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: bcPhnksmsCi51yECEl7BMEM4q6VRW3SDDa62ADnJxkkZjNpIotPK6WsSC78iPjox2yqVcHddwd
 CHB/XJ2LDbpGyCYxdvbVskQobPSm4rlxwxcmBAz321VHTpL51GtCYAVHmvadx+kDCIwwBjC2GS
 7iqrHGD0UYO2+9ZTKoJAjybO7URLWCyu6Hkp05kkoD4fftmWH+bV+Z1QtDJfOVEoLlrSrH9d5x
 qhlYww+INXV/cv3wmOYl+VAK0pBqCM+olKGYVT+p8NGLfCViDm2/nZNFrWOVGWNd3WTiam4xAP
 nFE=
From:   Damien Le Moal <Damien.LeMoal@wdc.com>
To:     Kanchan Joshi <joshi.k@samsung.com>,
        "axboe@kernel.dk" <axboe@kernel.dk>,
        "viro@zeniv.linux.org.uk" <viro@zeniv.linux.org.uk>,
        "bcrl@kvack.org" <bcrl@kvack.org>
CC:     "linux-fsdevel@vger.kernel.org" <linux-fsdevel@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "linux-aio@kvack.org" <linux-aio@kvack.org>,
        "io-uring@vger.kernel.org" <io-uring@vger.kernel.org>,
        "linux-block@vger.kernel.org" <linux-block@vger.kernel.org>,
        "selvakuma.s1@samsung.com" <selvakuma.s1@samsung.com>,
        "nj.shetty@samsung.com" <nj.shetty@samsung.com>,
        "javier.gonz@samsung.com" <javier.gonz@samsung.com>
Subject: Re: [PATCH 3/3] io_uring: add support for zone-append
Thread-Topic: [PATCH 3/3] io_uring: add support for zone-append
Thread-Index: AQHWRMyUe8chR0VkUU2K4ON9AWdJKQ==
Date:   Thu, 18 Jun 2020 07:39:58 +0000
Message-ID: <CY4PR04MB37510E916B6F243D189B4EB0E79B0@CY4PR04MB3751.namprd04.prod.outlook.com>
References: <1592414619-5646-1-git-send-email-joshi.k@samsung.com>
 <CGME20200617172713epcas5p352f2907a12bd4ee3c97be1c7d8e1569e@epcas5p3.samsung.com>
 <1592414619-5646-4-git-send-email-joshi.k@samsung.com>
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: f65a2c01-f6e3-4afb-792a-08d8135acd68
X-MS-Exchange-CrossTenant-originalarrivaltime: 18 Jun 2020 07:39:58.5536
 (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: Xu+c02sjtkU4uNRAUzrDGz4amiK/SWUQT2WqRy/iT0My/ThA9LxU38rnOa5spqA44VLf5sQLeqP1Tx5wHrrocA==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: CY4PR0401MB3586
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 2020/06/18 2:27, Kanchan Joshi wrote:=0A=
> From: Selvakumar S <selvakuma.s1@samsung.com>=0A=
> =0A=
> Introduce three new opcodes for zone-append -=0A=
> =0A=
>    IORING_OP_ZONE_APPEND     : non-vectord, similiar to IORING_OP_WRITE=
=0A=
>    IORING_OP_ZONE_APPENDV    : vectored, similar to IORING_OP_WRITEV=0A=
>    IORING_OP_ZONE_APPEND_FIXED : append using fixed-buffers=0A=
> =0A=
> Repurpose cqe->flags to return zone-relative offset.=0A=
> =0A=
> Signed-off-by: SelvaKumar S <selvakuma.s1@samsung.com>=0A=
> Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>=0A=
> Signed-off-by: Nitesh Shetty <nj.shetty@samsung.com>=0A=
> Signed-off-by: Javier Gonzalez <javier.gonz@samsung.com>=0A=
> ---=0A=
>  fs/io_uring.c                 | 72 +++++++++++++++++++++++++++++++++++++=
++++--=0A=
>  include/uapi/linux/io_uring.h |  8 ++++-=0A=
>  2 files changed, 77 insertions(+), 3 deletions(-)=0A=
> =0A=
> diff --git a/fs/io_uring.c b/fs/io_uring.c=0A=
> index 155f3d8..c14c873 100644=0A=
> --- a/fs/io_uring.c=0A=
> +++ b/fs/io_uring.c=0A=
> @@ -649,6 +649,10 @@ struct io_kiocb {=0A=
>  	unsigned long		fsize;=0A=
>  	u64			user_data;=0A=
>  	u32			result;=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	/* zone-relative offset for append, in bytes */=0A=
> +	u32			append_offset;=0A=
=0A=
this can overflow. u64 is needed.=0A=
=0A=
> +#endif=0A=
>  	u32			sequence;=0A=
>  =0A=
>  	struct list_head	link_list;=0A=
> @@ -875,6 +879,26 @@ static const struct io_op_def io_op_defs[] =3D {=0A=
>  		.hash_reg_file		=3D 1,=0A=
>  		.unbound_nonreg_file	=3D 1,=0A=
>  	},=0A=
> +	[IORING_OP_ZONE_APPEND] =3D {=0A=
> +		.needs_mm               =3D 1,=0A=
> +		.needs_file             =3D 1,=0A=
> +		.unbound_nonreg_file    =3D 1,=0A=
> +		.pollout		=3D 1,=0A=
> +	},=0A=
> +	[IORING_OP_ZONE_APPENDV] =3D {=0A=
> +	       .async_ctx              =3D 1,=0A=
> +	       .needs_mm               =3D 1,=0A=
> +	       .needs_file             =3D 1,=0A=
> +	       .hash_reg_file          =3D 1,=0A=
> +	       .unbound_nonreg_file    =3D 1,=0A=
> +	       .pollout			=3D 1,=0A=
> +	},=0A=
> +	[IORING_OP_ZONE_APPEND_FIXED] =3D {=0A=
> +	       .needs_file             =3D 1,=0A=
> +	       .hash_reg_file          =3D 1,=0A=
> +	       .unbound_nonreg_file    =3D 1,=0A=
> +	       .pollout			=3D 1,=0A=
> +	},=0A=
>  };=0A=
>  =0A=
>  static void io_wq_submit_work(struct io_wq_work **workptr);=0A=
> @@ -1285,7 +1309,16 @@ static void __io_cqring_fill_event(struct io_kiocb=
 *req, long res, long cflags)=0A=
>  	if (likely(cqe)) {=0A=
>  		WRITE_ONCE(cqe->user_data, req->user_data);=0A=
>  		WRITE_ONCE(cqe->res, res);=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +		if (req->opcode =3D=3D IORING_OP_ZONE_APPEND ||=0A=
> +				req->opcode =3D=3D IORING_OP_ZONE_APPENDV ||=0A=
> +				req->opcode =3D=3D IORING_OP_ZONE_APPEND_FIXED)=0A=
> +			WRITE_ONCE(cqe->res2, req->append_offset);=0A=
> +		else=0A=
> +			WRITE_ONCE(cqe->flags, cflags);=0A=
> +#else=0A=
>  		WRITE_ONCE(cqe->flags, cflags);=0A=
> +#endif=0A=
>  	} else if (ctx->cq_overflow_flushed) {=0A=
>  		WRITE_ONCE(ctx->rings->cq_overflow,=0A=
>  				atomic_inc_return(&ctx->cached_cq_overflow));=0A=
> @@ -1961,6 +1994,9 @@ static void io_complete_rw_common(struct kiocb *kio=
cb, long res)=0A=
>  static void io_complete_rw(struct kiocb *kiocb, long res, long res2)=0A=
>  {=0A=
>  	struct io_kiocb *req =3D container_of(kiocb, struct io_kiocb, rw.kiocb)=
;=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	req->append_offset =3D (u32)res2;=0A=
> +#endif=0A=
>  =0A=
>  	io_complete_rw_common(kiocb, res);=0A=
>  	io_put_req(req);=0A=
> @@ -1976,6 +2012,9 @@ static void io_complete_rw_iopoll(struct kiocb *kio=
cb, long res, long res2)=0A=
>  	if (res !=3D req->result)=0A=
>  		req_set_fail_links(req);=0A=
>  	req->result =3D res;=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	req->append_offset =3D (u32)res2;=0A=
> +#endif=0A=
>  	if (res !=3D -EAGAIN)=0A=
>  		WRITE_ONCE(req->iopoll_completed, 1);=0A=
>  }=0A=
> @@ -2408,7 +2447,8 @@ static ssize_t io_import_iovec(int rw, struct io_ki=
ocb *req,=0A=
>  	u8 opcode;=0A=
>  =0A=
>  	opcode =3D req->opcode;=0A=
> -	if (opcode =3D=3D IORING_OP_READ_FIXED || opcode =3D=3D IORING_OP_WRITE=
_FIXED) {=0A=
> +	if (opcode =3D=3D IORING_OP_READ_FIXED || opcode =3D=3D IORING_OP_WRITE=
_FIXED ||=0A=
> +			opcode =3D=3D IORING_OP_ZONE_APPEND_FIXED) {=0A=
>  		*iovec =3D NULL;=0A=
>  		return io_import_fixed(req, rw, iter);=0A=
>  	}=0A=
> @@ -2417,7 +2457,8 @@ static ssize_t io_import_iovec(int rw, struct io_ki=
ocb *req,=0A=
>  	if (req->buf_index && !(req->flags & REQ_F_BUFFER_SELECT))=0A=
>  		return -EINVAL;=0A=
>  =0A=
> -	if (opcode =3D=3D IORING_OP_READ || opcode =3D=3D IORING_OP_WRITE) {=0A=
> +	if (opcode =3D=3D IORING_OP_READ || opcode =3D=3D IORING_OP_WRITE ||=0A=
> +			opcode =3D=3D IORING_OP_ZONE_APPEND) {=0A=
>  		if (req->flags & REQ_F_BUFFER_SELECT) {=0A=
>  			buf =3D io_rw_buffer_select(req, &sqe_len, needs_lock);=0A=
>  			if (IS_ERR(buf)) {=0A=
> @@ -2704,6 +2745,9 @@ static int io_write(struct io_kiocb *req, bool forc=
e_nonblock)=0A=
>  		req->rw.kiocb.ki_flags &=3D ~IOCB_NOWAIT;=0A=
>  =0A=
>  	req->result =3D 0;=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	req->append_offset =3D 0;=0A=
> +#endif=0A=
>  	io_size =3D ret;=0A=
>  	if (req->flags & REQ_F_LINK_HEAD)=0A=
>  		req->result =3D io_size;=0A=
> @@ -2738,6 +2782,13 @@ static int io_write(struct io_kiocb *req, bool for=
ce_nonblock)=0A=
>  			__sb_writers_release(file_inode(req->file)->i_sb,=0A=
>  						SB_FREEZE_WRITE);=0A=
>  		}=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +		if (req->opcode =3D=3D IORING_OP_ZONE_APPEND ||=0A=
> +				req->opcode =3D=3D IORING_OP_ZONE_APPENDV ||=0A=
> +				req->opcode =3D=3D IORING_OP_ZONE_APPEND_FIXED)=0A=
> +			kiocb->ki_flags |=3D IOCB_ZONE_APPEND;=0A=
> +#endif=0A=
> +=0A=
>  		kiocb->ki_flags |=3D IOCB_WRITE;=0A=
>  =0A=
>  		if (!force_nonblock)=0A=
> @@ -4906,6 +4957,12 @@ static int io_req_defer_prep(struct io_kiocb *req,=
=0A=
>  	case IORING_OP_WRITEV:=0A=
>  	case IORING_OP_WRITE_FIXED:=0A=
>  	case IORING_OP_WRITE:=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	fallthrough;=0A=
> +	case IORING_OP_ZONE_APPEND:=0A=
> +	case IORING_OP_ZONE_APPENDV:=0A=
> +	case IORING_OP_ZONE_APPEND_FIXED:=0A=
> +#endif=0A=
>  		ret =3D io_write_prep(req, sqe, true);=0A=
>  		break;=0A=
>  	case IORING_OP_POLL_ADD:=0A=
> @@ -5038,6 +5095,12 @@ static void io_cleanup_req(struct io_kiocb *req)=
=0A=
>  	case IORING_OP_WRITEV:=0A=
>  	case IORING_OP_WRITE_FIXED:=0A=
>  	case IORING_OP_WRITE:=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	fallthrough;=0A=
> +	case IORING_OP_ZONE_APPEND:=0A=
> +	case IORING_OP_ZONE_APPENDV:=0A=
> +	case IORING_OP_ZONE_APPEND_FIXED:=0A=
> +#endif=0A=
>  		if (io->rw.iov !=3D io->rw.fast_iov)=0A=
>  			kfree(io->rw.iov);=0A=
>  		break;=0A=
> @@ -5086,6 +5149,11 @@ static int io_issue_sqe(struct io_kiocb *req, cons=
t struct io_uring_sqe *sqe,=0A=
>  		}=0A=
>  		ret =3D io_read(req, force_nonblock);=0A=
>  		break;=0A=
> +#ifdef CONFIG_BLK_DEV_ZONED=0A=
> +	case IORING_OP_ZONE_APPEND:=0A=
> +	case IORING_OP_ZONE_APPENDV:=0A=
> +	case IORING_OP_ZONE_APPEND_FIXED:=0A=
> +#endif=0A=
>  	case IORING_OP_WRITEV:=0A=
>  	case IORING_OP_WRITE_FIXED:=0A=
>  	case IORING_OP_WRITE:=0A=
> diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.=
h=0A=
> index 92c2269..6c8e932 100644=0A=
> --- a/include/uapi/linux/io_uring.h=0A=
> +++ b/include/uapi/linux/io_uring.h=0A=
> @@ -130,6 +130,9 @@ enum {=0A=
>  	IORING_OP_PROVIDE_BUFFERS,=0A=
>  	IORING_OP_REMOVE_BUFFERS,=0A=
>  	IORING_OP_TEE,=0A=
> +	IORING_OP_ZONE_APPEND,=0A=
> +	IORING_OP_ZONE_APPENDV,=0A=
> +	IORING_OP_ZONE_APPEND_FIXED,=0A=
>  =0A=
>  	/* this goes last, obviously */=0A=
>  	IORING_OP_LAST,=0A=
> @@ -157,7 +160,10 @@ enum {=0A=
>  struct io_uring_cqe {=0A=
>  	__u64	user_data;	/* sqe->data submission passed back */=0A=
>  	__s32	res;		/* result code for this event */=0A=
> -	__u32	flags;=0A=
> +	union {=0A=
> +		__u32	res2; /* res2 like aio, currently used for zone-append */=0A=
> +		__u32	flags;=0A=
> +	};=0A=
>  };=0A=
>  =0A=
>  /*=0A=
> =0A=
=0A=
=0A=
-- =0A=
Damien Le Moal=0A=
Western Digital Research=0A=