From: Jerome Marchand <jmarchan@redhat.com>
To: Dave Hansen <dave.hansen@intel.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] mm: add overcommit_kbytes sysctl variable
Date: Thu, 22 Aug 2013 11:25:31 +0200 [thread overview]
Message-ID: <5215D90B.2050008@redhat.com> (raw)
In-Reply-To: <5214E96B.3090009@intel.com>
On 08/21/2013 06:23 PM, Dave Hansen wrote:
> On 08/21/2013 08:22 AM, Jerome Marchand wrote:
>>>> Instead of introducing yet another tunable, why don't we just make the
>>>> ratio that comes in from the user more fine-grained?
>>>>
>>>> sysctl overcommit_ratio=0.2
>>>>
>>>> We change the internal 'sysctl_overcommit_ratio' to store tenths or
>>>> hundreths of a percent (or whatever), then parse the input as two
>>>> integers. I don't think we need fully correct floating point parsing
>>>> and rounding here, so it shouldn't be too much of a chore. It'd
>>>> probably end up being less code than you have as it stands.
>>>>
>> Now that I think about it, that could break user space. Sure write access
>> wouldn't be a problem (one can still write a plain integer), but a script
>> that reads a fractional value when it expects an integer might not be able
>> to cope with it.
>
> You're right. Something doing FOO=$(cat overcommit_ratio) and then
> trying do do arithmetic would just fail loudly. But, it would probably
> fail silently if we create another tunable that all of a sudden returns
> 0 (when the kernel is not _behaving_ like it is set to 0).
>
> I'm not sure there's a good way out of this without breakage (or at
> least confusing) of _some_ old scripts/programs. Either way has ups and
> downs.
>
> The existing dirty_ratio/bytes stuff just annoys me because I end up
> having to check two places whenever I go looking for it.
>
Right. Then we could just use some overcommit_fine_ratio internally and
overcommit_ratio would show and set a rounded value. I doubt that a script
that reads 80% would notice the difference if it is actually 79.5%.
We could also use overcommit_kbytes internally, but then overcommit_ratio
would fluctuate if RAM ram is added/removed (e.g. memory hotplug or baloon
driver). That might be a problem.
--
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:[~2013-08-22 9:25 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-08-19 15:17 [PATCH 1/2] mm: factor commit limit calculation Jerome Marchand
2013-08-19 15:17 ` [PATCH 2/2] mm: add overcommit_kbytes sysctl variable Jerome Marchand
2013-08-19 16:55 ` Dave Hansen
2013-08-20 8:58 ` Jerome Marchand
2013-08-21 15:22 ` Jerome Marchand
2013-08-21 16:23 ` Dave Hansen
2013-08-22 9:25 ` Jerome Marchand [this message]
2013-09-05 12:51 ` [PATCH 2/2 v2] mm: allow to set overcommit ratio more precisely Jerome Marchand
2013-09-05 14:41 ` Dave Hansen
2013-09-05 14:47 ` Jerome Marchand
2013-09-05 22:11 ` Pavel Machek
2013-09-06 8:38 ` Jerome Marchand
2013-09-06 14:11 ` [PATCH 2/2 v3] " Jerome Marchand
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=5215D90B.2050008@redhat.com \
--to=jmarchan@redhat.com \
--cc=dave.hansen@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
/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