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;
References: <20230706130930.64283-1-nmi@metaspace.dk>
 <20230706130930.64283-2-nmi@metaspace.dk>
 <51b660f3-8145-d35e-87b4-d9ac0623606d@kernel.org>
 <ZKdjVxMT/sVUA5BV@ovpn-8-34.pek2.redhat.com>
 <ZKuqt6QAXic3wuRX@infradead.org>
 <ZKvO+81b9fAx2L/r@ovpn-8-31.pek2.redhat.com>
 <ZKvQPAN9OkS3dZ4d@infradead.org> <87a5w3ymff.fsf@metaspace.dk>
 <ZK0RKkXFaQSotxVl@infradead.org> <875y6qzufc.fsf@metaspace.dk>
 <ZK0gcj4j/sRWx2Pl@infradead.org>
User-agent: mu4e 1.10.4; emacs 28.2.50
From:   "Andreas Hindborg (Samsung)" <nmi@metaspace.dk>
To:     Christoph Hellwig <hch@infradead.org>
Cc:     Ming Lei <ming.lei@redhat.com>,
        Damien Le Moal <dlemoal@kernel.org>,
        open list <linux-kernel@vger.kernel.org>,
        "open list:BLOCK LAYER" <linux-block@vger.kernel.org>,
        Minwoo Im <minwoo.im.dev@gmail.com>,
        Matias Bjorling <Matias.Bjorling@wdc.com>,
        gost.dev@samsung.com, Jens Axboe <axboe@kernel.dk>,
        Aravind Ramesh <Aravind.Ramesh@wdc.com>,
        Johannes Thumshirn <jth@kernel.org>,
        Hans Holmberg <Hans.Holmberg@wdc.com>
Subject: Re: [PATCH v6 1/3] ublk: add opcode offsets for DRV_IN/DRV_OUT
Date:   Tue, 11 Jul 2023 12:15:18 +0200
In-reply-to: <ZK0gcj4j/sRWx2Pl@infradead.org>
Message-ID: <871qhezr4d.fsf@metaspace.dk>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk


Christoph Hellwig <hch@infradead.org> writes:

> On Tue, Jul 11, 2023 at 11:02:15AM +0200, Andreas Hindborg (Samsung) wrote:
>> 
>> Christoph Hellwig <hch@infradead.org> writes:
>> 
>> > On Tue, Jul 11, 2023 at 08:23:40AM +0200, Andreas Hindborg (Samsung) wrote:
>> >> Yet most on-the-wire protocols for actual hardware does support this
>> >> some way or another.
>> >
>> > Supports what?  Passthrough?  No.
>> 
>> Both SCSI and NVMe has command identifier ranges reserved for vendor
>> specific commands. I would assume that one use of these is to implement
>> passthrough channels to a device for testing out new interfaces. Just
>> guessing though.
>
> Vendor specific commands is an entirely different concept from Linux
> passthrough requests.

And yet they are somewhat similar, in the sense that they allow the user
of a protocol to express semantics that is not captured in the
established protocol. Uring command passthrough -> request passthrough
-> vendor specific commands. They sort of map well in terms of what they
allow the user to achieve. Or did I misunderstand something completely?