| 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
| ||
|
Message-ID: <CAF12kFtQH3_EsPBnDgBL59tYZ6wintgt9yceS2tUiB1Pv8qjgQ@mail.gmail.com>
Date: Tue, 1 Mar 2022 14:27:50 +0800
From: Cixi Geng <gengcixi@...il.com>
To: Jonathan Cameron <Jonathan.Cameron@...wei.com>
Cc: Baolin Wang <baolin.wang7@...il.com>,
Orson Zhai <orsonzhai@...il.com>,
Chunyan Zhang <zhang.lyra@...il.com>, jic23@...nel.org,
Lars-Peter Clausen <lars@...afoo.de>,
Rob Herring <robh+dt@...nel.org>, lgirdwood@...il.com,
Mark Brown <broonie@...nel.org>,
朱玉明 (Yuming Zhu/11457)
<yuming.zhu1@...soc.com>, linux-iio@...r.kernel.org,
Devicetree List <devicetree@...r.kernel.org>,
LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/7] iio: adc: sc27xx: structure adjuststment and optimization
Jonathan Cameron <Jonathan.Cameron@...wei.com> 于2022年2月25日周五 18:19写道：
>
> On Wed, 23 Feb 2022 20:46:08 +0800
> Cixi Geng <gengcixi@...il.com> wrote:
>
> > Baolin Wang <baolin.wang7@...il.com> 于2022年2月10日周四 16:07写道：
> > >
> > > On Mon, Jan 24, 2022 at 4:07 PM Cixi Geng <gengcixi@...il.com> wrote:
> > > >
> > > > Baolin Wang <baolin.wang7@...il.com> 于2022年1月17日周一 14:15写道：
> > > > >
> > > > > On Thu, Jan 13, 2022 at 9:54 AM Cixi Geng <gengcixi@...il.com> wrote:
> > > > > >
> > > > > > Baolin Wang <baolin.wang7@...il.com> 于2022年1月7日周五 15:03写道：
> > > > > > >
> > > > > > > On Thu, Jan 6, 2022 at 9:00 PM Cixi Geng <gengcixi@...il.com> wrote:
> > > > > > > >
> > > > > > > > From: Cixi Geng <cixi.geng1@...soc.com>
> > > > > > > >
> > > > > > > > Introduce one variant device data structure to be compatible
> > > > > > > > with SC2731 PMIC since it has different scale and ratio calculation
> > > > > > > > and so on.
> > > > > > > >
> > > > > > > > Signed-off-by: Yuming Zhu <yuming.zhu1@...soc.com>
> > > > > > > > Signed-off-by: Cixi Geng <cixi.geng1@...soc.com>
> > > > > > > > ---
> > > > > > > > drivers/iio/adc/sc27xx_adc.c | 94 ++++++++++++++++++++++++++++++------
> > > > > > > > 1 file changed, 79 insertions(+), 15 deletions(-)
> > > > > > > >
> > > > > > > > diff --git a/drivers/iio/adc/sc27xx_adc.c b/drivers/iio/adc/sc27xx_adc.c
> > > > > > > > index aee076c8e2b1..d2712e54ee79 100644
> > > > > > > > --- a/drivers/iio/adc/sc27xx_adc.c
> > > > > > > > +++ b/drivers/iio/adc/sc27xx_adc.c
> > > > > > > > @@ -12,9 +12,9 @@
> > > > > > > > #include <linux/slab.h>
> > > > > > > >
> > > > > > > > /* PMIC global registers definition */
> > > > > > > > -#define SC27XX_MODULE_EN 0xc08
> > > > > > > > +#define SC2731_MODULE_EN 0xc08
> > > > > > > > #define SC27XX_MODULE_ADC_EN BIT(5)
> > > > > > > > -#define SC27XX_ARM_CLK_EN 0xc10
> > > > > > > > +#define SC2731_ARM_CLK_EN 0xc10
> > > > > > > > #define SC27XX_CLK_ADC_EN BIT(5)
> > > > > > > > #define SC27XX_CLK_ADC_CLK_EN BIT(6)
> > > > > > > >
> > > > > > > > @@ -78,6 +78,23 @@ struct sc27xx_adc_data {
> > > > > > > > int channel_scale[SC27XX_ADC_CHANNEL_MAX];
> > > > > > > > u32 base;
> > > > > > > > int irq;
> > > > > > > > + const struct sc27xx_adc_variant_data *var_data;
> > > > > > > > +};
> > > > > > > > +
> > > > > > > > +/*
> > > > > > > > + * Since different PMICs of SC27xx series can have different
> > > > > > > > + * address and ratio, we should save ratio config and base
> > > > > > > > + * in the device data structure.
> > > > > > > > + */
> > > > > > > > +struct sc27xx_adc_variant_data {
> > > > > > > > + u32 module_en;
> > > > > > > > + u32 clk_en;
> > > > > > > > + u32 scale_shift;
> > > > > > > > + u32 scale_mask;
> > > > > > > > + const struct sc27xx_adc_linear_graph *bscale_cal;
> > > > > > > > + const struct sc27xx_adc_linear_graph *sscale_cal;
> > > > > > > > + void (*init_scale)(struct sc27xx_adc_data *data);
> > > > > > > > + int (*get_ratio)(int channel, int scale);
> > > > > > > > };
> > > > > > > >
> > > > > > > > struct sc27xx_adc_linear_graph {
> > > > > > > > @@ -103,6 +120,16 @@ static struct sc27xx_adc_linear_graph small_scale_graph = {
> > > > > > > > 100, 341,
> > > > > > > > };
> > > > > > > >
> > > > > > > > +static const struct sc27xx_adc_linear_graph sc2731_big_scale_graph_calib = {
> > > > > > > > + 4200, 850,
> > > > > > > > + 3600, 728,
> > > > > > > > +};
> > > > > > > > +
> > > > > > > > +static const struct sc27xx_adc_linear_graph sc2731_small_scale_graph_calib = {
> > > > > > > > + 1000, 838,
> > > > > > > > + 100, 84,
> > > > > > > > +};
> > > > > > >
> > > > > > > The original big_scale_graph_calib and small_scale_graph_calib are for
> > > > > > > SC2731 PMIC, why add new structure definition for SC2731?
> > > > > > >
> > > > > > > > +
> > > > > > > > static const struct sc27xx_adc_linear_graph big_scale_graph_calib = {
> > > > > > > > 4200, 856,
> > > > > > > > 3600, 733,
> > > > > > > > @@ -130,11 +157,11 @@ static int sc27xx_adc_scale_calibration(struct sc27xx_adc_data *data,
> > > > > > > > size_t len;
> > > > > > > >
> > > > > > > > if (big_scale) {
> > > > > > > > - calib_graph = &big_scale_graph_calib;
> > > > > > > > + calib_graph = data->var_data->bscale_cal;
> > > > > > > > graph = &big_scale_graph;
> > > > > > > > cell_name = "big_scale_calib";
> > > > > > > > } else {
> > > > > > > > - calib_graph = &small_scale_graph_calib;
> > > > > > > > + calib_graph = data->var_data->sscale_cal;
> > > > > > > > graph = &small_scale_graph;
> > > > > > > > cell_name = "small_scale_calib";
> > > > > > > > }
> > > > > > > > @@ -160,7 +187,7 @@ static int sc27xx_adc_scale_calibration(struct sc27xx_adc_data *data,
> > > > > > > > return 0;
> > > > > > > > }
> > > > > > > >
> > > > > > > > -static int sc27xx_adc_get_ratio(int channel, int scale)
> > > > > > > > +static int sc2731_adc_get_ratio(int channel, int scale)
> > > > > > > > {
> > > > > > > > switch (channel) {
> > > > > > > > case 1:
> > > > > > > > @@ -185,6 +212,21 @@ static int sc27xx_adc_get_ratio(int channel, int scale)
> > > > > > > > return SC27XX_VOLT_RATIO(1, 1);
> > > > > > > > }
> > > > > > > >
> > > > > > > > +/*
> > > > > > > > + * According to the datasheet set specific value on some channel.
> > > > > > > > + */
> > > > > > > > +static void sc2731_adc_scale_init(struct sc27xx_adc_data *data)
> > > > > > > > +{
> > > > > > > > + int i;
> > > > > > > > +
> > > > > > > > + for (i = 0; i < SC27XX_ADC_CHANNEL_MAX; i++) {
> > > > > > > > + if (i == 5)
> > > > > > > > + data->channel_scale[i] = 1;
> > > > > > > > + else
> > > > > > > > + data->channel_scale[i] = 0;
> > > > > > > > + }
> > > > > > > > +}
> > > > > > >
> > > > > > > This is unnecessary I think, please see sc27xx_adc_write_raw() that
> > > > > > > can set the channel scale.
> > > > > > Did you mean that all the PMIC's scale_init function should put into
> > > > > > the sc27xx_adc_write_raw?
> > > > >
> > > > > No.
> > > > >
> > > > > > but the scale_init is all different by each PMIC, if implemented in
> > > > > > the write_raw, will add a lot of
> > > > > > if or switch_case branch
> > > > >
> > > > > What I mean is we should follow the original method to set the channel
> > > > > scale by iio_info. Please also refer to other drivers how ot handle
> > > > > the channel scale.
> > > > Hi Baolin, I understand the adc_write_raw() function is the method to set
> > > > channal scale for the userspace, we can change the channel scale by write
> > > > a value on a user code. did i understand right?
> > > > out scale_init is to set scale value when the driver probe stage, and I also
> > > > did not found other adc driver use the adc_write_raw() during the driver
> > > > initialization phase.
> > >
> > > Hi Jonathan,
> > >
> > > How do you think about the method in this patch to set the channel
> > > scale? Thanks.
> > >
> > Hi Jonathan,
> > Could you have a loot at this patch ,and give some advice about the
> > method to set the channel scale? Thanks very much.
>
> Hi, thanks for poking me on this - I'd missed the question buried deep in the thread!
>
> Anyhow, I don't quite follow the discussion but think it could be focused
> on one of 2 questions...
>
> 1) Does setting an initial default make sense?
> Yes, this is an acceptable thing to do if there is a particular set of defaults
> and there is no risk of regressions (i.e. the device wasn't previously supported
> with different defaults).
> 2) Should you use the write_raw callback to actually do the setting?
> Probably not as it has a set of parameters that don't make as much sense from within
> the driver. It 'might' make sense to have a common _set() function for this
> feature which is called both in this initialization case and from the write_raw()
> function however as that could do bounds checking etc in one common place.
> However, it is very simple here, so perhaps not necessary.
>
> Jonathan
>
Hi Jonathan, thanks for your comment !
And Baolin, I will send a new verision for the patches tto keep the
scale_init and
fix other issues . thanks!
> > > --
> > > Baolin Wang
>
Powered by blists - more mailing lists