Re: Resend Re: [PATCH v6] usb:serial:pl2303: add GPIOs interface on PL2303

[Johan Hovold <johan@kernel.org>]

On Fri, Aug 08, 2014 at 03:10:34AM +0800, Wang YanQing wrote:
> On Tue, Aug 05, 2014 at 03:54:08PM +0200, Johan Hovold wrote:
> > > > I noticed that setting direction to output and setting the gpio high has
> > > > no effect on the read-back value (i.e. I still read back 0) for my
> > > > pl2303hx (note that my device has no easily accessible gpios so I
> > > > haven't verified the actual state of the output pin).
> > > > 
> > > > What happens on your system? Is the read-back value still 0, even when
> > > > the GPIO output is actually high? Should we return the cached value in
> > > > this case?
> > > 
> > > If i set direction to output, then I could control gpio high and low
> > > by set 1 or 0, and the read-back value is 1 or 0 according to high and
> > > low(I test high and low by oscillscope)
> > > 
> > > I test it with my pl2303hx with only two gpios.
> > >
> > > Could you use usbmon to see whether the traffic is right according
> > > to comment in struct pl2303_gpio?
> > 
> > The traffic appears correct judging from the debug output (which I
> > trust). Output-enable is reflected in register 0x81, but the value
> > isn't.
> > 
> > What is the lsusb -v output for your device?
> 
> Bus 001 Device 005: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port.

You forgot the verbose flag (-v).
 
> It is strange your device doesn't work, I verify the control method by
> analyze usbmon output from linux host which has VirtualBox running 
> gpio test program, but I don't have right to distribute the gpio test
> program I think, so I can't help you to figure out why it doesn't work 
> for your device.

What do you use the gpio test program for? I thought you verified the
gpios with a scope?

Perhaps mine just does not support GPIOs? I don't know, but that's
partly why I asked for the lsusb output. Apparently there's a bunch of
different versions of these chips out there.

I'll see if I can find time to dissect my converter and try to access
the GPIO pins with the next version of the patch.

> > I suggest you just set the label to pl2303 until we have a valid
> > use-case that requires something more elaborate.
> 
> Ok, but pl2303-gpio maybe a better name?

No, not really. It's a gpio-chip label and is only used in that context
so a "-gpio" suffix adds no information.

Thanks,
Johan