Re: [PATCH 2/2] cpuidle: qcom: Add SPM register data for APQ8026 and MSM8226

From: Bartosz Dudziak <bartosz.dudziak@snejp.pl>
To: Stephan Gerhold <stephan@gerhold.net>
Cc: Rob Herring <robh+dt@kernel.org>, Andy Gross <agross@kernel.org>,
	Bjorn Andersson <bjorn.andersson@linaro.org>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Daniel Lezcano <daniel.lezcano@linaro.org>,
	David Sterba <dsterba@suse.com>, Jens Axboe <axboe@kernel.dk>,
	devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-arm-msm@vger.kernel.org, linux-pm@vger.kernel.org
Subject: Re: [PATCH 2/2] cpuidle: qcom: Add SPM register data for APQ8026 and MSM8226
Date: Tue, 25 May 2021 20:58:52 +0200	[thread overview]
Message-ID: <20210525185852.GA2416@PackardBell> (raw)
In-Reply-To: <YJ1u3syIquKRyuv2@gerhold.net>

Hi,

On Thu, May 13, 2021 at 08:24:30PM +0200, Stephan Gerhold wrote:
> Hi,
> 
> On Thu, May 13, 2021 at 05:01:50PM +0200, Bartosz Dudziak wrote:
> > Add APQ8026 and MSM8226 SoCs register data to SPM AVS Wrapper 2 (SAW2)
> > power controller driver.
> > 
> > Signed-off-by: Bartosz Dudziak <bartosz.dudziak@snejp.pl>
> > ---
> >  drivers/cpuidle/cpuidle-qcom-spm.c | 16 ++++++++++++++++
> >  1 file changed, 16 insertions(+)
> > 
> > diff --git a/drivers/cpuidle/cpuidle-qcom-spm.c b/drivers/cpuidle/cpuidle-qcom-spm.c
> > index adf91a6e4d..9711a98d68 100644
> > --- a/drivers/cpuidle/cpuidle-qcom-spm.c
> > +++ b/drivers/cpuidle/cpuidle-qcom-spm.c
> > @@ -87,6 +87,18 @@ static const struct spm_reg_data spm_reg_8974_8084_cpu  = {
> >  	.start_index[PM_SLEEP_MODE_SPC] = 3,
> >  };
> >  
> > +/* SPM register data for 8026, 8226 */
> > +static const struct spm_reg_data spm_reg_8x26_cpu  = {
> > +	.reg_offset = spm_reg_offset_v2_1,
> > +	.spm_cfg = 0x0,
> > +	.spm_dly = 0x3C102800,
> > +	.seq = { 0x60, 0x03, 0x60, 0x0B, 0x0F, 0x20, 0x10, 0x80, 0x30, 0x90,
> > +		0x5B, 0x60, 0x03, 0x60, 0x3B, 0x76, 0x76, 0x0B, 0x94, 0x5B,
> > +		0x80, 0x10, 0x26, 0x30, 0x0F },
> > +	.start_index[PM_SLEEP_MODE_STBY] = 0,
> > +	.start_index[PM_SLEEP_MODE_SPC] = 5,
> > +};
> > +
> >  static const u8 spm_reg_offset_v1_1[SPM_REG_NR] = {
> >  	[SPM_REG_CFG]		= 0x08,
> >  	[SPM_REG_SPM_CTL]	= 0x20,
> > @@ -259,6 +271,10 @@ static struct spm_driver_data *spm_get_drv(struct platform_device *pdev,
> >  }
> >  
> >  static const struct of_device_id spm_match_table[] = {
> > +	{ .compatible = "qcom,apq8026-saw2-v2.1-cpu",
> > +	  .data = &spm_reg_8x26_cpu },
> > +	{ .compatible = "qcom,msm8226-saw2-v2.1-cpu",
> > +	  .data = &spm_reg_8x26_cpu },
> 
> What is the reason for having a separate compatible for APQ8026?
> 
> If the difference between MSM8226 and APQ8026 is similar to other qcom
> SoCs (just lack of modem), both will end up using the same device tree
> include anyway. Then it's easier to have both use qcom,msm8226-saw2-v2.1-cpu.
> 
> Thanks,
> Stephan

You are right. There is no reason to have a seperate APQ8026 compatible because
it will share the MSM8226 device tree. I will send later a v2 patch with only
"qcom,msm8226-saw2-v2.1-cpu" option.

Thanks,
Bartosz 