From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Howells Subject: Re: [concept & "good taste" review] persistent store Date: Wed, 22 Dec 2010 00:32:24 +0000 Message-ID: <8362.1292977944@redhat.com> References: <1292977616.8743.87.camel@yhuang-dev> <1292892072.8743.70.camel@yhuang-dev> <1292813234.8743.66.camel@yhuang-dev> <4d0662e511688484b3@agluck-desktop.sc.intel.com> <4D0BEE1F.7020008@zytor.com> <20101219091752.GA16150@liondog.tnic> <17027.1292841992@redhat.com> <32058.1292926258@redhat.com> Return-path: Received: from mx1.redhat.com ([209.132.183.28]:11546 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752865Ab0LVAdV (ORCPT ); Tue, 21 Dec 2010 19:33:21 -0500 In-Reply-To: <1292977616.8743.87.camel@yhuang-dev> Sender: linux-arch-owner@vger.kernel.org List-ID: To: Huang Ying Cc: dhowells@redhat.com, "Luck, Tony" , Borislav Petkov , Tony Luck , Linus Torvalds , "H. Peter Anvin" , "linux-kernel@vger.kernel.org" , "linux-arch@vger.kernel.org" , "tglx@linutronix.de" , "mingo@elte.hu" , "greg@kroah.com" , "akpm@linux-foundation.org" , David Miller , Alan Cox , Jim Keniston , Kyungmin Park , Geert Uytterhoeven Huang Ying wrote: > The user space program (syslogd) is in my big picture, it will guarantee > an oops meesage actually go to disk via something like fsync. After > doing that, the user space program can erase the corresponding record in > persistent storage to free the space. So all in all, oops messages not > causing system panic or disk error will go to disk eventually and being > freed and will not use up the persistent storage. I see. So you rely on fsync() to hang forever if the message can't be written to disk because an oops killed the write path? David From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mx1.redhat.com ([209.132.183.28]:11546 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752865Ab0LVAdV (ORCPT ); Tue, 21 Dec 2010 19:33:21 -0500 From: David Howells In-Reply-To: <1292977616.8743.87.camel@yhuang-dev> References: <1292977616.8743.87.camel@yhuang-dev> <1292892072.8743.70.camel@yhuang-dev> <1292813234.8743.66.camel@yhuang-dev> <4d0662e511688484b3@agluck-desktop.sc.intel.com> <4D0BEE1F.7020008@zytor.com> <20101219091752.GA16150@liondog.tnic> <17027.1292841992@redhat.com> <32058.1292926258@redhat.com> Subject: Re: [concept & "good taste" review] persistent store Date: Wed, 22 Dec 2010 00:32:24 +0000 Message-ID: <8362.1292977944@redhat.com> Sender: linux-arch-owner@vger.kernel.org List-ID: To: Huang Ying Cc: dhowells@redhat.com, "Luck, Tony" , Borislav Petkov , Tony Luck , Linus Torvalds , "H. Peter Anvin" , "linux-kernel@vger.kernel.org" , "linux-arch@vger.kernel.org" , "tglx@linutronix.de" , "mingo@elte.hu" , "greg@kroah.com" , "akpm@linux-foundation.org" , David Miller , Alan Cox , Jim Keniston , Kyungmin Park , Geert Uytterhoeven Message-ID: <20101222003224.Jdt0tQ5X6v3CPIAk4HAIc-po2omNBewZQmxUHR43xyg@z> Huang Ying wrote: > The user space program (syslogd) is in my big picture, it will guarantee > an oops meesage actually go to disk via something like fsync. After > doing that, the user space program can erase the corresponding record in > persistent storage to free the space. So all in all, oops messages not > causing system panic or disk error will go to disk eventually and being > freed and will not use up the persistent storage. I see. So you rely on fsync() to hang forever if the message can't be written to disk because an oops killed the write path? David