From 21a091b970cdbcf3e8ff829234b51be6f9192766 Mon Sep 17 00:00:00 2001
From: Dylan Yudaken <dylany@fb.com>
Date: Tue, 30 Aug 2022 05:50:12 -0700
Subject: [PATCH] io_uring: signal registered eventfd to process deferred task
work
Some workloads rely on a registered eventfd (via
io_uring_register_eventfd(3)) in order to wake up and process the
io_uring.
In the case of a ring setup with IORING_SETUP_DEFER_TASKRUN, that eventfd
also needs to be signalled when there are tasks to run.
This changes an old behaviour which assumed 1 eventfd signal implied at
least 1 CQE, however only when this new flag is set (and so old users will
not notice). This should be expected with the IORING_SETUP_DEFER_TASKRUN
flag as it is not guaranteed that every task will result in a CQE.
Signed-off-by: Dylan Yudaken <dylany@fb.com>
Link: https://lore.kernel.org/r/20220830125013.570060-7-dylany@fb.com
[axboe: fold in call_rcu() serialization fix]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
include/linux/io_uring_types.h | 2 +
io_uring/io_uring.c | 84 ++++++++++++++++++++++++----------
2 files changed, 63 insertions(+), 23 deletions(-)
diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index d56ff21851680..aa4d90a538663 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -184,6 +184,8 @@ struct io_ev_fd {
struct eventfd_ctx *cq_ev_fd;
unsigned int eventfd_async: 1;
struct rcu_head rcu;
+ atomic_t refs;
+ atomic_t ops;
};
struct io_alloc_cache {
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 0fd03da951139..3a6badb799eeb 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -125,6 +125,11 @@ enum {
IO_CHECK_CQ_DROPPED_BIT,
};
+enum {
+ IO_EVENTFD_OP_SIGNAL_BIT,
+ IO_EVENTFD_OP_FREE_BIT,
+};
+
struct io_defer_entry {
struct list_head list;
struct io_kiocb *req;
@@ -478,33 +483,28 @@ static __cold void io_queue_deferred(struct io_ring_ctx *ctx)
}
}
-static void io_eventfd_put(struct rcu_head *rcu)
+
+static void io_eventfd_ops(struct rcu_head *rcu)
{
struct io_ev_fd *ev_fd = container_of(rcu, struct io_ev_fd, rcu);
+ int ops = atomic_xchg(&ev_fd->ops, 0);
+
+ if (ops & BIT(IO_EVENTFD_OP_SIGNAL_BIT))
+ eventfd_signal(ev_fd->cq_ev_fd, 1);
- eventfd_ctx_put(ev_fd->cq_ev_fd);
- kfree(ev_fd);
+ /* IO_EVENTFD_OP_FREE_BIT may not be set here depending on callback
+ * ordering in a race but if references are 0 we know we have to free
+ * it regardless.
+ */
+ if (atomic_dec_and_test(&ev_fd->refs)) {
+ eventfd_ctx_put(ev_fd->cq_ev_fd);
+ kfree(ev_fd);
+ }
}
static void io_eventfd_signal(struct io_ring_ctx *ctx)
{
- struct io_ev_fd *ev_fd;
- bool skip;
-
- spin_lock(&ctx->completion_lock);
- /*
- * Eventfd should only get triggered when at least one event has been
- * posted. Some applications rely on the eventfd notification count only
- * changing IFF a new CQE has been added to the CQ ring. There's no
- * depedency on 1:1 relationship between how many times this function is
- * called (and hence the eventfd count) and number of CQEs posted to the
- * CQ ring.
- */
- skip = ctx->cached_cq_tail == ctx->evfd_last_cq_tail;
- ctx->evfd_last_cq_tail = ctx->cached_cq_tail;
- spin_unlock(&ctx->completion_lock);
- if (skip)
- return;
+ struct io_ev_fd *ev_fd = NULL;
rcu_read_lock();
/*
@@ -522,13 +522,46 @@ static void io_eventfd_signal(struct io_ring_ctx *ctx)
goto out;
if (READ_ONCE(ctx->rings->cq_flags) & IORING_CQ_EVENTFD_DISABLED)
goto out;
+ if (ev_fd->eventfd_async && !io_wq_current_is_worker())
+ goto out;
- if (!ev_fd->eventfd_async || io_wq_current_is_worker())
+ if (likely(eventfd_signal_allowed())) {
eventfd_signal(ev_fd->cq_ev_fd, 1);
+ } else {
+ atomic_inc(&ev_fd->refs);
+ if (!atomic_fetch_or(BIT(IO_EVENTFD_OP_SIGNAL_BIT), &ev_fd->ops))
+ call_rcu(&ev_fd->rcu, io_eventfd_ops);
+ else
+ atomic_dec(&ev_fd->refs);
+ }
+
out:
rcu_read_unlock();
}
+static void io_eventfd_flush_signal(struct io_ring_ctx *ctx)
+{
+ bool skip;
+
+ spin_lock(&ctx->completion_lock);
+
+ /*
+ * Eventfd should only get triggered when at least one event has been
+ * posted. Some applications rely on the eventfd notification count
+ * only changing IFF a new CQE has been added to the CQ ring. There's
+ * no depedency on 1:1 relationship between how many times this
+ * function is called (and hence the eventfd count) and number of CQEs
+ * posted to the CQ ring.
+ */
+ skip = ctx->cached_cq_tail == ctx->evfd_last_cq_tail;
+ ctx->evfd_last_cq_tail = ctx->cached_cq_tail;
+ spin_unlock(&ctx->completion_lock);
+ if (skip)
+ return;
+
+ io_eventfd_signal(ctx);
+}
+
void __io_commit_cqring_flush(struct io_ring_ctx *ctx)
{
if (ctx->off_timeout_used || ctx->drain_active) {
@@ -540,7 +573,7 @@ void __io_commit_cqring_flush(struct io_ring_ctx *ctx)
spin_unlock(&ctx->completion_lock);
}
if (ctx->has_evfd)
- io_eventfd_signal(ctx);
+ io_eventfd_flush_signal(ctx);
}
static inline void io_cqring_ev_posted(struct io_ring_ctx *ctx)
@@ -1071,6 +1104,8 @@ static void io_req_local_work_add(struct io_kiocb *req)
if (ctx->flags & IORING_SETUP_TASKRUN_FLAG)
atomic_or(IORING_SQ_TASKRUN, &ctx->rings->sq_flags);
+ if (ctx->has_evfd)
+ io_eventfd_signal(ctx);
io_cqring_wake(ctx);
}
@@ -2474,6 +2509,8 @@ static int io_eventfd_register(struct io_ring_ctx *ctx, void __user *arg,
ev_fd->eventfd_async = eventfd_async;
ctx->has_evfd = true;
rcu_assign_pointer(ctx->io_ev_fd, ev_fd);
+ atomic_set(&ev_fd->refs, 1);
+ atomic_set(&ev_fd->ops, 0);
return 0;
}
@@ -2486,7 +2523,8 @@ static int io_eventfd_unregister(struct io_ring_ctx *ctx)
if (ev_fd) {
ctx->has_evfd = false;
rcu_assign_pointer(ctx->io_ev_fd, NULL);
- call_rcu(&ev_fd->rcu, io_eventfd_put);
+ if (!atomic_fetch_or(BIT(IO_EVENTFD_OP_FREE_BIT), &ev_fd->ops))
+ call_rcu(&ev_fd->rcu, io_eventfd_ops);
return 0;
}
--
GitLab