On Tue, Jun 02, 2009 at 07:13:20PM +0800, Wan ZongShun wrote:
> Dear Dmitry ,
>
> Thank you for giving me help again.
>
> 2009/6/2, Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx>:
> > Hi Wan,
> >
> > On Tue, Jun 02, 2009 at 02:03:16PM +0800, Wan ZongShun wrote:
> > > Dear Dmitry,
> > >
> > > Thank you for your clear patch,it looks much better than mine.
> > >
> > > For to make this driver work well,I have modified a piece of code
> > > and made a patch based on your patch given me.
> > >
> > > Hope to get some advice from you again. thanks a lot.
> > >
> > > ---
> > > w90p910 touch screen driver patch.
> > >
> > > Signed-off-by: Wan ZongShun <mcuos.com@xxxxxxxxx>
> > >
> > > --- linux-2.6.29/drivers/input/touchscreen/w90p910_ts.c
> > > +++ linux-2.6.30/drivers/input/touchscreen/w90p910_ts.c
> > > @@ -76,7 +76,7 @@
> > > __raw_writel(ADC_TSC_X, w90p910_ts->ts_reg + 0x04);
> > > ctlreg = __raw_readl(w90p910_ts->ts_reg);
> > > ctlreg |= ADC_SEMIAUTO | ADC_INT_EN | ADC_CONV;
> > > - ctlreg &= ~(ADC_WAITTRIG | WT_INT_EN);
> > > + ctlreg &= ~(WT_INT|ADC_WAITTRIG | WT_INT_EN);
> >
> > Do we really need to reset WT_INT bit? I would think it was a read-only
> > bit set by the hardware...
>
> From the w90p910 spec ,WT_INT should be clear by soft, and set by hardware,
> so I have to add code to clear it here.
>
I see.
> >
> > > __raw_writel(ctlreg, w90p910_ts->ts_reg);
> > >
> > > w90p910_ts->state = TS_WAIT_X_COORD;
> > > @@ -89,7 +89,7 @@
> > > __raw_writel(ADC_TSC_Y, w90p910_ts->ts_reg + 0x04);
> > > ctlreg = __raw_readl(w90p910_ts->ts_reg);
> > > ctlreg |= ADC_SEMIAUTO | ADC_INT_EN | ADC_CONV;
> > > - ctlreg &= ~(ADC_WAITTRIG | WT_INT_EN);
> > > + ctlreg &= ~(ADC_INT|ADC_WAITTRIG | WT_INT_EN);
> >
> > Same here with ADC_INT.
>
> From the w90p910 spec ,ADC_INT should be clear by soft, and set by hardware,
> so I have to add code to clear it here too.
>
OK.
> >
> > > __raw_writel(ctlreg, w90p910_ts->ts_reg);
> > >
> > > w90p910_ts->state = TS_WAIT_Y_COORD;
> > > @@ -110,21 +110,14 @@
> > > static irqreturn_t w90p910_ts_interrupt(int irq, void *dev_id)
> > > {
> > > struct w90p910_ts *w90p910_ts = dev_id;
> > > - unsigned long flags;
> > > + unsigned long flags, val;
> > >
> > > spin_lock_irqsave(&w90p910_ts->lock, flags);
> > >
> > > switch (w90p910_ts->state) {
> > > case TS_WAIT_NEW_PACKET:
> > > - if (__raw_readl(w90p910_ts->ts_reg + 0x04) & ADC_DOWN) {
> > > - w90p910_prepare_x_reading(w90p910_ts);
> > > - } else {
> > > - w90p910_report_event(w90p910_ts, false);
> > > - del_timer(&w90p910_ts->timer);
> > > - w90p910_prepare_next_packet(w90p910_ts);
> > > - }
> > > + w90p910_prepare_x_reading(w90p910_ts);
> >
> > OK, so it is not possible to get an interrupt when pen leaves the
> > screen, it only generates interrupts while we are touching it,
> > correct?
>
> Okay,you are right, according to the spec,only the interrupt can occur
> when touching it .
>
I see.
> >
> > > break;
> > > -
> > >
> > > case TS_WAIT_X_COORD:
> > > w90p910_prepare_y_reading(w90p910_ts);
> > > @@ -132,7 +125,15 @@
> > >
> > > case TS_WAIT_Y_COORD:
> > > w90p910_report_event(w90p910_ts, true);
> > > - w90p910_prepare_next_packet(w90p910_ts);
> > > +
> > > + /* When timer starts,it should disable all ts
> > > + * interrupt and change mode to wait trigger.
> > > + */
> > > + val = __raw_readl(w90p910_ts->ts_reg);
> > > + val |= ADC_WAITTRIG;
> > > + val &= ~(ADC_INT | ADC_INT_EN | WT_INT_EN);
> > > + __raw_writel(val, w90p910_ts->ts_reg);
> > > +
> >
> > I am not sure about this... are you positive that we can't continue
> > working the device in interrupt mode and have to resort to polling
> > after the first touch has been registered?
>
> After the first touch occur, with the help of the timer callback
> function, I want to know if the pen still keep down.and if it keep
> up,I will think the first touch has finished and I will close timer,
> change mode to wait trigger and enable the WT interrupt for waitting
> next touch, so next touch also rely on the WT interrupt.
>
> I assume a situation that I make a line by pen in my touchscreen and
> get all values of x and y to user regarding the line,for that,I need
> to press my pen and keep it down and moving for making the line,then I
> put up my pen,so the lines has been finished.
>
> I think the timer should be in one shot mode and only does the judge
> whether my pen still keep down or up. and if pen keep down, we can
> enable the ADC interrupt and continue to read values (x and y) of the
> line,contrarily(if pen up),I think the line has been finished and have
> to change to wait trigger mode and enable WT interrupt and wait next
> touch occur.
>
I see what are you trying to do here, but you you seem to be polling
your device every 40 ms. I wonder if we rely on interrupts to give us
the continuous stream of coordinates while the pen is touching the
screen and make timer to poll for "pen up" event every 200 msecs would
not give better results with lesser load on the system?
-
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
