From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx177.postini.com [74.125.245.177]) by kanga.kvack.org (Postfix) with SMTP id 1CF626B002C for ; Fri, 3 Feb 2012 12:28:27 -0500 (EST) From: =?UTF-8?q?Rados=C5=82aw=20Smogura?= Subject: [PATCH] mm: Change of refcounting method for compound page. Date: Fri, 3 Feb 2012 18:28:12 +0100 Message-Id: <1328290093-19294-1-git-send-email-mail@smogura.eu> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Sender: owner-linux-mm@kvack.org List-ID: To: linux-mm@kvack.org Cc: =?UTF-8?q?Rados=C5=82aw=20Smogura?= , Andrea Arcangeli , Hugh Dickins , Yongqiang Yang Hello, Currently, I try to enable huge pages for SHM (and generally for any page cache), during this I ran into problem with ref-countig of huge pages. Due to this I would like to propose patch for changing ref-counting for compound page and allowing auto destruction of such page. I tried to revert behavior from 2.6 kernels, and to make it concurrently safe, as well sophisticated and compatible with THP. I hope this removes risky way through call get_user_pages, too. For me enabling such ref-counting is really important to continue work, and to prevent huge changes in page cache managing. Regards, RadosA?aw Smogura (1): mm: Change of refcounting method for compound page. include/linux/mm.h | 94 +++++++++++----- include/linux/mm_types.h | 24 ++++- include/linux/pagemap.h | 1 - mm/huge_memory.c | 30 ++--- mm/internal.h | 46 -------- mm/memory.c | 2 +- mm/page_alloc.c | 2 + mm/swap.c | 272 ++++++++++++++++++++++++++++++---------------- 8 files changed, 279 insertions(+), 192 deletions(-) -- 1.7.3.4 -- 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: email@kvack.org