Re: [PATCH 4/4] usb: phy: omap-usb2: enable 960Mhz clock for omap5

[Felipe Balbi <balbi@ti.com>]

[-- Attachment #1: Type: text/plain, Size: 2858 bytes --]

On Wed, Sep 19, 2012 at 04:45:01PM +0200, Marc Kleine-Budde wrote:
> On 09/19/2012 01:56 PM, Felipe Balbi wrote:
> > On Wed, Sep 19, 2012 at 05:00:29PM +0530, Kishon Vijay Abraham I wrote:
> >> "usb_otg_ss_refclk960m" is needed by usb2 phy present in omap5. For
> >> omap4, the clk_get of this clock will fail since it does not have this
> >> clock.
> >>
> >> Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
> >> ---
> >>  Documentation/devicetree/bindings/usb/usb-phy.txt |    3 +++
> >>  drivers/usb/phy/omap-usb2.c                       |   28 ++++++++++++++++++++-
> >>  2 files changed, 30 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/Documentation/devicetree/bindings/usb/usb-phy.txt b/Documentation/devicetree/bindings/usb/usb-phy.txt
> >> index 7c5fd89..d5626de 100644
> >> --- a/Documentation/devicetree/bindings/usb/usb-phy.txt
> >> +++ b/Documentation/devicetree/bindings/usb/usb-phy.txt
> >> @@ -24,6 +24,9 @@ Required properties:
> >>  add the address of control module phy power register until a driver for
> >>  control module is added
> >>  
> >> +Optional properties:
> >> + - has960mhzclk: should be added if the phy needs 960mhz clock
> >> +
> >>  This is usually a subnode of ocp2scp to which it is connected.
> >>  
> >>  usb3phy@4a084400 {
> >> diff --git a/drivers/usb/phy/omap-usb2.c b/drivers/usb/phy/omap-usb2.c
> >> index d36c282..d6612ba 100644
> >> --- a/drivers/usb/phy/omap-usb2.c
> >> +++ b/drivers/usb/phy/omap-usb2.c
> >> @@ -146,6 +146,7 @@ static int __devinit omap_usb2_probe(struct platform_device *pdev)
> >>  	struct omap_usb			*phy;
> >>  	struct usb_otg			*otg;
> >>  	struct resource			*res;
> >> +	struct device_node		*np = pdev->dev.of_node;
> >>  
> >>  	phy = devm_kzalloc(&pdev->dev, sizeof(*phy), GFP_KERNEL);
> >>  	if (!phy) {
> >> @@ -190,6 +191,15 @@ static int __devinit omap_usb2_probe(struct platform_device *pdev)
> >>  	}
> >>  	clk_prepare(phy->wkupclk);
> >>  
> >> +	if (of_property_read_bool(np, "has960mhzclk")) {
> >> +		phy->optclk = devm_clk_get(phy->dev, "usb_otg_ss_refclk960m");
> >> +		if (IS_ERR(phy->optclk)) {
> >> +			dev_err(&pdev->dev, "unable to get refclk960m\n");
> >> +			return PTR_ERR(phy->optclk);
> >> +		}
> >> +		clk_prepare(phy->optclk);
> >> +	}
> > 
> > instead, can't you just always try to get the clock but ignore the error
> > if it fails ?
> 
> On imx we add no-op dummy clocks, so that the drivers will always find
> all needed clocks. If a clk_get fails it's an error.

It's an error from a clk API point of view, it doesn't mean device can't
work without that optional clock. Drivers are free to treat errors
however they like, even by ignoring it and defaulting to some other
behavior.

Adding dummy clocks just bloats the kernel with useless data IMHO.

my 2 cents.

-- 
balbi

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]