From: "Radosław Smogura" <mail@smogura.eu>
To: linux-mm@kvack.org
Cc: Andrea Arcangeli <aarcange@redhat.com>,
Hugh Dickins <hughd@google.com>,
Yongqiang Yang <xiaoqiangnk@gmail.com>
Subject: Re: [PATCH] mm: Change of refcounting method for compound page.
Date: Fri, 03 Feb 2012 20:05:27 +0100 [thread overview]
Message-ID: <1383050.EiyE3vO7UJ@localhost> (raw)
In-Reply-To: <1328290093-19294-2-git-send-email-mail@smogura.eu>
Dnia piątek, 3 lutego 2012 o 18:28:13 Radosław Smogura napisał(a):
> Compound pages are now refcounted in way allowing tracking of tail pages
> and automatically free of compound page when all references (counter)
> fell to zero. This in addition make get_page and get_page_unless_zero
> similar in work, as well put_page and put_page_unless_zero. In addition
> it makes procedures more friendly. One thing that should be taken, by
> developer, on account is to take care when page is putted or geted when
> compound lock is obtained, to avoid deadlocks. Locking is used to
> prevent concurrent compound split and only when page refcount goes from
> 0 to 1 or vice versa.
>
> Technically implementation uses 3rd element of compound page to store
> "tails usage counter". This counter is decremented when tail pages count
> goes to zero, and bumped when tail page is getted from zero usage
> (recovered) – this is to keep backward compatible usage of tail pages.
> If "tails usage counter" fell to zero head counter is decremented, if
> "tails usage counter" is increased to one the head count is increased,
> too. For compound pages without 3rd element (order of 1, two pages) 2nd
> page count is used in similar way as for higher order pages.
>
> Signed-off-by: Radosław Smogura <mail@smogura.eu>
> ---
> include/linux/mm.h | 94 ++++++++++++-----
> include/linux/mm_types.h | 24 ++++-
> include/linux/pagemap.h | 1 -
> mm/huge_memory.c | 25 +----
> mm/internal.h | 46 ---------
> mm/memory.c | 2 +-
> mm/page_alloc.c | 2 +
> mm/swap.c | 254
> +++++++++++++++++++++++++++++----------------- 8 files changed, 256
> insertions(+), 192 deletions(-)
<snip>
I would like to say huge sorry, I sent wrong patch (yesterday work). Please
discard previous patch, and consider patch I will send in reply to this email.
Not included change was introduced to allow proper concurrent work of get/put
page and split page .
Regards,
Radosław Smogura
--
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/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2012-02-03 19:05 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-02-03 17:28 Radosław Smogura
2012-02-03 17:28 ` Radosław Smogura
2012-02-03 19:05 ` Radosław Smogura [this message]
2012-02-03 19:14 ` Radosław Smogura
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=1383050.EiyE3vO7UJ@localhost \
--to=mail@smogura.eu \
--cc=aarcange@redhat.com \
--cc=hughd@google.com \
--cc=linux-mm@kvack.org \
--cc=xiaoqiangnk@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