linux-kernel - Re: [PATCH 2/2] iio: light: vcnl4000: Add ps high definition for vcnl4040

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] [day] [month] [year] [list]
Message-ID: <20231218181430.0ed55749@jic23-huawei>
Date: Mon, 18 Dec 2023 18:14:30 +0000
From: Jonathan Cameron <jic23@...nel.org>
To: Mårten Lindahl <martenli@...s.com>
Cc: linux-iio@...r.kernel.org, Lars-Peter Clausen <lars@...afoo.de>,
 kernel@...s.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] iio: light: vcnl4000: Add ps high definition for
 vcnl4040

On Mon, 18 Dec 2023 17:00:10 +0100
Mårten Lindahl <martenli@...s.com> wrote:

> On 12/18/23 16:19, Mårten Lindahl wrote:
> > On 12/17/23 15:15, Jonathan Cameron wrote:  
> >> On Fri, 15 Dec 2023 13:43:05 +0100
> >> Mårten Lindahl <marten.lindahl@...s.com> wrote:
> >>  
> >>> The vcnl4040 proximity sensor defaults to 12 bit data resolution, but
> >>> the chip also supports 16 bit data resolution, which is called 
> >>> proximity
> >>> high definition (PS_HD).
> >>>
> >>> Add read/write attribute for proximity resolution, and read attribute
> >>> for available proximity resolution values for the vcnl4040 chip.
> >>>
> >>> Signed-off-by: Mårten Lindahl <marten.lindahl@...s.com>  
> > Hi Jonathan!  
> >> I'll review this on basis the usecase is clear (see reply to cover 
> >> letter)  
> >
> > I'll skip this patch (see reply to cover letter comment)
> >
> > Your are right about the paired register manipulation. Better to 
> > read/write byte instead of word.  
> 
> Hi Jonathan!
> 
> I now remember why the register is read as a word (CONF1/CONF2). It is 
> addressed as one 16 bit register where CONF1 is the low 8 bit field and 
> CONF2 is the high bit field. It is the same address/code for both:
> 
> Register PS_CONF1 - COMMAND CODE: 0x03_L (0x03 DATA BYTE LOW)
> 
> Register PS_CONF2 - COMMAND CODE: 0x03_H (0x03 DATA BYTE HIGH)
> 
> Where in my case (PS_CONF2->PS_HD[3] is the same as PS_CONF1[11])
Ouch. That's a horrible way to define a register map.

Jonathan

> 
> Kind regards
> 
> Mårten
> 
> >
> > Kind regards
> >
> > Mårten
> >  
> >>
> >> The manipulation of the CONF1 and CONF2 registers in a pair is rather 
> >> odd given you
> >> only want to change one bit here.
> >>
> >> Why is that done?  
> >>> ---
> >>>   drivers/iio/light/vcnl4000.c | 87 
> >>> +++++++++++++++++++++++++++++++++++++++++++-
> >>>   1 file changed, 85 insertions(+), 2 deletions(-)
> >>>
> >>> diff --git a/drivers/iio/light/vcnl4000.c 
> >>> b/drivers/iio/light/vcnl4000.c
> >>> index fdf763a04b0b..2addff635b79 100644
> >>> --- a/drivers/iio/light/vcnl4000.c
> >>> +++ b/drivers/iio/light/vcnl4000.c
> >>> @@ -90,6 +90,7 @@
> >>>   #define VCNL4040_PS_CONF1_PS_SHUTDOWN    BIT(0)
> >>>   #define VCNL4040_PS_CONF2_PS_IT    GENMASK(3, 1) /* Proximity 
> >>> integration time */
> >>>   #define VCNL4040_CONF1_PS_PERS    GENMASK(5, 4) /* Proximity 
> >>> interrupt persistence setting */
> >>> +#define VCNL4040_PS_CONF2_PS_HD        BIT(11)    /* Proximity high 
> >>> definition */
> >>>   #define VCNL4040_PS_CONF2_PS_INT    GENMASK(9, 8) /* Proximity 
> >>> interrupt mode */
> >>>   #define VCNL4040_PS_CONF3_MPS        GENMASK(6, 5) /* Proximity 
> >>> multi pulse number */
> >>>   #define VCNL4040_PS_MS_LED_I        GENMASK(10, 8) /* Proximity 
> >>> current */
> >>> @@ -170,6 +171,11 @@ static const int vcnl4040_ps_calibbias_ua[][2] = {
> >>>       {0, 200000},
> >>>   };
> >>>   +static const int vcnl4040_ps_resolutions[2] = {
> >>> +    12,
> >>> +    16
> >>> +};
> >>> +
> >>>   static const int vcnl4040_als_persistence[] = {1, 2, 4, 8};
> >>>   static const int vcnl4040_ps_persistence[] = {1, 2, 3, 4};
> >>>   static const int vcnl4040_ps_oversampling_ratio[] = {1, 2, 4, 8};
> >>> @@ -880,6 +886,54 @@ static ssize_t 
> >>> vcnl4040_write_ps_calibbias(struct vcnl4000_data *data, int val)
> >>>       return ret;
> >>>   }
> >>>   +static ssize_t vcnl4040_read_ps_resolution(struct vcnl4000_data 
> >>> *data, int *val, int *val2)
> >>> +{
> >>> +    int ret;
> >>> +
> >>> +    ret = i2c_smbus_read_word_data(data->client, VCNL4200_PS_CONF1);  
> >> The field seems to be in PS_CONF2.  So you are reading a word and I 
> >> guess that
> >> gets you two registers.  Can we not do a byte read to get just CONF2?  
> >>> +    if (ret < 0)
> >>> +        return ret;
> >>> +
> >>> +    ret = FIELD_GET(VCNL4040_PS_CONF2_PS_HD, ret);
> >>> +    if (ret >= ARRAY_SIZE(vcnl4040_ps_resolutions))
> >>> +        return -EINVAL;
> >>> +
> >>> +    *val = vcnl4040_ps_resolutions[ret];
> >>> +    *val2 = 0;
> >>> +
> >>> +    return ret;
> >>> +}
> >>> +
> >>> +static ssize_t vcnl4040_write_ps_resolution(struct vcnl4000_data 
> >>> *data, int val)
> >>> +{
> >>> +    unsigned int i;
> >>> +    int ret;
> >>> +    u16 regval;
> >>> +
> >>> +    for (i = 0; i < ARRAY_SIZE(vcnl4040_ps_resolutions); i++) {
> >>> +        if (val == vcnl4040_ps_resolutions[i])
> >>> +            break;
> >>> +    }
> >>> +
> >>> +    if (i >= ARRAY_SIZE(vcnl4040_ps_resolutions))
> >>> +        return -EINVAL;
> >>> +
> >>> +    mutex_lock(&data->vcnl4000_lock);
> >>> +
> >>> +    ret = i2c_smbus_read_word_data(data->client, VCNL4200_PS_CONF1);
> >>> +    if (ret < 0)
> >>> +        goto out_unlock;
> >>> +
> >>> +    regval = (ret & ~VCNL4040_PS_CONF2_PS_HD);
> >>> +    regval |= FIELD_PREP(VCNL4040_PS_CONF2_PS_HD, i);
> >>> +    ret = i2c_smbus_write_word_data(data->client, VCNL4200_PS_CONF1,
> >>> +                    regval);
> >>> +
> >>> +out_unlock:
> >>> +    mutex_unlock(&data->vcnl4000_lock);
> >>> +    return ret;
> >>> +}  
> >> c),