Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
IronPort-SDR: GAF8GbCY2sMUDvs6oFDte2kMBtM77ZrZoggvsxxBn7dItF39O849VXdbaKEJslbY9JF+p+2xf8
 A+FKED8ex8sCU7BqJhU62Y5q07XogkBucjBTcbkrXJgHTCkUXsnPBN5hNWvJWat7TblslnM8ak
 5FXGJO5HbCklT7XUkIKaYZtaF1bauuY8cN7aFVDbdKCE4FjU3xes54tWGGQt24kUtId9HZuh9T
 3d2AaYREH2Ak+8Ma9+4gqaLZDnwNCNFHWmM+m0GOERNLEgWh7Z/N+k36C+j9hSWjOc0z8/XYxw
 MQ8=
From:   Chaitanya Kulkarni <Chaitanya.Kulkarni@wdc.com>
To:     Niklas Cassel <Niklas.Cassel@wdc.com>
CC:     Christoph Hellwig <hch@lst.de>, Sagi Grimberg <sagi@grimberg.me>,
        "linux-nvme@lists.infradead.org" <linux-nvme@lists.infradead.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 2/2] nvmet: remove workarounds for gcc bug wrt unnamed
 fields in initializers
Thread-Topic: [PATCH 2/2] nvmet: remove workarounds for gcc bug wrt unnamed
 fields in initializers
Thread-Index: AQHWRX1ZyGByOih9fUSrbaAjQMq7Ig==
Date:   Thu, 18 Jun 2020 17:29:00 +0000
Message-ID: <BYAPR04MB4965AA7B57AC6B6E19BCBBA5869B0@BYAPR04MB4965.namprd04.prod.outlook.com>
References: <20200618143241.1056800-1-niklas.cassel@wdc.com>
 <20200618143241.1056800-2-niklas.cassel@wdc.com>
 <BYAPR04MB49657026BADC613CA83CB896869B0@BYAPR04MB4965.namprd04.prod.outlook.com>
 <20200618161509.GA1059668@localhost.localdomain>
Accept-Language: en-US
Content-Language: en-US
wdcipoutbound: EOP-TRUE
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-Network-Message-Id: 385760c0-5ad9-4477-7bc5-08d813ad16de
X-MS-Exchange-CrossTenant-originalarrivaltime: 18 Jun 2020 17:29:00.5315
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: b61c8803-16f3-4c35-9b17-6f65f441df86
X-MS-Exchange-CrossTenant-mailboxtype: HOSTED
X-MS-Exchange-CrossTenant-userprincipalname: TYtemYXtOtzOq6Jg4HAgEjweYlr84bxJuPxJc4zoENwriFwFbxUk+8qIO3pQzdIBrVNEByNolvkbSjQ4brf3hsgKKmXGIqQ9SR/tcuHDRz0=
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BYAPR04MB5032
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

I'm  not against the code cleanup and it always welcome.=0A=
Please also have a look at other comment.=0A=
=0A=
>> What is the issue with existing code that we need this patch for ?=0A=
>>=0A=
> =0A=
> Hello Chaitanya,=0A=
> =0A=
> This is just a cleanup patch, no functional change intended.=0A=
> =0A=
I can see that.=0A=
> It simply performs the initialization at declaration time, which is how w=
e=0A=
> usually initialize a subset of all fields.=0A=
Absolutely not true in case nvme subsystem.=0A=
> =0A=
> This is also how it was originally done, but this was changed to a=0A=
> non-standard way in order to workaround a compiler bug.=0A=
> =0A=
and the existing code matches the pattern present in the repo no need to =
=0A=
change if it is not causing any problem.=0A=
> Since the compiler bug is no longer relevant, we can go back to the=0A=
> standard way of doing things.=0A=
Is there any problem with the code now which mandates this change ? =0A=
which I don't see any.=0A=
> =0A=
> Performing initialization in a uniform way makes it easier to read and=0A=
> comprehend the code, especially for people unfamiliar with it, since=0A=
> it follows the same pattern used in other places of the kernel.=0A=
> =0A=
This is absolutely not uniform way infact what you are proposing is true =
=0A=
then we need to change all existing function which does not follow this =0A=
pattern, have a look at the following list.=0A=
=0A=
In NVMe subsystem case there are more than 10 functions which are on the =
=0A=
similar line of this function and doesn't initialize the variable at the =
=0A=
time declaration.=0A=
=0A=
Please have a look the code :-=0A=
nvmf_reg_read32=0A=
nvmf_reg_read64=0A=
nvmf_reg_write32=0A=
nvmf_connect_admin_queue=0A=
nvmf_connect_io_queue=0A=
nvme_identify_ctrl=0A=
nvme_identify_ns_descs=0A=
nvme_identify_ns_list=0A=
nvme_identify_ns=0A=
nvme_features=0A=
nvme_get_log=0A=
nvme_toggle_streams=0A=
nvme_get_stream_params=0A=
=0A=
Also here :-=0A=
nvme_user_cmd=0A=
nvme_user_cmd64=0A=
=0A=
Last two are an exception of copy_from_user() call before initialization =
=0A=
case, but we can do copy from user from caller and pass that as an =0A=
argument if we really want to follow the declare-init pattern.=0A=
=0A=
In several places we don't follow this pattern when function is compact =0A=
and it looks ugly for larger structures such as this example. this is=0A=
exactly the reason {} used in nvme subsystem.=0A=
=0A=
> Just reading e.g. struct rdma_conn_param  param =3D { }; one assumes that=
=0A=
> all fields will be zero initialized.. reading futher down in the function=
=0A=
=0A=
No this is standard style and used in nvme/host/core.c everywhere =0A=
nothing wrong with this at all, please have a look at the author.=0A=
=0A=
> you realize that this function actually does initialize the struct..=0A=
> which causes a mental hiccup, so you need to do a mental pipeline flush=
=0A=
> and go back and read the code from the beginning. This only happens with=
=0A=
> patterns that deviate from the standard way of doing things.=0A=
=0A=
The function is small enough for such hiccup which follows the pattern =0A=
which we have it in the tree there is nothing wrong.=0A=
=0A=
> =0A=
> Kind regards,=0A=
> Niklas=0A=
> =0A=
=0A=
=0A=