Skip to contentSkip to conversation

tty: Add driver unthrottle in ioctl(...,TCFLSH,..).

PATCH In review Ilya Zykov 3 participants 5 messages Nov 27, 2012 – Jan 16, 2013 TTY layer and serial drivers

lore
Share & tools
lore

Ilya Zykov

author
#0 lore raw

PATCH v4 1/1 tty: Add driver unthrottle in ioctl(...,TCFLSH,..).

Sorry. More correct.

Regression 'tty: fix "IRQ45: nobody cared"' Regression commit 7b292b4bf9a9d6098440d85616d6ca4c608b8304

Function reset_buffer_flags() also invoked during the ioctl(...,TCFLSH,..).
At the time of request we can have full buffers and throttled driver too. If we don't unthrottle driver, we can get forever throttled driver, because, after request, we will have empty buffers and throttled driver and there is no place to unthrottle driver.
It simple reproduce with "pty" pair then one side sleep on tty->write_wait, and other side do ioctl(...,TCFLSH,..). Then there is no place to do writers wake up.

---

drivers/tty/tty_ioctl.c +6 -2
index 12b1fa0..4071a8f 100644
@@ -1096,12 +1096,16 @@ int tty_perform_flush(struct tty_struct *tty, unsigned long arg)
 	ld = tty_ldisc_ref_wait(tty);
 	switch (arg) {
 	case TCIFLUSH:
-		if (ld && ld->ops->flush_buffer)
+		if (ld && ld->ops->flush_buffer) {
 			ld->ops->flush_buffer(tty);
+			tty_unthrottle(tty);
+		}
 		break;
 	case TCIOFLUSH:
-		if (ld && ld->ops->flush_buffer)
+		if (ld && ld->ops->flush_buffer) {
 			ld->ops->flush_buffer(tty);
+			tty_unthrottle(tty);
+		}
 		/* fall through */
 	case TCOFLUSH:
 		tty_driver_flush_buffer(tty);
#1 lore raw
Quoted 1 line · Ilya Zykov

Sorry. More correct.

In what way? Should I wait for the 6th version? :)

thanks,

greg k-h

author
#2 lore raw
Quoted 9 lines · Greg Kroah-Hartman
Quoted 1 line · Ilya Zykov

Sorry. More correct.

In what way? Should I wait for the 6th version? :)

thanks,

greg k-h

No, if only you will accept:
[PATCH]tty: Incorrect use tty_ldisc_flush() in TTY drivers.

It can be done another way, simple revert:
'tty: fix "IRQ45: nobody cared"'
commit 7b292b4bf9a9d6098440d85616d6ca4c608b8304

#3 lore raw
Quoted 7 lines

No, if only you will accept:
[PATCH]tty: Incorrect use tty_ldisc_flush() in TTY drivers.

It can be done another way, simple revert:
'tty: fix "IRQ45: nobody cared"'
commit 7b292b4bf9a9d6098440d85616d6ca4c608b8304

NAK that revert - that swaps a minor glitch you've discovered that does want fixing for a nasty bug.

author
#4 lore raw

PATCH 1/1 tty: Add driver unthrottle in ioctl(...,TCFLSH,..).

Regression 'tty: fix "IRQ45: nobody cared"' Regression commit 7b292b4bf9a9d6098440d85616d6ca4c608b8304

Function reset_buffer_flags() also invoked during the ioctl(...,TCFLSH,..).
At the time of request we can have full buffers and throttled driver too. If we don't unthrottle driver, we can get forever throttled driver, because, after request, we will have empty buffers and throttled driver and there is no place to unthrottle driver.
It simple reproduce with "pty" pair then one side sleep on tty->write_wait, and other side do ioctl(...,TCFLSH,..). Then there is no place to do writers wake up.

---
 drivers/tty/tty_ioctl.c |    8 ++++++--
 1 files changed, 6 insertions(+), 2 deletions(-)
drivers/tty/tty_ioctl.c +6 -2
index 8481b29..cc0fc52 100644
@@ -1096,12 +1096,16 @@ int tty_perform_flush(struct tty_struct *tty, unsigned long arg)
 	ld = tty_ldisc_ref_wait(tty);
 	switch (arg) {
 	case TCIFLUSH:
-		if (ld && ld->ops->flush_buffer)
+		if (ld && ld->ops->flush_buffer) {
 			ld->ops->flush_buffer(tty);
+			tty_unthrottle(tty);
+		}
 		break;
 	case TCIOFLUSH:
-		if (ld && ld->ops->flush_buffer)
+		if (ld && ld->ops->flush_buffer) {
 			ld->ops->flush_buffer(tty);
+			tty_unthrottle(tty);
+		}
 		/* fall through */
 	case TCOFLUSH:
 		tty_driver_flush_buffer(tty);