[U-Boot] [PATCH v2 7/7] Tegra30: Add/enable Cardhu build (T30 reference board)

[Tom Warren]

Stephen,

On Mon, Dec 3, 2012 at 5:53 PM, Stephen Warren <swarren at wwwdotorg.org> wrote:
> On 12/03/2012 04:45 PM, Tom Warren wrote:
>> This build is stripped down. It boots to the command prompt.
>> GPIO is the only peripheral supported. Others TBD.
>>
>> include/configs/tegra-common.h now holds common config options
>> for Tegra SoCs.
>
>> diff --git a/board/nvidia/cardhu/cardhu.c b/board/nvidia/cardhu/cardhu.c
>
>> +void pinmux_init(void)
>> +{
>> +     pinmux_config_table(tegra3_pinmux_common,
>> +             ARRAY_SIZE(tegra3_pinmux_common));
>> +
>> +     pinmux_config_table(unused_pins_lowpower,
>> +             ARRAY_SIZE(unused_pins_lowpower));
>> +}
>
> Oh, I see. Is any of the pinmux configuration actually needed for just
> the UART to work? I suppose there's not much harm just setting up the
> whole configuration from the start rather than adding it piece-meal.

Yep, see previous reply for an explanation of why it's done in one shot.

>
>> diff --git a/board/nvidia/cardhu/pinmux-config-cardhu.h b/board/nvidia/cardhu/pinmux-config-cardhu.h
>
>> +static struct pingroup_config tegra3_pinmux_common[] = {
>
>> +#ifdef CONFIG_SND_HDA_CODEC_REALTEK
>> +     DEFAULT_PINMUX(SPDIF_IN, DAP2, DOWN, NORMAL, INPUT),
>> +#else
>> +     DEFAULT_PINMUX(SPDIF_IN, SPDIF, NORMAL, NORMAL, INPUT),
>> +#endif
>
> I'm pretty sure that define will never be enabled; it was some temporary
> test thing that should be removed from our downstream kernel, which is
> where I assume you got this table from.

No, this came from one of our internal T30 U-Boot branches. Not sure
where the OD (original dev) got it from.

I'll remove the REALTEK parts.

>
>> diff --git a/include/configs/cardhu.h b/include/configs/cardhu.h
>
>> diff --git a/include/configs/tegra-common.h b/include/configs/tegra-common.h
>
>> +#define CONFIG_SYS_CPU_OSC_FREQUENCY 1000000 /* Set CPU clock to 1GHz */
>
> Is that correct for Tegra30?

It was originally in tegra20-common.h, and was moved into
tegra-common.h in a batch with other thought-to-be-non-HW-specific
config defines.

Looking at the code, though, I don't see it used anywhere. It can be
removed, either in V3 of this patchset or in a clean-up round of
patches later once T30 is in.

>
>> +#define V_NS16550_CLK                        216000000       /* 216MHz (pllp_out0) */
>
> I thought PLL_P ran at 408MHz on Tegra30? The kernel certainly sets it
> up that way.

See my previous reply. In the internal U-Boot repo I ported from, PLLP
was initially set to 216MHz, then sped up to 408MHz. When this first
round of patches is in, I can address going to 408MHz first thing.

>
> In this patch, you also should move MEM_LAYOUT_ENV_SETTINGS from
> tegra-common-post.h into tegraNN-common.h, or something like that.

I can do that, or we can wait until those settings are going to be
used, i.e. once I've added peripheral support so a kernel can be found
and loaded and executed.

Tom