[PATCH 4/4] n_tty: Buffer work should not reschedule itself
From: Peter Hurley <hidden>
Date: 2013-06-15 11:29:35
Also in:
lkml
Subsystem:
the rest, tty layer and serial drivers · Maintainers:
Linus Torvalds, Greg Kroah-Hartman, Jiri Slaby
Although the driver-side input path must update the available buffer space, it should not reschedule itself. If space is still available and the flip buffers are not empty, flush_to_ldisc() will loop again. Signed-off-by: Peter Hurley <redacted> --- drivers/tty/n_tty.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-)
diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
index 0646165..4bf0fc0 100644
--- a/drivers/tty/n_tty.c
+++ b/drivers/tty/n_tty.c@@ -118,14 +118,14 @@ static inline int tty_put_user(struct tty_struct *tty, unsigned char x, * n_tty_set_room - receive space * @tty: terminal * - * Sets tty->receive_room to reflect the currently available space + * Updates tty->receive_room to reflect the currently available space * in the input buffer, and re-schedules the flip buffer work if space * just became available. * * Locks: Concurrent update is protected with read_lock */ -static void n_tty_set_room(struct tty_struct *tty) +static int set_room(struct tty_struct *tty) { struct n_tty_data *ldata = tty->disc_data; int left;
@@ -155,8 +155,13 @@ static void n_tty_set_room(struct tty_struct *tty) raw_spin_unlock_irqrestore(&ldata->read_lock, flags); + return left && !old_left; +} + +static void n_tty_set_room(struct tty_struct *tty) +{ /* Did this open up the receive buffer? We may need to flip */ - if (left && !old_left) { + if (set_room(tty)) { WARN_RATELIMIT(tty->port->itty == NULL, "scheduling with invalid itty\n"); /* see if ldisc has been killed - if so, this means that
@@ -1459,7 +1464,7 @@ static void n_tty_receive_buf(struct tty_struct *tty, const unsigned char *cp, tty->ops->flush_chars(tty); } - n_tty_set_room(tty); + set_room(tty); if ((!ldata->icanon && (ldata->read_cnt >= ldata->minimum_to_wake)) || L_EXTPROC(tty)) {
--
1.8.1.2