Re: [PATCH 03/21] spi: spi-geni-qcom: Use OPP API to set clk/perf state

[Rajendra Nayak]

On 4/9/2020 11:50 PM, Matthias Kaehlcke wrote:
> Hi Rajendra,
>
> On Wed, Apr 08, 2020 at 07:16:29PM +0530, Rajendra Nayak wrote:
> > geni spi needs to express a perforamnce state requirement on CX
> > depending on the frequency of the clock rates. Use OPP table from
> > DT to register with OPP framework and use dev_pm_opp_set_rate() to
> > set the clk/perf state.
> >
> > Signed-off-by: Rajendra Nayak <rnayak@xxxxxxxxxxxxxx>
> > Cc: Alok Chauhan <alokc@xxxxxxxxxxxxxx>
> > Cc: Akash Asthana <akashast@xxxxxxxxxxxxxx>
> > Cc: linux-spi@xxxxxxxxxxxxxxx
> > ---
> >   drivers/spi/spi-geni-qcom.c | 14 +++++++++++---
> >   1 file changed, 11 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/spi/spi-geni-qcom.c b/drivers/spi/spi-geni-qcom.c
> > index c397242..ce387dc 100644
> > --- a/drivers/spi/spi-geni-qcom.c
> > +++ b/drivers/spi/spi-geni-qcom.c
> > @@ -7,6 +7,7 @@
> >   #include <linux/log2.h>
> >   #include <linux/module.h>
> >   #include <linux/platform_device.h>
> > +#include <linux/pm_opp.h>
> >   #include <linux/pm_runtime.h>
> >   #include <linux/qcom-geni-se.h>
> >   #include <linux/spi/spi.h>
> > @@ -95,7 +96,6 @@ static int get_spi_clk_cfg(unsigned int speed_hz,
> >   {
> >   	unsigned long sclk_freq;
> >   	unsigned int actual_hz;
> > -	struct geni_se *se = &mas->se;
> >   	int ret;
> >   
> >   	ret = geni_se_clk_freq_match(&mas->se,
> > @@ -112,9 +112,9 @@ static int get_spi_clk_cfg(unsigned int speed_hz,
> >   
> >   	dev_dbg(mas->dev, "req %u=>%u sclk %lu, idx %d, div %d\n", speed_hz,
> >   				actual_hz, sclk_freq, *clk_idx, *clk_div);
> > -	ret = clk_set_rate(se->clk, sclk_freq);
> > +	ret = dev_pm_opp_set_rate(mas->dev, sclk_freq);
> >   	if (ret)
> > -		dev_err(mas->dev, "clk_set_rate failed %d\n", ret);
> > +		dev_err(mas->dev, "dev_pm_opp_set_rate failed %d\n", ret);
> >   	return ret;
> >   }
> >   
> > @@ -553,6 +553,7 @@ static int spi_geni_probe(struct platform_device *pdev)
> >   	if (!spi)
> >   		return -ENOMEM;
> >   
> > +
> >   	platform_set_drvdata(pdev, spi);
> >   	mas = spi_master_get_devdata(spi);
> >   	mas->irq = irq;
> > @@ -561,6 +562,8 @@ static int spi_geni_probe(struct platform_device *pdev)
> >   	mas->se.wrapper = dev_get_drvdata(dev->parent);
> >   	mas->se.base = base;
> >   	mas->se.clk = clk;
> > +	mas->se.opp = dev_pm_opp_set_clkname(&pdev->dev, "se");
>
> As commented on the serial patch, it seems an error check is needed
> and the OPP table saved in 'struct geni_se' is never used.

right, I explained about the OPP table in the other patch.

> > +	dev_pm_opp_of_add_table(&pdev->dev);
>
> This function could also fail for multiple reasons, so the return value
> should be checked.
>
>  From patch "[01/21] opp: Manage empty OPP tables with clk handle" it seems
> ignoring errors is intended to be able to operate when no OPP table is
> specified. But even with that you want to return in case of certain errors,
> like an invalid OPP table, out of memory or -EPROBE_DEFER.

Thats correct, my intention of not checking the errors was in cases where its
optional, and I did not want to break anything with existing DT. I will go back and
check if its indeed possible for it to return a -EPROBE_DEFER and others which
I should handle.

--
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
of Code Aurora Forum, hosted by The Linux Foundation