io_uring: prolong tctx_task_work() with flushing

io_submit_flush_completions() may enqueue linked requests for task_work
execution, so don't leave tctx_task_work() right after the tw list is
exhausted, but try to flush and then retry.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/0755d4c2c36301447c63bdd4146c10477cea4249.1630539342.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
Pavel Begunkov 2021-09-02 00:38:23 +01:00 committed by Jens Axboe
parent 636378535a
commit 8d4ad41e3e
1 changed files with 3 additions and 0 deletions

View File

@ -2102,6 +2102,9 @@ static void tctx_task_work(struct callback_head *cb)
while (1) { while (1) {
struct io_wq_work_node *node; struct io_wq_work_node *node;
if (!tctx->task_list.first && locked && ctx->submit_state.compl_nr)
io_submit_flush_completions(ctx);
spin_lock_irq(&tctx->task_lock); spin_lock_irq(&tctx->task_lock);
node = tctx->task_list.first; node = tctx->task_list.first;
INIT_WQ_LIST(&tctx->task_list); INIT_WQ_LIST(&tctx->task_list);