[PATCH 0/9] tty: drop broken alt-speed handling

[PATCH 0/9] tty: drop broken alt-speed handling

[Johan Hovold]

Setting an alt-speed using TIOCSSERIAL and SPD flags has been deprecated
since v2.1.69 and has been broken for all tty drivers but serial-core
since v3.10 and commit 6865ff222cca ("TTY: do not warn about setting
speed via SPD_*") without anyone noticing (for four years).

Instead of reverting the offending commit, lets get rid of this legacy
code while adding back a warning about the SPD flags being deprecated to
the drivers that once implemented it. Note that drivers implementing
SPD_CUST will continue to support using a custom divisor.

Also note that serial-core did not rely on the TTY layer for SPD
handling and continues to support it while warning about deprecation
(since 2003 at least).

Greg, I suggest you take it all trough the TTY tree even if merging
through multiple trees and applying the final patch once everything is
upstream would be an option. Also the irda clean up does not depend on
the rest of series as the code implementing SPD handling was ifdefed
out.

Johan


Johan Hovold (9):
  serial: rate limit custom-speed deprecation notice
  USB: serial: ftdi_sio: simplify TIOCSSERIAL flag logic
  USB: serial: ftdi_sio: remove broken alt-speed handling
  tty: simserial: drop unused alt_speed handling
  tty: amiserial: drop broken alt-speed support
  tty: cyclades: drop broken alt-speed support
  tty: rocket: drop broken alt-speed support
  tty: ircomm: remove dead and broken ioctl code
  tty: drop unused alt_speed from tty_struct

 arch/ia64/hp/sim/simserial.c       |  13 -----
 drivers/tty/amiserial.c            |  23 ++------
 drivers/tty/cyclades.c             |  21 ++++----
 drivers/tty/rocket.c               |  27 +++-------
 drivers/tty/serial/serial_core.c   |   5 +-
 drivers/usb/serial/ftdi_sio.c      |  67 +++++------------------
 include/linux/tty.h                |   1 -
 net/irda/ircomm/ircomm_tty_ioctl.c | 107 +------------------------------------
 8 files changed, 34 insertions(+), 230 deletions(-)

-- 
2.13.0

[PATCH 1/9] serial: rate limit custom-speed deprecation notice

[Johan Hovold]

Contrary to what a comment claimed, the ASYNC_SPD flags and custom
divisor can be set by a non-privileged user so rate limit the
deprecation notice as was intended.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/tty/serial/serial_core.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/tty/serial/serial_core.c b/drivers/tty/serial/serial_core.c
index 13bfd5dcffce..f534a40aebde 100644
--- a/drivers/tty/serial/serial_core.c
+++ b/drivers/tty/serial/serial_core.c
@@ -954,11 +954,10 @@ static int uart_set_info(struct tty_struct *tty, struct tty_port *port,
 		    old_custom_divisor != uport->custom_divisor) {
 			/*
 			 * If they're setting up a custom divisor or speed,
-			 * instead of clearing it, then bitch about it. No
-			 * need to rate-limit; it's CAP_SYS_ADMIN only.
+			 * instead of clearing it, then bitch about it.
 			 */
 			if (uport->flags & UPF_SPD_MASK) {
-				dev_notice(uport->dev,
+				dev_notice_ratelimited(uport->dev,
 				       "%s sets custom speed on %s. This is deprecated.\n",
 				      current->comm,
 				      tty_name(port->tty));
-- 
2.13.0

[PATCH 2/9] USB: serial: ftdi_sio: simplify TIOCSSERIAL flag logic

[Johan Hovold]

Simplify TIOCSSERIAL flag logic somewhat.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/usb/serial/ftdi_sio.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
index aba74f817dc6..df5c45a4b1d7 100644
--- a/drivers/usb/serial/ftdi_sio.c
+++ b/drivers/usb/serial/ftdi_sio.c
@@ -1505,8 +1505,7 @@ static int set_serial_info(struct tty_struct *tty,
 	/* Do error checking and permission checking */
 
 	if (!capable(CAP_SYS_ADMIN)) {
-		if (((new_serial.flags & ~ASYNC_USR_MASK) !=
-		     (priv->flags & ~ASYNC_USR_MASK))) {
+		if ((new_serial.flags ^ priv->flags) & ~ASYNC_USR_MASK) {
 			mutex_unlock(&priv->cfg_lock);
 			return -EPERM;
 		}
@@ -1530,8 +1529,7 @@ static int set_serial_info(struct tty_struct *tty,
 check_and_exit:
 	write_latency_timer(port);
 
-	if ((old_priv.flags & ASYNC_SPD_MASK) !=
-	     (priv->flags & ASYNC_SPD_MASK)) {
+	if ((priv->flags ^ old_priv.flags) & ASYNC_SPD_MASK) {
 		if ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
 			tty->alt_speed = 57600;
 		else if ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
@@ -1543,10 +1541,9 @@ static int set_serial_info(struct tty_struct *tty,
 		else
 			tty->alt_speed = 0;
 	}
-	if (((old_priv.flags & ASYNC_SPD_MASK) !=
-	     (priv->flags & ASYNC_SPD_MASK)) ||
-	    (((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_CUST) &&
-	     (old_priv.custom_divisor != priv->custom_divisor))) {
+	if ((priv->flags ^ old_priv.flags) & ASYNC_SPD_MASK ||
+			((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_CUST &&
+			 priv->custom_divisor != old_priv.custom_divisor)) {
 		change_speed(tty, port);
 		mutex_unlock(&priv->cfg_lock);
 	}
-- 
2.13.0

[PATCH 3/9] USB: serial: ftdi_sio: remove broken alt-speed handling

[Johan Hovold]

Remove the broken alt_speed code, and warn when trying to set the line
speed using TIOCSSERIAL and SPD flags.

The use of SPD flags to set the line speed has been deprecated since
v2.1.69 and support for alt_speed (e.g. "warp") has even been removed
from TTY core in v3.10 by commit 6865ff222cca ("TTY: do not warn about
setting speed via SPD_*"), effectively breaking all driver
implementations of this except for serial core.

Also remove the verbose and outdated comment on how to set baud rates.

Note that setting a custom divisor will continue to work with the
caveat that 38400 must again be selected every time the divisor is
changed since v2.6.24 and commit 669a6db1037e ("USB: ftd_sio: cleanups
and updates for new termios work") which started reporting back the
actual baud rate used.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/usb/serial/ftdi_sio.c | 56 +++++++------------------------------------
 1 file changed, 9 insertions(+), 47 deletions(-)

diff --git a/drivers/usb/serial/ftdi_sio.c b/drivers/usb/serial/ftdi_sio.c
index df5c45a4b1d7..1cec03799cdf 100644
--- a/drivers/usb/serial/ftdi_sio.c
+++ b/drivers/usb/serial/ftdi_sio.c
@@ -1244,42 +1244,13 @@ static __u32 get_ftdi_divisor(struct tty_struct *tty,
 	int div_okay = 1;
 	int baud;
 
-	/*
-	 * The logic involved in setting the baudrate can be cleanly split into
-	 * 3 steps.
-	 * 1. Standard baud rates are set in tty->termios->c_cflag
-	 * 2. If these are not enough, you can set any speed using alt_speed as
-	 * follows:
-	 *    - set tty->termios->c_cflag speed to B38400
-	 *    - set your real speed in tty->alt_speed; it gets ignored when
-	 *      alt_speed==0, (or)
-	 *    - call TIOCSSERIAL ioctl with (struct serial_struct) set as
-	 *	follows:
-	 *      flags & ASYNC_SPD_MASK == ASYNC_SPD_[HI, VHI, SHI, WARP],
-	 *	this just sets alt_speed to (HI: 57600, VHI: 115200,
-	 *	SHI: 230400, WARP: 460800)
-	 * ** Steps 1, 2 are done courtesy of tty_get_baud_rate
-	 * 3. You can also set baud rate by setting custom divisor as follows
-	 *    - set tty->termios->c_cflag speed to B38400
-	 *    - call TIOCSSERIAL ioctl with (struct serial_struct) set as
-	 *	follows:
-	 *      o flags & ASYNC_SPD_MASK == ASYNC_SPD_CUST
-	 *      o custom_divisor set to baud_base / your_new_baudrate
-	 * ** Step 3 is done courtesy of code borrowed from serial.c
-	 *    I should really spend some time and separate + move this common
-	 *    code to serial.c, it is replicated in nearly every serial driver
-	 *    you see.
-	 */
-
-	/* 1. Get the baud rate from the tty settings, this observes
-	      alt_speed hack */
-
 	baud = tty_get_baud_rate(tty);
 	dev_dbg(dev, "%s - tty_get_baud_rate reports speed %d\n", __func__, baud);
 
-	/* 2. Observe async-compatible custom_divisor hack, update baudrate
-	   if needed */
-
+	/*
+	 * Observe deprecated async-compatible custom_divisor hack, update
+	 * baudrate if needed.
+	 */
 	if (baud == 38400 &&
 	    ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_CUST) &&
 	     (priv->custom_divisor)) {
@@ -1288,8 +1259,6 @@ static __u32 get_ftdi_divisor(struct tty_struct *tty,
 			__func__, priv->custom_divisor, baud);
 	}
 
-	/* 3. Convert baudrate to device-specific divisor */
-
 	if (!baud)
 		baud = 9600;
 	switch (priv->chip_type) {
@@ -1529,21 +1498,14 @@ static int set_serial_info(struct tty_struct *tty,
 check_and_exit:
 	write_latency_timer(port);
 
-	if ((priv->flags ^ old_priv.flags) & ASYNC_SPD_MASK) {
-		if ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
-			tty->alt_speed = 57600;
-		else if ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
-			tty->alt_speed = 115200;
-		else if ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_SHI)
-			tty->alt_speed = 230400;
-		else if ((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_WARP)
-			tty->alt_speed = 460800;
-		else
-			tty->alt_speed = 0;
-	}
 	if ((priv->flags ^ old_priv.flags) & ASYNC_SPD_MASK ||
 			((priv->flags & ASYNC_SPD_MASK) == ASYNC_SPD_CUST &&
 			 priv->custom_divisor != old_priv.custom_divisor)) {
+
+		/* warn about deprecation unless clearing */
+		if (priv->flags & ASYNC_SPD_MASK)
+			dev_warn_ratelimited(&port->dev, "use of SPD flags is deprecated\n");
+
 		change_speed(tty, port);
 		mutex_unlock(&priv->cfg_lock);
 	}
-- 
2.13.0

[PATCH 4/9] tty: simserial: drop unused alt_speed handling

[Johan Hovold]

This driver was setting the deprecated and broken alt_speed based on port
flags, but never provided a means to change the flags or to actually
change the speed.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 arch/ia64/hp/sim/simserial.c | 13 -------------
 1 file changed, 13 deletions(-)

diff --git a/arch/ia64/hp/sim/simserial.c b/arch/ia64/hp/sim/simserial.c
index de8cba121013..70d52e9bb575 100644
--- a/arch/ia64/hp/sim/simserial.c
+++ b/arch/ia64/hp/sim/simserial.c
@@ -387,19 +387,6 @@ static int activate(struct tty_port *port, struct tty_struct *tty)
 	}
 
 	state->xmit.head = state->xmit.tail = 0;
-
-	/*
-	 * Set up the tty->alt_speed kludge
-	 */
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
-		tty->alt_speed = 57600;
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
-		tty->alt_speed = 115200;
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_SHI)
-		tty->alt_speed = 230400;
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_WARP)
-		tty->alt_speed = 460800;
-
 errout:
 	local_irq_restore(flags);
 	return retval;
-- 
2.13.0

[PATCH 5/9] tty: amiserial: drop broken alt-speed support

[Johan Hovold]

Setting an alt_speed using the ASYNC_SPD flags has been deprecated since
v2.1.69, and has been broken since v3.10 and commit 6865ff222cca ("TTY:
do not warn about setting speed via SPD_*") without anyone noticing.

Drop the broken alt-speed handling altogether, and add a ratelimited
warning about using TIOCCSERIAL to change speed as being deprecated.

Note that using ASYNC_SPD_CUST is still supported.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/tty/amiserial.c | 23 +++--------------------
 1 file changed, 3 insertions(+), 20 deletions(-)

diff --git a/drivers/tty/amiserial.c b/drivers/tty/amiserial.c
index dea16bb8c46a..9820e20993db 100644
--- a/drivers/tty/amiserial.c
+++ b/drivers/tty/amiserial.c
@@ -570,18 +570,6 @@ static int startup(struct tty_struct *tty, struct serial_state *info)
 	info->xmit.head = info->xmit.tail = 0;
 
 	/*
-	 * Set up the tty->alt_speed kludge
-	 */
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
-		tty->alt_speed = 57600;
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
-		tty->alt_speed = 115200;
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_SHI)
-		tty->alt_speed = 230400;
-	if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_WARP)
-		tty->alt_speed = 460800;
-
-	/*
 	 * and set the speed of the serial port
 	 */
 	change_speed(tty, info, NULL);
@@ -1084,14 +1072,9 @@ static int set_serial_info(struct tty_struct *tty, struct serial_state *state,
 check_and_exit:
 	if (tty_port_initialized(port)) {
 		if (change_spd) {
-			if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
-				tty->alt_speed = 57600;
-			if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
-				tty->alt_speed = 115200;
-			if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_SHI)
-				tty->alt_speed = 230400;
-			if ((port->flags & ASYNC_SPD_MASK) == ASYNC_SPD_WARP)
-				tty->alt_speed = 460800;
+			/* warn about deprecation unless clearing */
+			if (new_serial.flags & ASYNC_SPD_MASK)
+				dev_warn_ratelimited(tty->dev, "use of SPD flags is deprecated\n");
 			change_speed(tty, state, NULL);
 		}
 	} else
-- 
2.13.0

[PATCH 6/9] tty: cyclades: drop broken alt-speed support

[Johan Hovold]

Setting an alt_speed using the ASYNC_SPD flags has been deprecated since
v2.1.69, and has been broken since v3.10 and commit 6865ff222cca ("TTY:
do not warn about setting speed via SPD_*") without anyone noticing.

Drop the broken alt-speed handling altogether, and add a ratelimited
warning about using TIOCCSERIAL to to change speed as being deprecated.

Note that using ASYNC_SPD_CUST is still supported.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/tty/cyclades.c | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/drivers/tty/cyclades.c b/drivers/tty/cyclades.c
index 104f09c58163..d272bc4e7fb5 100644
--- a/drivers/tty/cyclades.c
+++ b/drivers/tty/cyclades.c
@@ -1975,18 +1975,6 @@ static void cy_set_line_char(struct cyclades_port *info, struct tty_struct *tty)
 	cflag = tty->termios.c_cflag;
 	iflag = tty->termios.c_iflag;
 
-	/*
-	 * Set up the tty->alt_speed kludge
-	 */
-	if ((info->port.flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
-		tty->alt_speed = 57600;
-	if ((info->port.flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
-		tty->alt_speed = 115200;
-	if ((info->port.flags & ASYNC_SPD_MASK) == ASYNC_SPD_SHI)
-		tty->alt_speed = 230400;
-	if ((info->port.flags & ASYNC_SPD_MASK) == ASYNC_SPD_WARP)
-		tty->alt_speed = 460800;
-
 	card = info->card;
 	channel = info->line - card->first_line;
 
@@ -2295,12 +2283,16 @@ cy_set_serial_info(struct cyclades_port *info, struct tty_struct *tty,
 		struct serial_struct __user *new_info)
 {
 	struct serial_struct new_serial;
+	int old_flags;
 	int ret;
 
 	if (copy_from_user(&new_serial, new_info, sizeof(new_serial)))
 		return -EFAULT;
 
 	mutex_lock(&info->port.mutex);
+
+	old_flags = info->port.flags;
+
 	if (!capable(CAP_SYS_ADMIN)) {
 		if (new_serial.close_delay != info->port.close_delay ||
 				new_serial.baud_base != info->baud ||
@@ -2332,6 +2324,11 @@ cy_set_serial_info(struct cyclades_port *info, struct tty_struct *tty,
 
 check_and_exit:
 	if (tty_port_initialized(&info->port)) {
+		if ((new_serial.flags ^ old_flags) & ASYNC_SPD_MASK) {
+			/* warn about deprecation unless clearing */
+			if (new_serial.flags & ASYNC_SPD_MASK)
+				dev_warn_ratelimited(tty->dev, "use of SPD flags is deprecated\n");
+		}
 		cy_set_line_char(info, tty);
 		ret = 0;
 	} else {
-- 
2.13.0

[PATCH 7/9] tty: rocket: drop broken alt-speed support

[Johan Hovold]

Setting an alt_speed using the ROCKET_SPD flags has been deprecated
since v2.1.69, and has been broken since commit 6865ff222cca ("TTY: do
not warn about setting speed via SPD_*") without anyone noticing.

To make things worse commit 6df3526b6649 ("rocket: first pass at termios
reporting") in v2.6.25 started reporting back the actual baud rate used,
something which also required 38400 to again be set whenever changing a
SPD flag.

Drop the broken alt-speed handling altogether, and add a ratelimited
warning about using TIOCCSERIAL to change speed as being deprecated.

Note that the rocket driver has never supported using a custom divisor
(ASYNC_SPD_CUST equivalent).

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 drivers/tty/rocket.c | 27 ++++++---------------------
 1 file changed, 6 insertions(+), 21 deletions(-)

diff --git a/drivers/tty/rocket.c b/drivers/tty/rocket.c
index b51a877da986..20d79a6007d5 100644
--- a/drivers/tty/rocket.c
+++ b/drivers/tty/rocket.c
@@ -947,18 +947,6 @@ static int rp_open(struct tty_struct *tty, struct file *filp)
 
 		tty_port_set_initialized(&info->port, 1);
 
-		/*
-		 * Set up the tty->alt_speed kludge
-		 */
-		if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_HI)
-			tty->alt_speed = 57600;
-		if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_VHI)
-			tty->alt_speed = 115200;
-		if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_SHI)
-			tty->alt_speed = 230400;
-		if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_WARP)
-			tty->alt_speed = 460800;
-
 		configure_r_port(tty, info, NULL);
 		if (C_BAUD(tty)) {
 			sSetDTR(cp);
@@ -1219,23 +1207,20 @@ static int set_config(struct tty_struct *tty, struct r_port *info,
 			return -EPERM;
 		}
 		info->flags = ((info->flags & ~ROCKET_USR_MASK) | (new_serial.flags & ROCKET_USR_MASK));
-		configure_r_port(tty, info, NULL);
 		mutex_unlock(&info->port.mutex);
 		return 0;
 	}
 
+	if ((new_serial.flags ^ info->flags) & ROCKET_SPD_MASK) {
+		/* warn about deprecation, unless clearing */
+		if (new_serial.flags & ROCKET_SPD_MASK)
+			dev_warn_ratelimited(tty->dev, "use of SPD flags is deprecated\n");
+	}
+
 	info->flags = ((info->flags & ~ROCKET_FLAGS) | (new_serial.flags & ROCKET_FLAGS));
 	info->port.close_delay = new_serial.close_delay;
 	info->port.closing_wait = new_serial.closing_wait;
 
-	if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_HI)
-		tty->alt_speed = 57600;
-	if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_VHI)
-		tty->alt_speed = 115200;
-	if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_SHI)
-		tty->alt_speed = 230400;
-	if ((info->flags & ROCKET_SPD_MASK) == ROCKET_SPD_WARP)
-		tty->alt_speed = 460800;
 	mutex_unlock(&info->port.mutex);
 
 	configure_r_port(tty, info, NULL);
-- 
2.13.0

[PATCH 8/9] tty: ircomm: remove dead and broken ioctl code

[Johan Hovold]

Remove three ifdefed and broken implementations of TIOCSSERIAL and
TIOCGICOUNT, and parity handling in set_termios which had suffered
severe bit rot.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 net/irda/ircomm/ircomm_tty_ioctl.c | 107 +------------------------------------
 1 file changed, 1 insertion(+), 106 deletions(-)

diff --git a/net/irda/ircomm/ircomm_tty_ioctl.c b/net/irda/ircomm/ircomm_tty_ioctl.c
index f18070118d05..171c3dee760e 100644
--- a/net/irda/ircomm/ircomm_tty_ioctl.c
+++ b/net/irda/ircomm/ircomm_tty_ioctl.c
@@ -97,33 +97,7 @@ static void ircomm_tty_change_speed(struct ircomm_tty_cb *self,
 		self->settings.flow_control &= ~IRCOMM_RTS_CTS_IN;
 	}
 	tty_port_set_check_carrier(&self->port, ~cflag & CLOCAL);
-#if 0
-	/*
-	 * Set up parity check flag
-	 */
-
-	if (I_INPCK(self->tty))
-		driver->read_status_mask |= LSR_FE | LSR_PE;
-	if (I_BRKINT(driver->tty) || I_PARMRK(driver->tty))
-		driver->read_status_mask |= LSR_BI;
-
-	/*
-	 * Characters to ignore
-	 */
-	driver->ignore_status_mask = 0;
-	if (I_IGNPAR(driver->tty))
-		driver->ignore_status_mask |= LSR_PE | LSR_FE;
-
-	if (I_IGNBRK(self->tty)) {
-		self->ignore_status_mask |= LSR_BI;
-		/*
-		 * If we're ignore parity and break indicators, ignore
-		 * overruns too. (For real raw support).
-		 */
-		if (I_IGNPAR(self->tty))
-			self->ignore_status_mask |= LSR_OE;
-	}
-#endif
+
 	self->settings.data_format = cval;
 
 	ircomm_param_request(self, IRCOMM_DATA_FORMAT, FALSE);
@@ -271,67 +245,6 @@ static int ircomm_tty_get_serial_info(struct ircomm_tty_cb *self,
 static int ircomm_tty_set_serial_info(struct ircomm_tty_cb *self,
 				      struct serial_struct __user *new_info)
 {
-#if 0
-	struct serial_struct new_serial;
-	struct ircomm_tty_cb old_state, *state;
-
-	if (copy_from_user(&new_serial,new_info,sizeof(new_serial)))
-		return -EFAULT;
-
-
-	state = self
-	old_state = *self;
-
-	if (!capable(CAP_SYS_ADMIN)) {
-		if ((new_serial.baud_base != state->settings.data_rate) ||
-		    (new_serial.close_delay != state->close_delay) ||
-		    ((new_serial.flags & ~ASYNC_USR_MASK) !=
-		     (self->flags & ~ASYNC_USR_MASK)))
-			return -EPERM;
-		state->flags = ((state->flags & ~ASYNC_USR_MASK) |
-				 (new_serial.flags & ASYNC_USR_MASK));
-		self->flags = ((self->flags & ~ASYNC_USR_MASK) |
-			       (new_serial.flags & ASYNC_USR_MASK));
-		/* self->custom_divisor = new_serial.custom_divisor; */
-		goto check_and_exit;
-	}
-
-	/*
-	 * OK, past this point, all the error checking has been done.
-	 * At this point, we start making changes.....
-	 */
-
-	if (self->settings.data_rate != new_serial.baud_base) {
-		self->settings.data_rate = new_serial.baud_base;
-		ircomm_param_request(self, IRCOMM_DATA_RATE, TRUE);
-	}
-
-	self->close_delay = new_serial.close_delay * HZ/100;
-	self->closing_wait = new_serial.closing_wait * HZ/100;
-	/* self->custom_divisor = new_serial.custom_divisor; */
-
-	self->flags = ((self->flags & ~ASYNC_FLAGS) |
-		       (new_serial.flags & ASYNC_FLAGS));
-	self->tty->low_latency = (self->flags & ASYNC_LOW_LATENCY) ? 1 : 0;
-
- check_and_exit:
-
-	if (tty_port_initialized(self)) {
-		if (((old_state.flags & ASYNC_SPD_MASK) !=
-		     (self->flags & ASYNC_SPD_MASK)) ||
-		    (old_driver.custom_divisor != driver->custom_divisor)) {
-			if ((driver->flags & ASYNC_SPD_MASK) == ASYNC_SPD_HI)
-				driver->tty->alt_speed = 57600;
-			if ((driver->flags & ASYNC_SPD_MASK) == ASYNC_SPD_VHI)
-				driver->tty->alt_speed = 115200;
-			if ((driver->flags & ASYNC_SPD_MASK) == ASYNC_SPD_SHI)
-				driver->tty->alt_speed = 230400;
-			if ((driver->flags & ASYNC_SPD_MASK) == ASYNC_SPD_WARP)
-				driver->tty->alt_speed = 460800;
-			ircomm_tty_change_speed(driver);
-		}
-	}
-#endif
 	return 0;
 }
 
@@ -367,24 +280,6 @@ int ircomm_tty_ioctl(struct tty_struct *tty,
 
 	case TIOCGICOUNT:
 		pr_debug("%s(), TIOCGICOUNT not impl!\n", __func__);
-#if 0
-		save_flags(flags); cli();
-		cnow = driver->icount;
-		restore_flags(flags);
-		p_cuser = (struct serial_icounter_struct __user *) arg;
-		if (put_user(cnow.cts, &p_cuser->cts) ||
-		    put_user(cnow.dsr, &p_cuser->dsr) ||
-		    put_user(cnow.rng, &p_cuser->rng) ||
-		    put_user(cnow.dcd, &p_cuser->dcd) ||
-		    put_user(cnow.rx, &p_cuser->rx) ||
-		    put_user(cnow.tx, &p_cuser->tx) ||
-		    put_user(cnow.frame, &p_cuser->frame) ||
-		    put_user(cnow.overrun, &p_cuser->overrun) ||
-		    put_user(cnow.parity, &p_cuser->parity) ||
-		    put_user(cnow.brk, &p_cuser->brk) ||
-		    put_user(cnow.buf_overrun, &p_cuser->buf_overrun))
-			return -EFAULT;
-#endif
 		return 0;
 	default:
 		ret = -ENOIOCTLCMD;  /* ioctls which we must ignore */
-- 
2.13.0

[PATCH 9/9] tty: drop unused alt_speed from tty_struct

[Johan Hovold]

Drop the now unused alt_speed field from struct tty_struct.

Setting an alt_speed using the ASYNC_SPD flags has been deprecated since
v2.1.69, and has been broken for all tty drivers but serial-core since
v3.10 and commit 6865ff222cca ("TTY: do not warn about setting speed via
SPD_*") without anyone noticing.

Note that serial-core still supports changing speed using TIOCSSERIAL
and SPD flags (including "alt-speeds"), but also warns about it being
deprecated since pre-git.

Signed-off-by: Johan Hovold <johan@kernel.org>
---
 include/linux/tty.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/include/linux/tty.h b/include/linux/tty.h
index eccb4ec30a8a..585cf2b5ca94 100644
--- a/include/linux/tty.h
+++ b/include/linux/tty.h
@@ -316,7 +316,6 @@ struct tty_struct {
 
 	struct tty_struct *link;
 	struct fasync_struct *fasync;
-	int alt_speed;		/* For magic substitution of 38400 bps */
 	wait_queue_head_t write_wait;
 	wait_queue_head_t read_wait;
 	struct work_struct hangup_work;
-- 
2.13.0

Re: [PATCH 8/9] tty: ircomm: remove dead and broken ioctl code

[Alan Cox]

On Tue,  6 Jun 2017 12:54:40 +0200
Johan Hovold <johan@kernel.org> wrote:

> Remove three ifdefed and broken implementations of TIOCSSERIAL and
> TIOCGICOUNT, and parity handling in set_termios which had suffered
> severe bit rot.

I would be amazed if the IRDA code still works. It's not been tested
properly for years and it never followed the tty rules properly in the
first place - so this looks good, although moving IRDA into staging
and /dev/null would IMHO be far better.

IRDA is dead, and IR remotes are handled via completely different code.

Alan

Re: [PATCH 0/9] tty: drop broken alt-speed handling

[Alan Cox]

On Tue,  6 Jun 2017 12:54:32 +0200
Johan Hovold <johan@kernel.org> wrote:

> Setting an alt-speed using TIOCSSERIAL and SPD flags has been deprecated
> since v2.1.69 and has been broken for all tty drivers but serial-core
> since v3.10 and commit 6865ff222cca ("TTY: do not warn about setting
> speed via SPD_*") without anyone noticing (for four years).

We have the ability to set the speed directly via termios so I doubt
anyone is using the remaining hacks either but I guess someone may. All
looks good to me.

Reviewed-by: Alan Cox <alan@linux.intel.com>

Alan

Re: [PATCH 0/9] tty: drop broken alt-speed handling

[Andy Shevchenko]

On Tue, Jun 6, 2017 at 1:54 PM, Johan Hovold <johan@kernel.org> wrote:
> Setting an alt-speed using TIOCSSERIAL and SPD flags has been deprecated
> since v2.1.69 and has been broken for all tty drivers but serial-core
> since v3.10 and commit 6865ff222cca ("TTY: do not warn about setting
> speed via SPD_*") without anyone noticing (for four years).
>
> Instead of reverting the offending commit, lets get rid of this legacy
> code while adding back a warning about the SPD flags being deprecated to
> the drivers that once implemented it. Note that drivers implementing
> SPD_CUST will continue to support using a custom divisor.
>
> Also note that serial-core did not rely on the TTY layer for SPD
> handling and continues to support it while warning about deprecation
> (since 2003 at least).
>
> Greg, I suggest you take it all trough the TTY tree even if merging
> through multiple trees and applying the final patch once everything is
> upstream would be an option. Also the irda clean up does not depend on
> the rest of series as the code implementing SPD handling was ifdefed
> out.

What a nice clean up!

Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>

>
> Johan
>
>
> Johan Hovold (9):
>   serial: rate limit custom-speed deprecation notice
>   USB: serial: ftdi_sio: simplify TIOCSSERIAL flag logic
>   USB: serial: ftdi_sio: remove broken alt-speed handling
>   tty: simserial: drop unused alt_speed handling
>   tty: amiserial: drop broken alt-speed support
>   tty: cyclades: drop broken alt-speed support
>   tty: rocket: drop broken alt-speed support
>   tty: ircomm: remove dead and broken ioctl code
>   tty: drop unused alt_speed from tty_struct
>
>  arch/ia64/hp/sim/simserial.c       |  13 -----
>  drivers/tty/amiserial.c            |  23 ++------
>  drivers/tty/cyclades.c             |  21 ++++----
>  drivers/tty/rocket.c               |  27 +++-------
>  drivers/tty/serial/serial_core.c   |   5 +-
>  drivers/usb/serial/ftdi_sio.c      |  67 +++++------------------
>  include/linux/tty.h                |   1 -
>  net/irda/ircomm/ircomm_tty_ioctl.c | 107 +------------------------------------
>  8 files changed, 34 insertions(+), 230 deletions(-)
>
> --
> 2.13.0
>



-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH 8/9] tty: ircomm: remove dead and broken ioctl code

[Greg Kroah-Hartman]

On Tue, Jun 06, 2017 at 12:27:04PM +0100, Alan Cox wrote:
> On Tue,  6 Jun 2017 12:54:40 +0200
> Johan Hovold <johan@kernel.org> wrote:
> 
> > Remove three ifdefed and broken implementations of TIOCSSERIAL and
> > TIOCGICOUNT, and parity handling in set_termios which had suffered
> > severe bit rot.
> 
> I would be amazed if the IRDA code still works. It's not been tested
> properly for years and it never followed the tty rules properly in the
> first place - so this looks good, although moving IRDA into staging
> and /dev/null would IMHO be far better.
> 
> IRDA is dead, and IR remotes are handled via completely different code.

I agree, and we should move it into staging and out of the tree, I'll
send netdev some patches for that...

thanks,

greg k-h

Re: [PATCH 0/9] tty: drop broken alt-speed handling

[Greg Kroah-Hartman]

On Tue, Jun 06, 2017 at 12:54:32PM +0200, Johan Hovold wrote:
> Setting an alt-speed using TIOCSSERIAL and SPD flags has been deprecated
> since v2.1.69 and has been broken for all tty drivers but serial-core
> since v3.10 and commit 6865ff222cca ("TTY: do not warn about setting
> speed via SPD_*") without anyone noticing (for four years).
> 
> Instead of reverting the offending commit, lets get rid of this legacy
> code while adding back a warning about the SPD flags being deprecated to
> the drivers that once implemented it. Note that drivers implementing
> SPD_CUST will continue to support using a custom divisor.
> 
> Also note that serial-core did not rely on the TTY layer for SPD
> handling and continues to support it while warning about deprecation
> (since 2003 at least).
> 
> Greg, I suggest you take it all trough the TTY tree even if merging
> through multiple trees and applying the final patch once everything is
> upstream would be an option. Also the irda clean up does not depend on
> the rest of series as the code implementing SPD handling was ifdefed
> out.

I've taken these all through my tty tree, nice work, thanks for the
cleanups.

greg k-h

Re: [PATCH 8/9] tty: ircomm: remove dead and broken ioctl code

[Johan Hovold]

On Tue, Jun 13, 2017 at 11:48:41AM +0200, Greg Kroah-Hartman wrote:
> On Tue, Jun 06, 2017 at 12:27:04PM +0100, Alan Cox wrote:
> > On Tue,  6 Jun 2017 12:54:40 +0200
> > Johan Hovold <johan@kernel.org> wrote:
> > 
> > > Remove three ifdefed and broken implementations of TIOCSSERIAL and
> > > TIOCGICOUNT, and parity handling in set_termios which had suffered
> > > severe bit rot.
> > 
> > I would be amazed if the IRDA code still works. It's not been tested
> > properly for years and it never followed the tty rules properly in the
> > first place - so this looks good, although moving IRDA into staging
> > and /dev/null would IMHO be far better.
> > 
> > IRDA is dead, and IR remotes are handled via completely different code.
> 
> I agree, and we should move it into staging and out of the tree, I'll
> send netdev some patches for that...

Sounds good, but please consider applying this patch before moving to
staging as people have been modifying this dead code as part of
kernel-wide updates even though hasn't even compiled for quite some
time.

Thanks,
Johan

Re: [PATCH 8/9] tty: ircomm: remove dead and broken ioctl code

[Johan Hovold]

On Tue, Jun 13, 2017 at 11:57:14AM +0200, Johan Hovold wrote:
> On Tue, Jun 13, 2017 at 11:48:41AM +0200, Greg Kroah-Hartman wrote:
> > On Tue, Jun 06, 2017 at 12:27:04PM +0100, Alan Cox wrote:
> > > On Tue,  6 Jun 2017 12:54:40 +0200
> > > Johan Hovold <johan@kernel.org> wrote:
> > > 
> > > > Remove three ifdefed and broken implementations of TIOCSSERIAL and
> > > > TIOCGICOUNT, and parity handling in set_termios which had suffered
> > > > severe bit rot.
> > > 
> > > I would be amazed if the IRDA code still works. It's not been tested
> > > properly for years and it never followed the tty rules properly in the
> > > first place - so this looks good, although moving IRDA into staging
> > > and /dev/null would IMHO be far better.
> > > 
> > > IRDA is dead, and IR remotes are handled via completely different code.
> > 
> > I agree, and we should move it into staging and out of the tree, I'll
> > send netdev some patches for that...
> 
> Sounds good, but please consider applying this patch before moving to
> staging as people have been modifying this dead code as part of
> kernel-wide updates even though hasn't even compiled for quite some
> time.

That is, just that which you ended up doing. :)

Thanks,
Johan

Re: [PATCH 8/9] tty: ircomm: remove dead and broken ioctl code

[Greg Kroah-Hartman]

On Tue, Jun 13, 2017 at 11:57:14AM +0200, Johan Hovold wrote:
> On Tue, Jun 13, 2017 at 11:48:41AM +0200, Greg Kroah-Hartman wrote:
> > On Tue, Jun 06, 2017 at 12:27:04PM +0100, Alan Cox wrote:
> > > On Tue,  6 Jun 2017 12:54:40 +0200
> > > Johan Hovold <johan@kernel.org> wrote:
> > > 
> > > > Remove three ifdefed and broken implementations of TIOCSSERIAL and
> > > > TIOCGICOUNT, and parity handling in set_termios which had suffered
> > > > severe bit rot.
> > > 
> > > I would be amazed if the IRDA code still works. It's not been tested
> > > properly for years and it never followed the tty rules properly in the
> > > first place - so this looks good, although moving IRDA into staging
> > > and /dev/null would IMHO be far better.
> > > 
> > > IRDA is dead, and IR remotes are handled via completely different code.
> > 
> > I agree, and we should move it into staging and out of the tree, I'll
> > send netdev some patches for that...
> 
> Sounds good, but please consider applying this patch before moving to
> staging as people have been modifying this dead code as part of
> kernel-wide updates even though hasn't even compiled for quite some
> time.

I already applied it :)

I'll wait for 4.13-rc1 before doing it (if I remember...)

thanks,

greg k-h