From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-block-owner@vger.kernel.org>
Received: from bombadil.infradead.org ([65.50.211.133]:40606 "EHLO
        bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
        with ESMTP id S1751398AbdJPLo5 (ORCPT
        <rfc822;linux-block@vger.kernel.org>);
        Mon, 16 Oct 2017 07:44:57 -0400
Date: Mon, 16 Oct 2017 04:44:51 -0700
From: Christoph Hellwig <hch@infradead.org>
To: Ilya Dryomov <idryomov@gmail.com>
Cc: Christoph Hellwig <hch@infradead.org>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        Hannes Reinecke <hare@suse.com>, Jens Axboe <axboe@fb.com>,
        linux-block@vger.kernel.org
Subject: Re: [PATCH v2 2/2] block: cope with WRITE ZEROES failing in
 blkdev_issue_zeroout()
Message-ID: <20171016114451.GA14591@infradead.org>
References: <1507129396-20458-1-git-send-email-idryomov@gmail.com>
 <1507129396-20458-3-git-send-email-idryomov@gmail.com>
 <20171005171327.GB31208@infradead.org>
 <CAOi1vP_7EaNOBXup--HPPnOOvKKi_6_ksXUqE5YmGAicjz0oWA@mail.gmail.com>
 <20171006120526.GA19292@infradead.org>
 <CAOi1vP9P6TrVpMQr9C+YU07uT2dKCZLBfrnMWYpsqYzTQcxBzQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <CAOi1vP9P6TrVpMQr9C+YU07uT2dKCZLBfrnMWYpsqYzTQcxBzQ@mail.gmail.com>
Sender: linux-block-owner@vger.kernel.org
List-Id: linux-block@vger.kernel.org

On Fri, Oct 06, 2017 at 02:31:20PM +0200, Ilya Dryomov wrote:
> This would unconditionally overwrite any WRITE ZEROS error.  If we get
> e.g. -EIO, and manual zeroing is not allowed, I don't think we want to
> return -EOPNOTSUPP?
> 
> Returning -EOPNOTSUPP to mean "can't zero using either method" doesn't
> make sense to me, because manual zeroing is always supported, just not
> always allowed.

Then thow the throw bdev_write_zeroes_sectors check back in:

	if (ret && try_write_zeroes) {
                if (!(flags & BLKDEV_ZERO_NOFALLBACK))
                        try_write_zeroes = false;
                        goto retry;
                }
		if (!bdev_write_zeroes_sectors(bdev))
	                ret = -EOPNOTSUPP;
        }

The important bit is that the structure in the current patch where the
bdev_write_zeroes_sectors check is on the same level as the method
selection is extremely confusing.