Re: [PATCH for-5.0] aio-posix: signal-proof fdmon-io_uring

["Philippe Mathieu-Daudé" <philmd@redhat.com>]

On 4/8/20 11:11 AM, Stefan Hajnoczi wrote:
> The io_uring_enter(2) syscall returns with errno=EINTR when interrupted
> by a signal.  Retry the syscall in this case.
> 
> It's essential to do this in the io_uring_submit_and_wait() case.  My
> interpretation of the Linux v5.5 io_uring_enter(2) code is that it
> shouldn't affect the io_uring_submit() case, but there is no guarantee
> this will always be the case.

This is how io_uring_enter() is documented indeed (no EINTR).

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>

> Let's check for -EINTR around both APIs.
> 
> Note that the liburing APIs have -errno return values.
> 
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
>   util/fdmon-io_uring.c | 10 ++++++++--
>   1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/util/fdmon-io_uring.c b/util/fdmon-io_uring.c
> index b4d6109f20..d5a80ed6fb 100644
> --- a/util/fdmon-io_uring.c
> +++ b/util/fdmon-io_uring.c
> @@ -88,7 +88,10 @@ static struct io_uring_sqe *get_sqe(AioContext *ctx)
>       }
>   
>       /* No free sqes left, submit pending sqes first */
> -    ret = io_uring_submit(ring);
> +    do {
> +        ret = io_uring_submit(ring);
> +    } while (ret == -EINTR);
> +
>       assert(ret > 1);
>       sqe = io_uring_get_sqe(ring);
>       assert(sqe);
> @@ -282,7 +285,10 @@ static int fdmon_io_uring_wait(AioContext *ctx, AioHandlerList *ready_list,
>   
>       fill_sq_ring(ctx);
>   
> -    ret = io_uring_submit_and_wait(&ctx->fdmon_io_uring, wait_nr);
> +    do {
> +        ret = io_uring_submit_and_wait(&ctx->fdmon_io_uring, wait_nr);
> +    } while (ret == -EINTR);
> +
>       assert(ret >= 0);
>   
>       return process_cq_ring(ctx, ready_list);
>