Received-SPF: pass (google.com: domain of linux-kernel+bounces-68902-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Fri, 16 Feb 2024 15:47:42 +0000
From: Jonathan Cameron <jic23@kernel.org>
To: Vasileios Amoiridis <vassilisamir@gmail.com>
Cc: lars@metafoo.de, ang.iglesiasg@gmail.com,
 andriy.shevchenko@linux.intel.com, 579lpy@gmail.com,
 linux-iio@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] drivers: iio: pressure: Add SPI support for BMP38x and
 BMP390
Message-ID: <20240216154742.685bd875@jic23-huawei>
In-Reply-To: <20240216132644.GA4236@vamoiridPC>
References: <20240215164332.506736-1-vassilisamir@gmail.com>
	<20240216111834.73287ab0@jic23-huawei>
	<20240216132644.GA4236@vamoiridPC>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

On Fri, 16 Feb 2024 14:26:44 +0100
Vasileios Amoiridis <vassilisamir@gmail.com> wrote:

> On Fri, Feb 16, 2024 at 11:18:34AM +0000, Jonathan Cameron wrote:
> > On Thu, 15 Feb 2024 17:43:32 +0100
> > Vasileios Amoiridis <vassilisamir@gmail.com> wrote:
> >   
> > > According to the datasheet of BMP38x and BMP390 devices, in SPI
> > > operation, the first byte that returns after a read operation is
> > > garbage and it needs to be dropped and return the rest of the
> > > bytes.  
> > 
> > Make it clear in the patch title that this is a fix and add a fixes tag.
> >   
> 
> The original support for SPI was added 8 years ago. Should I include that commit
> of 8 years ago in the fixes tag or just use a the word "fixes" with the rest of the
> title?
> 
Original git commit for the fixes tag.  Lets us know this wants to go in all stable kernels.
Also fixes in the title.


> > > +	ssize_t status;
> > > +	u8 buf;
> > > +
> > > +	memcpy(&buf, reg, reg_size);
> > > +	buf |= 0x80;  
> > 
> > Can you use regmap_bus read_flag_mask for this?  Seems to apply to 
> > all devices supported. + that's common for spi regmaps
> >  
> 
> Yes I noticed it yesterday in my tests that this was missing and it actually
> applies to all the devices. So the read_flag_mask should be added to both
> regmap_bus structs. 

It's there sort of indirectly for the bmp280 - the register addresses all happen
to include that bit, then it is cleared explicitly for the other direction.


> > 
> > Mind you I note the bmp280_regmap_spi_write() is masking the bit out which seems
> > backwards  - all the registers are defined with the bit set for that part
> > but not the 380.  Ah well - not part of this fix even if it's odd.
> > 

> > > diff --git a/drivers/iio/pressure/bmp280.h b/drivers/iio/pressure/bmp280.h
> > > index 4012387d7956..ca482b7e4295 100644
> > > --- a/drivers/iio/pressure/bmp280.h
> > > +++ b/drivers/iio/pressure/bmp280.h
> > > @@ -191,6 +191,8 @@
> > >  #define BMP380_TEMP_SKIPPED		0x800000
> > >  #define BMP380_PRESS_SKIPPED		0x800000
> > >  
> > > +#define BMP380_SPI_MAX_REG_COUNT_READ   3  
> > This doesn't seem useful as only used in one place.  
> 
> Could this define be moved in the bmp280-spi.c file or to not even use a define?
Not use it. Don't see how it is helpful. Just check that the
thing will fit in the array using an ARRAY_SIZE()...
> 
> > > +
> > >  /* BMP280 specific registers */
> > >  #define BMP280_REG_HUMIDITY_LSB		0xFE
> > >  #define BMP280_REG_HUMIDITY_MSB		0xFD  
> >