[RESEND PATCH] iio: adc: at91: don't use the last converted data register

Ludovic Desroches ludovic.desroches at atmel.com
Mon Sep 8 00:11:57 PDT 2014


On Wed, Sep 03, 2014 at 10:35:06AM +0200, Alexandre Belloni wrote:
> On 03/09/2014 at 08:34:29 +0200, Ludovic Desroches wrote :
> > If touchscreen mode is enabled and a conversion is requested on another
> > channel, the result in the last converted data register can be a
> > touchscreen relative value. Starting a conversion involves to do a
> > conversion for all active channel. It starts with ADC channels and ends
> > with touchscreen channels. Then if ADC_LCD register is not read quickly,
> > its content may be a touchscreen conversion.
> > To remove this temporal constraint, the conversion value is taken from
> > the channel data register.
> > 
> > Signed-off-by: Ludovic Desroches <ludovic.desroches at atmel.com>
> 
> I wondering whether we should get rid of drdy in the trigger path too.

Yes we can, then we have to use other interrupts. It could be
interesting because when conversion is done, it is done on all active
channels and touchscreen (if touchscreen mode is enabled of course).
Enabling drdy will cause to receive an interrupt after each conversion
(usually we are not fast enough to process irqs one by one, when we get
the first drdy irq and when we read the status register, others
conversions have already been completed). So getting rid of drdy is a
good idea but choosing the appropriate interrupt is a bit more
complicated:
- if using touchscreen with pressure measurement, we can rely on prdy
- if using touchscreen without pressure measurement, we can rely on yrdy
- if not using touchscreen, we can rely on EOCx where x is the higher
  channel number.

> Maybe this can be done in a follow-up patch. Else,
> 
> Acked-by: Alexandre Belloni <alexandre.belloni at free-electrons.com>
> 
> > ---
> > 
> > Sorry for the noise, there was a typo in the CC list.
> > 
> >  drivers/iio/adc/at91_adc.c |   14 +++++++++-----
> >  1 file changed, 9 insertions(+), 5 deletions(-)
> > 
> > diff --git a/drivers/iio/adc/at91_adc.c b/drivers/iio/adc/at91_adc.c
> > index 3b5bacd..dde22cb 100644
> > --- a/drivers/iio/adc/at91_adc.c
> > +++ b/drivers/iio/adc/at91_adc.c
> > @@ -196,6 +196,7 @@ struct at91_adc_state {
> >  	bool			done;
> >  	int			irq;
> >  	u16			last_value;
> > +	int			chnb;
> >  	struct mutex		lock;
> >  	u8			num_channels;
> >  	void __iomem		*reg_base;
> > @@ -274,7 +275,7 @@ void handle_adc_eoc_trigger(int irq, struct iio_dev *idev)
> >  		disable_irq_nosync(irq);
> >  		iio_trigger_poll(idev->trig, iio_get_time_ns());
> >  	} else {
> > -		st->last_value = at91_adc_readl(st, AT91_ADC_LCDR);
> > +		st->last_value = at91_adc_readl(st, AT91_ADC_CHAN(st, st->chnb));
> >  		st->done = true;
> >  		wake_up_interruptible(&st->wq_data_avail);
> >  	}
> > @@ -348,10 +349,11 @@ static irqreturn_t at91_adc_rl_interrupt(int irq, void *private)
> >  	struct iio_dev *idev = private;
> >  	struct at91_adc_state *st = iio_priv(idev);
> >  	u32 status = at91_adc_readl(st, st->registers->status_register);
> > +	const u32 eoc_mask = (1 << st->num_channels) - 1;
> >  	unsigned int reg;
> >  
> >  	status &= at91_adc_readl(st, AT91_ADC_IMR);
> > -	if (status & st->registers->drdy_mask)
> > +	if (status & eoc_mask)
> >  		handle_adc_eoc_trigger(irq, idev);
> >  
> >  	if (status & AT91RL_ADC_IER_PEN) {
> > @@ -413,12 +415,13 @@ static irqreturn_t at91_adc_9x5_interrupt(int irq, void *private)
> >  	struct iio_dev *idev = private;
> >  	struct at91_adc_state *st = iio_priv(idev);
> >  	u32 status = at91_adc_readl(st, st->registers->status_register);
> > +	const u32 eoc_mask = (1 << st->num_channels) - 1;
> >  	const uint32_t ts_data_irq_mask =
> >  		AT91_ADC_IER_XRDY |
> >  		AT91_ADC_IER_YRDY |
> >  		AT91_ADC_IER_PRDY;
> >  
> > -	if (status & st->registers->drdy_mask)
> > +	if (status & eoc_mask)
> >  		handle_adc_eoc_trigger(irq, idev);
> >  
> >  	if (status & AT91_ADC_IER_PEN) {
> > @@ -689,9 +692,10 @@ static int at91_adc_read_raw(struct iio_dev *idev,
> >  	case IIO_CHAN_INFO_RAW:
> >  		mutex_lock(&st->lock);
> >  
> > +		st->chnb = chan->channel;
> >  		at91_adc_writel(st, AT91_ADC_CHER,
> >  				AT91_ADC_CH(chan->channel));
> > -		at91_adc_writel(st, AT91_ADC_IER, st->registers->drdy_mask);
> > +		at91_adc_writel(st, AT91_ADC_IER, 1 << chan->channel);
> >  		at91_adc_writel(st, AT91_ADC_CR, AT91_ADC_START);
> >  
> >  		ret = wait_event_interruptible_timeout(st->wq_data_avail,
> > @@ -708,7 +712,7 @@ static int at91_adc_read_raw(struct iio_dev *idev,
> >  
> >  		at91_adc_writel(st, AT91_ADC_CHDR,
> >  				AT91_ADC_CH(chan->channel));
> > -		at91_adc_writel(st, AT91_ADC_IDR, st->registers->drdy_mask);
> > +		at91_adc_writel(st, AT91_ADC_IDR, 1 << chan->channel);
> >  
> >  		st->last_value = 0;
> >  		st->done = false;
> > -- 
> > 1.7.9.5
> > 
> 
> -- 
> Alexandre Belloni, Free Electrons
> Embedded Linux, Kernel and Android engineering
> http://free-electrons.com



More information about the linux-arm-kernel mailing list