Re: [PATCH v2 2/7] dax: change bdev_dax_supported() to support boolean returns

From: Dan Williams <dan.j.williams@intel.com>
To: Dave Chinner <david@fromorbit.com>
Cc: "Darrick J. Wong" <darrick.wong@oracle.com>,
	Mike Snitzer <snitzer@redhat.com>,
	linux-nvdimm <linux-nvdimm@lists.01.org>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	linux-xfs <linux-xfs@vger.kernel.org>,
	device-mapper development <dm-devel@redhat.com>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>
Subject: Re: [PATCH v2 2/7] dax: change bdev_dax_supported() to support boolean returns
Date: Thu, 31 May 2018 18:57:33 -0700	[thread overview]
Message-ID: <CAPcyv4g6dv9sG9dR-iY+aEfSi2r3ey1dXGbX-nexKsN72t0QMw@mail.gmail.com> (raw)
In-Reply-To: <20180601012657.GQ10363@dastard>

On Thu, May 31, 2018 at 6:26 PM, Dave Chinner <david@fromorbit.com> wrote:
> On Thu, May 31, 2018 at 12:13:32PM -0700, Darrick J. Wong wrote:
>> On Tue, May 29, 2018 at 04:01:14PM -0600, Ross Zwisler wrote:
>> > On Tue, May 29, 2018 at 02:25:10PM -0700, Darrick J. Wong wrote:
>> > > On Tue, May 29, 2018 at 01:51:01PM -0600, Ross Zwisler wrote:
>> > > > From: Dave Jiang <dave.jiang@intel.com>
>> > > >
>> > > > The function return values are confusing with the way the function is
>> > > > named. We expect a true or false return value but it actually returns
>> > > > 0/-errno.  This makes the code very confusing. Changing the return values
>> > > > to return a bool where if DAX is supported then return true and no DAX
>> > > > support returns false.
>> > > >
>> > > > Signed-off-by: Dave Jiang <dave.jiang@intel.com>
>> > > > Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com>
>> > >
>> > > Looks ok, do you want me to pull the first two patches through the xfs
>> > > tree?
>> > >
>> > > Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
>> >
>> > Thanks for the review.
>> >
>> > I'm not sure what's best.  If you do that then Mike will need to have a DM
>> > branch for the rest of the series based on your stable commits, yea?
>> >
>> > Mike what would you prefer?
>>
>> I /was/ about to say that I would pull in the first two patches, but now
>> I can't get xfs to mount with pmem at all, and have no way of testing
>> this...?
>
> I have similar problems, too, but:
>
> $ ndctl list
> [
>   {
>     "dev":"namespace1.0",
>     "mode":"raw",
>     "size":8589934592,
>     "sector_size":512,
>     "blockdev":"pmem1"
>   },
>   {
>     "dev":"namespace0.0",
>     "mode":"raw",
>     "size":8589934592,
>     "sector_size":512,
>     "blockdev":"pmem0"
>   }
> ]
> $ sudo ndctl create-namespace -f -e namespace0.0 --mode=fsdax
>   Error: operation failed, region0 fsdax mode not available
>
> failed to reconfigure namespace: Invalid argument
> $
>
> I can't make head or tail of what is going wrong here - how am I
> supposed to debug this and get it working again?
>
> FWIW, XFS+DAX used to just work on this setup (I hadn't even
> installed ndctl until this morning!) but after changing the kernel
> it no longer works. That would make it a regression, yes?

This commit caused the behavior change:

    569d0365f571 dax: require 'struct page' by default for filesystem dax

The justification is in that patch, but the short summary is we killed
off "pageless" dax because it had so many incomplete holes and
surprise behaviors. It needed to die on the path to making dax not
experimental, i.e. to close safety holes, and be feature complete for
all the ways userspace expects to use mappings (direct-io, fork,
poison handling, etc).

I suspect your kernel does not have CONFIG_ZONE_DEVICE enabled which
has the following dependencies:

        depends on MEMORY_HOTPLUG
        depends on MEMORY_HOTREMOVE
        depends on SPARSEMEM_VMEMMAP

I've created a task to go improve ndctl's error reporting for
troubleshooting this failure.