Re: [PATCH v5 4/7] iio: pressure: bmp280: Use sleep and forced mode for oneshot captures

[Andy Shevchenko]

On Wed, Sep 04, 2024 at 05:17:29PM +0300, Andy Shevchenko wrote:
> On Wed, Sep 04, 2024 at 12:24:27PM +0200, Vasileios Amoiridis wrote:
> > On Tue, Sep 03, 2024 at 05:26:38PM +0300, Andy Shevchenko wrote:
> > > On Mon, Sep 02, 2024 at 08:42:19PM +0200, Vasileios Amoiridis wrote:

...

> > > > +	if (!(reg & BMP380_STATUS_DRDY_PRESS_MASK) ||
> > > > +	    !(reg & BMP380_STATUS_DRDY_TEMP_MASK)) {
> > > > +		dev_err(data->dev, "Measurement cycle didn't complete.\n");
> > > > +		return -EBUSY;
> > > > +	}
> > > 
> > > Alternatively
> > > 
> > > 	if (!((reg & BMP380_STATUS_DRDY_PRESS_MASK) &&
> > > 	    !(reg & BMP380_STATUS_DRDY_TEMP_MASK)) {
> > > 		dev_err(data->dev, "Measurement cycle didn't complete.\n");
> > > 		return -EBUSY;
> > > 	}
> > 
> > Why would I use && instead of || ? I just need one of the 2 to be true
> > (one of the 2 measurements is not complete) and I can trigger the error
> > action.
> 
> Oh, I messed up the logic inversion, but wouldn't it be simpler to read
> "we return busy if neither press nor temp drdy bit set"?
> 
> 	if (!((reg & BMP380_STATUS_DRDY_PRESS_MASK) && (reg & BMP380_STATUS_DRDY_TEMP_MASK))) {
> 		dev_err(data->dev, "Measurement cycle didn't complete.\n");
> 		return -EBUSY;
> 	}
> 
> (I left long line for the better understanding of my point, you may break it to
>  two if needed)

Scratch below, it needs more thinking...

> With that, you even may have
> 
> #define BMP380_STATUS_DRDY_PRESS_AND_TEMP_MASK ...
> 
> 	if (!(reg & BMP380_STATUS_DRDY_PRESS_AND_TEMP_MASK)) {

Maybe ^, but I have no time to dive into this, you got the idea I believe.

> 		dev_err(data->dev, "Measurement cycle didn't complete.\n");
> 		return -EBUSY;
> 	}
> 
> which makes it all obvious.

-- 
With Best Regards,
Andy Shevchenko