Re: [PATCH v2 1/8] i2c: i2c-qcom-geni: Provide support for ACPI

[Lee Jones <lee.jones@xxxxxxxxxx>]

On Fri, 07 Jun 2019, Ard Biesheuvel wrote:

> On Fri, 7 Jun 2019 at 10:29, Lee Jones <lee.jones@xxxxxxxxxx> wrote:
> >
> > Add a match table to allow automatic probing of ACPI device
> > QCOM0220.  Ignore clock attainment errors.  Set default clock
> > frequency value.
> >
> > Signed-off-by: Lee Jones <lee.jones@xxxxxxxxxx>
> > ---
> >  drivers/i2c/busses/i2c-qcom-geni.c | 19 +++++++++++++++++--
> >  1 file changed, 17 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c
> > index db075bc0d952..0fa93b448e8d 100644
> > --- a/drivers/i2c/busses/i2c-qcom-geni.c
> > +++ b/drivers/i2c/busses/i2c-qcom-geni.c
> > @@ -1,6 +1,7 @@
> >  // SPDX-License-Identifier: GPL-2.0
> >  // Copyright (c) 2017-2018, The Linux Foundation. All rights reserved.
> >
> > +#include <linux/acpi.h>
> >  #include <linux/clk.h>
> >  #include <linux/dma-mapping.h>
> >  #include <linux/err.h>
> > @@ -483,6 +484,12 @@ static const struct i2c_algorithm geni_i2c_algo = {
> >         .functionality  = geni_i2c_func,
> >  };
> >
> > +static const struct acpi_device_id geni_i2c_acpi_match[] = {
> > +       { "QCOM0220"},
> > +       { },
> > +};
> > +MODULE_DEVICE_TABLE(acpi, geni_i2c_acpi_match);
> > +
> 
> We usually put #ifdef CONFIG_ACPI/#endif around these, otherwise you
> end up with acpi:XXXX modaliases even though ACPI is not compiled in.

No problem.

> >  static int geni_i2c_probe(struct platform_device *pdev)
> >  {
> >         struct geni_i2c_dev *gi2c;
> > @@ -502,7 +509,7 @@ static int geni_i2c_probe(struct platform_device *pdev)
> >                 return PTR_ERR(gi2c->se.base);
> >
> >         gi2c->se.clk = devm_clk_get(&pdev->dev, "se");
> 
> Can we avoid this call altogether in ACPI mode? Also, please use

I'm trying not to place all non-ACPI specific callers into if ()
statements.  The tabbing becomes ridiculous in some places.  A great
deal of these calls are requesting optional resources too, so it's
better to simply ignore the returning error in the cases where
non-optional resources (such as this one) are requested, since it has
the least impact on the existing code.

> 'has_acpi_companion()' to test whether we are probing via ACPI.

Sure.

> > -       if (IS_ERR(gi2c->se.clk)) {
> > +       if (IS_ERR(gi2c->se.clk) && !ACPI_HANDLE(&pdev->dev)) {
> 
> 
> >                 ret = PTR_ERR(gi2c->se.clk);
> >                 dev_err(&pdev->dev, "Err getting SE Core clk %d\n", ret);
> >                 return ret;
> > @@ -510,12 +517,19 @@ static int geni_i2c_probe(struct platform_device *pdev)
> >
> >         ret = device_property_read_u32(&pdev->dev, "clock-frequency",
> >                                                         &gi2c->clk_freq_out);
> > -       if (ret) {
> > +       if (ret && !ACPI_HANDLE(&pdev->dev)) {
> >                 dev_info(&pdev->dev,
> >                         "Bus frequency not specified, default to 100kHz.\n");
> >                 gi2c->clk_freq_out = KHZ(100);
> >         }
> >
> > +       if (ACPI_HANDLE(&pdev->dev)) {
> > +               ACPI_COMPANION_SET(&gi2c->adap.dev, ACPI_COMPANION(&pdev->dev));
> > +
> > +               /* Using default, same as the !ACPI case above */
> > +               gi2c->clk_freq_out = KHZ(100);
> > +       }
> > +
> 
> You are overriding the speed to 100 kHz even if the ACPI device has a
> "clock-frequency" property.

Will look at this.

Thanks Ard.

-- 
Lee Jones [李琼斯]
Linaro Services Technical Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog