Re: [PATCH v4 6/6] io_uring: add support for zone-append

From: Pavel Begunkov <asml.silence@gmail.com>
To: Kanchan Joshi <joshi.k@samsung.com>,
	axboe@kernel.dk, viro@zeniv.linux.org.uk, bcrl@kvack.org
Cc: willy@infradead.org, hch@infradead.org, Damien.LeMoal@wdc.com,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-aio@kvack.org, io-uring@vger.kernel.org,
	linux-block@vger.kernel.org, linux-api@vger.kernel.org,
	SelvaKumar S <selvakuma.s1@samsung.com>,
	Nitesh Shetty <nj.shetty@samsung.com>,
	Javier Gonzalez <javier.gonz@samsung.com>
Subject: Re: [PATCH v4 6/6] io_uring: add support for zone-append
Date: Thu, 30 Jul 2020 18:57:47 +0300	[thread overview]
Message-ID: <2d3c8287-55cf-0150-acd6-19feb9e85771@gmail.com> (raw)
In-Reply-To: <1595605762-17010-7-git-send-email-joshi.k@samsung.com>

On 24/07/2020 18:49, Kanchan Joshi wrote:
> From: SelvaKumar S <selvakuma.s1@samsung.com>
> 
> Repurpose [cqe->res, cqe->flags] into cqe->res64 (signed) to report
> 64bit written-offset for zone-append. The appending-write which requires
> reporting written-location (conveyed by IOCB_ZONE_APPEND flag) is
> ensured not to be a short-write; this avoids the need to report
> number-of-bytes-copied.
> append-offset is returned by lower-layer to io-uring via ret2 of
> ki_complete interface. Make changes to collect it and send to user-space
> via cqe->res64.
> 
> Signed-off-by: SelvaKumar S <selvakuma.s1@samsung.com>
> Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>
> Signed-off-by: Nitesh Shetty <nj.shetty@samsung.com>
> Signed-off-by: Javier Gonzalez <javier.gonz@samsung.com>
> ---
>  fs/io_uring.c                 | 49 ++++++++++++++++++++++++++++++++++++-------
>  include/uapi/linux/io_uring.h |  9 ++++++--
>  2 files changed, 48 insertions(+), 10 deletions(-)
> 
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index 7809ab2..6510cf5 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
...
> @@ -1244,8 +1254,15 @@ static bool io_cqring_overflow_flush(struct io_ring_ctx *ctx, bool force)
>  		req->flags &= ~REQ_F_OVERFLOW;
>  		if (cqe) {
>  			WRITE_ONCE(cqe->user_data, req->user_data);
> -			WRITE_ONCE(cqe->res, req->result);
> -			WRITE_ONCE(cqe->flags, req->cflags);
> +			if (unlikely(req->flags & REQ_F_ZONE_APPEND)) {
> +				if (likely(req->result > 0))
> +					WRITE_ONCE(cqe->res64, req->rw.append_offset);
> +				else
> +					WRITE_ONCE(cqe->res64, req->result);
> +			} else {
> +				WRITE_ONCE(cqe->res, req->result);
> +				WRITE_ONCE(cqe->flags, req->cflags);
> +			}
>  		} else {
>  			WRITE_ONCE(ctx->rings->cq_overflow,
>  				atomic_inc_return(&ctx->cached_cq_overflow));
> @@ -1284,8 +1301,15 @@ static void __io_cqring_fill_event(struct io_kiocb *req, long res, long cflags)
>  	cqe = io_get_cqring(ctx);
>  	if (likely(cqe)) {
>  		WRITE_ONCE(cqe->user_data, req->user_data);
> -		WRITE_ONCE(cqe->res, res);
> -		WRITE_ONCE(cqe->flags, cflags);
> +		if (unlikely(req->flags & REQ_F_ZONE_APPEND)) {
> +			if (likely(res > 0))
> +				WRITE_ONCE(cqe->res64, req->rw.append_offset);

1. as I mentioned before, that's not not nice to ignore @cflags
2. that's not the right place for opcode specific handling
3. it doesn't work with overflowed reqs, see the final else below

For this scheme, I'd pass @append_offset as an argument. That should
also remove this extra if from the fast path, which Jens mentioned.

> +			else
> +				WRITE_ONCE(cqe->res64, res);
> +		} else {
> +			WRITE_ONCE(cqe->res, res);
> +			WRITE_ONCE(cqe->flags, cflags);
> +		}
>  	} else if (ctx->cq_overflow_flushed) {
>  		WRITE_ONCE(ctx->rings->cq_overflow,
>  				atomic_inc_return(&ctx->cached_cq_overflow));
> @@ -1943,7 +1967,7 @@ static inline void req_set_fail_links(struct io_kiocb *req)
>  		req->flags |= REQ_F_FAIL_LINK;
>  }
>  

-- 
Pavel Begunkov