[PATCH 0/9] ARM: Kirkwood: Convert to pinctrl

[Jamie Lentin]

On Sun, 28 Oct 2012, Andrew Lunn wrote:

> On Wed, Oct 24, 2012 at 04:53:45PM +0200, Andrew Lunn wrote:
>> This patchset converts all DT kirkwood boards to using pinctrl.
>
> Thanks for the feedback everybody gave.
>
> Progress has been made:
>
> The bug causing the lockups has been found, fixed and passed upstream.
>
> I factored out the repeated pinmux definitions for sata0, sata1, spi,
> twsi0, uart0, uart1.
>
> I fixed a few types reported by testers.
>
> I found a solution to many of the gpio_request() failures. The ones
> for controlling power to SATA or USB i've replaced with a regulator in
> DT. The regulator will grab the GPIO, set it to output, and driver it
> high/low as required. I just hope we don't have an ordering issue.
> The regulator framework seems to load quite early and probes all the
> regulators. This fails, because pinctrl has not yet loaded, so the
> gpio framework returns EPROBE_DEFER. This is not fatal to the
> regulator framework, it tries again later. But this later is after
> most of the rest of the drivers have loaded. So we might get into
> issues where uboot has not enabled power to USB/SATA, the driver is
> probed and fails because of the lack of power, and once things of gone
> wrong, we turn the power on.
>
> I still expect problems with GPIOs used for power off. I've not yet
> touched them. Again, a regulator makes sense, but there is no 'out of
> the box' regulator type or property which adds itself to pm_power_off.
>
> The last problem is:
>
>        /* Set NAS to turn back on after a power failure */
>        dnskw_gpio_register(37, "dnskw:power:recover", 1);
>
> I think its a bit of a stretch calling this a regulator.

Definitely.

> Since the current patchset is 18 patches, plus a few other bug fixes
> and dependencies, i'm not going to post them all here. I think we are
> still at the smoke/no-smoke stage of testing. Please pull from:
>
> git://github.com/lunn/linux.git v3.7-rc2-pinctrl-v2
>
> and let me know if things work, what typo's need fixing, where the
> smoke comes out, etc.

No smoke came out, the SATA ports were both powered up. Beyond what's 
already known about, no problems.

>
>      Thanks
> 	Andrew
>

-- 
Jamie Lentin