linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Michal Hocko <mhocko@kernel.org>
To: Yafang Shao <laoar.shao@gmail.com>
Cc: jack@suse.cz, akpm@linux-foundation.org, hannes@cmpxchg.org,
	vdavydov.dev@gmail.com, jlayton@redhat.com, nborisov@suse.com,
	tytso@mit.edu, mawilcox@microsoft.com, linux-mm@kvack.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3] mm: introduce validity check on vm dirtiness settings
Date: Tue, 26 Sep 2017 12:25:32 +0200	[thread overview]
Message-ID: <20170926102532.culqxb45xwzafomj@dhcp22.suse.cz> (raw)
In-Reply-To: <1505861015-11919-1-git-send-email-laoar.shao@gmail.com>

On Wed 20-09-17 06:43:35, Yafang Shao wrote:
> we can find the logic in domain_dirty_limits() that
> when dirty bg_thresh is bigger than dirty thresh,
> bg_thresh will be set as thresh * 1 / 2.
> 	if (bg_thresh >= thresh)
> 		bg_thresh = thresh / 2;
> 
> But actually we can set vm background dirtiness bigger than
> vm dirtiness successfully. This behavior may mislead us.
> We'd better do this validity check at the beginning.

This is an admin only interface. You can screw setting this up even
when you keep consistency between the background and direct limits. In
general we do not try to be clever for these knobs because we _expect_
admins to do sane things. Why is this any different and why do we need
to add quite some code to handle one particular corner case?

To be honest I am not entirely sure this is worth the code and the
future maintenance burden.

> Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
> ---
>  Documentation/sysctl/vm.txt |  6 +++
>  mm/page-writeback.c         | 92 +++++++++++++++++++++++++++++++++++++++++----
>  2 files changed, 90 insertions(+), 8 deletions(-)
> 
> diff --git a/Documentation/sysctl/vm.txt b/Documentation/sysctl/vm.txt
> index 9baf66a..0bab85d 100644
> --- a/Documentation/sysctl/vm.txt
> +++ b/Documentation/sysctl/vm.txt
> @@ -156,6 +156,9 @@ read.
>  Note: the minimum value allowed for dirty_bytes is two pages (in bytes); any
>  value lower than this limit will be ignored and the old configuration will be
>  retained.
> +Note: the value of dirty_bytes also cannot be set lower than
> +dirty_background_bytes or the amount of memory corresponding to
> +dirty_background_ratio.
>  
>  ==============================================================
>  
> @@ -176,6 +179,9 @@ generating disk writes will itself start writing out dirty data.
>  
>  The total available memory is not equal to total system memory.
>  
> +Note: dirty_ratio cannot be set lower than dirty_background_ratio or
> +ratio corresponding to dirty_background_bytes.
> +
>  ==============================================================
>  
>  dirty_writeback_centisecs
> diff --git a/mm/page-writeback.c b/mm/page-writeback.c
> index 0b9c5cb..fadb1d7 100644
> --- a/mm/page-writeback.c
> +++ b/mm/page-writeback.c
> @@ -511,15 +511,71 @@ bool node_dirty_ok(struct pglist_data *pgdat)
>  	return nr_pages <= limit;
>  }
>  
> +static bool vm_dirty_settings_valid(void)
> +{
> +	bool ret = true;
> +	unsigned long bytes;
> +
> +	if (vm_dirty_ratio > 0) {
> +		if (dirty_background_ratio >= vm_dirty_ratio) {
> +			ret = false;
> +			goto out;
> +		}
> +
> +		bytes = global_dirtyable_memory() * PAGE_SIZE / 100 *
> +				vm_dirty_ratio;
> +		if (dirty_background_bytes >= bytes) {
> +			ret = false;
> +			goto out;
> +		}
> +	}
> +
> +	if (vm_dirty_bytes > 0) {
> +		if (dirty_background_bytes >= vm_dirty_bytes) {
> +			ret = false;
> +			goto out;
> +		}
> +
> +		bytes = global_dirtyable_memory() * PAGE_SIZE / 100 *
> +				dirty_background_ratio;
> +
> +		if (bytes >= vm_dirty_bytes) {
> +			ret = false;
> +			goto out;
> +		}
> +	}
> +
> +	if ((vm_dirty_bytes == 0 && vm_dirty_ratio == 0) ||
> +		(dirty_background_bytes == 0 && dirty_background_ratio == 0))
> +		ret = false;
> +
> +out:
> +	return ret;
> +}
> +
>  int dirty_background_ratio_handler(struct ctl_table *table, int write,
>  		void __user *buffer, size_t *lenp,
>  		loff_t *ppos)
>  {
>  	int ret;
> +	int old_ratio = dirty_background_ratio;
>  
>  	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
> -	if (ret == 0 && write)
> -		dirty_background_bytes = 0;
> +
> +	/* When dirty_background_ratio is 0 and dirty_background_bytes isn't 0,
> +	 * it's not correct to set dirty_background_bytes to 0 if we reset
> +	 * dirty_background_ratio to 0.
> +	 * So do nothing if the new ratio is not different.
> +	 */
> +	if (ret == 0 && write && dirty_background_ratio != old_ratio) {
> +		if (vm_dirty_settings_valid())
> +			dirty_background_bytes = 0;
> +		else {
> +			dirty_background_ratio = old_ratio;
> +			ret = -EINVAL;
> +		}
> +	}
> +
>  	return ret;
>  }
>  
> @@ -528,10 +584,20 @@ int dirty_background_bytes_handler(struct ctl_table *table, int write,
>  		loff_t *ppos)
>  {
>  	int ret;
> +	unsigned long old_bytes = dirty_background_bytes;
>  
>  	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
> -	if (ret == 0 && write)
> -		dirty_background_ratio = 0;
> +
> +	/* the reson is same as above */
> +	if (ret == 0 && write && dirty_background_bytes != old_bytes) {
> +		if (vm_dirty_settings_valid())
> +			dirty_background_ratio = 0;
> +		else {
> +			dirty_background_bytes = old_bytes;
> +			ret = -EINVAL;
> +		}
> +	}
> +
>  	return ret;
>  }
>  
> @@ -544,8 +610,13 @@ int dirty_ratio_handler(struct ctl_table *table, int write,
>  
>  	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
>  	if (ret == 0 && write && vm_dirty_ratio != old_ratio) {
> -		writeback_set_ratelimit();
> -		vm_dirty_bytes = 0;
> +		if (vm_dirty_settings_valid()) {
> +			writeback_set_ratelimit();
> +			vm_dirty_bytes = 0;
> +		} else {
> +			vm_dirty_ratio = old_ratio;
> +			ret = -EINVAL;
> +		}
>  	}
>  	return ret;
>  }
> @@ -559,8 +630,13 @@ int dirty_bytes_handler(struct ctl_table *table, int write,
>  
>  	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
>  	if (ret == 0 && write && vm_dirty_bytes != old_bytes) {
> -		writeback_set_ratelimit();
> -		vm_dirty_ratio = 0;
> +		if (vm_dirty_settings_valid()) {
> +			writeback_set_ratelimit();
> +			vm_dirty_ratio = 0;
> +		} else {
> +			vm_dirty_bytes = old_bytes;
> +			ret = -EINVAL;
> +		}
>  	}
>  	return ret;
>  }
> -- 
> 1.8.3.1
> 

-- 
Michal Hocko
SUSE Labs

--
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>

  reply	other threads:[~2017-09-26 10:25 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-19 22:43 Yafang Shao
2017-09-26 10:25 ` Michal Hocko [this message]
2017-09-26 11:06   ` Yafang Shao
2017-09-26 11:26     ` Michal Hocko
2017-09-26 11:45       ` Yafang Shao
2017-09-26 11:54         ` Michal Hocko
2017-09-26 13:33           ` Jan Kara

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=20170926102532.culqxb45xwzafomj@dhcp22.suse.cz \
    --to=mhocko@kernel.org \
    --cc=akpm@linux-foundation.org \
    --cc=hannes@cmpxchg.org \
    --cc=jack@suse.cz \
    --cc=jlayton@redhat.com \
    --cc=laoar.shao@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mawilcox@microsoft.com \
    --cc=nborisov@suse.com \
    --cc=tytso@mit.edu \
    --cc=vdavydov.dev@gmail.com \
    /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