From mboxrd@z Thu Jan 1 00:00:00 1970 From: Toshi Kani Subject: Re: [PATCH v2 5/5] dax: handle media errors in dax_do_io Date: Fri, 15 Apr 2016 13:08:28 -0600 Message-ID: <1460747308.4597.9.camel@hpe.com> References: <1459303190-20072-1-git-send-email-vishal.l.verma@intel.com> <1459303190-20072-6-git-send-email-vishal.l.verma@intel.com> <1460739288.3012.3.camel@intel.com> <1460741821.3012.11.camel@intel.com> <1460746909.4597.7.camel@hpe.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Cc: "axboe@fb.com" , "jack@suse.cz" , "david@fromorbit.com" , "linux-kernel@vger.kernel.org" , "xfs@oss.sgi.com" , "hch@infradead.org" , "linux-mm@kvack.org" , "linux-block@vger.kernel.org" , "viro@zeniv.linux.org.uk" , "linux-nvdimm@ml01.01.org" , "linux-fsdevel@vger.kernel.org" , "akpm@linux-foundation.org" , "linux-ext4@vger.kernel.org" , "Wilcox, Matthew R" To: Dan Williams , Jeff Moyer Return-path: In-Reply-To: <1460746909.4597.7.camel@hpe.com> Sender: owner-linux-mm@kvack.org List-Id: linux-ext4.vger.kernel.org On Fri, 2016-04-15 at 13:01 -0600, Toshi Kani wrote: > On Fri, 2016-04-15 at 11:17 -0700, Dan Williams wrote: > >=20 > > On Fri, Apr 15, 2016 at 11:06 AM, Jeff Moyer wrot= e: > > >=20 > > > Dan Williams writes: > > > =C2=A0 > > > > > > There's a lot of special casing here, so you might consider > > > > > > adding comments. > > > > > Correct - maybe we should reconsider wrapper-izing this? :) > > > > Another option is just to skip dax_do_io() and this special casin= g > > > > fallback entirely if errors are present.=C2=A0=C2=A0I.e. only att= empt > > > > dax_do_io when: IS_DAX() && gendisk->bb && bb->count =3D=3D 0. > > > > > > So, if there's an error anywhere on the device, penalize all I/O (n= ot > > > just writes, and not just on sectors that are bad)?=C2=A0=C2=A0I'm = not sure > > > that's a great plan, either. > > >=20 > > If errors are rare how much are we actually losing in practice? > > Moreover, we're going to do the full badblocks lookup anyway when we > > call ->direct_access().=C2=A0=C2=A0If we had that information earlier= we can > > avoid this fallback dance. > > A system running with DAX may have active data set in NVDIMM lager than > RAM size. =C2=A0In this case, falling back to non-DAX will allocate pag= e cache > for the data, which will saturate the system with memory pressure. Oh, sorry, we are still in DIO path. =C2=A0Falling back to DIO should not= cause this issue. -Toshi -- 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 From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay.sgi.com (relay3.corp.sgi.com [198.149.34.15]) by oss.sgi.com (Postfix) with ESMTP id 8D03A7D01 for ; Fri, 15 Apr 2016 14:17:04 -0500 (CDT) Received: from cuda.sgi.com (cuda3.sgi.com [192.48.176.15]) by relay3.corp.sgi.com (Postfix) with ESMTP id E9061AC001 for ; Fri, 15 Apr 2016 12:17:03 -0700 (PDT) Received: from g4t3426.houston.hp.com (g4t3426.houston.hp.com [15.201.208.54]) by cuda.sgi.com with ESMTP id NZhtVQZpMFm1uyLl (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO) for ; Fri, 15 Apr 2016 12:17:02 -0700 (PDT) Message-ID: <1460747308.4597.9.camel@hpe.com> Subject: Re: [PATCH v2 5/5] dax: handle media errors in dax_do_io From: Toshi Kani Date: Fri, 15 Apr 2016 13:08:28 -0600 In-Reply-To: <1460746909.4597.7.camel@hpe.com> References: <1459303190-20072-1-git-send-email-vishal.l.verma@intel.com> <1459303190-20072-6-git-send-email-vishal.l.verma@intel.com> <1460739288.3012.3.camel@intel.com> <1460741821.3012.11.camel@intel.com> <1460746909.4597.7.camel@hpe.com> Mime-Version: 1.0 List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64 Errors-To: xfs-bounces@oss.sgi.com Sender: xfs-bounces@oss.sgi.com To: Dan Williams , Jeff Moyer Cc: "hch@infradead.org" , "jack@suse.cz" , "linux-nvdimm@ml01.01.org" , "linux-kernel@vger.kernel.org" , "xfs@oss.sgi.com" , "axboe@fb.com" , "linux-mm@kvack.org" , "linux-block@vger.kernel.org" , "viro@zeniv.linux.org.uk" , "linux-fsdevel@vger.kernel.org" , "akpm@linux-foundation.org" , "linux-ext4@vger.kernel.org" , "Wilcox, Matthew R" T24gRnJpLCAyMDE2LTA0LTE1IGF0IDEzOjAxIC0wNjAwLCBUb3NoaSBLYW5pIHdyb3RlOgo+IE9u IEZyaSwgMjAxNi0wNC0xNSBhdCAxMToxNyAtMDcwMCwgRGFuIFdpbGxpYW1zIHdyb3RlOgo+ID4g Cj4gPiBPbiBGcmksIEFwciAxNSwgMjAxNiBhdCAxMTowNiBBTSwgSmVmZiBNb3llciA8am1veWVy QHJlZGhhdC5jb20+IHdyb3RlOgo+ID4gPiAKPiA+ID4gRGFuIFdpbGxpYW1zIDxkYW4uai53aWxs aWFtc0BpbnRlbC5jb20+IHdyaXRlczoKPiA+ID4gwqAKPiA+ID4gPiA+ID4gVGhlcmUncyBhIGxv dCBvZiBzcGVjaWFsIGNhc2luZyBoZXJlLCBzbyB5b3UgbWlnaHQgY29uc2lkZXIKPiA+ID4gPiA+ ID4gYWRkaW5nIGNvbW1lbnRzLgo+ID4gPiA+ID4gQ29ycmVjdCAtIG1heWJlIHdlIHNob3VsZCBy ZWNvbnNpZGVyIHdyYXBwZXItaXppbmcgdGhpcz8gOikKPiA+ID4gPiBBbm90aGVyIG9wdGlvbiBp cyBqdXN0IHRvIHNraXAgZGF4X2RvX2lvKCkgYW5kIHRoaXMgc3BlY2lhbCBjYXNpbmcKPiA+ID4g PiBmYWxsYmFjayBlbnRpcmVseSBpZiBlcnJvcnMgYXJlIHByZXNlbnQuwqDCoEkuZS4gb25seSBh dHRlbXB0Cj4gPiA+ID4gZGF4X2RvX2lvIHdoZW46IElTX0RBWCgpICYmIGdlbmRpc2stPmJiICYm IGJiLT5jb3VudCA9PSAwLgo+ID4gPgo+ID4gPiBTbywgaWYgdGhlcmUncyBhbiBlcnJvciBhbnl3 aGVyZSBvbiB0aGUgZGV2aWNlLCBwZW5hbGl6ZSBhbGwgSS9PIChub3QKPiA+ID4ganVzdCB3cml0 ZXMsIGFuZCBub3QganVzdCBvbiBzZWN0b3JzIHRoYXQgYXJlIGJhZCk/wqDCoEknbSBub3Qgc3Vy ZQo+ID4gPiB0aGF0J3MgYSBncmVhdCBwbGFuLCBlaXRoZXIuCj4gPiA+IAo+ID4gSWYgZXJyb3Jz IGFyZSByYXJlIGhvdyBtdWNoIGFyZSB3ZSBhY3R1YWxseSBsb3NpbmcgaW4gcHJhY3RpY2U/Cj4g PiBNb3Jlb3Zlciwgd2UncmUgZ29pbmcgdG8gZG8gdGhlIGZ1bGwgYmFkYmxvY2tzIGxvb2t1cCBh bnl3YXkgd2hlbiB3ZQo+ID4gY2FsbCAtPmRpcmVjdF9hY2Nlc3MoKS7CoMKgSWYgd2UgaGFkIHRo YXQgaW5mb3JtYXRpb24gZWFybGllciB3ZSBjYW4KPiA+IGF2b2lkIHRoaXMgZmFsbGJhY2sgZGFu Y2UuCj4KPiBBIHN5c3RlbSBydW5uaW5nIHdpdGggREFYIG1heSBoYXZlIGFjdGl2ZSBkYXRhIHNl dCBpbiBOVkRJTU0gbGFnZXIgdGhhbgo+IFJBTSBzaXplLiDCoEluIHRoaXMgY2FzZSwgZmFsbGlu ZyBiYWNrIHRvIG5vbi1EQVggd2lsbCBhbGxvY2F0ZSBwYWdlIGNhY2hlCj4gZm9yIHRoZSBkYXRh LCB3aGljaCB3aWxsIHNhdHVyYXRlIHRoZSBzeXN0ZW0gd2l0aCBtZW1vcnkgcHJlc3N1cmUuCgpP aCwgc29ycnksIHdlIGFyZSBzdGlsbCBpbiBESU8gcGF0aC4gwqBGYWxsaW5nIGJhY2sgdG8gRElP IHNob3VsZCBub3QgY2F1c2UKdGhpcyBpc3N1ZS4KCi1Ub3NoaQoKX19fX19fX19fX19fX19fX19f X19fX19fX19fX19fX19fX19fX19fX19fX19fX18KeGZzIG1haWxpbmcgbGlzdAp4ZnNAb3NzLnNn aS5jb20KaHR0cDovL29zcy5zZ2kuY29tL21haWxtYW4vbGlzdGluZm8veGZzCg== From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-oi0-f71.google.com (mail-oi0-f71.google.com [209.85.218.71]) by kanga.kvack.org (Postfix) with ESMTP id 3AD286B0253 for ; Fri, 15 Apr 2016 15:17:03 -0400 (EDT) Received: by mail-oi0-f71.google.com with SMTP id h201so188703325oib.1 for ; Fri, 15 Apr 2016 12:17:03 -0700 (PDT) Received: from g4t3426.houston.hp.com (g4t3426.houston.hp.com. [15.201.208.54]) by mx.google.com with ESMTPS id v64si17025363oif.138.2016.04.15.12.17.02 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 15 Apr 2016 12:17:02 -0700 (PDT) Message-ID: <1460747308.4597.9.camel@hpe.com> Subject: Re: [PATCH v2 5/5] dax: handle media errors in dax_do_io From: Toshi Kani Date: Fri, 15 Apr 2016 13:08:28 -0600 In-Reply-To: <1460746909.4597.7.camel@hpe.com> References: <1459303190-20072-1-git-send-email-vishal.l.verma@intel.com> <1459303190-20072-6-git-send-email-vishal.l.verma@intel.com> <1460739288.3012.3.camel@intel.com> <1460741821.3012.11.camel@intel.com> <1460746909.4597.7.camel@hpe.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: owner-linux-mm@kvack.org List-ID: To: Dan Williams , Jeff Moyer Cc: "axboe@fb.com" , "jack@suse.cz" , "david@fromorbit.com" , "linux-kernel@vger.kernel.org" , "xfs@oss.sgi.com" , "hch@infradead.org" , "linux-mm@kvack.org" , "linux-block@vger.kernel.org" , "viro@zeniv.linux.org.uk" , "linux-nvdimm@ml01.01.org" , "linux-fsdevel@vger.kernel.org" , "akpm@linux-foundation.org" , "linux-ext4@vger.kernel.org" , "Wilcox, Matthew R" On Fri, 2016-04-15 at 13:01 -0600, Toshi Kani wrote: > On Fri, 2016-04-15 at 11:17 -0700, Dan Williams wrote: > > > > On Fri, Apr 15, 2016 at 11:06 AM, Jeff Moyer wrote: > > > > > > Dan Williams writes: > > > A > > > > > > There's a lot of special casing here, so you might consider > > > > > > adding comments. > > > > > Correct - maybe we should reconsider wrapper-izing this? :) > > > > Another option is just to skip dax_do_io() and this special casing > > > > fallback entirely if errors are present.A A I.e. only attempt > > > > dax_do_io when: IS_DAX() && gendisk->bb && bb->count == 0. > > > > > > So, if there's an error anywhere on the device, penalize all I/O (not > > > just writes, and not just on sectors that are bad)?A A I'm not sure > > > that's a great plan, either. > > > > > If errors are rare how much are we actually losing in practice? > > Moreover, we're going to do the full badblocks lookup anyway when we > > call ->direct_access().A A If we had that information earlier we can > > avoid this fallback dance. > > A system running with DAX may have active data set in NVDIMM lager than > RAM size. A In this case, falling back to non-DAX will allocate page cache > for the data, which will saturate the system with memory pressure. Oh, sorry, we are still in DIO path. A Falling back to DIO should not cause this issue. -Toshi -- 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 From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932193AbcDOTb6 (ORCPT ); Fri, 15 Apr 2016 15:31:58 -0400 Received: from g4t3426.houston.hp.com ([15.201.208.54]:44273 "EHLO g4t3426.houston.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752710AbcDOTRD (ORCPT ); Fri, 15 Apr 2016 15:17:03 -0400 Message-ID: <1460747308.4597.9.camel@hpe.com> Subject: Re: [PATCH v2 5/5] dax: handle media errors in dax_do_io From: Toshi Kani To: Dan Williams , Jeff Moyer Cc: "axboe@fb.com" , "jack@suse.cz" , "david@fromorbit.com" , "linux-kernel@vger.kernel.org" , "xfs@oss.sgi.com" , "hch@infradead.org" , "linux-mm@kvack.org" , "linux-block@vger.kernel.org" , "viro@zeniv.linux.org.uk" , "linux-nvdimm@ml01.01.org" , "linux-fsdevel@vger.kernel.org" , "akpm@linux-foundation.org" , "linux-ext4@vger.kernel.org" , "Wilcox, Matthew R" Date: Fri, 15 Apr 2016 13:08:28 -0600 In-Reply-To: <1460746909.4597.7.camel@hpe.com> References: <1459303190-20072-1-git-send-email-vishal.l.verma@intel.com> <1459303190-20072-6-git-send-email-vishal.l.verma@intel.com> <1460739288.3012.3.camel@intel.com> <1460741821.3012.11.camel@intel.com> <1460746909.4597.7.camel@hpe.com> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.18.5.2 (3.18.5.2-1.fc23) Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2016-04-15 at 13:01 -0600, Toshi Kani wrote: > On Fri, 2016-04-15 at 11:17 -0700, Dan Williams wrote: > > > > On Fri, Apr 15, 2016 at 11:06 AM, Jeff Moyer wrote: > > > > > > Dan Williams writes: > > >   > > > > > > There's a lot of special casing here, so you might consider > > > > > > adding comments. > > > > > Correct - maybe we should reconsider wrapper-izing this? :) > > > > Another option is just to skip dax_do_io() and this special casing > > > > fallback entirely if errors are present.  I.e. only attempt > > > > dax_do_io when: IS_DAX() && gendisk->bb && bb->count == 0. > > > > > > So, if there's an error anywhere on the device, penalize all I/O (not > > > just writes, and not just on sectors that are bad)?  I'm not sure > > > that's a great plan, either. > > > > > If errors are rare how much are we actually losing in practice? > > Moreover, we're going to do the full badblocks lookup anyway when we > > call ->direct_access().  If we had that information earlier we can > > avoid this fallback dance. > > A system running with DAX may have active data set in NVDIMM lager than > RAM size.  In this case, falling back to non-DAX will allocate page cache > for the data, which will saturate the system with memory pressure. Oh, sorry, we are still in DIO path.  Falling back to DIO should not cause this issue. -Toshi