[U-Boot] [PATCH] tegra: pinmux: fix FUNCMUX_NDFLASH_KBC_8_BIT

[Lucas Stach <dev@lynxeye.de>]

Am Dienstag, den 22.01.2013, 09:24 -0700 schrieb Stephen Warren:
> On 01/21/2013 05:20 PM, Lucas Stach wrote:
> > Even the 8bit case needs KBCB configured, as pin D7 is located in this
> > pingroup. Also pingroup ATC seems to come out of reset with config set
> > to NAND, so we need to explictly configure some other function to this
> > group in order to avoid clashing settings.
> 
> > diff --git a/arch/arm/cpu/tegra20-common/funcmux.c b/arch/arm/cpu/tegra20-common/funcmux.c
> 
> > @@ -266,17 +266,25 @@ int funcmux_select(enum periph_id id, int config)
> >  			break;
> >  		case FUNCMUX_NDFLASH_KBC_8_BIT:
> ...
> > +			/*
> > +			 * configure pingroup ATC to something unrelated to
> > +			 * avoid ATC overriding KBC
> > +			 */
> > +			pinmux_set_func(PINGRP_ATC, PMUX_FUNC_GMI);
> > +
> 
> This gets a bit dangerous; what if pingroup ATC was already configured
> for some function other than NAND or GMI? This code will then break that
> setting. I would suggest one of the following alternatives:
> 
> 1) Use the new pinmux_avoid_func() function implemented in the patch
> that I just sent.
> 
> 2) Move Tegra20 over to the new board-wide pinmux style that Tegra30
> uses, where the entire pinmux is initialized in one shot. This will
> completely avoid any kind of uninitialized pinmux settings, and to some
> extent is the only sensible thing to do on a device like Tegra which has
> the potential for conflicts like this patch tries to avoid.
> 
I'll take a look on how much work it is to implement option #2. If it
isn't too much and I find some time in this U-Boot release cycle, I'm
very much inclined to do this the ultimately right way.

Regards,
Lucas