[PATCH v2 0/2] Add power control for pcie-histb driver

[PATCH v2 0/2] Add power control for pcie-histb driver

[Shawn Guo]

The first patch is a small fix which is useful for the second one.
This version basically follows Fabio's suggestion to reimplement the
first version patch, using a regulator that can be backed by GPIO to
control the power to PCIe port.

Shawn Guo (2):
  PCI: histb: fix error path of histb_pcie_host_enable()
  PCI: histb: add an optional regulator for PCIe port power control

 .../bindings/pci/hisilicon-histb-pcie.txt          |  1 +
 drivers/pci/dwc/pcie-histb.c                       | 28 ++++++++++++++++++----
 2 files changed, 25 insertions(+), 4 deletions(-)

-- 
1.9.1

[PATCH v2 1/2] PCI: histb: fix error path of histb_pcie_host_enable()

[Shawn Guo]

If clk_prepare_enable() call fails on a particular clock, we should not
call clk_disable_unprepare() on this clock, but on the clocks that
succeed from clk_prepare_enable() previously.

Signed-off-by: Shawn Guo <shawn.guo-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
---
 drivers/pci/dwc/pcie-histb.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/pci/dwc/pcie-histb.c b/drivers/pci/dwc/pcie-histb.c
index 33b01b734d7d..6395394be5b4 100644
--- a/drivers/pci/dwc/pcie-histb.c
+++ b/drivers/pci/dwc/pcie-histb.c
@@ -279,13 +279,12 @@ static int histb_pcie_host_enable(struct pcie_port *pp)
 	return 0;
 
 err_aux_clk:
-	clk_disable_unprepare(hipcie->aux_clk);
-err_pipe_clk:
 	clk_disable_unprepare(hipcie->pipe_clk);
-err_sys_clk:
+err_pipe_clk:
 	clk_disable_unprepare(hipcie->sys_clk);
-err_bus_clk:
+err_sys_clk:
 	clk_disable_unprepare(hipcie->bus_clk);
+err_bus_clk:
 
 	return ret;
 }
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 2/2] PCI: histb: add an optional regulator for PCIe port power control

[Shawn Guo]

The power supply to PCIe port are often controlled by GPIO on some board
designs.  Let's add an optional regulator which can be backed by GPIO to
control the power.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
---
 .../bindings/pci/hisilicon-histb-pcie.txt           |  1 +
 drivers/pci/dwc/pcie-histb.c                        | 21 +++++++++++++++++++++
 2 files changed, 22 insertions(+)

diff --git a/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt b/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt
index c84bc027930b..f995664e6d7f 100644
--- a/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt
+++ b/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt
@@ -34,6 +34,7 @@ Required properties
 
 Optional properties:
 - reset-gpios: The gpio to generate PCIe PERST# assert and deassert signal.
+- vpcie-supply: Should specify the regulator in charge of PCIe port power.
 - phys: List of phandle and phy mode specifier, should be 0.
 - phy-names: Must be "phy".
 
diff --git a/drivers/pci/dwc/pcie-histb.c b/drivers/pci/dwc/pcie-histb.c
index 6395394be5b4..8eb3028432b3 100644
--- a/drivers/pci/dwc/pcie-histb.c
+++ b/drivers/pci/dwc/pcie-histb.c
@@ -64,6 +64,7 @@ struct histb_pcie {
 	struct reset_control *bus_reset;
 	void __iomem *ctrl;
 	int reset_gpio;
+	struct regulator *vpcie;
 };
 
 static u32 histb_pcie_readl(struct histb_pcie *histb_pcie, u32 reg)
@@ -230,6 +231,9 @@ static void histb_pcie_host_disable(struct histb_pcie *hipcie)
 
 	if (gpio_is_valid(hipcie->reset_gpio))
 		gpio_set_value_cansleep(hipcie->reset_gpio, 0);
+
+	if (hipcie->vpcie)
+		regulator_disable(hipcie->vpcie);
 }
 
 static int histb_pcie_host_enable(struct pcie_port *pp)
@@ -240,6 +244,14 @@ static int histb_pcie_host_enable(struct pcie_port *pp)
 	int ret;
 
 	/* power on PCIe device if have */
+	if (hipcie->vpcie) {
+		ret = regulator_enable(hipcie->vpcie);
+		if (ret) {
+			dev_err(dev, "failed to enable regulator: %d\n", ret);
+			return ret;
+		}
+	}
+
 	if (gpio_is_valid(hipcie->reset_gpio))
 		gpio_set_value_cansleep(hipcie->reset_gpio, 1);
 
@@ -285,6 +297,8 @@ static int histb_pcie_host_enable(struct pcie_port *pp)
 err_sys_clk:
 	clk_disable_unprepare(hipcie->bus_clk);
 err_bus_clk:
+	if (hipcie->vpcie)
+		regulator_disable(hipcie->vpcie);
 
 	return ret;
 }
@@ -334,6 +348,13 @@ static int histb_pcie_probe(struct platform_device *pdev)
 		return PTR_ERR(pci->dbi_base);
 	}
 
+	hipcie->vpcie = devm_regulator_get_optional(dev, "vpcie");
+	if (IS_ERR(hipcie->vpcie)) {
+		if (PTR_ERR(hipcie->vpcie) == -EPROBE_DEFER)
+			return -EPROBE_DEFER;
+		hipcie->vpcie = NULL;
+	}
+
 	hipcie->reset_gpio = of_get_named_gpio_flags(np,
 				"reset-gpios", 0, &of_flags);
 	if (of_flags & OF_GPIO_ACTIVE_LOW)
-- 
1.9.1

Re: [PATCH v2 2/2] PCI: histb: add an optional regulator for PCIe port power control

[Rob Herring]

On Tue, Jan 23, 2018 at 11:45:37PM +0800, Shawn Guo wrote:
> The power supply to PCIe port are often controlled by GPIO on some board
> designs.  Let's add an optional regulator which can be backed by GPIO to
> control the power.
> 
> Signed-off-by: Shawn Guo <shawn.guo-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>
> ---
>  .../bindings/pci/hisilicon-histb-pcie.txt           |  1 +

Acked-by: Rob Herring <robh-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>

>  drivers/pci/dwc/pcie-histb.c                        | 21 +++++++++++++++++++++
>  2 files changed, 22 insertions(+)
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 0/2] Add power control for pcie-histb driver

[Shawn Guo]

On Tue, Jan 23, 2018 at 11:45:35PM +0800, Shawn Guo wrote:
> The first patch is a small fix which is useful for the second one.
> This version basically follows Fabio's suggestion to reimplement the
> first version patch, using a regulator that can be backed by GPIO to
> control the power to PCIe port.
> 
> Shawn Guo (2):
>   PCI: histb: fix error path of histb_pcie_host_enable()
>   PCI: histb: add an optional regulator for PCIe port power control
> 
>  .../bindings/pci/hisilicon-histb-pcie.txt          |  1 +
>  drivers/pci/dwc/pcie-histb.c                       | 28 ++++++++++++++++++----
>  2 files changed, 25 insertions(+), 4 deletions(-)

Hi Bjorn,

Did you get a chance to look at the series?

Shawn

Re: [PATCH v2 0/2] Add power control for pcie-histb driver

[Bjorn Helgaas]

[+cc Lorenzo]

On Thu, Mar 01, 2018 at 09:22:46AM +0800, Shawn Guo wrote:
> On Tue, Jan 23, 2018 at 11:45:35PM +0800, Shawn Guo wrote:
> > The first patch is a small fix which is useful for the second one.
> > This version basically follows Fabio's suggestion to reimplement the
> > first version patch, using a regulator that can be backed by GPIO to
> > control the power to PCIe port.
> > 
> > Shawn Guo (2):
> >   PCI: histb: fix error path of histb_pcie_host_enable()
> >   PCI: histb: add an optional regulator for PCIe port power control
> > 
> >  .../bindings/pci/hisilicon-histb-pcie.txt          |  1 +
> >  drivers/pci/dwc/pcie-histb.c                       | 28 ++++++++++++++++++----
> >  2 files changed, 25 insertions(+), 4 deletions(-)
> 
> Hi Bjorn,
> 
> Did you get a chance to look at the series?

I cc'd Lorenzo, who is now taking care of this area.

If/when you repost this, please incorporate Rob's ack and capitalize
the first word of your summary ("Fix", "Add") to match the drivers/pci
convention.

Also, the original post used "PCI: hisi:", but here you use
"PCI: histb:".  It probably does make sense to distinguish
"hisi" from "histb", but I want it to be a conscious choice
so future patches can be consistent with it.

Bjorn

Re: [PATCH v2 2/2] PCI: histb: add an optional regulator for PCIe port power control

[Bjorn Helgaas]

On Tue, Jan 23, 2018 at 11:45:37PM +0800, Shawn Guo wrote:
> The power supply to PCIe port are often controlled by GPIO on some board
> designs.  Let's add an optional regulator which can be backed by GPIO to
> control the power.
> 
> Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
> ---
>  .../bindings/pci/hisilicon-histb-pcie.txt           |  1 +
>  drivers/pci/dwc/pcie-histb.c                        | 21 +++++++++++++++++++++
>  2 files changed, 22 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt b/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt
> index c84bc027930b..f995664e6d7f 100644
> --- a/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt
> +++ b/Documentation/devicetree/bindings/pci/hisilicon-histb-pcie.txt
> @@ -34,6 +34,7 @@ Required properties
>  
>  Optional properties:
>  - reset-gpios: The gpio to generate PCIe PERST# assert and deassert signal.
> +- vpcie-supply: Should specify the regulator in charge of PCIe port power.

s/Should specify the/The/

>  - phys: List of phandle and phy mode specifier, should be 0.
>  - phy-names: Must be "phy".
>  
> diff --git a/drivers/pci/dwc/pcie-histb.c b/drivers/pci/dwc/pcie-histb.c
> index 6395394be5b4..8eb3028432b3 100644
> --- a/drivers/pci/dwc/pcie-histb.c
> +++ b/drivers/pci/dwc/pcie-histb.c
> @@ -64,6 +64,7 @@ struct histb_pcie {
>  	struct reset_control *bus_reset;
>  	void __iomem *ctrl;
>  	int reset_gpio;
> +	struct regulator *vpcie;
>  };
>  
>  static u32 histb_pcie_readl(struct histb_pcie *histb_pcie, u32 reg)
> @@ -230,6 +231,9 @@ static void histb_pcie_host_disable(struct histb_pcie *hipcie)
>  
>  	if (gpio_is_valid(hipcie->reset_gpio))
>  		gpio_set_value_cansleep(hipcie->reset_gpio, 0);
> +
> +	if (hipcie->vpcie)
> +		regulator_disable(hipcie->vpcie);
>  }
>  
>  static int histb_pcie_host_enable(struct pcie_port *pp)
> @@ -240,6 +244,14 @@ static int histb_pcie_host_enable(struct pcie_port *pp)
>  	int ret;
>  
>  	/* power on PCIe device if have */
> +	if (hipcie->vpcie) {
> +		ret = regulator_enable(hipcie->vpcie);
> +		if (ret) {
> +			dev_err(dev, "failed to enable regulator: %d\n", ret);
> +			return ret;
> +		}
> +	}
> +
>  	if (gpio_is_valid(hipcie->reset_gpio))
>  		gpio_set_value_cansleep(hipcie->reset_gpio, 1);
>  
> @@ -285,6 +297,8 @@ static int histb_pcie_host_enable(struct pcie_port *pp)
>  err_sys_clk:
>  	clk_disable_unprepare(hipcie->bus_clk);
>  err_bus_clk:
> +	if (hipcie->vpcie)
> +		regulator_disable(hipcie->vpcie);
>  
>  	return ret;
>  }
> @@ -334,6 +348,13 @@ static int histb_pcie_probe(struct platform_device *pdev)
>  		return PTR_ERR(pci->dbi_base);
>  	}
>  
> +	hipcie->vpcie = devm_regulator_get_optional(dev, "vpcie");
> +	if (IS_ERR(hipcie->vpcie)) {
> +		if (PTR_ERR(hipcie->vpcie) == -EPROBE_DEFER)
> +			return -EPROBE_DEFER;
> +		hipcie->vpcie = NULL;
> +	}
> +
>  	hipcie->reset_gpio = of_get_named_gpio_flags(np,
>  				"reset-gpios", 0, &of_flags);
>  	if (of_flags & OF_GPIO_ACTIVE_LOW)
> -- 
> 1.9.1
> 

Re: [PATCH v2 0/2] Add power control for pcie-histb driver

[Lorenzo Pieralisi]

On Thu, Mar 01, 2018 at 09:08:22AM -0600, Bjorn Helgaas wrote:
> [+cc Lorenzo]
> 
> On Thu, Mar 01, 2018 at 09:22:46AM +0800, Shawn Guo wrote:
> > On Tue, Jan 23, 2018 at 11:45:35PM +0800, Shawn Guo wrote:
> > > The first patch is a small fix which is useful for the second one.
> > > This version basically follows Fabio's suggestion to reimplement the
> > > first version patch, using a regulator that can be backed by GPIO to
> > > control the power to PCIe port.
> > > 
> > > Shawn Guo (2):
> > >   PCI: histb: fix error path of histb_pcie_host_enable()
> > >   PCI: histb: add an optional regulator for PCIe port power control
> > > 
> > >  .../bindings/pci/hisilicon-histb-pcie.txt          |  1 +
> > >  drivers/pci/dwc/pcie-histb.c                       | 28 ++++++++++++++++++----
> > >  2 files changed, 25 insertions(+), 4 deletions(-)
> > 
> > Hi Bjorn,
> > 
> > Did you get a chance to look at the series?
> 
> I cc'd Lorenzo, who is now taking care of this area.
> 
> If/when you repost this, please incorporate Rob's ack and capitalize
> the first word of your summary ("Fix", "Add") to match the drivers/pci
> convention.
> 
> Also, the original post used "PCI: hisi:", but here you use
> "PCI: histb:".  It probably does make sense to distinguish
> "hisi" from "histb", but I want it to be a conscious choice
> so future patches can be consistent with it.

I suspect that using histb makes more sense but I will wait Shawn's
feedback on this, this series was on my radar anyway but thanks
for the flag up.

Lorenzo

Re: [PATCH v2 0/2] Add power control for pcie-histb driver

[Shawn Guo]

On Thu, Mar 01, 2018 at 06:11:19PM +0000, Lorenzo Pieralisi wrote:
> On Thu, Mar 01, 2018 at 09:08:22AM -0600, Bjorn Helgaas wrote:
> > [+cc Lorenzo]
> > 
> > On Thu, Mar 01, 2018 at 09:22:46AM +0800, Shawn Guo wrote:
> > > On Tue, Jan 23, 2018 at 11:45:35PM +0800, Shawn Guo wrote:
> > > > The first patch is a small fix which is useful for the second one.
> > > > This version basically follows Fabio's suggestion to reimplement the
> > > > first version patch, using a regulator that can be backed by GPIO to
> > > > control the power to PCIe port.
> > > > 
> > > > Shawn Guo (2):
> > > >   PCI: histb: fix error path of histb_pcie_host_enable()
> > > >   PCI: histb: add an optional regulator for PCIe port power control
> > > > 
> > > >  .../bindings/pci/hisilicon-histb-pcie.txt          |  1 +
> > > >  drivers/pci/dwc/pcie-histb.c                       | 28 ++++++++++++++++++----
> > > >  2 files changed, 25 insertions(+), 4 deletions(-)
> > > 
> > > Hi Bjorn,
> > > 
> > > Did you get a chance to look at the series?
> > 
> > I cc'd Lorenzo, who is now taking care of this area.
> > 
> > If/when you repost this, please incorporate Rob's ack and capitalize
> > the first word of your summary ("Fix", "Add") to match the drivers/pci
> > convention.
> > 
> > Also, the original post used "PCI: hisi:", but here you use
> > "PCI: histb:".  It probably does make sense to distinguish
> > "hisi" from "histb", but I want it to be a conscious choice
> > so future patches can be consistent with it.
> 
> I suspect that using histb makes more sense but I will wait Shawn's
> feedback on this,

Yes, we want to distinguish the driver pcie-histb.c from pcie-hisi.c,
and will consistently use "PCI: histb:" for this driver.

> this series was on my radar anyway but thanks
> for the flag up.

Cool.  I will keep you posted on v3.

Thanks, Bjorn, Lorenzo.

Shawn