tty: stop using "delayed_work" in the tty layer

Using delayed-work for tty flip buffers ends up causing us to wait for
the next tick to complete some actions.  That's usually not all that
noticeable, but for certain latency-critical workloads it ends up being
totally unacceptable.

As an extreme case of this, passing a token back-and-forth over a pty
will take two ticks per iteration, so even just a thousand iterations
will take 8 seconds assuming a common 250Hz configuration.

Avoiding the whole delayed work issue brings that ping-pong test-case
down to 0.009s on my machine.

In more practical terms, this latency has been a performance problem for
things like dive computer simulators (simulating the serial interface
using the ptys) and for other environments (Alan mentions a CP/M emulator).
Reported-by: Jef Driesen <jefdriesen@telenet.be>
Acked-by: Greg KH <gregkh@suse.de>
Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/tty/tty_buffer.c

@@ -322,7 +322,7 @@ void tty_schedule_flip(struct tty_struct *tty)
 	if (tty->buf.tail != NULL)
 		tty->buf.tail->commit = tty->buf.tail->used;
 	spin_unlock_irqrestore(&tty->buf.lock, flags);
-	schedule_delayed_work(&tty->buf.work, 1);
+	schedule_work(&tty->buf.work);
 }
 EXPORT_SYMBOL(tty_schedule_flip);
 
@@ -402,7 +402,7 @@ EXPORT_SYMBOL_GPL(tty_prepare_flip_string_flags);
 static void flush_to_ldisc(struct work_struct *work)
 {
 	struct tty_struct *tty =
-		container_of(work, struct tty_struct, buf.work.work);
+		container_of(work, struct tty_struct, buf.work);
 	unsigned long 	flags;
 	struct tty_ldisc *disc;
 
@@ -443,7 +443,7 @@ static void flush_to_ldisc(struct work_struct *work)
 			if (test_bit(TTY_FLUSHPENDING, &tty->flags))
 				break;
 			if (!tty->receive_room || seen_tail) {
-				schedule_delayed_work(&tty->buf.work, 1);
+				schedule_work(&tty->buf.work);
 				break;
 			}
 			if (count > tty->receive_room)
@@ -481,7 +481,7 @@ static void flush_to_ldisc(struct work_struct *work)
  */
 void tty_flush_to_ldisc(struct tty_struct *tty)
 {
-	flush_delayed_work(&tty->buf.work);
+	flush_work(&tty->buf.work);
 }
 
 /**
@@ -506,9 +506,9 @@ void tty_flip_buffer_push(struct tty_struct *tty)
 	spin_unlock_irqrestore(&tty->buf.lock, flags);
 
 	if (tty->low_latency)
-		flush_to_ldisc(&tty->buf.work.work);
+		flush_to_ldisc(&tty->buf.work);
 	else
-		schedule_delayed_work(&tty->buf.work, 1);
+		schedule_work(&tty->buf.work);
 }
 EXPORT_SYMBOL(tty_flip_buffer_push);
 
@@ -529,6 +529,6 @@ void tty_buffer_init(struct tty_struct *tty)
 	tty->buf.tail = NULL;
 	tty->buf.free = NULL;
 	tty->buf.memory_used = 0;
-	INIT_DELAYED_WORK(&tty->buf.work, flush_to_ldisc);
+	INIT_WORK(&tty->buf.work, flush_to_ldisc);
 }
 

drivers/tty/tty_ldisc.c

@@ -529,7 +529,7 @@ static void tty_ldisc_restore(struct tty_struct *tty, struct tty_ldisc *old)
 static int tty_ldisc_halt(struct tty_struct *tty)
 {
 	clear_bit(TTY_LDISC, &tty->flags);
-	return cancel_delayed_work_sync(&tty->buf.work);
+	return cancel_work_sync(&tty->buf.work);
 }
 
 /**
@@ -542,7 +542,7 @@ static void tty_ldisc_flush_works(struct tty_struct *tty)
 {
 	flush_work_sync(&tty->hangup_work);
 	flush_work_sync(&tty->SAK_work);
-	flush_delayed_work_sync(&tty->buf.work);
+	flush_work_sync(&tty->buf.work);
 }
 
 /**
@@ -722,9 +722,9 @@ int tty_set_ldisc(struct tty_struct *tty, int ldisc)
 	/* Restart the work queue in case no characters kick it off. Safe if
 	   already running */
 	if (work)
-		schedule_delayed_work(&tty->buf.work, 1);
+		schedule_work(&tty->buf.work);
 	if (o_work)
-		schedule_delayed_work(&o_tty->buf.work, 1);
+		schedule_work(&o_tty->buf.work);
 	mutex_unlock(&tty->ldisc_mutex);
 	tty_unlock();
 	return retval;
@@ -830,12 +830,12 @@ void tty_ldisc_hangup(struct tty_struct *tty)
 
 	/*
 	 * this is like tty_ldisc_halt, but we need to give up
-	 * the BTM before calling cancel_delayed_work_sync,
-	 * which may need to wait for another function taking the BTM
+	 * the BTM before calling cancel_work_sync, which may
+	 * need to wait for another function taking the BTM
 	 */
 	clear_bit(TTY_LDISC, &tty->flags);
 	tty_unlock();
-	cancel_delayed_work_sync(&tty->buf.work);
+	cancel_work_sync(&tty->buf.work);
 	mutex_unlock(&tty->ldisc_mutex);
 
 	tty_lock();

include/linux/kbd_kern.h

@@ -159,7 +159,7 @@ static inline void con_schedule_flip(struct tty_struct *t)
 	if (t->buf.tail != NULL)
 		t->buf.tail->commit = t->buf.tail->used;
 	spin_unlock_irqrestore(&t->buf.lock, flags);
-	schedule_delayed_work(&t->buf.work, 0);
+	schedule_work(&t->buf.work);
 }
 
 #endif

include/linux/tty.h

@@ -82,7 +82,7 @@ struct tty_buffer {
 
 
 struct tty_bufhead {
-	struct delayed_work work;
+	struct work_struct work;
 	spinlock_t lock;
 	struct tty_buffer *head;	/* Queue head */
 	struct tty_buffer *tail;	/* Active buffer */