Skip to contentSkip to conversation

usb-serial/ftdi_sio: fix DTR/RTS line modes

PATCH In review Daniel Mack 5 participants 10 messages Jun 3, 2010 – Jun 6, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 usb-serial/ftdi_sio: fix DTR/RTS line modes

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

Cc Greg Kroah-Hartman <[email protected]>
Cc Johan Hovold <[email protected]>
Cc Alan Cox <[email protected]>
---
 drivers/usb/serial/ftdi_sio.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/usb/serial/ftdi_sio.c +4 -0
index 050211a..79dd1ae 100644
@@ -2005,6 +2005,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 				"urb failed to set to rts/cts flow control\n");
 		}
 
+		/* raise DTR/RTS */
+		set_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	} else {
 		/*
 		 * Xon/Xoff code
@@ -2052,6 +2054,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 			}
 		}
 
+		/* lower DTR/RTS */
+		clear_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	}
 	return;
 }
Signature · 1 line
1.7.1
#1 lore raw
Quoted 11 lines · Daniel Mack

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

Quoted 5 lines
Cc Greg Kroah-Hartman <[email protected]>
Cc Johan Hovold <[email protected]>
Cc Alan Cox <[email protected]>

Oops. I forgot to Cc: [email protected].
This is in fact broken since 2.6.31-something.

Quoted 29 lines
---
 drivers/usb/serial/ftdi_sio.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/usb/serial/ftdi_sio.c +4 -0
index 050211a..79dd1ae 100644
@@ -2005,6 +2005,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 				"urb failed to set to rts/cts flow control\n");
 		}
 
+		/* raise DTR/RTS */
+		set_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	} else {
 		/*
 		 * Xon/Xoff code
@@ -2052,6 +2054,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 			}
 		}
 
+		/* lower DTR/RTS */
+		clear_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	}
 	return;
 }
Signature · 1 line
1.7.1
#2 lore raw
Quoted 12 lines · Daniel Mack
Quoted 11 lines · Daniel Mack

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

Quoted 8 lines
Quoted 5 lines
Cc Greg Kroah-Hartman <[email protected]>
Cc Johan Hovold <[email protected]>
Cc Alan Cox <[email protected]>

Oops. I forgot to Cc: [email protected].
This is in fact broken since 2.6.31-something.

Thanks for letting me know, I'll add that marking to the patch when I queue it up.

thanks,

greg k-h

#3 lore raw
In reply to Daniel Mack
Quoted 42 lines · Daniel Mack
Quoted 11 lines · Daniel Mack

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

Quoted 5 lines
Cc Greg Kroah-Hartman <[email protected]>
Cc Johan Hovold <[email protected]>
Cc Alan Cox <[email protected]>

Oops. I forgot to Cc: [email protected].
This is in fact broken since 2.6.31-something.

Quoted 21 lines
---
 drivers/usb/serial/ftdi_sio.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
(quoted diff) +0 -0

b/drivers/usb/serial/ftdi_sio.c index 050211a..79dd1ae 100644

drivers/usb/serial/ftdi_sio.c +2 -0
@@ -2005,6 +2005,8 @@ static void ftdi_set_termios(struct tty_struct
 *tty, "urb failed to set to rts/cts flow control\n");
 		}
 
+		/* raise DTR/RTS */
+		set_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	} else {
 		/*
 		 * Xon/Xoff code
@@ -2052,6 +2054,8 @@ static void ftdi_set_termios(struct tty_struct
 *tty, }
 		}
 
Quoted 6 lines

drivers/usb/serial/ftdi_sio.c static void ftdi_set_termios(struct tty_struct *tty, · line 2061

Quoted 5 lines
drivers/usb/serial/ftdi_sio.c +2 -0
+		/* lower DTR/RTS */
+		clear_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	}
 	return;
 }

How soon can this get in? I ask because I have been crippled in all attempts to use an ftdi adaptor to talk to an old legacy machine I have for about that long now. xon/xoff seems completely broken, and 7wire seems to scramble things to the point where no null modem cable I have made (and I have a book on them) will allow the 7wire flow controls to function.

Signature · 6 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
"Being against torture ought to be sort of a multipartisan thing."
-- Karl Lehenbauer, as amended by Jeff Daiell, a Libertarian
#4 lore raw
Quoted 5 lines · Gene Heskett

How soon can this get in? I ask because I have been crippled in all attempts to use an ftdi adaptor to talk to an old legacy machine I have for about that long now. xon/xoff seems completely broken, and 7wire seems to scramble things to the point where no null modem cable I have made (and I have a book on them) will allow the 7wire flow controls to function.

Give it a week or so to get out to all the stable kernel releases.

thanks,

greg k-h

#5 lore raw
Quoted 8 lines · Greg KH
Quoted 6 lines · Gene Heskett

How soon can this get in? I ask because I have been crippled in all attempts to use an ftdi adaptor to talk to an old legacy machine I have for about that long now. xon/xoff seems completely broken, and 7wire seems to scramble things to the point where no null modem cable I have made (and I have a book on them) will allow the 7wire flow controls to function.

Quoted 6 lines

Give it a week or so to get out to all the stable kernel releases.

thanks,

greg k-h

Thanks Greg.

Signature · 6 lines
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Well, the handwriting is on the floor.
		-- Joe E. Lewis
#6 lore raw
In reply to Daniel Mack
Quoted 45 lines · Daniel Mack
Quoted 11 lines · Daniel Mack

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

Quoted 5 lines
Cc Greg Kroah-Hartman <[email protected]>
Cc Johan Hovold <[email protected]>
Cc Alan Cox <[email protected]>

Oops. I forgot to Cc: [email protected].
This is in fact broken since 2.6.31-something.

Quoted 21 lines
---
 drivers/usb/serial/ftdi_sio.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/usb/serial/ftdi_sio.c +2 -0
index 050211a..79dd1ae 100644
@@ -2005,6 +2005,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 				"urb failed to set to rts/cts flow control\n");
 		}
 
+		/* raise DTR/RTS */
+		set_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	} else {
 		/*
 		 * Xon/Xoff code
@@ -2052,6 +2054,8 @@ static void ftdi_set_termios(struct tty_struct *tty,
 			}
 		}
 
Quoted 9 lines
Quoted 8 lines
drivers/usb/serial/ftdi_sio.c +2 -0
+		/* lower DTR/RTS */
+		clear_mctrl(port, TIOCM_DTR | TIOCM_RTS);
 	}
 	return;
 }
Signature · 1 line
1.7.1

--

Hmm, I tried this patch on plain 2.6.34 but still have a problem. As you indicated since 2.6.31-something something is broken.

I'm using a receive only device (http://rfxcom.com/receivers.htm)
which works fine until the system becomes more busy, usually it runs almost idle and all is fine.

However, when this happens, it appears the received characters get buffered somewhere as nothing seems to get through or is delayed for many minutes.
When I then =write= a character to the serial device
echo > /dev/ttyUSB0
the data is received in sudden burst, seems no data is actually lost just seriously delayed.

Any ideas ?

#7 lore raw
Quoted 15 lines · Ruud Linders

Hmm, I tried this patch on plain 2.6.34 but still have a problem. As you indicated since 2.6.31-something something is broken.

I'm using a receive only device (http://rfxcom.com/receivers.htm)
which works fine until the system becomes more busy, usually it runs almost idle and all is fine.

However, when this happens, it appears the received characters get buffered somewhere as nothing seems to get through or is delayed for many minutes.
When I then =write= a character to the serial device
echo > /dev/ttyUSB0
the data is received in sudden burst, seems no data is actually lost just seriously delayed.

Quoted 1 line

Any ideas ?

Hmm, no, sorry. Things work fine for me again now.

Did you try reverting the commit I blamed in the commit log? If this doesn't help, I can only suggest you run a bisect to see which commit between 2.6.31 and 2.6.34 broke it for you.

Daniel

#8 lore raw
In reply to Daniel Mack
Quoted 10 lines · On Thu, 3 Jun 2010 13:55:02 +0200 Daniel Mack <[email protected]> wrote:

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

If you've got hardware the other end expecting RTS/CTS signals you really ought to set CRTSCTS. However you patch makes complete sense in terms of which way it is best to end up.

#9 lore raw
Quoted 14 lines · Alan Cox
Quoted 10 lines · On Thu, 3 Jun 2010 13:55:02 +0200 Daniel Mack <[email protected]> wrote:

Call set_mctrl() and clear_mctrl() according to the flow control mode selected. This makes serial communication for FT232 connected devices work when CRTSCTS is not set.

This fixes a regression introduced by 4175f3e31 ("tty_port: If we are opened non blocking we still need to raise the carrier"). This patch calls the low-level driver's dtr_rts() function which consequently sets TIOCM_DTR | TIOCM_RTS. A later call to set_termios() without CRTSCTS in cflags, however, does not reset these bits, and so data is not actually sent out on the serial wire.

Quoted 2 lines

If you've got hardware the other end expecting RTS/CTS signals you really ought to set CRTSCTS.

I know. It is, however, a regression in my case, and there might be other cases where RTS/CTS aren't connected at all. The core of the problem is that the driver sets hardware flow control, even though the userspace asked for not doing that.

Quoted 2 lines

However you patch makes complete sense in
terms of which way it is best to end up.

And it shouldn't break anything I'd say.

Thanks,
Daniel