From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail138.messagelabs.com (mail138.messagelabs.com [216.82.249.35]) by kanga.kvack.org (Postfix) with ESMTP id 0A0C56B0169 for ; Sat, 20 Aug 2011 14:45:31 -0400 (EDT) Received: by eyh6 with SMTP id 6so2492756eyh.20 for ; Sat, 20 Aug 2011 11:45:29 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20110820132701.GA29322@redhat.com> References: <20110820132701.GA29322@redhat.com> Date: Sun, 21 Aug 2011 00:15:28 +0530 Message-ID: Subject: Re: what protects page lru list? From: Rajesh Ghanekar Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Sender: owner-linux-mm@kvack.org List-ID: To: Johannes Weiner Cc: linux-mm@kvack.org Thanks Johannes. It makes sense. I was searching for any code in which pages is on free list and being operated with lru_lock. But as you said only pages on LRU lists are operated by lru_lock. I am facing a panic in __rmqueue at list_del where the struct page is corrupted or probably the free_list is corrupted. The kernel is 2.6.32.43.xxx from SLES11SP1. I will first see if any of the other code (proprietary) running is not broken. - Rajesh On Sat, Aug 20, 2011 at 6:57 PM, Johannes Weiner wrote= : > On Sat, Aug 20, 2011 at 02:11:53AM +0530, Rajesh Ghanekar wrote: >> Hi, >> =A0 =A0I am confused with what protects page->lru? Is it both zone->lru_= lock or >> zone->lock? I can see it being protected either by lru_lock or lock. > > It's not so much about page->lru but the actual list the page is > linked to. > > The zone's lists of unallocated pages are protected by zone->lock, > while the LRU lists with the pages for userspace are protected by > zone->lru_lock. > -- 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