From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <mpatocka@redhat.com>
Date: Mon, 31 Oct 2016 12:36:16 -0400 (EDT)
From: Mikulas Patocka <mpatocka@redhat.com>
To: Christoph Hellwig <hch@infradead.org>
cc: Bart Van Assche <bart.vanassche@sandisk.com>, Jens Axboe <axboe@kernel.dk>,
        linux-block@vger.kernel.org, dm-devel@redhat.com,
        "Alasdair G. Kergon" <agk@redhat.com>,
        Mike Snitzer <snitzer@redhat.com>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        Chaitanya Kulkarni <Chaitanya.Kulkarni@wdc.com>
Subject: Re: REQ_OP for zeroing, was Re: [dm-devel] [PATCH 1/4] brd: handle
 misaligned discard
In-Reply-To: <20161028131457.GA15656@infradead.org>
Message-ID: <alpine.LRH.2.02.1610311233550.27707@file01.intranet.prod.int.rdu2.redhat.com>
References: <20161024155756.GA48306@redhat.com> <20161025130712.GA12717@infradead.org> <20161025143719.GA51266@redhat.com> <710b07a3-9091-6935-37c4-ea1dcedcab4f@kernel.dk> <alpine.LRH.2.02.1610261624100.12287@file01.intranet.prod.int.rdu2.redhat.com>
 <alpine.LRH.2.02.1610261625580.12287@file01.intranet.prod.int.rdu2.redhat.com> <338b76f0-40d1-904f-66e0-f6455c5287e7@sandisk.com> <alpine.LRH.2.02.1610261733190.25893@file01.intranet.prod.int.rdu2.redhat.com> <20161026215052.GA11679@infradead.org>
 <alpine.LRH.2.02.1610280741350.22676@file01.intranet.prod.int.rdu2.redhat.com> <20161028131457.GA15656@infradead.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
List-ID: <linux-block@vger.kernel.org>



On Fri, 28 Oct 2016, Christoph Hellwig wrote:

> [adding Chaitanya to Cc]
> 
> On Fri, Oct 28, 2016 at 07:43:41AM -0400, Mikulas Patocka wrote:
> > We could detect if the REQ_OP_WRITE_SAME command contains all zeroes and 
> > if it does, turn it into "Write Zeroes" or TRIM command (if the device 
> > guarantees zeroing on trim). If it doesn't contain all zeroes and the 
> > device doesn't support non-zero WRITE SAME, then reject it.
> 
> I don't like this because it's very inefficient - we have to allocate
> a payload first and then compare the whole payload for very operation.
> 
> > Or maybe we could add a new command REQ_OP_WRITE_ZEROES - I'm not sure 
> > which of these two possibilities is better.
> 
> Chaitanya actually did an initial prototype implementation of this for
> NVMe that he shared with me.  I liked it a lot and I think he'll be
> ready to post it in a few days.  Now that we have the REQ_OP* values
> instead of mapping different command types to flags it's actually
> surprisingly easy to add new block layer operations.

OK - when it is in the kernel, let me know, so that I can write device 
mapper support for that.

We should remove the flag "discard_zeroes_data" afterwards, because it is 
unreliable and impossible to support correctly in the device mapper.

Mikulas