Re: [PATCH for-next] io_uring: ensure IOSQE_ASYNC file table grabbing works, with SQPOLL

[Pavel Begunkov <asml.silence@gmail.com>]

On 08/09/2020 20:48, Jens Axboe wrote:
> Fd instantiating commands like IORING_OP_ACCEPT now work with SQPOLL, but
> we have an error in grabbing that if IOSQE_ASYNC is set. Ensure we assign
> the ring fd/file appropriately so we can defer grab them.

IIRC, for fcheck() in io_grab_files() to work it should be under fdget(),
that isn't the case with SQPOLL threads. Am I mistaken?

And it looks strange that the following snippet will effectively disable
such requests.

fd = dup(ring_fd)
close(ring_fd)
ring_fd = fd

> 
> Reported-by: Josef Grieb <josef.grieb@gmail.com>
> Signed-off-by: Jens Axboe <axboe@kernel.dk>
> ---
>  fs/io_uring.c | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/io_uring.c b/fs/io_uring.c
> index 80913973337a..e21a7a9c6a59 100644
> --- a/fs/io_uring.c
> +++ b/fs/io_uring.c
> @@ -6757,7 +6757,7 @@ static enum sq_ret __io_sq_thread(struct io_ring_ctx *ctx,
>  
>  	mutex_lock(&ctx->uring_lock);
>  	if (likely(!percpu_ref_is_dying(&ctx->refs)))
> -		ret = io_submit_sqes(ctx, to_submit, NULL, -1);
> +		ret = io_submit_sqes(ctx, to_submit, ctx->ring_file, ctx->ring_fd);
>  	mutex_unlock(&ctx->uring_lock);
>  
>  	if (!io_sqring_full(ctx) && wq_has_sleeper(&ctx->sqo_sq_wait))
> @@ -8966,6 +8966,11 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p,
>  		goto err;
>  	}
>  
> +	if (p->flags & IORING_SETUP_SQPOLL) {
> +		ctx->ring_fd = fd;
> +		ctx->ring_file = file;
> +	}
> +
>  	ret = io_sq_offload_create(ctx, p);
>  	if (ret)
>  		goto err;
> 

-- 
Pavel Begunkov