[U-Boot] [PATCH 2/9] Tegra: T30: Add AVP (arm720t) files

[Tom Warren]

Stephen,

On Thu, Sep 13, 2012 at 1:02 PM, Stephen Warren <swarren at wwwdotorg.org> wrote:
> On 09/12/2012 04:10 PM, Tom Warren wrote:
>
> Patch descriptions would be nice.

Sure, sorry. Not sure how much more info I can add beyond what's in
the commit msg, though, at least for this patch.

>
>> diff --git a/arch/arm/cpu/arm720t/tegra30/cpu.c b/arch/arm/cpu/arm720t/tegra30/cpu.c
>
> There's quite a bit of Tegra20-support in this file. Can this file be
> shared with Tegra20 rather than forked and enhanced?
>
>> +/* Returns 1 if the current CPU executing is a Cortex-A9, else 0 */
>> +int cpu_is_cortexa9(void)
>> +{
>> +     u32 id = readb(NV_PA_PG_UP_BASE + PG_UP_TAG_0);
>> +     return id == (PG_UP_TAG_0_PID_CPU & 0xff);
>> +}
>
> Hmm. Given this is support for the AVP/COP running SPL, shouldn't this
> always be true? I thought Allen's SPL patches had cleaned this up.

Copied from tegra20/cpu.c - didn't notice it never gets called (same
for Tegra20). So it's vestigial and can be removed.

>
>> +static void enable_cpu_power_rail(void)
>> +{
>> +     struct pmc_ctlr *pmc = (struct pmc_ctlr *)NV_PA_PMC_BASE;
>> +     u32 reg;
>> +
>> +     debug("enable_cpu_power_rail entry\n");
>> +     reg = readl(&pmc->pmc_cntrl);
>> +     reg |= CPUPWRREQ_OE;
>> +     writel(reg, &pmc->pmc_cntrl);
>> +
>> +     /*
>> +      * Pulse PWRREQ via I2C.  We need to find out what this is
>> +      * doing, tidy up the code and maybe find a better place for it.
>> +      */
>> +     tegra_i2c_ll_write_addr(0x005a, 0x0002);
>> +     tegra_i2c_ll_write_data(0x2328, 0x0a02);
>> +     udelay(1000);
>> +     tegra_i2c_ll_write_data(0x0127, 0x0a02);
>> +     udelay(10 * 1000);
>
> Those functions access the DVC I2C controller's register space, so
> presumably they're doing I2C accesses. Not all boards use the same PMIC,
> so it seems like we really do need to factor this out.

It's in the original internal T30 repo for Cardhu, with a comment from
Simon Glass that I edited somewhat. I haven't tried removing it to see
if the board still boots.

>
>> +     /*
>> +      * The TI PMU65861C needs a 3.75ms delay between enabling
>> +      * the power rail and enabling the CPU clock.  This delay
>> +      * between SM1EN and SM1 is for switching time + the ramp
>> +      * up of the voltage to the CPU (VDD_CPU from PMU). We use 0xf00 as
>> +      * is is ARM-friendly (can fit in a single ARMv4T mov immmediate
>> +      * instruction).
>> +      */
>> +     udelay(3840);
>
> The Cardhu board at least does not use the TPS65861. At the very least
> the comment isn't quite right. Is this code needed?
>

No idea - it's in our internal T30 bringup repo, as well as Simon's. I
can try removing it and see if we power up consistently.

>> diff --git a/arch/arm/cpu/arm720t/tegra30/spl.c b/arch/arm/cpu/arm720t/tegra30/spl.c
>
>> +void board_init_f(ulong dummy)
>> +{
>> +     board_init_uart_f();
>> +
>> +     /* Initialize periph GPIOs */
>> +#ifdef CONFIG_SPI_UART_SWITCH
>> +     gpio_early_init_uart();
>> +#else
>> +     gpio_config_uart();
>> +#endif
>
> Didn't we have patches to get rid of that mess and just use the same
> function consistently across all boards, or was that only discussed and
> never actually implemented?

Someone on the list talked about a cleanup (Lucas? Thierry?), but,
AFAIK, that never happened, or I would have put it in tegra/next. At
the very least, it's not needed in Tegra30/spl.c, so I'll remove
it/clean it up.

Thanks,

Tom