Wine and serial port AGAIN

[Pavel Troller]

Hi!
  As a technician, I need often to use some form of serial port communication.
  My recent experience is with a program for programming TYT radios from China.
They supply a simple windows app, which allows to program all the features of
the radio, which cannot be accomplished using the radio keyboard only.
  The program works perfectly in wine, with one exception - a physical comms
with the radio. It complains that it doesn't receive a response from the radio,
while the radio crashes - it stops working, when a programming attempt is made,
and has to be power-cycled.
  The radio is connected using a USB cable, which contains just a regular 
PL-2303 USB to serial converter, which is perfectly supported in Linux. 
Port assignment for wine is also done, proof of which is, that the radio 
crashes upon the programming request. The data volume to be programmed is
small, just a few hundreds of bytes. It's really stupid that I can prepare my
data in the program only, and then I need to ask my colleague with real windows
to perform a programming job :-(.
  The program can be freely downloaded from the TYT pages:
  http://www.tyt888.com/WebEditor/UploadFile/201096144939483.rar
  but a real radio is needed to debug the comms problem. I've captured the
console output with WINEDEBUG=+comm, which is attached. It contains a log of
three consecutive attempts to program the radio. I'm ready to do more debugging,
but I don't know, how to proceed now.
  With regards, Pavel.

[Uwe Bonnes]

> > > > > "Pavel" == Pavel Troller  writes:

Pavel> Hi!  As a technician, I need often to use some form of serial
    Pavel> port communication.  My recent experience is with a program for
    Pavel> programming TYT radios from China.  They supply a simple windows
    Pavel> app, which allows to program all the features of the radio, which
    Pavel> cannot be accomplished using the radio keyboard only.  The
    Pavel> program works perfectly in wine, with one exception - a physical
    Pavel> comms with the radio. It complains that it doesn't receive a
    Pavel> response from the radio, while the radio crashes - it stops
    Pavel> working, when a programming attempt is made, and has to be
    Pavel> power-cycled.

By programming you mean volatile programming of some parameter and not flash
programming?
> trace:comm:get_irq_info TIOCGICOUNT err Invalid argument

Go to dlls/ntdll/serial.c  get_irq_info() and and add TRACES for the
parameters to find the invalid argument.

A run with +comm,+relay can also be helpfull. Look for the messagebox with
the "It complains that it doesn't receive a response from the radio" and
look what call before fails. If it is the get_irq_info above, we are on the
right track...

Bye

[Uwe Bonnes]

> > > > > "Pavel" == Pavel Troller  writes:

Pavel> Hi!  As a technician, I need often to use some form of serial
    Pavel> port communication.  My recent experience is with a program for
...
    Pavel> power-cycled.  The radio is connected using a USB cable, which
    Pavel> contains just a regular PL-2303 USB to serial converter, which is
    Pavel> perfectly supported in Linux.  Port assignment for wine is also

It seems that the PL2303 driver doesn't implement TIOCGICOUNT
/usr/src/linux/usb/serial/pl2303.c
static int pl2303_ioctl(struct tty_struct *tty, struct file *file,
                        unsigned int cmd, unsigned long arg)
{
        struct serial_struct ser;
        struct usb_serial_port *port = tty->driver_data;
        dbg("%s (%d) cmd = 0x%04x", __func__, port->number, cmd);

switch (cmd) {
        case TIOCGSERIAL:
                memset(&ser, 0, sizeof ser);
                ser.type = PORT_16654;
                ser.line = port->serial->minor;
                ser.port = port->number;
                ser.baud_base = 460800;

if (copy_to_user((void __user *)arg, &ser, sizeof ser))
                        return -EFAULT;

return 0;

case TIOCMIWAIT:
                dbg("%s (%d) TIOCMIWAIT", __func__,  port->number);
                return wait_modem_info(port, arg);
        default:
                dbg("%s not supported = 0x%04x", __func__, cmd);
                break;
        }
        return -ENOIOCTLCMD;
}
Perhaps
http://www.mp3car.com/vbulletin/linux/82704-iguidance-2-1-3-wine-7.html has
some hints.

Bye

[Wolfram Sang]

On 17/12/10 11:21, Uwe Bonnes wrote:
> > > > > > "Pavel" == Pavel Troller  writes:
> Pavel> Hi!  As a technician, I need often to use some form of serial
> Pavel> port communication.  My recent experience is with a program for
>
> ...
>     Pavel> power-cycled.  The radio is connected using a USB cable, which
>     Pavel> contains just a regular PL-2303 USB to serial converter, which is
>     Pavel> perfectly supported in Linux.  Port assignment for wine is also
>
> It seems that the PL2303 driver doesn't implement TIOCGICOUNT
> /usr/src/linux/usb/serial/pl2303.c

Well, no, this ioctl gets handled by the tty-layer. Yes, the driver needs
to define a get_icount-routine which it lacks.

Regards,

Wolfram

[Vijay Kiran Kamuju]

On Fri, Dec 17, 2010 at 6:42 AM, Wolfram Sang  wrote:
> On 17/12/10 11:21, Uwe Bonnes wrote:
> > > > > > > "Pavel" == Pavel Troller  writes:
> > Pavel> Hi!  As a technician, I need often to use some form of serial
> > Pavel> port communication.  My recent experience is with a program
>
> for
> > ...
> >     Pavel> power-cycled.  The radio is connected using a USB cable, which
> >     Pavel> contains just a regular PL-2303 USB to serial converter, which
>
> is
> > Pavel> perfectly supported in Linux.  Port assignment for wine is
>
> also
> > It seems that the PL2303 driver doesn't implement TIOCGICOUNT
> > /usr/src/linux/usb/serial/pl2303.c
>
> Well, no, this ioctl gets handled by the tty-layer. Yes, the driver needs
> to define a get_icount-routine which it lacks.

I think this is introduced in kernel 2.6.37 (not yet released)
> Regards,
>
> Wolfram

[Wolfram Sang]

> > It seems that the PL2303 driver doesn't implement TIOCGICOUNT
> > /usr/src/linux/usb/serial/pl2303.c
>
> Well, no, this ioctl gets handled by the tty-layer. Yes, the driver
> needs to define a get_icount-routine which it lacks.
>
> I think this is introduced in kernel 2.6.37 (not yet released)

So, its commit-id would be? ;) (And I can't see anything even in
linux-next...) Anyway, this driver-talk became offtopic for this list.

Regards,

Wolfram

[Vijay Kiran Kamuju]

On Fri, Dec 17, 2010 at 7:28 AM, Wolfram Sang  wrote:
> > > It seems that the PL2303 driver doesn't implement TIOCGICOUNT
> > > /usr/src/linux/usb/serial/pl2303.c
> >
> > Well, no, this ioctl gets handled by the tty-layer. Yes, the driver
> > needs to define a get_icount-routine which it lacks.
> >
> > I think this is introduced in kernel 2.6.37 (not yet released)
>
> So, its commit-id would be? ;) (And I can't see anything even in
> linux-next...) Anyway, this driver-talk became offtopic for this list.

OT:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...

This is kernel related :P
> Regards,
>
> Wolfram

[Wolfram Sang]

> OT:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
>
> This is kernel related :P

Ehrm, where does that add a handler for pl2303.c?

[Vijay Kiran Kamuju]

On Fri, Dec 17, 2010 at 8:42 AM, Wolfram Sang  wrote:
> > OT:
>
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> >
>
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > This is kernel related :P
>
> Ehrm, where does that add a handler for pl2303.c?

I just made a patch for it.
Its not there in the kernel yet.
I need someone to test it.

attaching the c file

[Pavel Troller]

> On Fri, Dec 17, 2010 at 8:42 AM, Wolfram Sang  wrote:
> > > OT:
> >
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > >
> >
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > > This is kernel related :P
> >
> > Ehrm, where does that add a handler for pl2303.c?
>
> I just made a patch for it.
> Its not there in the kernel yet.
> I need someone to test it.

Hi Vijay!
  IT WORKS!!! It REALLY works! Sorry for yelling, but it's really working!
I've used the 2.6.36 version, which Vijay sent me privately, after adding a
missing semicolon it compiled cleanly and since the first attempt, the program
writes the radio PERFECTLY! It's even better than on windows, because on my
colleague's computer, attempt to program the radio always gives error for
the first time and one must click "Retry". The second attempt is then OK. Here,
it works without any glitch!
  So, really many thanks to Vijay and all the experts here, which found the
real cause of this problem! And sorry for blaming wine, it was not its fault!

With regards,
  Pavel

[Vijay Kiran Kamuju]

Hi All,

Can anyone tell me how to get that into the mainline kernel ASAP.
Or Anyone can send it on my behalf to lkml list to get it into main line

Thanks,
Vijay

On Sat, Dec 18, 2010 at 2:20 AM, Pavel Troller  wrote:
> > On Fri, Dec 17, 2010 at 8:42 AM, Wolfram Sang
>
> wrote:
> > > > OT:
>
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > > >
>
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > > > This is kernel related :P
> > >
> > > Ehrm, where does that add a handler for pl2303.c?
> >
> > I just made a patch for it.
> > Its not there in the kernel yet.
> > I need someone to test it.
>
> Hi Vijay!
>  IT WORKS!!! It REALLY works! Sorry for yelling, but it's really working!
> I've used the 2.6.36 version, which Vijay sent me privately, after adding a
> missing semicolon it compiled cleanly and since the first attempt, the
> program
> writes the radio PERFECTLY! It's even better than on windows, because on my
> colleague's computer, attempt to program the radio always gives error for
> the first time and one must click "Retry". The second attempt is then OK.
> Here,
> it works without any glitch!
>  So, really many thanks to Vijay and all the experts here, which found the
> real cause of this problem! And sorry for blaming wine, it was not its
> fault!
>
> With regards,
>   Pavel

[Pavel Troller]

Hi!
  It would be very good! I'm voting even for including the 2.6.36 version into
the next .36 patch (probably .36.3).
  I can contact Pavel Machek, which has a lot of drivers in the kernel, but
if anybody here could push it to the sources directly, it would be better.

Regards, Pavel
> Hi All,
>
> Can anyone tell me how to get that into the mainline kernel ASAP.
> Or Anyone can send it on my behalf to lkml list to get it into main line
>
> Thanks,
> Vijay
>
> On Sat, Dec 18, 2010 at 2:20 AM, Pavel Troller  wrote:
> > > On Fri, Dec 17, 2010 at 8:42 AM, Wolfram Sang
> >
> > wrote:
> > > > > OT:
> >
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > > > >
> >
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=...
> > > > > This is kernel related :P
> > > >
> > > > Ehrm, where does that add a handler for pl2303.c?
> > >
> > > I just made a patch for it.
> > > Its not there in the kernel yet.
> > > I need someone to test it.
> >
> > Hi Vijay!
> >  IT WORKS!!! It REALLY works! Sorry for yelling, but it's really working!
> > I've used the 2.6.36 version, which Vijay sent me privately, after adding a
> > missing semicolon it compiled cleanly and since the first attempt, the
> > program
> > writes the radio PERFECTLY! It's even better than on windows, because on my
> > colleague's computer, attempt to program the radio always gives error for
> > the first time and one must click "Retry". The second attempt is then OK.
> > Here,
> > it works without any glitch!
> >  So, really many thanks to Vijay and all the experts here, which found the
> > real cause of this problem! And sorry for blaming wine, it was not its
> > fault!
> >
> > With regards,
> >   Pavel

[Wolfram Sang]

> It would be very good! I'm voting even for including the 2.6.36 version into
> the next .36 patch (probably .36.3).

The dot releases are for backported bug-fixes only. This is a new 
functionality, so it won't be in 2.6.36. Same for 2.6.37, which is in 
the stabilization phase now, usually no new features here, too. In fact, 
the patch should be send to linux-usb soon, so it will be included in 
the merge-window for 2.6.38 (please, no discussion about this workflow 
on wine-devel ;))
> I can contact Pavel Machek, which has a lot of drivers in the kernel, but
> if anybody here could push it to the sources directly, it would be better.

No need to do this. Vijay just needs to

1) read Documentation/SubmittingPatches
2) consult the MAINTAINERS file (or use scripts/get_maintainer.pl)
    (you can set me on CC, too)
3) send the patch
4) wait for comments
5) if not applied goto 3

No magic involved :)

Regards,

Wolfram

[Uwe Bonnes]

> > > > > "Wolfram" == Wolfram Sang  writes:

...
    Wolfram> No need to do this. Vijay just needs to

Wolfram> 1) read Documentation/SubmittingPatches 2) consult the
    Wolfram> MAINTAINERS file (or use scripts/get_maintainer.pl) (you can
    Wolfram> set me on CC, too) 3) send the patch 4) wait for comments 5) if
    Wolfram> not applied goto 3

Wolfram> No magic involved :)

Only patience and perstistance needed ...

And hopefully more feedback than on wine-devel ...

[Pavel Troller]

Hi!
> > > > > > "Wolfram" == Wolfram Sang  writes:
>
> ...
>     Wolfram> No need to do this. Vijay just needs to
> Wolfram> 1) read Documentation/SubmittingPatches 2) consult the
> Wolfram> MAINTAINERS file (or use scripts/get_maintainer.pl) (you can
> Wolfram> set me on CC, too) 3) send the patch 4) wait for comments 5) if
> Wolfram> not applied goto 3
>
> Wolfram> No magic involved :)
>
> Only patience and perstistance needed ...

Not a problem here. Thanks to Vijay, I can keep my kernels with his individual
patches until they are included in the mainstream. The only thing I'm thinking
about now is, how to make this information more publicly available, to help
the others to find the patches. Do you think that this mailing list is enough ?
> And hopefully more feedback than on wine-devel ...

Of course...

With regards, Pavel
> -- 
> Uwe Bonnes                [e-mail]
>
> Institut fuer Kernphysik  Schlossgartenstrasse 9  64289 Darmstadt
> --------- Tel. [phone] -------- Fax. [phone] ----------

[Ferenc Wagner]

Vijay Kiran Kamuju  writes:
> Can anyone tell me how to get that into the mainline kernel ASAP.
> Or Anyone can send it on my behalf to lkml list to get it into main line

Why don't you send the patch to [e-mail]?