1 files changed, 237 insertions, 102 deletions

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 155f3d830ddb..32b0064f806e 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -541,6 +541,7 @@ enum {
 	REQ_F_NO_FILE_TABLE_BIT,
 	REQ_F_QUEUE_TIMEOUT_BIT,
 	REQ_F_WORK_INITIALIZED_BIT,
+	REQ_F_TASK_PINNED_BIT,
 
 	/* not a real bit, just to check we're not overflowing the space */
 	__REQ_F_LAST_BIT,
@@ -598,10 +599,13 @@ enum {
 	REQ_F_QUEUE_TIMEOUT	= BIT(REQ_F_QUEUE_TIMEOUT_BIT),
 	/* io_wq_work is initialized */
 	REQ_F_WORK_INITIALIZED	= BIT(REQ_F_WORK_INITIALIZED_BIT),
+	/* req->task is refcounted */
+	REQ_F_TASK_PINNED	= BIT(REQ_F_TASK_PINNED_BIT),
 };
 
 struct async_poll {
 	struct io_poll_iocb	poll;
+	struct io_poll_iocb	*double_poll;
 	struct io_wq_work	work;
 };
 
@@ -887,6 +891,7 @@ static int __io_sqe_files_update(struct io_ring_ctx *ctx,
 				 struct io_uring_files_update *ip,
 				 unsigned nr_args);
 static int io_grab_files(struct io_kiocb *req);
+static void io_complete_rw_common(struct kiocb *kiocb, long res);
 static void io_cleanup_req(struct io_kiocb *req);
 static int io_file_get(struct io_submit_state *state, struct io_kiocb *req,
 		       int fd, struct file **out_file, bool fixed);
@@ -910,6 +915,21 @@ struct sock *io_uring_get_socket(struct file *file)
 }
 EXPORT_SYMBOL(io_uring_get_socket);
 
+static void io_get_req_task(struct io_kiocb *req)
+{
+	if (req->flags & REQ_F_TASK_PINNED)
+		return;
+	get_task_struct(req->task);
+	req->flags |= REQ_F_TASK_PINNED;
+}
+
+/* not idempotent -- it doesn't clear REQ_F_TASK_PINNED */
+static void __io_put_req_task(struct io_kiocb *req)
+{
+	if (req->flags & REQ_F_TASK_PINNED)
+		put_task_struct(req->task);
+}
+
 static void io_file_put_work(struct work_struct *work);
 
 /*
@@ -1045,8 +1065,6 @@ static inline void io_req_work_grab_env(struct io_kiocb *req,
 		}
 		spin_unlock(&current->fs->lock);
 	}
-	if (!req->work.task_pid)
-		req->work.task_pid = task_pid_vnr(current);
 }
 
 static inline void io_req_work_drop_env(struct io_kiocb *req)
@@ -1079,6 +1097,8 @@ static inline void io_prep_async_work(struct io_kiocb *req,
 {
 	const struct io_op_def *def = &io_op_defs[req->opcode];
 
+	io_req_init_async(req);
+
 	if (req->flags & REQ_F_ISREG) {
 		if (def->hash_reg_file)
 			io_wq_hash_work(&req->work, file_inode(req->file));
@@ -1256,6 +1276,7 @@ static bool io_cqring_overflow_flush(struct io_ring_ctx *ctx, bool force)
 	if (cqe) {
 		clear_bit(0, &ctx->sq_check_overflow);
 		clear_bit(0, &ctx->cq_check_overflow);
+		ctx->rings->sq_flags &= ~IORING_SQ_CQ_OVERFLOW;
 	}
 	spin_unlock_irqrestore(&ctx->completion_lock, flags);
 	io_cqring_ev_posted(ctx);
@@ -1293,6 +1314,7 @@ static void __io_cqring_fill_event(struct io_kiocb *req, long res, long cflags)
 		if (list_empty(&ctx->cq_overflow_list)) {
 			set_bit(0, &ctx->sq_check_overflow);
 			set_bit(0, &ctx->cq_check_overflow);
+			ctx->rings->sq_flags |= IORING_SQ_CQ_OVERFLOW;
 		}
 		req->flags |= REQ_F_OVERFLOW;
 		refcount_inc(&req->refs);
@@ -1398,9 +1420,7 @@ static void __io_req_aux_free(struct io_kiocb *req)
 	kfree(req->io);
 	if (req->file)
 		io_put_file(req, req->file, (req->flags & REQ_F_FIXED_FILE));
-	if (req->task)
-		put_task_struct(req->task);
-
+	__io_put_req_task(req);
 	io_req_work_drop_env(req);
 }
 
@@ -1727,6 +1747,26 @@ static int io_put_kbuf(struct io_kiocb *req)
 	return cflags;
 }
 
+static void io_iopoll_queue(struct list_head *again)
+{
+	struct io_kiocb *req;
+
+	do {
+		req = list_first_entry(again, struct io_kiocb, list);
+		list_del(&req->list);
+
+		/* shouldn't happen unless io_uring is dying, cancel reqs */
+		if (unlikely(!current->mm)) {
+			io_complete_rw_common(&req->rw.kiocb, -EAGAIN);
+			io_put_req(req);
+			continue;
+		}
+
+		refcount_inc(&req->refs);
+		io_queue_async_work(req);
+	} while (!list_empty(again));
+}
+
 /*
  * Find and free completed poll iocbs
  */
@@ -1735,12 +1775,21 @@ static void io_iopoll_complete(struct io_ring_ctx *ctx, unsigned int *nr_events,
 {
 	struct req_batch rb;
 	struct io_kiocb *req;
+	LIST_HEAD(again);
+
+	/* order with ->result store in io_complete_rw_iopoll() */
+	smp_rmb();
 
 	rb.to_free = rb.need_iter = 0;
 	while (!list_empty(done)) {
 		int cflags = 0;
 
 		req = list_first_entry(done, struct io_kiocb, list);
+		if (READ_ONCE(req->result) == -EAGAIN) {
+			req->iopoll_completed = 0;
+			list_move_tail(&req->list, &again);
+			continue;
+		}
 		list_del(&req->list);
 
 		if (req->flags & REQ_F_BUFFER_SELECTED)
@@ -1758,18 +1807,9 @@ static void io_iopoll_complete(struct io_ring_ctx *ctx, unsigned int *nr_events,
 	if (ctx->flags & IORING_SETUP_SQPOLL)
 		io_cqring_ev_posted(ctx);
 	io_free_req_many(ctx, &rb);
-}
-
-static void io_iopoll_queue(struct list_head *again)
-{
-	struct io_kiocb *req;
 
-	do {
-		req = list_first_entry(again, struct io_kiocb, list);
-		list_del(&req->list);
-		refcount_inc(&req->refs);
-		io_queue_async_work(req);
-	} while (!list_empty(again));
+	if (!list_empty(&again))
+		io_iopoll_queue(&again);
 }
 
 static int io_do_iopoll(struct io_ring_ctx *ctx, unsigned int *nr_events,
@@ -1777,7 +1817,6 @@ static int io_do_iopoll(struct io_ring_ctx *ctx, unsigned int *nr_events,
 {
 	struct io_kiocb *req, *tmp;
 	LIST_HEAD(done);
-	LIST_HEAD(again);
 	bool spin;
 	int ret;
 
@@ -1803,13 +1842,6 @@ static int io_do_iopoll(struct io_ring_ctx *ctx, unsigned int *nr_events,
 		if (!list_empty(&done))
 			break;
 
-		if (req->result == -EAGAIN) {
-			list_move_tail(&req->list, &again);
-			continue;
-		}
-		if (!list_empty(&again))
-			break;
-
 		ret = kiocb->ki_filp->f_op->iopoll(kiocb, spin);
 		if (ret < 0)
 			break;
@@ -1822,9 +1854,6 @@ static int io_do_iopoll(struct io_ring_ctx *ctx, unsigned int *nr_events,
 	if (!list_empty(&done))
 		io_iopoll_complete(ctx, nr_events, &done);
 
-	if (!list_empty(&again))
-		io_iopoll_queue(&again);
-
 	return ret;
 }
 
@@ -1973,11 +2002,13 @@ static void io_complete_rw_iopoll(struct kiocb *kiocb, long res, long res2)
 	if (kiocb->ki_flags & IOCB_WRITE)
 		kiocb_end_write(req);
 
-	if (res != req->result)
+	if (res != -EAGAIN && res != req->result)
 		req_set_fail_links(req);
-	req->result = res;
-	if (res != -EAGAIN)
-		WRITE_ONCE(req->iopoll_completed, 1);
+
+	WRITE_ONCE(req->result, res);
+	/* order with io_poll_complete() checking ->result */
+	smp_wmb();
+	WRITE_ONCE(req->iopoll_completed, 1);
 }
 
 /*
@@ -2650,8 +2681,8 @@ copy_iov:
 		}
 	}
 out_free:
-	kfree(iovec);
-	req->flags &= ~REQ_F_NEED_CLEANUP;
+	if (!(req->flags & REQ_F_NEED_CLEANUP))
+		kfree(iovec);
 	return ret;
 }
 
@@ -2773,8 +2804,8 @@ copy_iov:
 		}
 	}
 out_free:
-	req->flags &= ~REQ_F_NEED_CLEANUP;
-	kfree(iovec);
+	if (!(req->flags & REQ_F_NEED_CLEANUP))
+		kfree(iovec);
 	return ret;
 }
 
@@ -3524,6 +3555,7 @@ static int io_sendmsg_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 	if (req->flags & REQ_F_NEED_CLEANUP)
 		return 0;
 
+	io->msg.msg.msg_name = &io->msg.addr;
 	io->msg.iov = io->msg.fast_iov;
 	ret = sendmsg_copy_msghdr(&io->msg.msg, sr->msg, sr->msg_flags,
 					&io->msg.iov);
@@ -3705,6 +3737,7 @@ static int __io_compat_recvmsg_copy_hdr(struct io_kiocb *req,
 
 static int io_recvmsg_copy_hdr(struct io_kiocb *req, struct io_async_ctx *io)
 {
+	io->msg.msg.msg_name = &io->msg.addr;
 	io->msg.iov = io->msg.fast_iov;
 
 #ifdef CONFIG_COMPAT
@@ -3813,10 +3846,16 @@ static int io_recvmsg(struct io_kiocb *req, bool force_nonblock)
 
 		ret = __sys_recvmsg_sock(sock, &kmsg->msg, req->sr_msg.msg,
 						kmsg->uaddr, flags);
-		if (force_nonblock && ret == -EAGAIN)
-			return io_setup_async_msg(req, kmsg);
+		if (force_nonblock && ret == -EAGAIN) {
+			ret = io_setup_async_msg(req, kmsg);
+			if (ret != -EAGAIN)
+				kfree(kbuf);
+			return ret;
+		}
 		if (ret == -ERESTARTSYS)
 			ret = -EINTR;
+		if (kbuf)
+			kfree(kbuf);
 	}
 
 	if (kmsg && kmsg->iov != kmsg->fast_iov)
@@ -4045,6 +4084,29 @@ struct io_poll_table {
 	int error;
 };
 
+static int io_req_task_work_add(struct io_kiocb *req, struct callback_head *cb)
+{
+	struct task_struct *tsk = req->task;
+	struct io_ring_ctx *ctx = req->ctx;
+	int ret, notify = TWA_RESUME;
+
+	/*
+	 * SQPOLL kernel thread doesn't need notification, just a wakeup.
+	 * If we're not using an eventfd, then TWA_RESUME is always fine,
+	 * as we won't have dependencies between request completions for
+	 * other kernel wait conditions.
+	 */
+	if (ctx->flags & IORING_SETUP_SQPOLL)
+		notify = 0;
+	else if (ctx->cq_ev_fd)
+		notify = TWA_SIGNAL;
+
+	ret = task_work_add(tsk, cb, notify);
+	if (!ret)
+		wake_up_process(tsk);
+	return ret;
+}
+
 static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,
 			   __poll_t mask, task_work_func_t func)
 {
@@ -4068,13 +4130,13 @@ static int __io_async_wake(struct io_kiocb *req, struct io_poll_iocb *poll,
 	 * of executing it. We can't safely execute it anyway, as we may not
 	 * have the needed state needed for it anyway.
 	 */
-	ret = task_work_add(tsk, &req->task_work, true);
+	ret = io_req_task_work_add(req, &req->task_work);
 	if (unlikely(ret)) {
 		WRITE_ONCE(poll->canceled, true);
 		tsk = io_wq_get_task(req->ctx->io_wq);
-		task_work_add(tsk, &req->task_work, true);
+		task_work_add(tsk, &req->task_work, 0);
+		wake_up_process(tsk);
 	}
-	wake_up_process(tsk);
 	return 1;
 }
 
@@ -4098,9 +4160,9 @@ static bool io_poll_rewait(struct io_kiocb *req, struct io_poll_iocb *poll)
 	return false;
 }
 
-static void io_poll_remove_double(struct io_kiocb *req)
+static void io_poll_remove_double(struct io_kiocb *req, void *data)
 {
-	struct io_poll_iocb *poll = (struct io_poll_iocb *) req->io;
+	struct io_poll_iocb *poll = data;
 
 	lockdep_assert_held(&req->ctx->completion_lock);
 
@@ -4120,7 +4182,7 @@ static void io_poll_complete(struct io_kiocb *req, __poll_t mask, int error)
 {
 	struct io_ring_ctx *ctx = req->ctx;
 
-	io_poll_remove_double(req);
+	io_poll_remove_double(req, req->io);
 	req->poll.done = true;
 	io_cqring_fill_event(req, error ? error : mangle_poll(mask));
 	io_commit_cqring(ctx);
@@ -4163,21 +4225,21 @@ static int io_poll_double_wake(struct wait_queue_entry *wait, unsigned mode,
 			       int sync, void *key)
 {
 	struct io_kiocb *req = wait->private;
-	struct io_poll_iocb *poll = (struct io_poll_iocb *) req->io;
+	struct io_poll_iocb *poll = req->apoll->double_poll;
 	__poll_t mask = key_to_poll(key);
 
 	/* for instances that support it check for an event match first: */
 	if (mask && !(mask & poll->events))
 		return 0;
 
-	if (req->poll.head) {
+	if (poll && poll->head) {
 		bool done;
 
-		spin_lock(&req->poll.head->lock);
-		done = list_empty(&req->poll.wait.entry);
+		spin_lock(&poll->head->lock);
+		done = list_empty(&poll->wait.entry);
 		if (!done)
-			list_del_init(&req->poll.wait.entry);
-		spin_unlock(&req->poll.head->lock);
+			list_del_init(&poll->wait.entry);
+		spin_unlock(&poll->head->lock);
 		if (!done)
 			__io_async_wake(req, poll, mask, io_poll_task_func);
 	}
@@ -4197,7 +4259,8 @@ static void io_init_poll_iocb(struct io_poll_iocb *poll, __poll_t events,
 }
 
 static void __io_queue_proc(struct io_poll_iocb *poll, struct io_poll_table *pt,
-			    struct wait_queue_head *head)
+			    struct wait_queue_head *head,
+			    struct io_poll_iocb **poll_ptr)
 {
 	struct io_kiocb *req = pt->req;
 
@@ -4208,7 +4271,7 @@ static void __io_queue_proc(struct io_poll_iocb *poll, struct io_poll_table *pt,
 	 */
 	if (unlikely(poll->head)) {
 		/* already have a 2nd entry, fail a third attempt */
-		if (req->io) {
+		if (*poll_ptr) {
 			pt->error = -EINVAL;
 			return;
 		}
@@ -4220,7 +4283,7 @@ static void __io_queue_proc(struct io_poll_iocb *poll, struct io_poll_table *pt,
 		io_init_poll_iocb(poll, req->poll.events, io_poll_double_wake);
 		refcount_inc(&req->refs);
 		poll->wait.private = req;
-		req->io = (void *) poll;
+		*poll_ptr = poll;
 	}
 
 	pt->error = 0;
@@ -4232,8 +4295,31 @@ static void io_async_queue_proc(struct file *file, struct wait_queue_head *head,
 			       struct poll_table_struct *p)
 {
 	struct io_poll_table *pt = container_of(p, struct io_poll_table, pt);
+	struct async_poll *apoll = pt->req->apoll;
 
-	__io_queue_proc(&pt->req->apoll->poll, pt, head);
+	__io_queue_proc(&apoll->poll, pt, head, &apoll->double_poll);
+}
+
+static void io_sq_thread_drop_mm(struct io_ring_ctx *ctx)
+{
+	struct mm_struct *mm = current->mm;
+
+	if (mm) {
+		kthread_unuse_mm(mm);
+		mmput(mm);
+	}
+}
+
+static int io_sq_thread_acquire_mm(struct io_ring_ctx *ctx,
+				   struct io_kiocb *req)
+{
+	if (io_op_defs[req->opcode].needs_mm && !current->mm) {
+		if (unlikely(!mmget_not_zero(ctx->sqo_mm)))
+			return -EFAULT;
+		kthread_use_mm(ctx->sqo_mm);
+	}
+
+	return 0;
 }
 
 static void io_async_task_func(struct callback_head *cb)
@@ -4261,20 +4347,27 @@ static void io_async_task_func(struct callback_head *cb)
 		}
 	}
 
+	io_poll_remove_double(req, apoll->double_poll);
 	spin_unlock_irq(&ctx->completion_lock);
 
 	/* restore ->work in case we need to retry again */
 	if (req->flags & REQ_F_WORK_INITIALIZED)
 		memcpy(&req->work, &apoll->work, sizeof(req->work));
+	kfree(apoll->double_poll);
 	kfree(apoll);
 
 	if (!canceled) {
 		__set_current_state(TASK_RUNNING);
+		if (io_sq_thread_acquire_mm(ctx, req)) {
+			io_cqring_add_event(req, -EFAULT);
+			goto end_req;
+		}
 		mutex_lock(&ctx->uring_lock);
 		__io_queue_sqe(req, NULL);
 		mutex_unlock(&ctx->uring_lock);
 	} else {
 		io_cqring_ev_posted(ctx);
+end_req:
 		req_set_fail_links(req);
 		io_double_put_req(req);
 	}
@@ -4348,7 +4441,6 @@ static bool io_arm_poll_handler(struct io_kiocb *req)
 	struct async_poll *apoll;
 	struct io_poll_table ipt;
 	__poll_t mask, ret;
-	bool had_io;
 
 	if (!req->file || !file_can_poll(req->file))
 		return false;
@@ -4360,14 +4452,13 @@ static bool io_arm_poll_handler(struct io_kiocb *req)
 	apoll = kmalloc(sizeof(*apoll), GFP_ATOMIC);
 	if (unlikely(!apoll))
 		return false;
+	apoll->double_poll = NULL;
 
 	req->flags |= REQ_F_POLLED;
 	if (req->flags & REQ_F_WORK_INITIALIZED)
 		memcpy(&apoll->work, &req->work, sizeof(req->work));
-	had_io = req->io != NULL;
 
-	get_task_struct(current);
-	req->task = current;
+	io_get_req_task(req);
 	req->apoll = apoll;
 	INIT_HLIST_NODE(&req->hash_node);
 
@@ -4383,13 +4474,11 @@ static bool io_arm_poll_handler(struct io_kiocb *req)
 	ret = __io_arm_poll_handler(req, &apoll->poll, &ipt, mask,
 					io_async_wake);
 	if (ret) {
-		ipt.error = 0;
-		/* only remove double add if we did it here */
-		if (!had_io)
-			io_poll_remove_double(req);
+		io_poll_remove_double(req, apoll->double_poll);
 		spin_unlock_irq(&ctx->completion_lock);
 		if (req->flags & REQ_F_WORK_INITIALIZED)
 			memcpy(&req->work, &apoll->work, sizeof(req->work));
+		kfree(apoll->double_poll);
 		kfree(apoll);
 		return false;
 	}
@@ -4420,11 +4509,13 @@ static bool io_poll_remove_one(struct io_kiocb *req)
 	bool do_complete;
 
 	if (req->opcode == IORING_OP_POLL_ADD) {
-		io_poll_remove_double(req);
+		io_poll_remove_double(req, req->io);
 		do_complete = __io_poll_remove_one(req, &req->poll);
 	} else {
 		struct async_poll *apoll = req->apoll;
 
+		io_poll_remove_double(req, apoll->double_poll);
+
 		/* non-poll requests have submit ref still */
 		do_complete = __io_poll_remove_one(req, &apoll->poll);
 		if (do_complete) {
@@ -4437,6 +4528,7 @@ static bool io_poll_remove_one(struct io_kiocb *req)
 			if (req->flags & REQ_F_WORK_INITIALIZED)
 				memcpy(&req->work, &apoll->work,
 				       sizeof(req->work));
+			kfree(apoll->double_poll);
 			kfree(apoll);
 		}
 	}
@@ -4537,7 +4629,7 @@ static void io_poll_queue_proc(struct file *file, struct wait_queue_head *head,
 {
 	struct io_poll_table *pt = container_of(p, struct io_poll_table, pt);
 
-	__io_queue_proc(&pt->req->poll, pt, head);
+	__io_queue_proc(&pt->req->poll, pt, head, (struct io_poll_iocb **) &pt->req->io);
 }
 
 static int io_poll_add_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
@@ -4555,8 +4647,7 @@ static int io_poll_add_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe
 	events = READ_ONCE(sqe->poll_events);
 	poll->events = demangle_poll(events) | EPOLLERR | EPOLLHUP;
 
-	get_task_struct(current);
-	req->task = current;
+	io_get_req_task(req);
 	return 0;
 }
 
@@ -4646,7 +4737,9 @@ static int io_timeout_remove_prep(struct io_kiocb *req,
 {
 	if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
 		return -EINVAL;
-	if (sqe->flags || sqe->ioprio || sqe->buf_index || sqe->len)
+	if (unlikely(req->flags & (REQ_F_FIXED_FILE | REQ_F_BUFFER_SELECT)))
+		return -EINVAL;
+	if (sqe->ioprio || sqe->buf_index || sqe->len)
 		return -EINVAL;
 
 	req->timeout.addr = READ_ONCE(sqe->addr);
@@ -4772,7 +4865,7 @@ static int io_async_cancel_one(struct io_ring_ctx *ctx, void *sqe_addr)
 	enum io_wq_cancel cancel_ret;
 	int ret = 0;
 
-	cancel_ret = io_wq_cancel_cb(ctx->io_wq, io_cancel_cb, sqe_addr);
+	cancel_ret = io_wq_cancel_cb(ctx->io_wq, io_cancel_cb, sqe_addr, false);
 	switch (cancel_ret) {
 	case IO_WQ_CANCEL_OK:
 		ret = 0;
@@ -4824,8 +4917,9 @@ static int io_async_cancel_prep(struct io_kiocb *req,
 {
 	if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
 		return -EINVAL;
-	if (sqe->flags || sqe->ioprio || sqe->off || sqe->len ||
-	    sqe->cancel_flags)
+	if (unlikely(req->flags & (REQ_F_FIXED_FILE | REQ_F_BUFFER_SELECT)))
+		return -EINVAL;
+	if (sqe->ioprio || sqe->off || sqe->len || sqe->cancel_flags)
 		return -EINVAL;
 
 	req->cancel.addr = READ_ONCE(sqe->addr);
@@ -4843,7 +4937,9 @@ static int io_async_cancel(struct io_kiocb *req)
 static int io_files_update_prep(struct io_kiocb *req,
 				const struct io_uring_sqe *sqe)
 {
-	if (sqe->flags || sqe->ioprio || sqe->rw_flags)
+	if (unlikely(req->flags & (REQ_F_FIXED_FILE | REQ_F_BUFFER_SELECT)))
+		return -EINVAL;
+	if (sqe->ioprio || sqe->rw_flags)
 		return -EINVAL;
 
 	req->files_update.offset = READ_ONCE(sqe->off);
@@ -5308,9 +5404,6 @@ static int io_issue_sqe(struct io_kiocb *req, const struct io_uring_sqe *sqe,
 	if ((ctx->flags & IORING_SETUP_IOPOLL) && req->file) {
 		const bool in_async = io_wq_current_is_worker();
 
-		if (req->result == -EAGAIN)
-			return -EAGAIN;
-
 		/* workqueue context doesn't hold uring_lock, grab it now */
 		if (in_async)
 			mutex_lock(&ctx->uring_lock);
@@ -5637,6 +5730,7 @@ fail_req:
 		 * Never try inline submit of IOSQE_ASYNC is set, go straight
 		 * to async execution.
 		 */
+		io_req_init_async(req);
 		req->work.flags |= IO_WQ_WORK_CONCURRENT;
 		io_queue_async_work(req);
 	} else {
@@ -5817,17 +5911,14 @@ static int io_init_req(struct io_ring_ctx *ctx, struct io_kiocb *req,
 	req->flags = 0;
 	/* one is dropped after submission, the other at completion */
 	refcount_set(&req->refs, 2);
-	req->task = NULL;
+	req->task = current;
 	req->result = 0;
 
 	if (unlikely(req->opcode >= IORING_OP_LAST))
 		return -EINVAL;
 
-	if (io_op_defs[req->opcode].needs_mm && !current->mm) {
-		if (unlikely(!mmget_not_zero(ctx->sqo_mm)))
-			return -EFAULT;
-		kthread_use_mm(ctx->sqo_mm);
-	}
+	if (unlikely(io_sq_thread_acquire_mm(ctx, req)))
+		return -EFAULT;
 
 	sqe_flags = READ_ONCE(sqe->flags);
 	/* enforce forwards compatibility on users */
@@ -5936,16 +6027,6 @@ fail_req:
 	return submitted;
 }
 
-static inline void io_sq_thread_drop_mm(struct io_ring_ctx *ctx)
-{
-	struct mm_struct *mm = current->mm;
-
-	if (mm) {
-		kthread_unuse_mm(mm);
-		mmput(mm);
-	}
-}
-
 static int io_sq_thread(void *data)
 {
 	struct io_ring_ctx *ctx = data;
@@ -5979,7 +6060,7 @@ static int io_sq_thread(void *data)
 		 * If submit got -EBUSY, flag us as needing the application
 		 * to enter the kernel to reap and flush events.
 		 */
-		if (!to_submit || ret == -EBUSY) {
+		if (!to_submit || ret == -EBUSY || need_resched()) {
 			/*
 			 * Drop cur_mm before scheduling, we can't hold it for
 			 * long periods (or over schedule()). Do this before
@@ -5995,7 +6076,7 @@ static int io_sq_thread(void *data)
 			 * more IO, we should wait for the application to
 			 * reap events and wake us up.
 			 */
-			if (!list_empty(&ctx->poll_list) ||
+			if (!list_empty(&ctx->poll_list) || need_resched() ||
 			    (!time_after(jiffies, timeout) && ret != -EBUSY &&
 			    !percpu_ref_is_dying(&ctx->refs))) {
 				if (current->task_works)
@@ -6021,9 +6102,9 @@ static int io_sq_thread(void *data)
 			}
 
 			/* Tell userspace we may need a wakeup call */
+			spin_lock_irq(&ctx->completion_lock);
 			ctx->rings->sq_flags |= IORING_SQ_NEED_WAKEUP;
-			/* make sure to read SQ tail after writing flags */
-			smp_mb();
+			spin_unlock_irq(&ctx->completion_lock);
 
 			to_submit = io_sqring_entries(ctx);
 			if (!to_submit || ret == -EBUSY) {
@@ -6041,13 +6122,17 @@ static int io_sq_thread(void *data)
 				schedule();
 				finish_wait(&ctx->sqo_wait, &wait);
 
+				spin_lock_irq(&ctx->completion_lock);
 				ctx->rings->sq_flags &= ~IORING_SQ_NEED_WAKEUP;
+				spin_unlock_irq(&ctx->completion_lock);
 				ret = 0;
 				continue;
 			}
 			finish_wait(&ctx->sqo_wait, &wait);
 
+			spin_lock_irq(&ctx->completion_lock);
 			ctx->rings->sq_flags &= ~IORING_SQ_NEED_WAKEUP;
+			spin_unlock_irq(&ctx->completion_lock);
 		}
 
 		mutex_lock(&ctx->uring_lock);
@@ -6146,15 +6231,23 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events,
 	do {
 		prepare_to_wait_exclusive(&ctx->wait, &iowq.wq,
 						TASK_INTERRUPTIBLE);
+		/* make sure we run task_work before checking for signals */
 		if (current->task_works)
 			task_work_run();
-		if (io_should_wake(&iowq, false))
-			break;
-		schedule();
 		if (signal_pending(current)) {
+			if (current->jobctl & JOBCTL_TASK_WORK) {
+				spin_lock_irq(&current->sighand->siglock);
+				current->jobctl &= ~JOBCTL_TASK_WORK;
+				recalc_sigpending();
+				spin_unlock_irq(&current->sighand->siglock);
+				continue;
+			}
 			ret = -EINTR;
 			break;
 		}
+		if (io_should_wake(&iowq, false))
+			break;
+		schedule();
 	} while (1);
 	finish_wait(&ctx->wait, &iowq.wq);
 
@@ -6626,6 +6719,7 @@ static int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg,
 		for (i = 0; i < nr_tables; i++)
 			kfree(ctx->file_data->table[i].files);
 
+		percpu_ref_exit(&ctx->file_data->refs);
 		kfree(ctx->file_data->table);
 		kfree(ctx->file_data);
 		ctx->file_data = NULL;
@@ -6778,8 +6872,10 @@ static int __io_sqe_files_update(struct io_ring_ctx *ctx,
 			}
 			table->files[index] = file;
 			err = io_sqe_file_register(ctx, file, i);
-			if (err)
+			if (err) {
+				fput(file);
 				break;
+			}
 		}
 		nr_args--;
 		done++;
@@ -7275,9 +7371,6 @@ static void io_ring_ctx_free(struct io_ring_ctx *ctx)
 	io_mem_free(ctx->sq_sqes);
 
 	percpu_ref_exit(&ctx->refs);
-	if (ctx->account_mem)
-		io_unaccount_mem(ctx->user,
-				ring_pages(ctx->sq_entries, ctx->cq_entries));
 	free_uid(ctx->user);
 	put_cred(ctx->creds);
 	kfree(ctx->cancel_hash);
@@ -7331,7 +7424,17 @@ static void io_ring_exit_work(struct work_struct *work)
 	if (ctx->rings)
 		io_cqring_overflow_flush(ctx, true);
 
-	wait_for_completion(&ctx->ref_comp);
+	/*
+	 * If we're doing polled IO and end up having requests being
+	 * submitted async (out-of-line), then completions can come in while
+	 * we're waiting for refs to drop. We need to reap these manually,
+	 * as nobody else will be looking for them.
+	 */
+	while (!wait_for_completion_timeout(&ctx->ref_comp, HZ/20)) {
+		io_iopoll_reap_events(ctx);
+		if (ctx->rings)
+			io_cqring_overflow_flush(ctx, true);
+	}
 	io_ring_ctx_free(ctx);
 }
 
@@ -7352,6 +7455,16 @@ static void io_ring_ctx_wait_and_kill(struct io_ring_ctx *ctx)
 	if (ctx->rings)
 		io_cqring_overflow_flush(ctx, true);
 	idr_for_each(&ctx->personality_idr, io_remove_personalities, ctx);
+
+	/*
+	 * Do this upfront, so we won't have a grace period where the ring
+	 * is closed but resources aren't reaped yet. This can cause
+	 * spurious failure in setting up a new ring.
+	 */
+	if (ctx->account_mem)
+		io_unaccount_mem(ctx->user,
+				ring_pages(ctx->sq_entries, ctx->cq_entries));
+
 	INIT_WORK(&ctx->exit_work, io_ring_exit_work);
 	queue_work(system_wq, &ctx->exit_work);
 }
@@ -7365,9 +7478,22 @@ static int io_uring_release(struct inode *inode, struct file *file)
 	return 0;
 }
 
+static bool io_wq_files_match(struct io_wq_work *work, void *data)
+{
+	struct files_struct *files = data;
+
+	return work->files == files;
+}
+
 static void io_uring_cancel_files(struct io_ring_ctx *ctx,
 				  struct files_struct *files)
 {
+	if (list_empty_careful(&ctx->inflight_list))
+		return;
+
+	/* cancel all at once, should be faster than doing it one by one*/
+	io_wq_cancel_cb(ctx->io_wq, io_wq_files_match, files, true);
+
 	while (!list_empty_careful(&ctx->inflight_list)) {
 		struct io_kiocb *cancel_req = NULL, *req;
 		DEFINE_WAIT(wait);
@@ -7398,6 +7524,7 @@ static void io_uring_cancel_files(struct io_ring_ctx *ctx,
 			if (list_empty(&ctx->cq_overflow_list)) {
 				clear_bit(0, &ctx->sq_check_overflow);
 				clear_bit(0, &ctx->cq_check_overflow);
+				ctx->rings->sq_flags &= ~IORING_SQ_CQ_OVERFLOW;
 			}
 			spin_unlock_irq(&ctx->completion_lock);
 
@@ -7423,6 +7550,14 @@ static void io_uring_cancel_files(struct io_ring_ctx *ctx,
 	}
 }
 
+static bool io_cancel_task_cb(struct io_wq_work *work, void *data)
+{
+	struct io_kiocb *req = container_of(work, struct io_kiocb, work);
+	struct task_struct *task = data;
+
+	return req->task == task;
+}
+
 static int io_uring_flush(struct file *file, void *data)
 {
 	struct io_ring_ctx *ctx = file->private_data;
@@ -7433,7 +7568,7 @@ static int io_uring_flush(struct file *file, void *data)
 	 * If the task is going away, cancel work it may have pending
 	 */
 	if (fatal_signal_pending(current) || (current->flags & PF_EXITING))
-		io_wq_cancel_pid(ctx->io_wq, task_pid_vnr(current));
+		io_wq_cancel_cb(ctx->io_wq, io_cancel_task_cb, current, true);
 
 	return 0;
 }