Re: Sony Clie with USB sync

[Lars Weissflog <L rs-w de>]

Hi David,

thanks for hooking on here. As I indicated, that was my quick'n'dirty
notes taken at the time I got my clie to work, must be over a year
since. Most of the tips that helped me came from you, didn't want to
question you as the expert. 

Daniel, you should have a look into the howtos David quotes below. Let
us know how far you get, maybe more Ideas come to my mind if you post
the right hint, waking my tired brain up.... 

Good night.
Lars

On Mon, 2003-08-18 at 21:32, David A. Desrosiers wrote:
> > I had problems there to, I have to open the /var/log/messages and wait
> > until the first few lines of the connect-message appear. then I hit enter
> > to confirm the pre-entered pilot-xfer command in another term and then it
> > connects. I had some help on this list regarding this problem and I can
> > live with it now, but it's not totally okay yet.
> 
> 	This is how USB works. The hardware has to see a physical
> (electrical) connection between Palm and cradle (plugging your Palm into the
> cradle is not enough to connect it to the desktop, just like putting a light
> bulb in a socket isn't enough to complete the circuit and provide light), by
> hitting HotSync. Once you make that electrical connection, the device sends
> data across the bus, the kernel sees the data, and binds to the device,
> through the visor driver (visor.o). There is no way to hit Sync on the
> desktop, and have a Palm-in-the-cradle "wake up" and begin syncronizing.
> 
> 	There is a workaround in place in pilot-link to sleep() until the
> port becomes available, for a maximum of 5 seconds. You can hit sync on the
> Palm first, or launch pilot-xfer first, and it will just hook up. It works
> with J-Pilot and gnome-pilot in my tests here.
> 
Though there is your workaround, I sometimes have to try several times
until I get pilot-xfer to connect, and I am doing it by "listening" to
my laptop: Shortly after the clie has initiated the sync, i.e. when the
log shows it tries to connect, the harddisk gets active for a moment as
the module is loaded (I guess). This is the right moment to initiate the
pilot-xfer. Anyway, this is dirty but reliable ;-))


> 	Also, you can't run pilot-xfer at the same time as gnome-pilot
> (gpilotd) is running. You have to kill off gpilotd before pilot-xfer
> communicate with the Palm handheld. They can't both bind to the device and
> listen for incoming connections at the same time.
> 
> > 6. satisfy all deps on libs * install various packages: * long list,
> >    complete it next time libpanel-applet-dev
> 
> 	pilot-link doesn't require any of these dev packages.
> 
agreed. I noted this because at the time compiling pilot-xfer I also
compiled the gnome-pilot stuff, so one could put number 6 further down
the list.

> 	More detail here:
> 	http://howto.pilot-link.org/evosync/d.html
> 
> > 7. remove pilot-link folder after each failed configure / build, unpack it
> >    again and do a new configure as above until it is compiled.
> 
> 	This isn't necessary, just fix the dependancy and re-run configure
> again, and it will detect the new dependancy (if installed correctly) and
> continue onward.
> 
> > 8. #> make install as root to install new pilot-link tools
> 
> 	Test first, make sure it works locally (in ./src) BEFORE installing
> it into a production location. Once you verify that it does indeed work,
> install it as per your normal method of installing source-built software
> into the proper $prefix.
> 
> > 9. check (locate) all pilot-xfer and check versions; remove old versions
> >    and set links to new version
> 
> 	Or don't, and leave both versions installed. One in /usr (typically
> from your distribution) and one in /usr/local, and adjust your $PATH to
> $prefix/bin accordingly, to call the appropriate one first.
> 
> > 10. make sure usb is complete and working kernel should have: usb, uhci,
> >     usbserial, usbmouse, hid, usbcore, usbdevfs, visor at least
> 
> 	usbmouse? hid? Those aren't needed. README.usb goes over this
> (linked below). usbmouse and hid are both input mechanisms for peripherals,
> and not needed to verify that your USB-connected Palm handheld is working.
> 
You are right, this whole usb-stuff is simply all I need to run my clie
and mouse. and maybe there is even a little to much, but it caused no
harm, so I have it in.


> > 11. modprobing visor and check with lsmod for at least: uhci, usbcore,
> >     usbserial, visor
> 
> 	usb-uhci, not uhci. uhci on its own has had some reported issues,
> and switching to usb-uhci (or usb-ohci, depending on your hardware) seems to
> be a bit more forgiving with Palm handheld devices using USB connections.
> 
I'm running uhci. Didn't try to switch over to usb-uhci yet. NETARUS:
Never touch a running system ;-)


> > 14. set permissions to 777 on /dev/ttyUSB1
> 
> 	You only need 0666, 0777 is a bit too permissive.
> 
> > 15. use pilot-xfer -p /dev/ttyUSB1 -l to check for connection at all.
> >     Won't work if uhci is not loaded though visor loads without it. Check
> >     for uhci!!!!
> 
> 	Or usb-uhci. Make sure visor is a _module_, not monolithic.
> 
> > 17. configure gnome-pilot as above (delete and reconfigure after each
> >     failed try) use this:  ./configure -prefix=/usr
> >     -with-pisock=/usr/local
> 
> 	http://howto.pilot-link.org/evosync/e.html
> 
> > 40. launch evolution and run a hotsync. should work.
> 
> 	Except the Address and Memo conduits, those have issues right now..
> 
> > Puh. This was what I noted when I brought my SJ30 to talk to Evo. Maybe it
> > helps.
> 
> 	http://pilot-link.org/README.usb
> 
> 
> d.
> 
> _______________________________________________
> gnome-pilot-list mailing list
> gnome-pilot-list gnome org
> http://mail.gnome.org/mailman/listinfo/gnome-pilot-list
-- 
Lars  Weissflog         
Droegestrasse 7    
D-22305 Hamburg
Germany
Fon:        +49-(0)40-69208923
Mobile:  +49-(0)172-3816263 
email:   L rs-w de