Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220405194747.2386619-1-jane.chu@oracle.com> <20220405194747.2386619-5-jane.chu@oracle.com>
In-Reply-To: <20220405194747.2386619-5-jane.chu@oracle.com>
From:   Dan Williams <dan.j.williams@intel.com>
Date:   Mon, 11 Apr 2022 21:57:36 -0700
Message-ID: <CAPcyv4jpOss6hzPgM913v_QsZ+PB6Jzo1WV=YdUvnKZiwtfjiA@mail.gmail.com>
Subject: Re: [PATCH v7 4/6] dax: add DAX_RECOVERY flag and .recovery_write dev_pgmap_ops
To:     Jane Chu <jane.chu@oracle.com>
Cc:     david <david@fromorbit.com>, "Darrick J. Wong" <djwong@kernel.org>,
        Christoph Hellwig <hch@infradead.org>,
        Vishal L Verma <vishal.l.verma@intel.com>,
        Dave Jiang <dave.jiang@intel.com>,
        Alasdair Kergon <agk@redhat.com>,
        Mike Snitzer <snitzer@redhat.com>,
        device-mapper development <dm-devel@redhat.com>,
        "Weiny, Ira" <ira.weiny@intel.com>,
        Matthew Wilcox <willy@infradead.org>,
        Vivek Goyal <vgoyal@redhat.com>,
        linux-fsdevel <linux-fsdevel@vger.kernel.org>,
        Linux NVDIMM <nvdimm@lists.linux.dev>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-xfs <linux-xfs@vger.kernel.org>, X86 ML <x86@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Apr 5, 2022 at 12:48 PM Jane Chu <jane.chu@oracle.com> wrote:
>
> Introduce DAX_RECOVERY flag to dax_direct_access(). The flag is
> not set by default in dax_direct_access() such that the helper
> does not translate a pmem range to kernel virtual address if the
> range contains uncorrectable errors.  When the flag is set,
> the helper ignores the UEs and return kernel virtual adderss so
> that the caller may get on with data recovery via write.

It strikes me that there is likely never going to be any other flags
to dax_direct_access() and what this option really is an access type.
I also find code changes like this error prone to read:

 -       rc = dax_direct_access(iter->iomap.dax_dev, pgoff, 1, &kaddr, NULL);
 +       rc = dax_direct_access(iter->iomap.dax_dev, pgoff, 1, 0, &kaddr, NULL);

...i.e. without looking at the prototype, which option is the nr_pages
and which is the flags?

So how about change 'int flags' to 'enum dax_access_mode mode' where
dax_access_mode is:

/**
 * enum dax_access_mode - operational mode for dax_direct_access()
 * @DAX_ACCESS: nominal access, fail / trim access on encountering poison
 * @DAX_RECOVERY_WRITE: ignore poison and provide a pointer suitable
for use with dax_recovery_write()
 */
enum dax_access_mode {
    DAX_ACCESS,
    DAX_RECOVERY_WRITE,
};

Then the conversions look like this:

 -       rc = dax_direct_access(iter->iomap.dax_dev, pgoff, 1, &kaddr, NULL);
 +       rc = dax_direct_access(iter->iomap.dax_dev, pgoff, 1,
DAX_ACCESS, &kaddr, NULL);

...and there's less chance of confusion with the @nr_pages argument.