mel Gorman
2021-09-21 20:46 ` Dave Chinner
2021-09-22 17:52   ` Mel Gorman
2021-09-29 10:09 [PATCH 0/5] Remove dependency on congestion_wait in mm/ v2 Mel Gorman
2021-09-29 10:09 ` [PATCH 1/5] mm/vmscan: Throttle reclaim until some writeback completes if congested Mel Gorman

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=20210922121620.4716-1-hdanton@sina.com \
    --to=hdanton@sina.com \
    --cc=david@fromorbit.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mgorman@techsingularity.net \
    --cc=neilb@suse.de \
    --subject='Re: [PATCH 1/5] mm/vmscan: Throttle reclaim until some writeback completes if congested' \
    /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

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.