[PATCH 1/2] platform/x86/intel/pmc: Add resume_fixup callback

[PATCH 1/2] platform/x86/intel/pmc: Add resume_fixup callback

[David E. Box]

Add a resume_fixup callback to perform platform specific fixups during
resume from suspend.

Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
 drivers/platform/x86/intel/pmc/core.c | 3 +++
 drivers/platform/x86/intel/pmc/core.h | 2 ++
 2 files changed, 5 insertions(+)

diff --git a/drivers/platform/x86/intel/pmc/core.c b/drivers/platform/x86/intel/pmc/core.c
index da6e7206d38b..0fcd1cb7264b 100644
--- a/drivers/platform/x86/intel/pmc/core.c
+++ b/drivers/platform/x86/intel/pmc/core.c
@@ -1229,6 +1229,9 @@ static __maybe_unused int pmc_core_resume(struct device *dev)
 	const struct pmc_bit_map **maps = pmcdev->map->lpm_sts;
 	int offset = pmcdev->map->lpm_status_offset;
 
+	if (pmcdev->resume_fixup)
+		pmcdev->resume_fixup();
+
 	/* Check if the syspend used S0ix */
 	if (pm_suspend_via_firmware())
 		return 0;
diff --git a/drivers/platform/x86/intel/pmc/core.h b/drivers/platform/x86/intel/pmc/core.h
index 9ca9b9746719..b58458d00bd3 100644
--- a/drivers/platform/x86/intel/pmc/core.h
+++ b/drivers/platform/x86/intel/pmc/core.h
@@ -327,6 +327,7 @@ struct pmc_reg_map {
  * @lpm_en_modes:	Array of enabled modes from lowest to highest priority
  * @lpm_req_regs:	List of substate requirements
  * @core_configure:	Function pointer to configure the platform
+ * @resume_fixup:	Function to perform fixups during resume
  *
  * pmc_dev contains info about power management controller device.
  */
@@ -345,6 +346,7 @@ struct pmc_dev {
 	int lpm_en_modes[LPM_MAX_NUM_MODES];
 	u32 *lpm_req_regs;
 	void (*core_configure)(struct pmc_dev *pmcdev);
+	void (*resume_fixup)(void);
 };
 
 extern const struct pmc_bit_map msr_map[];
-- 
2.34.1

[PATCH 2/2] platform/x86/intel/pmc/mtl: Put devices in D3 during resume

[David E. Box]

An earlier commit placed some driverless devices in D3 during boot so that
they don't block package cstate entry. Also place these devices in D3 after
resume from suspend.

Fixes: 336ba968d3e3 ("platform/x86/intel/pmc/mtl: Put GNA/IPU/VPU devices in D3")
Signed-off-by: David E. Box <david.e.box@linux.intel.com>
---
 drivers/platform/x86/intel/pmc/mtl.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/platform/x86/intel/pmc/mtl.c b/drivers/platform/x86/intel/pmc/mtl.c
index e8cc156412ce..d87c4597c6d4 100644
--- a/drivers/platform/x86/intel/pmc/mtl.c
+++ b/drivers/platform/x86/intel/pmc/mtl.c
@@ -68,11 +68,8 @@ static void mtl_set_device_d3(unsigned int device)
 	}
 }
 
-void mtl_core_init(struct pmc_dev *pmcdev)
+static void mtl_fixup(void)
 {
-	pmcdev->map = &mtl_reg_map;
-	pmcdev->core_configure = mtl_core_configure;
-
 	/*
 	 * Set power state of select devices that do not have drivers to D3
 	 * so that they do not block Package C entry.
@@ -81,3 +78,13 @@ void mtl_core_init(struct pmc_dev *pmcdev)
 	mtl_set_device_d3(MTL_IPU_PCI_DEV);
 	mtl_set_device_d3(MTL_VPU_PCI_DEV);
 }
+
+void mtl_core_init(struct pmc_dev *pmcdev)
+{
+	pmcdev->map = &mtl_reg_map;
+	pmcdev->core_configure = mtl_core_configure;
+
+	mtl_fixup();
+
+	pmcdev->resume_fixup = mtl_fixup;
+}
-- 
2.34.1

Re: [PATCH 2/2] platform/x86/intel/pmc/mtl: Put devices in D3 during resume

[Ilpo Järvinen]

On Fri, 2 Jun 2023, David E. Box wrote:

> An earlier commit placed some driverless devices in D3 during boot so that
> they don't block package cstate entry. Also place these devices in D3 after
> resume from suspend.
> 
> Fixes: 336ba968d3e3 ("platform/x86/intel/pmc/mtl: Put GNA/IPU/VPU devices in D3")
> Signed-off-by: David E. Box <david.e.box@linux.intel.com>
> ---
>  drivers/platform/x86/intel/pmc/mtl.c | 15 +++++++++++----
>  1 file changed, 11 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/platform/x86/intel/pmc/mtl.c b/drivers/platform/x86/intel/pmc/mtl.c
> index e8cc156412ce..d87c4597c6d4 100644
> --- a/drivers/platform/x86/intel/pmc/mtl.c
> +++ b/drivers/platform/x86/intel/pmc/mtl.c
> @@ -68,11 +68,8 @@ static void mtl_set_device_d3(unsigned int device)
>  	}
>  }
>  
> -void mtl_core_init(struct pmc_dev *pmcdev)
> +static void mtl_fixup(void)
>  {
> -	pmcdev->map = &mtl_reg_map;
> -	pmcdev->core_configure = mtl_core_configure;
> -
>  	/*
>  	 * Set power state of select devices that do not have drivers to D3
>  	 * so that they do not block Package C entry.
> @@ -81,3 +78,13 @@ void mtl_core_init(struct pmc_dev *pmcdev)
>  	mtl_set_device_d3(MTL_IPU_PCI_DEV);
>  	mtl_set_device_d3(MTL_VPU_PCI_DEV);

I'd prefer the function be called something related to d3 / power state / 
or some along those lines rather than something obscure such as 
mtl_fixup(). And you can move the comment to be a function comment now.

>  }
> +
> +void mtl_core_init(struct pmc_dev *pmcdev)
> +{
> +	pmcdev->map = &mtl_reg_map;
> +	pmcdev->core_configure = mtl_core_configure;
> +
> +	mtl_fixup();
> +
> +	pmcdev->resume_fixup = mtl_fixup;

I'm a bit on the edge here whether this is a good approach in long-term or 
if it would be better to just provide a way for the platform file to 
replace entire .resume() (for this task it's obviously enough but it 
feels a bit hacky to hook into one fixed place on resume path).

static __maybe_unused int pmc_core_resume(struct device *dev)
{
	if (pmcdev->resume)
		return pmcdev->resume();
	else
		return pmc_core_resume_common();
}

where pmc_core_resume_common() contains the current pmc_core_resume() 
contents.

mtl_resume() would just call the d3 func and the common resume functions.


-- 
 i.

Re: [PATCH 2/2] platform/x86/intel/pmc/mtl: Put devices in D3 during resume

[David E. Box]

On Mon, 2023-06-05 at 14:59 +0300, Ilpo Järvinen wrote:
> On Fri, 2 Jun 2023, David E. Box wrote:
> 
> > An earlier commit placed some driverless devices in D3 during boot so that
> > they don't block package cstate entry. Also place these devices in D3 after
> > resume from suspend.
> > 
> > Fixes: 336ba968d3e3 ("platform/x86/intel/pmc/mtl: Put GNA/IPU/VPU devices in
> > D3")
> > Signed-off-by: David E. Box <david.e.box@linux.intel.com>
> > ---
> >  drivers/platform/x86/intel/pmc/mtl.c | 15 +++++++++++----
> >  1 file changed, 11 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/platform/x86/intel/pmc/mtl.c
> > b/drivers/platform/x86/intel/pmc/mtl.c
> > index e8cc156412ce..d87c4597c6d4 100644
> > --- a/drivers/platform/x86/intel/pmc/mtl.c
> > +++ b/drivers/platform/x86/intel/pmc/mtl.c
> > @@ -68,11 +68,8 @@ static void mtl_set_device_d3(unsigned int device)
> >         }
> >  }
> >  
> > -void mtl_core_init(struct pmc_dev *pmcdev)
> > +static void mtl_fixup(void)
> >  {
> > -       pmcdev->map = &mtl_reg_map;
> > -       pmcdev->core_configure = mtl_core_configure;
> > -
> >         /*
> >          * Set power state of select devices that do not have drivers to D3
> >          * so that they do not block Package C entry.
> > @@ -81,3 +78,13 @@ void mtl_core_init(struct pmc_dev *pmcdev)
> >         mtl_set_device_d3(MTL_IPU_PCI_DEV);
> >         mtl_set_device_d3(MTL_VPU_PCI_DEV);
> 
> I'd prefer the function be called something related to d3 / power state / 
> or some along those lines rather than something obscure such as 
> mtl_fixup(). And you can move the comment to be a function comment now.

Okay.

> 
> >  }
> > +
> > +void mtl_core_init(struct pmc_dev *pmcdev)
> > +{
> > +       pmcdev->map = &mtl_reg_map;
> > +       pmcdev->core_configure = mtl_core_configure;
> > +
> > +       mtl_fixup();
> > +
> > +       pmcdev->resume_fixup = mtl_fixup;
> 
> I'm a bit on the edge here whether this is a good approach in long-term or 
> if it would be better to just provide a way for the platform file to 
> replace entire .resume() (for this task it's obviously enough but it 
> feels a bit hacky to hook into one fixed place on resume path).
> 
> static __maybe_unused int pmc_core_resume(struct device *dev)
> {
>         if (pmcdev->resume)
>                 return pmcdev->resume();
>         else
>                 return pmc_core_resume_common();
> }
> 
> where pmc_core_resume_common() contains the current pmc_core_resume() 
> contents.
> 
> mtl_resume() would just call the d3 func and the common resume functions.

Yeah, makes sense. There are several conditional tasks in the current resume
code, so trying to have a fixed place in there to call a platform specific
workaround may not be possible if we need to use this again for a different
reason.

David

> 
>