PPP over ttyUSB (visor.o, Treo)

PPP over ttyUSB (visor.o, Treo)

[Jan Kasprzak]

	Hello @subscribers,

does anybody have working PPP over ttyUSB device in 2.6 kernels?
When I connect the Handspring Treo to the 2.6.0-test11 machine,
I am able to synchronize it, but not to run PPP to it (it can work
as modem). I am able to dial the modem connection using AT commands,
but I am not able to run PPP over it. Firstly the chat(1) program
complains about not being able to get terminal attributes (TCGETATTR,
from strace output), and then pppd(8) complains about not being able
to set terminal attributes (TCSETATTR from the strace output).
In 2.4 the same setup works OK.

	Any hints? (Kernel config and other details are available
on request). Thanks,

-Yenya

-- 
| Jan "Yenya" Kasprzak  <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839      Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
| http://www.fi.muni.cz/~kas/   Czech Linux Homepage: http://www.linux.cz/ |
|  I actually have a lot of admiration and respect for the PATA knowledge  |
| embedded in drivers/ide. But I would never call it pretty:) -Jeff Garzik |

Re: PPP over ttyUSB (visor.o, Treo)

[Stian Jordet]

ons, 10.12.2003 kl. 16.55 skrev Jan Kasprzak:
> 	Hello @subscribers,
> 
> does anybody have working PPP over ttyUSB device in 2.6 kernels?
> When I connect the Handspring Treo to the 2.6.0-test11 machine,
> I am able to synchronize it, but not to run PPP to it (it can work
> as modem). I am able to dial the modem connection using AT commands,
> but I am not able to run PPP over it. Firstly the chat(1) program
> complains about not being able to get terminal attributes (TCGETATTR,
> from strace output), and then pppd(8) complains about not being able
> to set terminal attributes (TCSETATTR from the strace output).
> In 2.4 the same setup works OK.
> 
> 	Any hints? (Kernel config and other details are available
> on request). Thanks,

Try to run pppd with "nodetach". Don't ask me why, but it works for me.

Best regards,
Stian

Re: PPP over ttyUSB (visor.o, Treo)

[Greg KH]

On Wed, Dec 10, 2003 at 04:55:41PM +0100, Jan Kasprzak wrote:
> 	Hello @subscribers,
> 
> does anybody have working PPP over ttyUSB device in 2.6 kernels?
> When I connect the Handspring Treo to the 2.6.0-test11 machine,
> I am able to synchronize it, but not to run PPP to it (it can work
> as modem). I am able to dial the modem connection using AT commands,
> but I am not able to run PPP over it. Firstly the chat(1) program
> complains about not being able to get terminal attributes (TCGETATTR,
> from strace output), and then pppd(8) complains about not being able
> to set terminal attributes (TCSETATTR from the strace output).
> In 2.4 the same setup works OK.
> 
> 	Any hints? (Kernel config and other details are available
> on request). Thanks,

Can you try the patch below?  I think it will fix the problem.

thanks,

greg k-h



diff -Nru a/drivers/usb/serial/usb-serial.c b/drivers/usb/serial/usb-serial.c
--- a/drivers/usb/serial/usb-serial.c	Wed Dec 10 13:28:31 2003
+++ b/drivers/usb/serial/usb-serial.c	Wed Dec 10 13:28:31 2003
@@ -493,12 +493,15 @@
 	return retval;
 }
 
-static void __serial_close(struct usb_serial_port *port, struct file *filp)
+static void serial_close(struct tty_struct *tty, struct file * filp)
 {
-	if (!port->open_count) {
-		dbg ("%s - port not opened", __FUNCTION__);
+	struct usb_serial_port *port = (struct usb_serial_port *) tty->driver_data;
+	struct usb_serial *serial = get_usb_serial (port, __FUNCTION__);
+
+	if (!serial)
 		return;
-	}
+
+	dbg("%s - port %d", __FUNCTION__, port->number);
 
 	--port->open_count;
 	if (port->open_count <= 0) {
@@ -506,30 +509,18 @@
 		 * port is being closed by the last owner */
 		port->serial->type->close(port, filp);
 		port->open_count = 0;
+
+		if (port->tty) {
+			if (port->tty->driver_data)
+				port->tty->driver_data = NULL;
+			port->tty = NULL;
+		}
 	}
 
 	module_put(port->serial->type->owner);
 	kobject_put(&port->serial->kobj);
 }
 
-static void serial_close(struct tty_struct *tty, struct file * filp)
-{
-	struct usb_serial_port *port = (struct usb_serial_port *) tty->driver_data;
-	struct usb_serial *serial = get_usb_serial (port, __FUNCTION__);
-
-	if (!serial)
-		return;
-
-	dbg("%s - port %d", __FUNCTION__, port->number);
-
-	/* if disconnect beat us to the punch here, there's nothing to do */
-	if (tty && tty->driver_data) {
-		__serial_close(port, filp);
-		tty->driver_data = NULL;
-	}
-	port->tty = NULL;
-}
-
 static int serial_write (struct tty_struct * tty, int from_user, const unsigned char *buf, int count)
 {
 	struct usb_serial_port *port = (struct usb_serial_port *) tty->driver_data;
@@ -848,19 +839,6 @@
 	dbg ("%s - %s", __FUNCTION__, kobj->name);
 
 	serial = to_usb_serial(kobj);
-
-	/* fail all future close/read/write/ioctl/etc calls */
-	for (i = 0; i < serial->num_ports; ++i) {
-		port = serial->port[i];
-		if (port->tty != NULL) {
-			port->tty->driver_data = NULL;
-			while (port->open_count > 0) {
-				__serial_close(port, NULL);
-			}
-			port->tty = NULL;
-		}
-	}
-
 	serial_shutdown (serial);
 
 	/* return the minor range that this device had */
@@ -1242,7 +1220,7 @@
 	/* register all of the individual ports with the driver core */
 	for (i = 0; i < num_ports; ++i) {
 		port = serial->port[i];
-		port->dev.parent = &serial->dev->dev;
+		port->dev.parent = &interface->dev;
 		port->dev.driver = NULL;
 		port->dev.bus = &usb_serial_bus_type;
 		port->dev.release = &port_release;

Re: PPP over ttyUSB (visor.o, Treo)

[Stian Jordet]

ons, 10.12.2003 kl. 22.28 skrev Greg KH:
> Can you try the patch below?  I think it will fix the problem.

Fixes it for me. Thanks :)

Best regards,
Stian

Re: PPP over ttyUSB (visor.o, Treo)

[Stian Jordet]

tor, 11.12.2003 kl. 02.22 skrev Stian Jordet:
> ons, 10.12.2003 kl. 22.28 skrev Greg KH:
> > Can you try the patch below?  I think it will fix the problem.
> 
> Fixes it for me. Thanks :)
> 
Uhm.. I was a bit too fast. It fixed the problem, okay, but it makes the
kernel spit out a lot of these messages:


Dec 11 02:29:40 chevrolet kernel: usb 1-1: USB disconnect, address 7
Dec 11 02:29:40 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
assigned address 8
Dec 11 02:29:40 chevrolet kernel: input: USB HID v1.10 Keyboard
[Logitech USB Receiver] on usb-0000:00:11.2-1
Dec 11 02:29:40 chevrolet kernel: input: USB HID v1.10 Mouse [Logitech
USB Receiver] on usb-0000:00:11.2-1
Dec 11 02:33:59 chevrolet kernel: usb 1-1: USB disconnect, address 8
Dec 11 02:33:59 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
assigned address 9
Dec 11 02:33:59 chevrolet kernel: input: USB HID v1.10 Keyboard
[Logitech USB Receiver] on usb-0000:00:11.2-1
Dec 11 02:33:59 chevrolet kernel: input: USB HID v1.10 Mouse [Logitech
USB Receiver] on usb-0000:00:11.2-1
Dec 11 02:34:15 chevrolet kernel: usb 1-1: USB disconnect, address 9
Dec 11 02:34:16 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
assigned address 10
Dec 11 02:34:16 chevrolet kernel: input: USB HID v1.10 Keyboard
[Logitech USB Receiver] on usb-0000:00:11.2-1
Dec 11 02:34:16 chevrolet kernel: input: USB HID v1.10 Mouse [Logitech
USB Receiver] on usb-0000:00:11.2-1
Dec 11 02:35:15 chevrolet kernel: usb 1-1: USB disconnect, address 10
Dec 11 02:35:15 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
assigned address 11
Dec 11 02:35:20 chevrolet kernel: usb 1-1: control timeout on ep0out

Untill finally this last line makes my mouse stop working :( Won't help
unplug/replug, have to restart to make it work again.. I'm using the
ftdi_sio driver. It might be a problem with that, but since pppd won't
work without the patch, I don't know...

Best regards,
Stian

Re: PPP over ttyUSB (visor.o, Treo)

[Greg KH]

On Thu, Dec 11, 2003 at 04:44:50AM +0100, Stian Jordet wrote:
> tor, 11.12.2003 kl. 02.22 skrev Stian Jordet:
> > ons, 10.12.2003 kl. 22.28 skrev Greg KH:
> > > Can you try the patch below?  I think it will fix the problem.
> > 
> > Fixes it for me. Thanks :)
> > 
> Uhm.. I was a bit too fast. It fixed the problem, okay, but it makes the
> kernel spit out a lot of these messages:
> 
> 
> Dec 11 02:29:40 chevrolet kernel: usb 1-1: USB disconnect, address 7
> Dec 11 02:29:40 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
> assigned address 8

This has nothing to do with the visor or other usb-serial drivers.  It
looks like you have either a flaky USB connection, or a power issue on
your USB hub.  Either way, the device keeps disconnecting itself
electronically (nothing Linux can do about that) and then reconnecting
itself.  Not good.

Is this connected to a powered hub?  If not, I'd recommend using one, or
getting a new keyboard/mouse as this is on the fritz.

thanks,

greg k-h

Re: PPP over ttyUSB (visor.o, Treo)

[Stian Jordet]

tor, 11.12.2003 kl. 07.44 skrev Greg KH:
> On Thu, Dec 11, 2003 at 04:44:50AM +0100, Stian Jordet wrote:
> > tor, 11.12.2003 kl. 02.22 skrev Stian Jordet:
> > > ons, 10.12.2003 kl. 22.28 skrev Greg KH:
> > > > Can you try the patch below?  I think it will fix the problem.
> > > 
> > > Fixes it for me. Thanks :)
> > > 
> > Uhm.. I was a bit too fast. It fixed the problem, okay, but it makes the
> > kernel spit out a lot of these messages:
> > 
> > 
> > Dec 11 02:29:40 chevrolet kernel: usb 1-1: USB disconnect, address 7
> > Dec 11 02:29:40 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
> > assigned address 8
> 
> This has nothing to do with the visor or other usb-serial drivers.  It
> looks like you have either a flaky USB connection, or a power issue on
> your USB hub.  Either way, the device keeps disconnecting itself
> electronically (nothing Linux can do about that) and then reconnecting
> itself.  Not good.
> 
> Is this connected to a powered hub?  If not, I'd recommend using one, or
> getting a new keyboard/mouse as this is on the fritz.

It isn't connected to a hub at all. I have never had any usb problems,
before I tried this patch. Now it's unusable :( What ever is the cause,
the patch triggers it. 

Best regards,
Stian

Re: PPP over ttyUSB (visor.o, Treo)

[Stian Jordet]

tor, 11.12.2003 kl. 15.23 skrev Stian Jordet:
> tor, 11.12.2003 kl. 07.44 skrev Greg KH:
> > On Thu, Dec 11, 2003 at 04:44:50AM +0100, Stian Jordet wrote:
> > > tor, 11.12.2003 kl. 02.22 skrev Stian Jordet:
> > > > ons, 10.12.2003 kl. 22.28 skrev Greg KH:
> > > > > Can you try the patch below?  I think it will fix the problem.
> > > > 
> > > > Fixes it for me. Thanks :)
> > > > 
> > > Uhm.. I was a bit too fast. It fixed the problem, okay, but it makes the
> > > kernel spit out a lot of these messages:
> > > 
> > > 
> > > Dec 11 02:29:40 chevrolet kernel: usb 1-1: USB disconnect, address 7
> > > Dec 11 02:29:40 chevrolet kernel: hub 1-0:1.0: new USB device on port 1,
> > > assigned address 8
> > 
> > This has nothing to do with the visor or other usb-serial drivers.  It
> > looks like you have either a flaky USB connection, or a power issue on
> > your USB hub.  Either way, the device keeps disconnecting itself
> > electronically (nothing Linux can do about that) and then reconnecting
> > itself.  Not good.
> > 
> > Is this connected to a powered hub?  If not, I'd recommend using one, or
> > getting a new keyboard/mouse as this is on the fritz.
> 
> It isn't connected to a hub at all. I have never had any usb problems,
> before I tried this patch. Now it's unusable :( What ever is the cause,
> the patch triggers it. 

I just tried 2.4.24-pre1, and I have the same behaviour, so I guess
either the ftdi_sio or my mouse is b0rked. Weird, though, since they are
connected on two different buses, and I have four other usb devices
connected as well, without problem.

Best regards,
Stian

Re: PPP over ttyUSB (visor.o, Treo)

[Greg KH]

On Thu, Dec 11, 2003 at 03:53:06PM +0100, Stian Jordet wrote:
> 
> I just tried 2.4.24-pre1, and I have the same behaviour, so I guess
> either the ftdi_sio or my mouse is b0rked. Weird, though, since they are
> connected on two different buses, and I have four other usb devices
> connected as well, without problem.

I really think you are exceeding the power limits for this hub, or you
have a flaky device.

thanks,

greg k-h

Re: PPP over ttyUSB (visor.o, Treo)

[Stian Jordet]

fre, 12.12.2003 kl. 22.15 skrev Greg KH:
> On Thu, Dec 11, 2003 at 03:53:06PM +0100, Stian Jordet wrote:
> > 
> > I just tried 2.4.24-pre1, and I have the same behaviour, so I guess
> > either the ftdi_sio or my mouse is b0rked. Weird, though, since they are
> > connected on two different buses, and I have four other usb devices
> > connected as well, without problem.
> I really think you are exceeding the power limits for this hub, or you
> have a flaky device.

I really understand if you are too busy answering stupid questions from
me. But I still hope someone can answer my one question.

I have six usb-devices: 3com Bluetooth, HP printer, Epson scanner,
Logitech keyboard/mouse, Kodak digital camera and a no-name memory card
reader. My motherboard have six onboard usb-ports.

The problem _only_ occur when I have a ppp-connection through the
ftdi_sio adapter. I tried disconnecting everything except the usb-serial
adapter and the Logitech. Same problem. Without an active
ppp-connection. No problems what-so-ever.

If I use a PS/2 keyboard and mouse, no problems with the rest of the usb
devices, even with a ppp connection.

Can my usb-serial converter be the flaky device, even though it's the
Logitech that is complaining, or can it be the motherboard? Since I
never-ever saw any error messages untill I recently switched my pl2303
against the ftdi_sio.

Best regards,
Stian

Re: PPP over ttyUSB (visor.o, Treo)

[Greg KH]

On Sat, Dec 13, 2003 at 01:04:48AM +0100, Stian Jordet wrote:
> 
> Can my usb-serial converter be the flaky device, even though it's the
> Logitech that is complaining, or can it be the motherboard? Since I
> never-ever saw any error messages untill I recently switched my pl2303
> against the ftdi_sio.

The ftdi_sio device could be sucking more power than the pl2303 device
from the USB bus.  That's what I would guess is happening.

greg k-h

Re: PPP over ttyUSB (visor.o, Treo)

[Jan Kasprzak]

Greg KH wrote:
: 
: Can you try the patch below?  I think it will fix the problem.
: 
	Well, it fixes the problem (pppd connects, and works OK),
but the kernel complains about badness in local_bh_enable.

	Full syslog is at http://www.fi.muni.cz/~kas/tmp/visor-mesages.txt,
kernel config is at http://www.fi.muni.cz/~kas/tmp/visor-kconfig.txt.
Any other info I should add?

	Thanks?

-Yenya

Dec 11 13:04:21 nisus kernel: PPP generic driver version 2.4.2
Dec 11 13:04:21 nisus pppd[1943]: pppd 2.4.1 started by root, uid 0
Dec 11 13:04:21 nisus pppd[1943]: Using interface ppp0
Dec 11 13:04:21 nisus pppd[1943]: Connect: ppp0 <--> /dev/ttyUSB1
Dec 11 13:04:21 nisus kernel: Badness in local_bh_enable at kernel/softirq.c:121Dec 11 13:04:21 nisus kernel: Call Trace:
Dec 11 13:04:21 nisus kernel:  [<c011e8b4>] local_bh_enable+0x70/0x72
Dec 11 13:04:21 nisus kernel:  [<de9e9ee2>] ppp_async_input+0x2de/0x583 [ppp_async]
Dec 11 13:04:21 nisus kernel:  [<de9e9396>] ppp_asynctty_receive+0x4b/0x8a [ppp_async]
Dec 11 13:04:21 nisus kernel:  [<c01c1924>] flush_to_ldisc+0x7d/0xce
Dec 11 13:04:21 nisus kernel:  [<de9dd7f5>] visor_read_bulk_callback+0x131/0x2a7 [visor]
Dec 11 13:04:21 nisus kernel:  [<c021e785>] usb_hcd_giveback_urb+0x25/0x39
Dec 11 13:04:21 nisus kernel:  [<de9d08a3>] uhci_finish_completion+0x48/0x5f [uhci_hcd]
Dec 11 13:04:21 nisus kernel:  [<c021e7cf>] usb_hcd_irq+0x36/0x5f
Dec 11 13:04:21 nisus kernel:  [<c010a6be>] handle_IRQ_event+0x3a/0x64
Dec 11 13:04:21 nisus kernel:  [<c010a97e>] do_IRQ+0x6f/0xdf
Dec 11 13:04:21 nisus kernel:  [<c0105000>] rest_init+0x0/0x27
Dec 11 13:04:21 nisus kernel:  [<c0109140>] common_interrupt+0x18/0x20
Dec 11 13:04:21 nisus kernel:  [<c0105000>] rest_init+0x0/0x27
Dec 11 13:04:21 nisus kernel:  [<de89d23e>] acpi_processor_idle+0xd5/0x1c7 [processor]
Dec 11 13:04:21 nisus kernel:  [<c0105000>] rest_init+0x0/0x27
Dec 11 13:04:21 nisus kernel:  [<c010709f>] cpu_idle+0x2c/0x35
Dec 11 13:04:21 nisus kernel:  [<c036a64d>] start_kernel+0x13f/0x14a
Dec 11 13:04:21 nisus kernel:  [<c036a3ea>] unknown_bootoption+0x0/0xfd
Dec 11 13:04:21 nisus kernel:
-- 
| Jan "Yenya" Kasprzak  <kas at {fi.muni.cz - work | yenya.net - private}> |
| GPG: ID 1024/D3498839      Fingerprint 0D99A7FB206605D7 8B35FCDE05B18A5E |
| http://www.fi.muni.cz/~kas/   Czech Linux Homepage: http://www.linux.cz/ |
|  I actually have a lot of admiration and respect for the PATA knowledge  |
| embedded in drivers/ide. But I would never call it pretty:) -Jeff Garzik |

Re: PPP over ttyUSB (visor.o, Treo)

[Greg KH]

On Thu, Dec 11, 2003 at 02:25:08PM +0100, Jan Kasprzak wrote:
> Greg KH wrote:
> : 
> : Can you try the patch below?  I think it will fix the problem.
> : 
> 	Well, it fixes the problem (pppd connects, and works OK),
> but the kernel complains about badness in local_bh_enable.
> 
> 	Full syslog is at http://www.fi.muni.cz/~kas/tmp/visor-mesages.txt,
> kernel config is at http://www.fi.muni.cz/~kas/tmp/visor-kconfig.txt.
> Any other info I should add?

Nah, this is a known problem, sorry.  The usb-serial core needs to have
the same kind of patch that the cdc-acm driver just had (change to use a
tasklet).  It's on my TODO list...

thanks,

greg k-h