[v2,0/3] Get rid of big array from imx pinctrl driver

Commit Message

Shawn Guo April 8, 2013, 6:09 a.m. UTC

On Sun, Apr 07, 2013 at 11:32:45AM -0300, Fabio Estevam wrote:
> On Sun, Apr 7, 2013 at 11:26 AM, Shawn Guo <[e-mail]> wrote:
> 
> > That's expected, as the pin ID gets re-numbered based on offset of
> > mux/conf register, so that we can determine the pin ID at runtime and
> > avoid encoding it in device tree.
> 
> but debug messages seems buggy now:
> 
Yes, you are right.  The change attached below should fix the problem.
I am fixing it up for all imx pinctrl drivers on imx/dt branch.

> Prior to the patch we had:
> 
> imx6q-pinctrl 20e0000.iomuxc: pinconf set pin MX6Q_PAD_EIM_D21
> 
> and now we have:
> 
> imx6q-pinctrl 20e0000.iomuxc: pinconf set pin MX6Q_PAD_EIM_A17
> 
> MX6Q_PAD_EIM_A17 has nothing to do with I2C, so we should print
> MX6Q_PAD_EIM_A17 when we are actually configuring MX6Q_PAD_EIM_D21.
> 
> Also, where does the 19 offset in the  imx6q_pads enum come from?

We choose to have the pin ID simply calculated from register offset.
For example, the mux register offset for MX6Q_PAD_SD2_DAT1 is 0x4c, and
the ID would be 0x4c / 4 = 19.

Shawn

--8<-----

Comments

Linus Walleij April 9, 2013, 8:36 a.m. UTC | #1

On Mon, Apr 8, 2013 at 8:09 AM, Shawn Guo <[e-mail]> wrote:

> --8<-----
>
> diff --git a/drivers/pinctrl/pinctrl-imx6q.c b/drivers/pinctrl/pinctrl-imx6q.c
> index f00f532..76dd9c4 100644
> --- a/drivers/pinctrl/pinctrl-imx6q.c
> +++ b/drivers/pinctrl/pinctrl-imx6q.c

Acked-by. I trust you to fix up and merge the outcome of this Shawn...

Yours,
Linus Walleij

Shawn Guo April 9, 2013, 11:57 a.m. UTC | #2

On Tue, Apr 09, 2013 at 10:36:59AM +0200, Linus Walleij wrote:
> On Mon, Apr 8, 2013 at 8:09 AM, Shawn Guo <[e-mail]> wrote:
> 
> > --8<-----
> >
> > diff --git a/drivers/pinctrl/pinctrl-imx6q.c b/drivers/pinctrl/pinctrl-imx6q.c
> > index f00f532..76dd9c4 100644
> > --- a/drivers/pinctrl/pinctrl-imx6q.c
> > +++ b/drivers/pinctrl/pinctrl-imx6q.c
> 
> Acked-by. I trust you to fix up and merge the outcome of this Shawn...

Thanks, Linus.  Thanks to Fabio too for noticing the bugs.

Shawn

Patch

diff --git a/drivers/pinctrl/pinctrl-imx6q.c b/drivers/pinctrl/pinctrl-imx6q.c
index f00f532..76dd9c4 100644
--- a/drivers/pinctrl/pinctrl-imx6q.c
+++ b/drivers/pinctrl/pinctrl-imx6q.c
@@ -23,6 +23,25 @@ 
 #include "pinctrl-imx.h"
 
 enum imx6q_pads {
+	MX6Q_PAD_RESERVE0 = 0,
+	MX6Q_PAD_RESERVE1 = 1,
+	MX6Q_PAD_RESERVE2 = 2,
+	MX6Q_PAD_RESERVE3 = 3,
+	MX6Q_PAD_RESERVE4 = 4,
+	MX6Q_PAD_RESERVE5 = 5,
+	MX6Q_PAD_RESERVE6 = 6,
+	MX6Q_PAD_RESERVE7 = 7,
+	MX6Q_PAD_RESERVE8 = 8,
+	MX6Q_PAD_RESERVE9 = 9,
+	MX6Q_PAD_RESERVE10 = 10,
+	MX6Q_PAD_RESERVE11 = 11,
+	MX6Q_PAD_RESERVE12 = 12,
+	MX6Q_PAD_RESERVE13 = 13,
+	MX6Q_PAD_RESERVE14 = 14,
+	MX6Q_PAD_RESERVE15 = 15,
+	MX6Q_PAD_RESERVE16 = 16,
+	MX6Q_PAD_RESERVE17 = 17,
+	MX6Q_PAD_RESERVE18 = 18,
 	MX6Q_PAD_SD2_DAT1 = 19,
 	MX6Q_PAD_SD2_DAT2 = 20,
 	MX6Q_PAD_SD2_DAT0 = 21,
@@ -224,6 +243,25 @@  enum imx6q_pads {
 
 /* Pad names for the pinmux subsystem */
 static const struct pinctrl_pin_desc imx6q_pinctrl_pads[] = {
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE0),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE1),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE2),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE3),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE4),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE5),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE6),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE7),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE8),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE9),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE10),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE11),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE12),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE13),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE14),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE15),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE16),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE17),
+	IMX_PINCTRL_PIN(MX6Q_PAD_RESERVE18),
 	IMX_PINCTRL_PIN(MX6Q_PAD_SD2_DAT1),
 	IMX_PINCTRL_PIN(MX6Q_PAD_SD2_DAT2),
 	IMX_PINCTRL_PIN(MX6Q_PAD_SD2_DAT0),