Hi,

What happens if you use "-P usb" ?

--HPS

> What happens if you use "-P usb" ?

# ldd /usr/local/bin/avrdude 
/usr/local/bin/avrdude:
 	libusb.so.2 => /usr/lib/libusb.so.2 (0x80068e000)
 	libm.so.5 => /lib/libm.so.5 (0x80079c000)
 	libreadline.so.8 => /lib/libreadline.so.8 (0x8008bb000)
 	libncurses.so.5.7 => /usr/local/lib/libncurses.so.5.7 (0x8009f7000)
 	libncurses.so.8 => /lib/libncurses.so.8 (0x800b14000)
 	libc.so.7 => /lib/libc.so.7 (0x800c5e000)
 	libtinfo.so.5.7 => /usr/local/lib/libtinfo.so.5.7 (0x800e91000)
# avrdude -c arduino -p m328p -P usb -v

avrdude: Version 5.10, compiled on Aug  3 2010 at 23:59:35
          Copyright (c) 2000-2005 Brian Dean, http://www.bdmicro.com/
          Copyright (c) 2007-2009 Joerg Wunsch

          System wide configuration file is "/usr/local/etc/avrdude.conf"
          User configuration file is "/root/.avrduderc"
          User configuration file does not exist or is not a regular file, skipping

          Using Port                    : usb
          Using Programmer              : arduino
avrdude: ser_open(): can't open device "usb": No such file or directory

--Marcin

As Marcin Cieslak wrote:

> Checked with the same hardware (dual-boot) and Microsoft
> Vista (with arduino-0018 IDE) and the board can be
> contacted and programmed without any problems.

Your check on Windows has been using a stock AVRDUDE (e.g., a WinAVR
compilation), too?  Or another tool?

IIRC, the Arduino bootloader requires some special tricks in order to
talk to it.  I think AVRDUDE v5.10 still lacks that feature.  Could
you try the SVN version of AVRDUDE?

(I don't see a GNATS ID in the subject.  Has this been actually filed
via send-pr?)

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)

I think you need to define one more parameter when using USB.

-t stk600

--HPS

As Hans Petter Selasky wrote:

> I think you need to define one more parameter when using USB.

The Arduino bootloader is /not/ an USB device from AVRDUDE's
point of view.  AVRDUDE wants to see it as a generic tty device
(/dev/cua*).
-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)

On Wednesday 04 August 2010 21:07:37 Joerg Wunsch wrote:
> As Hans Petter Selasky wrote:
> > I think you need to define one more parameter when using USB.
> 
> The Arduino bootloader is /not/ an USB device from AVRDUDE's
> point of view.  AVRDUDE wants to see it as a generic tty device
> (/dev/cua*).

And what USB protocol is it running? If raw, then the ID should be added to 
the ugensa driver instead.

--HPS

As Hans Petter Selasky wrote:

> > AVRDUDE wants to see it as a generic tty device
> > (/dev/cua*).
> 
> And what USB protocol is it running?

Whatever is hidden behind a /dev/cua*. ;-)  I suppose it's CDC.

AVRDUDE simply doesn't want to know that.  Those programmers where
AVRDUDE accepts a -P usb option are programmers where AVRDUDE talks
directly to using an application-specific protocol.  For all other
programmers, the parameter to -P must be a device node under /dev.

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)

On Wed, 4 Aug 2010, Joerg Wunsch wrote:

> As Marcin Cieslak wrote:
>
>> Checked with the same hardware (dual-boot) and Microsoft
>> Vista (with arduino-0018 IDE) and the board can be
>> contacted and programmed without any problems.
>
> Your check on Windows has been using a stock AVRDUDE (e.g., a WinAVR
> compilation), too?  Or another tool?

On Windows I have used arduino Java IDE. Arduino IDE distribution
actually includes a whole WinAVR stack including avrdude.exe,
but I didn't pay attention what is actually used.
I will reboot to Windows now and check the commandline invocation.

On FreeBSD arduino the Java IDE tries to use avrdude and fails the same
was as from commandline.

> IIRC, the Arduino bootloader requires some special tricks in order to
> talk to it.  I think AVRDUDE v5.10 still lacks that feature.  Could
> you try the SVN version of AVRDUDE?

AVRDUDE is mentioned for example here http://www.arduino.cc/playground/FreeBSD/CLI
as the tool to use.

I have tried "-c arduino" or "-c stk500v1" with trunk and I get still
the same affect as with 5.10 (Programmer timeout).

What maybe important: TX/RX LED on the board don't react at all
when trying to use avrdude (there is only a single blink
on firmware LED - which means bootloader start).

With Windows - TX/RX indicated clearly some activity.

> (I don't see a GNATS ID in the subject.  Has this been actually filed
> via send-pr?)

Hm, yes, sorry... You were copied on the original send-pr.

--Marcin

On Wednesday 04 August 2010 22:21:32 Marcin Cieslak wrote:
> On Wed, 4 Aug 2010, Joerg Wunsch wrote:
> > As Marcin Cieslak wrote:
> >> Checked with the same hardware (dual-boot) and Microsoft
> >> Vista (with arduino-0018 IDE) and the board can be
> >> contacted and programmed without any problems.
> > 
> > Your check on Windows has been using a stock AVRDUDE (e.g., a WinAVR
> > compilation), too?  Or another tool?
> 
> On Windows I have used arduino Java IDE. Arduino IDE distribution
> actually includes a whole WinAVR stack including avrdude.exe,
> but I didn't pay attention what is actually used.
> I will reboot to Windows now and check the commandline invocation.
> 
> On FreeBSD arduino the Java IDE tries to use avrdude and fails the same
> was as from commandline.
> 
> > IIRC, the Arduino bootloader requires some special tricks in order to
> > talk to it.  I think AVRDUDE v5.10 still lacks that feature.  Could
> > you try the SVN version of AVRDUDE?
> 
> AVRDUDE is mentioned for example here
> http://www.arduino.cc/playground/FreeBSD/CLI as the tool to use.
> 
> I have tried "-c arduino" or "-c stk500v1" with trunk and I get still
> the same affect as with 5.10 (Programmer timeout).
> 
> What maybe important: TX/RX LED on the board don't react at all
> when trying to use avrdude (there is only a single blink
> on firmware LED - which means bootloader start).
> 
> With Windows - TX/RX indicated clearly some activity.
> 
> > (I don't see a GNATS ID in the subject.  Has this been actually filed
> > via send-pr?)
> 

You can try to comment out:

        /* clear stall at first run */
        mtx_lock(&sc->sc_mtx);
        usbd_xfer_set_stall(sc->sc_xfer[UFTDI_BULK_DT_WR]);
        usbd_xfer_set_stall(sc->sc_xfer[UFTDI_BULK_DT_RD]);
        mtx_unlock(&sc->sc_mtx);

In uftdi_attach in sys/dev/usb/serial/uftdi.c.

--HPS

Steven Kreuzer and I had that type of problem when working on the 
devel/arduino port, but eventually ironed them out (AFAIK).

If you install the port and set build.verbose=true and 
upload.verbose=true in ~/.arduino/preferences.txt, you should be able to 
see what options it uses.  My hunch is a 19200 or 57600 baud rate, but I 
can't swear to that, and it's been some time since I tried it.

On Wed, 4 Aug 2010, Hans Petter Selasky wrote:

> You can try to comment out:
>
>        /* clear stall at first run */
>        mtx_lock(&sc->sc_mtx);
>        usbd_xfer_set_stall(sc->sc_xfer[UFTDI_BULK_DT_WR]);
>        usbd_xfer_set_stall(sc->sc_xfer[UFTDI_BULK_DT_RD]);
>        mtx_unlock(&sc->sc_mtx);
>
> In uftdi_attach in sys/dev/usb/serial/uftdi.c.

That did not fix it ...

--Marcin

The problem got solved. I got:

hw.usb.ucom.cons_unit=0

Resetting this value to -1 enabled the communication.

The problem was because ucom_get_data() didn't fetch
user-supplied data.

Please close this PR.

--Marcin