io_uring: optimise io_req_task_work_add()

Inline io_task_work_add() into io_req_task_work_add(). They both work
with a request, so keeping them separate doesn't make things much more
clear, but merging allows optimise it. Apart from small wins like not
reading req->ctx or not calculating @notify in the hot path, i.e. with
tctx->task_state set, it avoids doing wake_up_process() for every single
add, but only after actually done task_work_add().

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 6b805b1c003db4c55705d6a90a5228781397e961..7a6b2b31332840cbcabda64831cb15d6e98ba0ca 100644 (file)
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -1911,13 +1911,17 @@ static void tctx_task_work(struct callback_head *cb)
                cond_resched();
 }
 
-static int io_task_work_add(struct task_struct *tsk, struct io_kiocb *req,
-                           enum task_work_notify_mode notify)
+static int io_req_task_work_add(struct io_kiocb *req)
 {
+       struct task_struct *tsk = req->task;
        struct io_uring_task *tctx = tsk->io_uring;
+       enum task_work_notify_mode notify;
        struct io_wq_work_node *node, *prev;
        unsigned long flags;
-       int ret;
+       int ret = 0;
+
+       if (unlikely(tsk->flags & PF_EXITING))
+               return -ESRCH;
 
        WARN_ON_ONCE(!tctx);
 
@@ -1930,14 +1934,23 @@ static int io_task_work_add(struct task_struct *tsk, struct io_kiocb *req,
            test_and_set_bit(0, &tctx->task_state))
                return 0;
 
-       if (!task_work_add(tsk, &tctx->task_work, notify))
+       /*
+        * SQPOLL kernel thread doesn't need notification, just a wakeup. For
+        * all other cases, use TWA_SIGNAL unconditionally to ensure we're
+        * processing task_work. There's no reliable way to tell if TWA_RESUME
+        * will do the job.
+        */
+       notify = (req->ctx->flags & IORING_SETUP_SQPOLL) ? TWA_NONE : TWA_SIGNAL;
+
+       if (!task_work_add(tsk, &tctx->task_work, notify)) {
+               wake_up_process(tsk);
                return 0;
+       }
 
        /*
         * Slow path - we failed, find and delete work. if the work is not
         * in the list, it got run and we're fine.
         */
-       ret = 0;
        spin_lock_irqsave(&tctx->task_lock, flags);
        wq_list_for_each(node, prev, &tctx->task_list) {
                if (&req->io_task_work.node == node) {
@@ -1951,33 +1964,6 @@ static int io_task_work_add(struct task_struct *tsk, struct io_kiocb *req,
        return ret;
 }
 
-static int io_req_task_work_add(struct io_kiocb *req)
-{
-       struct task_struct *tsk = req->task;
-       struct io_ring_ctx *ctx = req->ctx;
-       enum task_work_notify_mode notify;
-       int ret;
-
-       if (tsk->flags & PF_EXITING)
-               return -ESRCH;
-
-       /*
-        * SQPOLL kernel thread doesn't need notification, just a wakeup. For
-        * all other cases, use TWA_SIGNAL unconditionally to ensure we're
-        * processing task_work. There's no reliable way to tell if TWA_RESUME
-        * will do the job.
-        */
-       notify = TWA_NONE;
-       if (!(ctx->flags & IORING_SETUP_SQPOLL))
-               notify = TWA_SIGNAL;
-
-       ret = io_task_work_add(tsk, req, notify);
-       if (!ret)
-               wake_up_process(tsk);
-
-       return ret;
-}
-
 static bool io_run_task_work_head(struct callback_head **work_head)
 {
        struct callback_head *work, *next;