[U-Boot] [PATCH] tegra: add Colibri T20 board support

[Lucas Stach]

Am Donnerstag, den 27.09.2012, 16:49 -0600 schrieb Stephen Warren:
> On 09/27/2012 03:52 PM, Lucas Stach wrote:
> > This adds board support for the Toradex Colibri T20 module.
> > 
> > Working functions:
> > - MMC boot
> > - USB boot
> > - Network
> > - NAND environment
> 
> > diff --git a/board/toradex/colibri_t20/colibri_t20.c b/board/toradex/colibri_t20/colibri_t20.c
> 
> > +#ifdef CONFIG_USB_EHCI_TEGRA
> > +void pin_mux_usb(void)
> > +{
> > +	/* USB 1 aka Tegra USB port 3 */
> > +	pinmux_tristate_disable(PINGRP_SPIG);
> > +
> > +	/* USB 3 aka Tegra USB port 2 */
> 
> Re: the comments:
> 
> The only reason other board files have HW port USB3 listed as DT alias
> usb0 is because HW port USB3 was the most useful port on those boards,
> and the Tegra USB driver only supported the first listed HW port. Now
> that the Tegra USB driver supports any/all of the HW ports, you could
> just arrange the aliases in DT so that HW USB1 == user-visible USB1, etc.
> 
I arranged them in that way so they match best with the connector label
on the carrier board.

> > +#ifdef CONFIG_TEGRA_NAND
> > +void pin_mux_nand(void)
> > +{
> > +	funcmux_select(PERIPH_ID_NDFLASH, FUNCMUX_NDFLASH_KBC_8_BIT);
> 
> I don't recall a patch to implement FUNCMUX_NDFLASH_KBC_8_BIT. Maybe I
> just missed it?
> 
No, you are right I missed to send it out. Will follow in a sec.

> > diff --git a/board/toradex/dts/tegra20-colibri.dts b/board/toradex/dts/tegra20-colibri.dts
> 
> This filename should exactly match the name in boards.cfg, so
> tegra20-colibri_t20.dts. That said, if Colibri is always Tegra20, I'd
> rather see the boards.cfg entry renamed to plain "colibri".
> 
I'll rename the dts file. Colibri is a family name and T30 hardware will
follow in the next few weeks. Not sure if I will get one at my hands to
do the upstream work.

> > +};
> > \ No newline at end of file
> 
> That's probably unintended?
> 
Yep.

> > diff --git a/include/configs/colibri_t20.h b/include/configs/colibri_t20.h
> 
> > +/* Environment in NAND, 64K is a bit excessive but erase block is 256K anyway */
> > +#define CONFIG_ENV_IS_IN_NAND
> > +#define CONFIG_ENV_OFFSET              (SZ_2M)
> > +#undef CONFIG_ENV_SIZE /* undef size from tegra20-common.h */
> > +#define CONFIG_ENV_SIZE                (SZ_64K)
> 
> Any particular reason to change the environment size? All the other
> boards that have eMMC put the environment at the end of the second eMMC
> boot-block even when NAND is also available. Is there a particular
> reason not to do so for Colibri too? Does Colibri boot (BCT+bootloader
> image) from eMMC or NAND?
> 
Colibri has no eMMC, it boots from NAND.