From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1030577AbeEYVgv (ORCPT ); Fri, 25 May 2018 17:36:51 -0400 Received: from mga14.intel.com ([192.55.52.115]:42439 "EHLO mga14.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1030258AbeEYVgs (ORCPT ); Fri, 25 May 2018 17:36:48 -0400 X-Amp-Result: UNKNOWN X-Amp-Original-Verdict: FILE UNKNOWN X-Amp-File-Uploaded: False X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.49,441,1520924400"; d="scan'208";a="43987511" Date: Fri, 25 May 2018 15:36:47 -0600 From: Ross Zwisler To: Mike Snitzer Cc: Ross Zwisler , Toshi Kani , dm-devel@redhat.com, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, linux-nvdimm@lists.01.org Subject: Re: [PATCH 4/7] dm: prevent DAX mounts if not supported Message-ID: <20180525213647.GA3521@linux.intel.com> Mail-Followup-To: Ross Zwisler , Mike Snitzer , Toshi Kani , dm-devel@redhat.com, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, linux-nvdimm@lists.01.org References: <20180525025518.11405-1-ross.zwisler@linux.intel.com> <20180525025518.11405-5-ross.zwisler@linux.intel.com> <20180525195410.GA11008@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180525195410.GA11008@redhat.com> User-Agent: Mutt/1.9.2 (2017-12-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, May 25, 2018 at 03:54:10PM -0400, Mike Snitzer wrote: > On Thu, May 24 2018 at 10:55pm -0400, > Ross Zwisler wrote: > > > Currently the code in dm_dax_direct_access() only checks whether the target > > type has a direct_access() operation defined, not whether the underlying > > block devices all support DAX. This latter property can be seen by looking > > at whether we set the QUEUE_FLAG_DAX request queue flag when creating the > > DM device. > > > > This is problematic if we have, for example, a dm-linear device made up of > > a PMEM namespace in fsdax mode followed by a ramdisk from BRD. > > QUEUE_FLAG_DAX won't be set on the dm-linear device's request queue, but > > we have a working direct_access() entry point and the first member of the > > dm-linear set *does* support DAX. > > > > This allows the user to create a filesystem on the dm-linear device, and > > then mount it with DAX. The filesystem's bdev_dax_supported() test will > > pass because it'll operate on the first member of the dm-linear device, > > which happens to be a fsdax PMEM namespace. > > > > All DAX I/O will then fail to that dm-linear device because the lack of > > QUEUE_FLAG_DAX prevents fs_dax_get_by_bdev() from working. This means that > > the struct dax_device isn't ever set in the filesystem, so > > dax_direct_access() will always return -EOPNOTSUPP. > > > > By failing out of dm_dax_direct_access() if QUEUE_FLAG_DAX isn't set we let > > the filesystem know we don't support DAX at mount time. The filesystem > > will then silently fall back and remove the dax mount option, causing it to > > work properly. > > > > Signed-off-by: Ross Zwisler > > Fixes: commit 545ed20e6df6 ("dm: add infrastructure for DAX support") > > --- > > drivers/md/dm.c | 5 ++--- > > 1 file changed, 2 insertions(+), 3 deletions(-) > > > > diff --git a/drivers/md/dm.c b/drivers/md/dm.c > > index 0a7b0107ca78..9728433362d1 100644 > > --- a/drivers/md/dm.c > > +++ b/drivers/md/dm.c > > @@ -1050,14 +1050,13 @@ static long dm_dax_direct_access(struct dax_device *dax_dev, pgoff_t pgoff, > > > > if (!ti) > > goto out; > > - if (!ti->type->direct_access) > > + if (!blk_queue_dax(md->queue)) > > goto out; > > len = max_io_len(sector, ti) / PAGE_SECTORS; > > if (len < 1) > > goto out; > > nr_pages = min(len, nr_pages); > > - if (ti->type->direct_access) > > - ret = ti->type->direct_access(ti, pgoff, nr_pages, kaddr, pfn); > > + ret = ti->type->direct_access(ti, pgoff, nr_pages, kaddr, pfn); > > So I followed all the rationale for this patch. But the last change > doesn't make any sense. We should still verify that the target has > ti->type->direct_access before calling it. So please reinstate that > check before calling it. You know that type has direct_access() via the blk_queue_dax() check. This tells you not only that the target has direct_access(), but also that you've successfully checked all members of that DM device and they all have working DAX I/O paths, etc. This is all done via the bdev_dax_supported() check and the rest of the code in dm_table_supports_dax() and device_supports_dax(). If this is too subtle I can add a comment or add the check back.