From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751765Ab2IZU4e (ORCPT ); Wed, 26 Sep 2012 16:56:34 -0400 Received: from zene.cmpxchg.org ([85.214.230.12]:58301 "EHLO zene.cmpxchg.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751677Ab2IZU4d (ORCPT ); Wed, 26 Sep 2012 16:56:33 -0400 Date: Wed, 26 Sep 2012 16:56:17 -0400 From: Johannes Weiner To: Michal Hocko Cc: Andrew Morton , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Hillf Danton , Mel Gorman , KAMEZAWA Hiroyuki , Andrea Arcangeli , David Rientjes Subject: Re: [PATCH] hugetlb: do not use vma_hugecache_offset for vma_prio_tree_foreach Message-ID: <20120926205617.GA2667@cmpxchg.org> References: <1344866141-27906-1-git-send-email-mhocko@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1344866141-27906-1-git-send-email-mhocko@suse.cz> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Aug 13, 2012 at 03:55:41PM +0200, Michal Hocko wrote: > 0c176d5 (mm: hugetlb: fix pgoff computation when unmapping page > from vma) fixed pgoff calculation but it has replaced it by > vma_hugecache_offset which is not approapriate for offsets used for > vma_prio_tree_foreach because that one expects index in page units > rather than in huge_page_shift. > Using vma_hugecache_offset is not incorrect because the pgoff will fit > into the same vmas but it is confusing so the standard PAGE_SHIFT based > index calculation is used instead. I do think it's incorrect. The resulting index may not be too big, but it can be too small: assume hpage size of 2M and the address to unmap to be 0x200000. This is regular page index 512 and hpage index 1. If you have a VMA that maps the file only starting at the second huge page, that VMAs vm_pgoff will be 512 but you ask for offset 1 and miss it even though it does map the page of interest. hugetlb_cow() will try to unmap, miss the vma, and retry the cow until the allocation succeeds or the skipped vma(s) go away. Unless I missed something, this should not be deferred as a cleanup. From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx127.postini.com [74.125.245.127]) by kanga.kvack.org (Postfix) with SMTP id 7960E6B0044 for ; Wed, 26 Sep 2012 16:56:31 -0400 (EDT) Date: Wed, 26 Sep 2012 16:56:17 -0400 From: Johannes Weiner Subject: Re: [PATCH] hugetlb: do not use vma_hugecache_offset for vma_prio_tree_foreach Message-ID: <20120926205617.GA2667@cmpxchg.org> References: <1344866141-27906-1-git-send-email-mhocko@suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1344866141-27906-1-git-send-email-mhocko@suse.cz> Sender: owner-linux-mm@kvack.org List-ID: To: Michal Hocko Cc: Andrew Morton , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Hillf Danton , Mel Gorman , KAMEZAWA Hiroyuki , Andrea Arcangeli , David Rientjes On Mon, Aug 13, 2012 at 03:55:41PM +0200, Michal Hocko wrote: > 0c176d5 (mm: hugetlb: fix pgoff computation when unmapping page > from vma) fixed pgoff calculation but it has replaced it by > vma_hugecache_offset which is not approapriate for offsets used for > vma_prio_tree_foreach because that one expects index in page units > rather than in huge_page_shift. > Using vma_hugecache_offset is not incorrect because the pgoff will fit > into the same vmas but it is confusing so the standard PAGE_SHIFT based > index calculation is used instead. I do think it's incorrect. The resulting index may not be too big, but it can be too small: assume hpage size of 2M and the address to unmap to be 0x200000. This is regular page index 512 and hpage index 1. If you have a VMA that maps the file only starting at the second huge page, that VMAs vm_pgoff will be 512 but you ask for offset 1 and miss it even though it does map the page of interest. hugetlb_cow() will try to unmap, miss the vma, and retry the cow until the allocation succeeds or the skipped vma(s) go away. Unless I missed something, this should not be deferred as a cleanup. -- 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: email@kvack.org