2024-02-14 21:18:46

by Konrad Dybcio

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On 14.02.2024 22:13, Abel Vesa wrote:
> Rather than setting up the core, obsrv and chnls in probe by using
> version specific conditionals, add a dedicated "get_core_resources"
> version specific op and move the acquiring in there.
>
> Signed-off-by: Abel Vesa <[email protected]>
> ---
> drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
> 1 file changed, 78 insertions(+), 33 deletions(-)
>
> diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
> index 23939c0d225f..489556467a4c 100644
> --- a/drivers/spmi/spmi-pmic-arb.c
> +++ b/drivers/spmi/spmi-pmic-arb.c
> @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
> */
> struct pmic_arb_ver_ops {
> const char *ver_str;
> + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
> int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
> int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
> /* spmi commands (read_cmd, write_cmd, cmd) functionality */
> @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
> return 0;
> }
>
> +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
> + void __iomem *core)
> +{
> + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> +
> + pmic_arb->wr_base = core;
> + pmic_arb->rd_base = core;
> +
> + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
> +
> + return 0;
> +}
> +
> static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
> {
> u32 *mapping_table;
> @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
> return apid;
> }
>
> +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
> +{
> + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> + struct device *dev = &pdev->dev;
> + struct resource *res;
> +
> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,

It's no longer indented to deep, no need to keep such aggressive wrapping

> + "obsrvr");
> + pmic_arb->rd_base = devm_ioremap(dev, res->start,
> + resource_size(res));
> + if (IS_ERR(pmic_arb->rd_base))
> + return PTR_ERR(pmic_arb->rd_base);
> +
> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> + "chnls");
> + pmic_arb->wr_base = devm_ioremap(dev, res->start,
> + resource_size(res));
> + if (IS_ERR(pmic_arb->wr_base))
> + return PTR_ERR(pmic_arb->wr_base);

Could probably make it "devm_platform_get_and_ioremap_resource "

Konrad


2024-02-14 21:36:42

by Abel Vesa

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On 24-02-14 22:18:33, Konrad Dybcio wrote:
> On 14.02.2024 22:13, Abel Vesa wrote:
> > Rather than setting up the core, obsrv and chnls in probe by using
> > version specific conditionals, add a dedicated "get_core_resources"
> > version specific op and move the acquiring in there.
> >
> > Signed-off-by: Abel Vesa <[email protected]>
> > ---
> > drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
> > 1 file changed, 78 insertions(+), 33 deletions(-)
> >
> > diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
> > index 23939c0d225f..489556467a4c 100644
> > --- a/drivers/spmi/spmi-pmic-arb.c
> > +++ b/drivers/spmi/spmi-pmic-arb.c
> > @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
> > */
> > struct pmic_arb_ver_ops {
> > const char *ver_str;
> > + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
> > int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
> > int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
> > /* spmi commands (read_cmd, write_cmd, cmd) functionality */
> > @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
> > return 0;
> > }
> >
> > +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
> > + void __iomem *core)
> > +{
> > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > +
> > + pmic_arb->wr_base = core;
> > + pmic_arb->rd_base = core;
> > +
> > + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
> > +
> > + return 0;
> > +}
> > +
> > static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
> > {
> > u32 *mapping_table;
> > @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
> > return apid;
> > }
> >
> > +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
> > +{
> > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > + struct device *dev = &pdev->dev;
> > + struct resource *res;
> > +
> > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
>
> It's no longer indented to deep, no need to keep such aggressive wrapping
>

The pmic_arb_get_obsrvr_chnls_v2 is used by both:
pmic_arb_get_core_resources_v2
pmic_arb_get_core_resources_v7

> > + "obsrvr");
> > + pmic_arb->rd_base = devm_ioremap(dev, res->start,
> > + resource_size(res));
> > + if (IS_ERR(pmic_arb->rd_base))
> > + return PTR_ERR(pmic_arb->rd_base);
> > +
> > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> > + "chnls");
> > + pmic_arb->wr_base = devm_ioremap(dev, res->start,
> > + resource_size(res));
> > + if (IS_ERR(pmic_arb->wr_base))
> > + return PTR_ERR(pmic_arb->wr_base);
>
> Could probably make it "devm_platform_get_and_ioremap_resource "

The reason this needs to stay as is is because of reason explained by
the following comment found in probe:

/*
* Please don't replace this with devm_platform_ioremap_resource() or
* devm_ioremap_resource(). These both result in a call to
* devm_request_mem_region() which prevents multiple mappings of this
* register address range. SoCs with PMIC arbiter v7 may define two
* arbiter devices, for the two physical SPMI interfaces, which share
* some register address ranges (i.e. "core", "obsrvr", and "chnls").
* Ensure that both devices probe successfully by calling devm_ioremap()
* which does not result in a devm_request_mem_region() call.
*/

Even though, AFAICT, there is no platform that adds a second node for
the second bus, currently, in mainline, we should probably allow the
"legacy" approach to still work.

>
> Konrad

2024-02-14 21:51:19

by Konrad Dybcio

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On 14.02.2024 22:36, Abel Vesa wrote:
> On 24-02-14 22:18:33, Konrad Dybcio wrote:
>> On 14.02.2024 22:13, Abel Vesa wrote:
>>> Rather than setting up the core, obsrv and chnls in probe by using
>>> version specific conditionals, add a dedicated "get_core_resources"
>>> version specific op and move the acquiring in there.
>>>
>>> Signed-off-by: Abel Vesa <[email protected]>
>>> ---
>>> drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
>>> 1 file changed, 78 insertions(+), 33 deletions(-)
>>>
>>> diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
>>> index 23939c0d225f..489556467a4c 100644
>>> --- a/drivers/spmi/spmi-pmic-arb.c
>>> +++ b/drivers/spmi/spmi-pmic-arb.c
>>> @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
>>> */
>>> struct pmic_arb_ver_ops {
>>> const char *ver_str;
>>> + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
>>> int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
>>> int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
>>> /* spmi commands (read_cmd, write_cmd, cmd) functionality */
>>> @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
>>> return 0;
>>> }
>>>
>>> +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
>>> + void __iomem *core)
>>> +{
>>> + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
>>> +
>>> + pmic_arb->wr_base = core;
>>> + pmic_arb->rd_base = core;
>>> +
>>> + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
>>> +
>>> + return 0;
>>> +}
>>> +
>>> static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
>>> {
>>> u32 *mapping_table;
>>> @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
>>> return apid;
>>> }
>>>
>>> +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
>>> +{
>>> + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
>>> + struct device *dev = &pdev->dev;
>>> + struct resource *res;
>>> +
>>> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
>>
>> It's no longer indented to deep, no need to keep such aggressive wrapping
>>
>
> The pmic_arb_get_obsrvr_chnls_v2 is used by both:
> pmic_arb_get_core_resources_v2
> pmic_arb_get_core_resources_v7

I meant line wrapping

>
>>> + "obsrvr");
>>> + pmic_arb->rd_base = devm_ioremap(dev, res->start,
>>> + resource_size(res));
>>> + if (IS_ERR(pmic_arb->rd_base))
>>> + return PTR_ERR(pmic_arb->rd_base);
>>> +
>>> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
>>> + "chnls");
>>> + pmic_arb->wr_base = devm_ioremap(dev, res->start,
>>> + resource_size(res));
>>> + if (IS_ERR(pmic_arb->wr_base))
>>> + return PTR_ERR(pmic_arb->wr_base);
>>
>> Could probably make it "devm_platform_get_and_ioremap_resource "
>
> The reason this needs to stay as is is because of reason explained by
> the following comment found in probe:
>
> /*
> * Please don't replace this with devm_platform_ioremap_resource() or
> * devm_ioremap_resource(). These both result in a call to
> * devm_request_mem_region() which prevents multiple mappings of this
> * register address range. SoCs with PMIC arbiter v7 may define two
> * arbiter devices, for the two physical SPMI interfaces, which share
> * some register address ranges (i.e. "core", "obsrvr", and "chnls").
> * Ensure that both devices probe successfully by calling devm_ioremap()
> * which does not result in a devm_request_mem_region() call.
> */
>
> Even though, AFAICT, there is no platform that adds a second node for
> the second bus, currently, in mainline, we should probably allow the
> "legacy" approach to still work.

OK right, let's keep it.

Konrad

2024-02-15 09:32:14

by Dmitry Baryshkov

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On Wed, 14 Feb 2024 at 23:36, Abel Vesa <[email protected]> wrote:
>
> On 24-02-14 22:18:33, Konrad Dybcio wrote:
> > On 14.02.2024 22:13, Abel Vesa wrote:
> > > Rather than setting up the core, obsrv and chnls in probe by using
> > > version specific conditionals, add a dedicated "get_core_resources"
> > > version specific op and move the acquiring in there.
> > >
> > > Signed-off-by: Abel Vesa <[email protected]>
> > > ---
> > > drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
> > > 1 file changed, 78 insertions(+), 33 deletions(-)
> > >
> > > diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
> > > index 23939c0d225f..489556467a4c 100644
> > > --- a/drivers/spmi/spmi-pmic-arb.c
> > > +++ b/drivers/spmi/spmi-pmic-arb.c
> > > @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
> > > */
> > > struct pmic_arb_ver_ops {
> > > const char *ver_str;
> > > + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
> > > int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
> > > int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
> > > /* spmi commands (read_cmd, write_cmd, cmd) functionality */
> > > @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
> > > return 0;
> > > }
> > >
> > > +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
> > > + void __iomem *core)
> > > +{
> > > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > > +
> > > + pmic_arb->wr_base = core;
> > > + pmic_arb->rd_base = core;
> > > +
> > > + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
> > > +
> > > + return 0;
> > > +}
> > > +
> > > static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
> > > {
> > > u32 *mapping_table;
> > > @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
> > > return apid;
> > > }
> > >
> > > +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
> > > +{
> > > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > > + struct device *dev = &pdev->dev;
> > > + struct resource *res;
> > > +
> > > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> >
> > It's no longer indented to deep, no need to keep such aggressive wrapping
> >
>
> The pmic_arb_get_obsrvr_chnls_v2 is used by both:
> pmic_arb_get_core_resources_v2
> pmic_arb_get_core_resources_v7
>
> > > + "obsrvr");
> > > + pmic_arb->rd_base = devm_ioremap(dev, res->start,
> > > + resource_size(res));
> > > + if (IS_ERR(pmic_arb->rd_base))
> > > + return PTR_ERR(pmic_arb->rd_base);
> > > +
> > > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> > > + "chnls");
> > > + pmic_arb->wr_base = devm_ioremap(dev, res->start,
> > > + resource_size(res));
> > > + if (IS_ERR(pmic_arb->wr_base))
> > > + return PTR_ERR(pmic_arb->wr_base);
> >
> > Could probably make it "devm_platform_get_and_ioremap_resource "
>
> The reason this needs to stay as is is because of reason explained by
> the following comment found in probe:
>
> /*
> * Please don't replace this with devm_platform_ioremap_resource() or
> * devm_ioremap_resource(). These both result in a call to
> * devm_request_mem_region() which prevents multiple mappings of this
> * register address range. SoCs with PMIC arbiter v7 may define two
> * arbiter devices, for the two physical SPMI interfaces, which share
> * some register address ranges (i.e. "core", "obsrvr", and "chnls").
> * Ensure that both devices probe successfully by calling devm_ioremap()
> * which does not result in a devm_request_mem_region() call.
> */
>
> Even though, AFAICT, there is no platform that adds a second node for
> the second bus, currently, in mainline, we should probably allow the
> "legacy" approach to still work.

If there were no DT files which used two SPMI devices, I think we
should drop this comment and use existing helpers. We must keep
compatibility with the existing DTs, not with the _possible_ device
trees.

>
> >
> > Konrad



--
With best wishes
Dmitry

2024-02-15 09:51:41

by Abel Vesa

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On 24-02-14 22:44:55, Konrad Dybcio wrote:
> On 14.02.2024 22:36, Abel Vesa wrote:
> > On 24-02-14 22:18:33, Konrad Dybcio wrote:
> >> On 14.02.2024 22:13, Abel Vesa wrote:
> >>> Rather than setting up the core, obsrv and chnls in probe by using
> >>> version specific conditionals, add a dedicated "get_core_resources"
> >>> version specific op and move the acquiring in there.
> >>>
> >>> Signed-off-by: Abel Vesa <[email protected]>
> >>> ---
> >>> drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
> >>> 1 file changed, 78 insertions(+), 33 deletions(-)
> >>>
> >>> diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
> >>> index 23939c0d225f..489556467a4c 100644
> >>> --- a/drivers/spmi/spmi-pmic-arb.c
> >>> +++ b/drivers/spmi/spmi-pmic-arb.c
> >>> @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
> >>> */
> >>> struct pmic_arb_ver_ops {
> >>> const char *ver_str;
> >>> + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
> >>> int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
> >>> int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
> >>> /* spmi commands (read_cmd, write_cmd, cmd) functionality */
> >>> @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
> >>> return 0;
> >>> }
> >>>
> >>> +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
> >>> + void __iomem *core)
> >>> +{
> >>> + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> >>> +
> >>> + pmic_arb->wr_base = core;
> >>> + pmic_arb->rd_base = core;
> >>> +
> >>> + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
> >>> +
> >>> + return 0;
> >>> +}
> >>> +
> >>> static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
> >>> {
> >>> u32 *mapping_table;
> >>> @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
> >>> return apid;
> >>> }
> >>>
> >>> +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
> >>> +{
> >>> + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> >>> + struct device *dev = &pdev->dev;
> >>> + struct resource *res;
> >>> +
> >>> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> >>
> >> It's no longer indented to deep, no need to keep such aggressive wrapping
> >>
> >
> > The pmic_arb_get_obsrvr_chnls_v2 is used by both:
> > pmic_arb_get_core_resources_v2
> > pmic_arb_get_core_resources_v7
>
> I meant line wrapping

Oh, ok. Will do.

>
> >
> >>> + "obsrvr");
> >>> + pmic_arb->rd_base = devm_ioremap(dev, res->start,
> >>> + resource_size(res));
> >>> + if (IS_ERR(pmic_arb->rd_base))
> >>> + return PTR_ERR(pmic_arb->rd_base);
> >>> +
> >>> + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> >>> + "chnls");
> >>> + pmic_arb->wr_base = devm_ioremap(dev, res->start,
> >>> + resource_size(res));
> >>> + if (IS_ERR(pmic_arb->wr_base))
> >>> + return PTR_ERR(pmic_arb->wr_base);
> >>
> >> Could probably make it "devm_platform_get_and_ioremap_resource "
> >
> > The reason this needs to stay as is is because of reason explained by
> > the following comment found in probe:
> >
> > /*
> > * Please don't replace this with devm_platform_ioremap_resource() or
> > * devm_ioremap_resource(). These both result in a call to
> > * devm_request_mem_region() which prevents multiple mappings of this
> > * register address range. SoCs with PMIC arbiter v7 may define two
> > * arbiter devices, for the two physical SPMI interfaces, which share
> > * some register address ranges (i.e. "core", "obsrvr", and "chnls").
> > * Ensure that both devices probe successfully by calling devm_ioremap()
> > * which does not result in a devm_request_mem_region() call.
> > */
> >
> > Even though, AFAICT, there is no platform that adds a second node for
> > the second bus, currently, in mainline, we should probably allow the
> > "legacy" approach to still work.
>
> OK right, let's keep it.
>
> Konrad

2024-02-15 14:08:37

by Abel Vesa

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On 24-02-15 11:30:23, Dmitry Baryshkov wrote:
> On Wed, 14 Feb 2024 at 23:36, Abel Vesa <[email protected]> wrote:
> >
> > On 24-02-14 22:18:33, Konrad Dybcio wrote:
> > > On 14.02.2024 22:13, Abel Vesa wrote:
> > > > Rather than setting up the core, obsrv and chnls in probe by using
> > > > version specific conditionals, add a dedicated "get_core_resources"
> > > > version specific op and move the acquiring in there.
> > > >
> > > > Signed-off-by: Abel Vesa <[email protected]>
> > > > ---
> > > > drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
> > > > 1 file changed, 78 insertions(+), 33 deletions(-)
> > > >
> > > > diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
> > > > index 23939c0d225f..489556467a4c 100644
> > > > --- a/drivers/spmi/spmi-pmic-arb.c
> > > > +++ b/drivers/spmi/spmi-pmic-arb.c
> > > > @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
> > > > */
> > > > struct pmic_arb_ver_ops {
> > > > const char *ver_str;
> > > > + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
> > > > int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
> > > > int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
> > > > /* spmi commands (read_cmd, write_cmd, cmd) functionality */
> > > > @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
> > > > return 0;
> > > > }
> > > >
> > > > +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
> > > > + void __iomem *core)
> > > > +{
> > > > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > > > +
> > > > + pmic_arb->wr_base = core;
> > > > + pmic_arb->rd_base = core;
> > > > +
> > > > + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
> > > > +
> > > > + return 0;
> > > > +}
> > > > +
> > > > static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
> > > > {
> > > > u32 *mapping_table;
> > > > @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
> > > > return apid;
> > > > }
> > > >
> > > > +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
> > > > +{
> > > > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > > > + struct device *dev = &pdev->dev;
> > > > + struct resource *res;
> > > > +
> > > > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> > >
> > > It's no longer indented to deep, no need to keep such aggressive wrapping
> > >
> >
> > The pmic_arb_get_obsrvr_chnls_v2 is used by both:
> > pmic_arb_get_core_resources_v2
> > pmic_arb_get_core_resources_v7
> >
> > > > + "obsrvr");
> > > > + pmic_arb->rd_base = devm_ioremap(dev, res->start,
> > > > + resource_size(res));
> > > > + if (IS_ERR(pmic_arb->rd_base))
> > > > + return PTR_ERR(pmic_arb->rd_base);
> > > > +
> > > > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> > > > + "chnls");
> > > > + pmic_arb->wr_base = devm_ioremap(dev, res->start,
> > > > + resource_size(res));
> > > > + if (IS_ERR(pmic_arb->wr_base))
> > > > + return PTR_ERR(pmic_arb->wr_base);
> > >
> > > Could probably make it "devm_platform_get_and_ioremap_resource "
> >
> > The reason this needs to stay as is is because of reason explained by
> > the following comment found in probe:
> >
> > /*
> > * Please don't replace this with devm_platform_ioremap_resource() or
> > * devm_ioremap_resource(). These both result in a call to
> > * devm_request_mem_region() which prevents multiple mappings of this
> > * register address range. SoCs with PMIC arbiter v7 may define two
> > * arbiter devices, for the two physical SPMI interfaces, which share
> > * some register address ranges (i.e. "core", "obsrvr", and "chnls").
> > * Ensure that both devices probe successfully by calling devm_ioremap()
> > * which does not result in a devm_request_mem_region() call.
> > */
> >
> > Even though, AFAICT, there is no platform that adds a second node for
> > the second bus, currently, in mainline, we should probably allow the
> > "legacy" approach to still work.
>
> If there were no DT files which used two SPMI devices, I think we
> should drop this comment and use existing helpers. We must keep
> compatibility with the existing DTs, not with the _possible_ device
> trees.

Sure.

Should I drop the qcom,bus-id from the driver as well? It is optional
after all.

>
> >
> > >
> > > Konrad
>
>
>
> --
> With best wishes
> Dmitry

2024-02-15 15:09:00

by Dmitry Baryshkov

[permalink] [raw]
Subject: Re: [PATCH RFC v3 3/4] spmi: pmic-arb: Make core resources acquiring a version operation

On Thu, 15 Feb 2024 at 15:32, Abel Vesa <[email protected]> wrote:
>
> On 24-02-15 11:30:23, Dmitry Baryshkov wrote:
> > On Wed, 14 Feb 2024 at 23:36, Abel Vesa <[email protected]> wrote:
> > >
> > > On 24-02-14 22:18:33, Konrad Dybcio wrote:
> > > > On 14.02.2024 22:13, Abel Vesa wrote:
> > > > > Rather than setting up the core, obsrv and chnls in probe by using
> > > > > version specific conditionals, add a dedicated "get_core_resources"
> > > > > version specific op and move the acquiring in there.
> > > > >
> > > > > Signed-off-by: Abel Vesa <[email protected]>
> > > > > ---
> > > > > drivers/spmi/spmi-pmic-arb.c | 111 ++++++++++++++++++++++++++++++-------------
> > > > > 1 file changed, 78 insertions(+), 33 deletions(-)
> > > > >
> > > > > diff --git a/drivers/spmi/spmi-pmic-arb.c b/drivers/spmi/spmi-pmic-arb.c
> > > > > index 23939c0d225f..489556467a4c 100644
> > > > > --- a/drivers/spmi/spmi-pmic-arb.c
> > > > > +++ b/drivers/spmi/spmi-pmic-arb.c
> > > > > @@ -203,6 +203,7 @@ struct spmi_pmic_arb {
> > > > > */
> > > > > struct pmic_arb_ver_ops {
> > > > > const char *ver_str;
> > > > > + int (*get_core_resources)(struct platform_device *pdev, void __iomem *core);
> > > > > int (*init_apid)(struct spmi_pmic_arb *pmic_arb, int index);
> > > > > int (*ppid_to_apid)(struct spmi_pmic_arb *pmic_arb, u16 ppid);
> > > > > /* spmi commands (read_cmd, write_cmd, cmd) functionality */
> > > > > @@ -956,6 +957,19 @@ static int pmic_arb_init_apid_min_max(struct spmi_pmic_arb *pmic_arb)
> > > > > return 0;
> > > > > }
> > > > >
> > > > > +static int pmic_arb_get_core_resources_v1(struct platform_device *pdev,
> > > > > + void __iomem *core)
> > > > > +{
> > > > > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > > > > +
> > > > > + pmic_arb->wr_base = core;
> > > > > + pmic_arb->rd_base = core;
> > > > > +
> > > > > + pmic_arb->max_periphs = PMIC_ARB_MAX_PERIPHS;
> > > > > +
> > > > > + return 0;
> > > > > +}
> > > > > +
> > > > > static int pmic_arb_init_apid_v1(struct spmi_pmic_arb *pmic_arb, int index)
> > > > > {
> > > > > u32 *mapping_table;
> > > > > @@ -1063,6 +1077,41 @@ static u16 pmic_arb_find_apid(struct spmi_pmic_arb *pmic_arb, u16 ppid)
> > > > > return apid;
> > > > > }
> > > > >
> > > > > +static int pmic_arb_get_obsrvr_chnls_v2(struct platform_device *pdev)
> > > > > +{
> > > > > + struct spmi_pmic_arb *pmic_arb = platform_get_drvdata(pdev);
> > > > > + struct device *dev = &pdev->dev;
> > > > > + struct resource *res;
> > > > > +
> > > > > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> > > >
> > > > It's no longer indented to deep, no need to keep such aggressive wrapping
> > > >
> > >
> > > The pmic_arb_get_obsrvr_chnls_v2 is used by both:
> > > pmic_arb_get_core_resources_v2
> > > pmic_arb_get_core_resources_v7
> > >
> > > > > + "obsrvr");
> > > > > + pmic_arb->rd_base = devm_ioremap(dev, res->start,
> > > > > + resource_size(res));
> > > > > + if (IS_ERR(pmic_arb->rd_base))
> > > > > + return PTR_ERR(pmic_arb->rd_base);
> > > > > +
> > > > > + res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> > > > > + "chnls");
> > > > > + pmic_arb->wr_base = devm_ioremap(dev, res->start,
> > > > > + resource_size(res));
> > > > > + if (IS_ERR(pmic_arb->wr_base))
> > > > > + return PTR_ERR(pmic_arb->wr_base);
> > > >
> > > > Could probably make it "devm_platform_get_and_ioremap_resource "
> > >
> > > The reason this needs to stay as is is because of reason explained by
> > > the following comment found in probe:
> > >
> > > /*
> > > * Please don't replace this with devm_platform_ioremap_resource() or
> > > * devm_ioremap_resource(). These both result in a call to
> > > * devm_request_mem_region() which prevents multiple mappings of this
> > > * register address range. SoCs with PMIC arbiter v7 may define two
> > > * arbiter devices, for the two physical SPMI interfaces, which share
> > > * some register address ranges (i.e. "core", "obsrvr", and "chnls").
> > > * Ensure that both devices probe successfully by calling devm_ioremap()
> > > * which does not result in a devm_request_mem_region() call.
> > > */
> > >
> > > Even though, AFAICT, there is no platform that adds a second node for
> > > the second bus, currently, in mainline, we should probably allow the
> > > "legacy" approach to still work.
> >
> > If there were no DT files which used two SPMI devices, I think we
> > should drop this comment and use existing helpers. We must keep
> > compatibility with the existing DTs, not with the _possible_ device
> > trees.
>
> Sure.
>
> Should I drop the qcom,bus-id from the driver as well? It is optional
> after all.

I think so. Let's drop it completely. And for the new sub-devices you
perfectly know the bus ID.

--
With best wishes
Dmitry