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: ano407McD//KZrd9gURjFVxGxyGJgF3K8iWvRkW3RSrSSom4szOGcUDmK5h8LPC9E7J7uBTIwe
 O6L6uLfv/Z5tnodipFGIjnSew+pA1QP8GcSA4u83rqht+4GDRdDggjx8Hn6Jy4RP01cEcx8qmU
 +b9S9XImnV9KhPEJEXnDdheiLDr5jKKXtdSvxchrSAEowE0N6cAIe6VIy/4rlHLrw9WNFSdRcu
 EM/bPnFaigKwuROS+w9kI9dGEYap90iO7E6BzA1e5gD1xXijqEVCh13Z5hueOWCAr4BgXqsSwa
 ef0=
From:   Damien Le Moal <Damien.LeMoal@wdc.com>
To:     Niklas Cassel <Niklas.Cassel@wdc.com>
CC:     Jonathan Corbet <corbet@lwn.net>, Jens Axboe <axboe@kernel.dk>,
        Keith Busch <kbusch@kernel.org>,
        Christoph Hellwig <hch@lst.de>,
        Sagi Grimberg <sagi@grimberg.me>,
        "James E.J. Bottomley" <jejb@linux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        "linux-doc@vger.kernel.org" <linux-doc@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "linux-block@vger.kernel.org" <linux-block@vger.kernel.org>,
        "linux-nvme@lists.infradead.org" <linux-nvme@lists.infradead.org>,
        "linux-scsi@vger.kernel.org" <linux-scsi@vger.kernel.org>
Subject: Re: [PATCH 1/2] block: add max_open_zones to blk-sysfs
Thread-Topic: [PATCH 1/2] block: add max_open_zones to blk-sysfs
Thread-Index: AQHWQ8jtlLWiiEumE0+QMSt8tXUnYw==
Date:   Fri, 3 Jul 2020 04:56:30 +0000
Message-ID: <CY4PR04MB37519F9526D1A28797271BF6E76A0@CY4PR04MB3751.namprd04.prod.outlook.com>
References: <20200616102546.491961-1-niklas.cassel@wdc.com>
 <20200616102546.491961-2-niklas.cassel@wdc.com>
 <CY4PR04MB3751C2C03ACCA263541DA348E76F0@CY4PR04MB3751.namprd04.prod.outlook.com>
 <20200702123755.GA609677@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: 1a93abd7-8f1f-477f-049e-08d81f0d73bd
X-MS-Exchange-CrossTenant-originalarrivaltime: 03 Jul 2020 04:56:30.8359
 (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: 4wdi2oI7Sgrnw4ELyS4rsFed9K68EVg4o72HyLPI7I+TXdFjgzrJeosfIZ8NHp3p2g5GA9kJ1b+3KPiCViZBpA==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: CY4PR0401MB3585
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On 2020/07/02 21:37, Niklas Cassel wrote:=0A=
> On Tue, Jun 30, 2020 at 01:49:41AM +0000, Damien Le Moal wrote:=0A=
>> On 2020/06/16 19:28, Niklas Cassel wrote:=0A=
>>> diff --git a/drivers/nvme/host/zns.c b/drivers/nvme/host/zns.c=0A=
>>> index c08f6281b614..af156529f3b6 100644=0A=
>>> --- a/drivers/nvme/host/zns.c=0A=
>>> +++ b/drivers/nvme/host/zns.c=0A=
>>> @@ -82,6 +82,7 @@ int nvme_update_zone_info(struct gendisk *disk, struc=
t nvme_ns *ns,=0A=
>>>  =0A=
>>>  	q->limits.zoned =3D BLK_ZONED_HM;=0A=
>>>  	blk_queue_flag_set(QUEUE_FLAG_ZONE_RESETALL, q);=0A=
>>> +	blk_queue_max_open_zones(q, le32_to_cpu(id->mor) + 1);=0A=
>>>  free_data:=0A=
>>>  	kfree(id);=0A=
>>>  	return status;=0A=
>>> diff --git a/drivers/scsi/sd_zbc.c b/drivers/scsi/sd_zbc.c=0A=
>>> index 183a20720da9..aa3564139b40 100644=0A=
>>> --- a/drivers/scsi/sd_zbc.c=0A=
>>> +++ b/drivers/scsi/sd_zbc.c=0A=
>>> @@ -717,6 +717,10 @@ int sd_zbc_read_zones(struct scsi_disk *sdkp, unsi=
gned char *buf)=0A=
>>>  	/* The drive satisfies the kernel restrictions: set it up */=0A=
>>>  	blk_queue_flag_set(QUEUE_FLAG_ZONE_RESETALL, q);=0A=
>>>  	blk_queue_required_elevator_features(q, ELEVATOR_F_ZBD_SEQ_WRITE);=0A=
>>> +	if (sdkp->zones_max_open =3D=3D U32_MAX)=0A=
>>> +		blk_queue_max_open_zones(q, 0);=0A=
>>> +	else=0A=
>>> +		blk_queue_max_open_zones(q, sdkp->zones_max_open);=0A=
>>=0A=
>> This is correct only for host-managed drives. Host-aware models define t=
he=0A=
>> "OPTIMAL NUMBER OF OPEN SEQUENTIAL WRITE PREFERRED ZONES" instead of a m=
aximum=0A=
>> number of open sequential write required zones.=0A=
>>=0A=
>> Since the standard does not actually explicitly define what the value of=
 the=0A=
>> maximum number of open sequential write required zones should be for a=
=0A=
>> host-aware drive, I would suggest to always have the max_open_zones valu=
e set to=0A=
>> 0 for host-aware disks.=0A=
> =0A=
> Isn't this already the case?=0A=
> =0A=
> At least according to the comments:=0A=
> =0A=
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/d=
rivers/scsi/sd_zbc.c?h=3Dv5.8-rc3#n555=0A=
> =0A=
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/d=
rivers/scsi/sd_zbc.c?h=3Dv5.8-rc3#n561=0A=
> =0A=
> We seem to set=0A=
> =0A=
> sdkp->zones_max_open =3D 0;=0A=
> =0A=
> for host-aware, and=0A=
> =0A=
> sdkp->zones_max_open =3D get_unaligned_be32(&buf[16]);=0A=
> =0A=
> for host-managed.=0A=
> =0A=
> So the blk_queue_max_open_zones(q, sdkp->zones_max_open) call in=0A=
> sd_zbc_read_zones() should already export this new sysfs property=0A=
> as 0 for host-aware disks.=0A=
=0A=
Oh, yes ! You are absolutely right. Forgot about that code :)=0A=
Please disregard this comment.=0A=
=0A=
> =0A=
> =0A=
> Kind regards,=0A=
> Niklas=0A=
> =0A=
>>=0A=
>>>  	nr_zones =3D round_up(sdkp->capacity, zone_blocks) >> ilog2(zone_bloc=
ks);=0A=
>>>  =0A=
>>>  	/* READ16/WRITE16 is mandatory for ZBC disks */=0A=
>>> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h=0A=
>>> index 8fd900998b4e..2f332f00501d 100644=0A=
>>> --- a/include/linux/blkdev.h=0A=
>>> +++ b/include/linux/blkdev.h=0A=
>>> @@ -520,6 +520,7 @@ struct request_queue {=0A=
>>>  	unsigned int		nr_zones;=0A=
>>>  	unsigned long		*conv_zones_bitmap;=0A=
>>>  	unsigned long		*seq_zones_wlock;=0A=
>>> +	unsigned int		max_open_zones;=0A=
>>>  #endif /* CONFIG_BLK_DEV_ZONED */=0A=
>>>  =0A=
>>>  	/*=0A=
>>> @@ -729,6 +730,17 @@ static inline bool blk_queue_zone_is_seq(struct re=
quest_queue *q,=0A=
>>>  		return true;=0A=
>>>  	return !test_bit(blk_queue_zone_no(q, sector), q->conv_zones_bitmap);=
=0A=
>>>  }=0A=
>>> +=0A=
>>> +static inline void blk_queue_max_open_zones(struct request_queue *q,=
=0A=
>>> +		unsigned int max_open_zones)=0A=
>>> +{=0A=
>>> +	q->max_open_zones =3D max_open_zones;=0A=
>>> +}=0A=
>>> +=0A=
>>> +static inline unsigned int queue_max_open_zones(const struct request_q=
ueue *q)=0A=
>>> +{=0A=
>>> +	return q->max_open_zones;=0A=
>>> +}=0A=
>>>  #else /* CONFIG_BLK_DEV_ZONED */=0A=
>>>  static inline unsigned int blk_queue_nr_zones(struct request_queue *q)=
=0A=
>>>  {=0A=
>>> @@ -744,6 +756,14 @@ static inline unsigned int blk_queue_zone_no(struc=
t request_queue *q,=0A=
>>>  {=0A=
>>>  	return 0;=0A=
>>>  }=0A=
>>> +static inline void blk_queue_max_open_zones(struct request_queue *q,=
=0A=
>>> +		unsigned int max_open_zones)=0A=
>>> +{=0A=
>>> +}=0A=
>>=0A=
>> Why is this one necessary ? For the !CONFIG_BLK_DEV_ZONED case, no drive=
r should=0A=
>> ever call this function.=0A=
> =0A=
> Will remove in v2.=0A=
> =0A=
=0A=
=0A=
-- =0A=
Damien Le Moal=0A=
Western Digital Research=0A=