Merge tag 'io_uring-5.12-2021-03-21' of git://git.kernel.dk/linux-block

Pull io_uring followup fixes from Jens Axboe:

 - The SIGSTOP change from Eric, so we properly ignore that for
   PF_IO_WORKER threads.

 - Disallow sending signals to PF_IO_WORKER threads in general, we're
   not interested in having them funnel back to the io_uring owning
   task.

 - Stable fix from Stefan, ensuring we properly break links for short
   send/sendmsg recv/recvmsg if MSG_WAITALL is set.

 - Catch and loop when needing to run task_work before a PF_IO_WORKER
   threads goes to sleep.

* tag 'io_uring-5.12-2021-03-21' of git://git.kernel.dk/linux-block:
  io_uring: call req_set_fail_links() on short send[msg]()/recv[msg]() with MSG_WAITALL
  io-wq: ensure task is running before processing task_work
  signal: don't allow STOP on PF_IO_WORKER threads
  signal: don't allow sending any signals to PF_IO_WORKER threads

fs/io-wq.c

@@ -386,13 +386,16 @@ static struct io_wq_work *io_get_next_work(struct io_wqe *wqe)
 	return NULL;
 }
 
-static void io_flush_signals(void)
+static bool io_flush_signals(void)
 {
 	if (unlikely(test_tsk_thread_flag(current, TIF_NOTIFY_SIGNAL))) {
+		__set_current_state(TASK_RUNNING);
 		if (current->task_works)
 			task_work_run();
 		clear_tsk_thread_flag(current, TIF_NOTIFY_SIGNAL);
+		return true;
 	}
+	return false;
 }
 
 static void io_assign_current_work(struct io_worker *worker,
@@ -499,7 +502,8 @@ static int io_wqe_worker(void *data)
 		}
 		__io_worker_idle(wqe, worker);
 		raw_spin_unlock_irq(&wqe->lock);
-		io_flush_signals();
+		if (io_flush_signals())
+			continue;
 		ret = schedule_timeout(WORKER_IDLE_TIMEOUT);
 		if (try_to_freeze() || ret)
 			continue;

fs/io_uring.c

@@ -4386,6 +4386,7 @@ static int io_sendmsg(struct io_kiocb *req, unsigned int issue_flags)
 	struct io_async_msghdr iomsg, *kmsg;
 	struct socket *sock;
 	unsigned flags;
+	int min_ret = 0;
 	int ret;
 
 	sock = sock_from_file(req->file);
@@ -4406,6 +4407,9 @@ static int io_sendmsg(struct io_kiocb *req, unsigned int issue_flags)
 	else if (issue_flags & IO_URING_F_NONBLOCK)
 		flags |= MSG_DONTWAIT;
 
+	if (flags & MSG_WAITALL)
+		min_ret = iov_iter_count(&kmsg->msg.msg_iter);
+
 	ret = __sys_sendmsg_sock(sock, &kmsg->msg, flags);
 	if ((issue_flags & IO_URING_F_NONBLOCK) && ret == -EAGAIN)
 		return io_setup_async_msg(req, kmsg);
@@ -4416,7 +4420,7 @@ static int io_sendmsg(struct io_kiocb *req, unsigned int issue_flags)
 	if (kmsg->free_iov)
 		kfree(kmsg->free_iov);
 	req->flags &= ~REQ_F_NEED_CLEANUP;
-	if (ret < 0)
+	if (ret < min_ret)
 		req_set_fail_links(req);
 	__io_req_complete(req, issue_flags, ret, 0);
 	return 0;
@@ -4429,6 +4433,7 @@ static int io_send(struct io_kiocb *req, unsigned int issue_flags)
 	struct iovec iov;
 	struct socket *sock;
 	unsigned flags;
+	int min_ret = 0;
 	int ret;
 
 	sock = sock_from_file(req->file);
@@ -4450,6 +4455,9 @@ static int io_send(struct io_kiocb *req, unsigned int issue_flags)
 	else if (issue_flags & IO_URING_F_NONBLOCK)
 		flags |= MSG_DONTWAIT;
 
+	if (flags & MSG_WAITALL)
+		min_ret = iov_iter_count(&msg.msg_iter);
+
 	msg.msg_flags = flags;
 	ret = sock_sendmsg(sock, &msg);
 	if ((issue_flags & IO_URING_F_NONBLOCK) && ret == -EAGAIN)
@@ -4457,7 +4465,7 @@ static int io_send(struct io_kiocb *req, unsigned int issue_flags)
 	if (ret == -ERESTARTSYS)
 		ret = -EINTR;
 
-	if (ret < 0)
+	if (ret < min_ret)
 		req_set_fail_links(req);
 	__io_req_complete(req, issue_flags, ret, 0);
 	return 0;
@@ -4609,6 +4617,7 @@ static int io_recvmsg(struct io_kiocb *req, unsigned int issue_flags)
 	struct socket *sock;
 	struct io_buffer *kbuf;
 	unsigned flags;
+	int min_ret = 0;
 	int ret, cflags = 0;
 	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
 
@@ -4640,6 +4649,9 @@ static int io_recvmsg(struct io_kiocb *req, unsigned int issue_flags)
 	else if (force_nonblock)
 		flags |= MSG_DONTWAIT;
 
+	if (flags & MSG_WAITALL)
+		min_ret = iov_iter_count(&kmsg->msg.msg_iter);
+
 	ret = __sys_recvmsg_sock(sock, &kmsg->msg, req->sr_msg.umsg,
 					kmsg->uaddr, flags);
 	if (force_nonblock && ret == -EAGAIN)
@@ -4653,7 +4665,7 @@ static int io_recvmsg(struct io_kiocb *req, unsigned int issue_flags)
 	if (kmsg->free_iov)
 		kfree(kmsg->free_iov);
 	req->flags &= ~REQ_F_NEED_CLEANUP;
-	if (ret < 0)
+	if (ret < min_ret || ((flags & MSG_WAITALL) && (kmsg->msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))))
 		req_set_fail_links(req);
 	__io_req_complete(req, issue_flags, ret, cflags);
 	return 0;
@@ -4668,6 +4680,7 @@ static int io_recv(struct io_kiocb *req, unsigned int issue_flags)
 	struct socket *sock;
 	struct iovec iov;
 	unsigned flags;
+	int min_ret = 0;
 	int ret, cflags = 0;
 	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
 
@@ -4699,6 +4712,9 @@ static int io_recv(struct io_kiocb *req, unsigned int issue_flags)
 	else if (force_nonblock)
 		flags |= MSG_DONTWAIT;
 
+	if (flags & MSG_WAITALL)
+		min_ret = iov_iter_count(&msg.msg_iter);
+
 	ret = sock_recvmsg(sock, &msg, flags);
 	if (force_nonblock && ret == -EAGAIN)
 		return -EAGAIN;
@@ -4707,7 +4723,7 @@ static int io_recv(struct io_kiocb *req, unsigned int issue_flags)
 out_free:
 	if (req->flags & REQ_F_BUFFER_SELECTED)
 		cflags = io_put_recv_kbuf(req);
-	if (ret < 0)
+	if (ret < min_ret || ((flags & MSG_WAITALL) && (msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))))
 		req_set_fail_links(req);
 	__io_req_complete(req, issue_flags, ret, cflags);
 	return 0;

kernel/signal.c

@@ -288,7 +288,8 @@ bool task_set_jobctl_pending(struct task_struct *task, unsigned long mask)
 			JOBCTL_STOP_SIGMASK | JOBCTL_TRAPPING));
 	BUG_ON((mask & JOBCTL_TRAPPING) && !(mask & JOBCTL_PENDING_MASK));
 
-	if (unlikely(fatal_signal_pending(task) || (task->flags & PF_EXITING)))
+	if (unlikely(fatal_signal_pending(task) ||
+		     (task->flags & (PF_EXITING | PF_IO_WORKER))))
 		return false;
 
 	if (mask & JOBCTL_STOP_SIGMASK)
@@ -833,6 +834,9 @@ static int check_kill_permission(int sig, struct kernel_siginfo *info,
 
 	if (!valid_signal(sig))
 		return -EINVAL;
+	/* PF_IO_WORKER threads don't take any signals */
+	if (t->flags & PF_IO_WORKER)
+		return -ESRCH;
 
 	if (!si_fromuser(info))
 		return 0;