On 10/6/22 2:42 PM, Pavel Begunkov wrote:
> Running local task_work requires taking uring_lock, for submit + wait we
> can try to run them right after submit while we still hold the lock and
> save one lock/unlokc pair. The optimisation was implemented in the first
> local tw patches but got dropped for simplicity.
>
> Suggested-by: Dylan Yudaken <dylany@xxxxxx>
> Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx>
> ---
> io_uring/io_uring.c | 12 ++++++++++--
> io_uring/io_uring.h | 7 +++++++
> 2 files changed, 17 insertions(+), 2 deletions(-)
>
> diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
> index 355fc1f3083d..b092473eca1d 100644
> --- a/io_uring/io_uring.c
> +++ b/io_uring/io_uring.c
> @@ -3224,8 +3224,16 @@ SYSCALL_DEFINE6(io_uring_enter, unsigned int, fd, u32, to_submit,
> mutex_unlock(&ctx->uring_lock);
> goto out;
> }
> - if ((flags & IORING_ENTER_GETEVENTS) && ctx->syscall_iopoll)
> - goto iopoll_locked;
> + if (flags & IORING_ENTER_GETEVENTS) {
> + if (ctx->syscall_iopoll)
> + goto iopoll_locked;
> + /*
> + * Ignore errors, we'll soon call io_cqring_wait() and
> + * it should handle ownership problems if any.
> + */
> + if (ctx->flags & IORING_SETUP_DEFER_TASKRUN)
> + (void)io_run_local_work_locked(ctx);
> + }
> mutex_unlock(&ctx->uring_lock);
> }
>
> diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h
> index e733d31f31d2..8504bc1f3839 100644
> --- a/io_uring/io_uring.h
> +++ b/io_uring/io_uring.h
> @@ -275,6 +275,13 @@ static inline int io_run_task_work_ctx(struct io_ring_ctx *ctx)
> return ret;
> }
>
> +static inline int io_run_local_work_locked(struct io_ring_ctx *ctx)
> +{
> + if (llist_empty(&ctx->work_llist))
> + return 0;
> + return __io_run_local_work(ctx, true);
> +}
Do you have pending patches that also use this? If not, maybe we
should just keep it in io_uring.c? If you do, then this looks fine
to me rather than needing to shuffle it later.
--
Jens Axboe
