Linux Keyspan USB serial driver ignoring XOFF

Hello,

I have narrowed down the issue I'm about to describe to keyspan.c; a Digi Edgeport/1 with identical configuration works fine.

I am configuring a Raspberry Pi running 4.19.66 (though keyspan.c hasn't changed since 2017) to talk to a real-live vt420. Configuring agetty with systemd worked easy enough, but I found that XON/XOFF wasn't working. stty -a shows ixon and ixoff as appropriate, but sending Ctrl-S (tested from multiple ways of sending) had no effect on output in bash, or scrolling output. (Emacs, though, recognized it as the start of a search, so I knew it was getting down the line.)

After a great deal of head-scratching on this, I went to look at the kernel source and found that keyspan.c does not appear to be honoring XOFF. I also have a Digi Edgeport/1 on hand (which uses io_ti.c), and when I swapped to that, everything worked fine - Ctrl-S caused the expected pause.

As far as I can tell, keyspan.c simply never implemented handling of XOFF, but you guys are the experts there.

I twiddled many, many settings with stty but could not find anything that would make it honor xoff/Ctrl-S with the Keyspan. Even as a regular user that would use that trick to pause scrolling, it had no effect. Different TERM settings also had no effect.

Output of stty -a:

speed 19200 baud; rows 24; columns 80; line = 0;
intr = ^C; quit = ^\; erase = ^?; kill = ^U; eof = ^D; eol = <undef>; eol2 = <undef>; swtch = <undef>; start = ^Q; stop = ^S; susp = ^Z; rprnt = ^R; werase = ^W; lnext = ^V; discard = ^O; min = 1; time = 0;
-parenb -parodd -cmspar cs8 hupcl -cstopb cread clocal -crtscts -ignbrk -brkint -ignpar -parmrk -inpck -istrip -inlcr -igncr icrnl ixon ixoff -iuclc -ixany -imaxbel -iutf8
opost -olcuc -ocrnl onlcr -onocr -onlret -ofill -ofdel nl0 cr0 tab0 bs0 vt0 ff0 isig icanon -iexten echo echoe echok -echonl -noflsh -xcase -tostop -echoprt echoctl echoke -flusho -extproc

Thanks,

John

That's correct, the driver does not support software flow control even if the hardware seems to have some support for assisted XON/XOFF.

Would you mind testing the below patch which may be enough to turn sw-flow control always on. If that works, I can probably find time to cook up a proper patch to make this configurable later this week.

Johan

From 55b46d78fe63f182923e4674659fa18f4624d6b8 Mon Sep 17 00:00:00 2001
From: Johan Hovold <[email protected]>
Date: Mon, 23 Sep 2019 12:14:56 +0200
Subject: [PATCH] USB: serial: keyspan: enable XON flow control

---
 drivers/usb/serial/keyspan.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

index d34779fe4a8d..934430cbdfc4 100644
@@ -2118,7 +2118,7 @@ static int keyspan_usa26_send_setup(struct usb_serial *serial,
 	msg.setLcr = 0xff;
 
 	msg.ctsFlowControl = (p_priv->flow_control == flow_cts);
-	msg.xonFlowControl = 0;
+	msg.xonFlowControl = 1;
 	msg.setFlowControl = 0xff;
 	msg.forwardingLength = 16;
 	msg.xonChar = 17;

@@ -2234,7 +2234,7 @@ static int keyspan_usa28_send_setup(struct usb_serial *serial,
 	msg.parity = 0;		/* XXX for now */
 
 	msg.ctsFlowControl = (p_priv->flow_control == flow_cts);
-	msg.xonFlowControl = 0;
+	msg.xonFlowControl = 1;
 
 	/* Do handshaking outputs, DTR is inverted relative to RTS */
 	msg.rts = p_priv->rts_state;

@@ -2388,7 +2388,7 @@ static int keyspan_usa49_send_setup(struct usb_serial *serial,
 	msg.setLcr = 0xff;
 
 	msg.ctsFlowControl = (p_priv->flow_control == flow_cts);
-	msg.xonFlowControl = 0;
+	msg.xonFlowControl = 1;
 	msg.setFlowControl = 0xff;
 
 	msg.forwardingLength = 16;

@@ -2690,7 +2690,7 @@ static int keyspan_usa67_send_setup(struct usb_serial *serial,
 	msg.setLcr = 0xff;
 
 	msg.ctsFlowControl = (p_priv->flow_control == flow_cts);
-	msg.xonFlowControl = 0;
+	msg.xonFlowControl = 1;
 	msg.setFlowControl = 0xff;
 	msg.forwardingLength = 16;
 	msg.xonChar = 17;

2.23.0

Hi Johan,

I gave it a try, but unfortunately that patch made no difference to the
behavior I was seeing. I'm happy to do whatever debugging may be
helpful.

Thanks!

- John