linux-kernel - Re: [PATCH v3 6/6] iio: pressure: Add triggered buffer support for BMP280 driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240324115546.695d3e33@jic23-huawei>
Date: Sun, 24 Mar 2024 11:55:46 +0000
From: Jonathan Cameron <jic23@...nel.org>
To: Vasileios Amoiridis <vassilisamir@...il.com>
Cc: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>, lars@...afoo.de,
 ang.iglesiasg@...il.com, mazziesaccount@...il.com, ak@...klinger.de,
 petre.rodan@...dimension.ro, phil@...pberrypi.com, 579lpy@...il.com,
 linus.walleij@...aro.org, semen.protsenko@...aro.org,
 linux-iio@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 6/6] iio: pressure: Add triggered buffer support for
 BMP280 driver

On Wed, 20 Mar 2024 18:46:02 +0100
Vasileios Amoiridis <vassilisamir@...il.com> wrote:

> On Wed, Mar 20, 2024 at 01:16:03PM +0200, Andy Shevchenko wrote:
> > On Tue, Mar 19, 2024 at 01:29:25AM +0100, Vasileios Amoiridis wrote:  
> > > BMP2xx, BMP3xx, and BMP5xx use consecutive buffers for their
> > > temperature, pressure and humidity readings. This facilitates
> > > the use of burst reads in order to acquire data much faster
> > > and in a different way from the one used in oneshot captures.
> > > 
> > > BMP085 and BMP180 use a completely different measurement
> > > process that is well defined and is used in their buffer_handler().  
> > 
> > ...
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP280_REG_TEMP_MSB,
> > > -			       data->buf, sizeof(data->buf));
> > > +			       data->buf, BMP280_NUM_TEMP_BYTES);  
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP280_REG_PRESS_MSB,
> > > -			       data->buf, sizeof(data->buf));
> > > +			       data->buf, BMP280_NUM_PRESS_BYTES);  
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP280_REG_HUMIDITY_MSB,
> > > -			       &data->be16, sizeof(data->be16));
> > > +			       &data->be16, BME280_NUM_HUMIDITY_BYTES);  
> >   
> > > -	adc_humidity = be16_to_cpu(data->be16);
> > > +	adc_humidity = get_unaligned_be16(&data->be16);  
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP380_REG_TEMP_XLSB,
> > > -			       data->buf, sizeof(data->buf));
> > > +			       data->buf, BMP280_NUM_TEMP_BYTES);  
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP380_REG_PRESS_XLSB,
> > > -			       data->buf, sizeof(data->buf));
> > > +			       data->buf, BMP280_NUM_PRESS_BYTES);  
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP580_REG_TEMP_XLSB, data->buf,
> > > -			       sizeof(data->buf));
> > > +			       BMP280_NUM_TEMP_BYTES);  
> >   
> > >  	ret = regmap_bulk_read(data->regmap, BMP580_REG_PRESS_XLSB, data->buf,
> > > -			       sizeof(data->buf));
> > > +			       BMP280_NUM_PRESS_BYTES);  
> > 
> > These smell to me as candidates to a separate patch with more explanation why.
> > (Yes, with the definitions you introduced.) But I leave it to Jonathan to
> > decide if we need to split.

The are somewhat confusing, though only when you start doing bulk reads
to do these makes sense - so I'm not sure how to do it as 2 patches.

Pity we don't have sizeof(be24) available.

> > 
> > ...
> > 
> > The below are applicable to the bmp280_buffer_handler(),
> > bmp380_buffer_handler() implementations as well.
> > 
> > ...
> >   
> > > +	/* Burst read data registers */
> > > +	ret = regmap_bulk_read(data->regmap, BMP580_REG_TEMP_XLSB,
> > > +			       data->buf, 6);  
> > 
> > Magic size.
> >   
> 
> Hi Andy,
> 
> Thank you again for your feedback. When I was writing it, it was 
> looking as a magic number to me as well but then I though that 
> since I put the comment above it could be obvious. Now that I see 
> it again, I think it was not a good idea and maybe some type of 
> definition like
> 
> 	#define BMP280_BURST_READ_NUM_BYTES 6
> 	#define BME280_BURST_READ_NUM_BYTES 8
I think these are sums of the other quantities.  Better to
express them as such if possible.

> 
> could look better and be more intuitive.
> 
> > ...
> >   
> > > +	/* Temperature calculations */
> > > +	memcpy(&chan_value, &data->buf[3], 3);  
> > 
> > _le24() + sign_extend32()?
> >   
> 
> In the next line from your comment the _le24 or _be24 takes place.
I think you can get that data directly rather than bouncing it via
a memcpy.
e.g.
adc_temp = FIELD_GET(BMP280_MEAS_TRIM_MASK, get_unaligned_be24(&data->buf[3]);

> If the sign_extend32() is needed here, shouldn't it also be used
> in all the oneshot captures as well?

Is this actually signed?  Compensated values are, but at this point
it's just a raw adc count.  So I'm not seeing why we'd sign extend it
(unless the device is returning a signed be24 - I haven't checked.)


> 
> > ...
> >   
> > > +	/* Pressure calculations */
> > > +	memcpy(&chan_value, &data->buf[0], 3);  
> > 
> > _le24() + sign_extend32()?
> > 
> > ...
> >   
> > >  	/*
> > > -	 * Maximum number of consecutive bytes read for a temperature or
> > > -	 * pressure measurement is 3.
> > > +	 * Maximum number of a burst read for temperature, pressure, humidity
> > > +	 * is 8 bytes.
> > >  	 */
> > > -	if (val_size > 3)
> > > +	if (val_size > 8)  
> > 
> > sizeof() / new definition for the buf[] size?
> >   
> 
> In a previous commit that I was fixing this SPI driver, Jonathan had mentioned
> that there is no need for a specific definition since it will only be used
> here so that's why I kept it as is.

Never trust me ;)  Size of the buf is sensible here.

Jonathan

> 
> Cheers,
> Vasilis
> > >  		return -EINVAL;  
> > 
> > -- 
> > With Best Regards,
> > Andy Shevchenko
> > 
> >