From mboxrd@z Thu Jan  1 00:00:00 1970
From: Stephen Warren <swarren@wwwdotorg.org>
Date: Thu, 17 Jan 2013 15:23:19 -0700
Subject: [U-Boot] [PATCH 1/7] Tegra114: Add arch-tegra114 include files
In-Reply-To: <CA+m5__+s-6AVTZnFHokG-9nVHWYAyOf=V306grQsB0_OCXHjug@mail.gmail.com>
References: <1358370848-29469-1-git-send-email-twarren@nvidia.com>
	<1358370848-29469-2-git-send-email-twarren@nvidia.com>
	<50F72A9B.80308@wwwdotorg.org>
	<CA+m5__+s-6AVTZnFHokG-9nVHWYAyOf=V306grQsB0_OCXHjug@mail.gmail.com>
Message-ID: <50F879D7.5090003@wwwdotorg.org>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

On 01/17/2013 10:51 AM, Tom Warren wrote:
> Stephen,
> 
> On Wed, Jan 16, 2013 at 3:32 PM, Stephen Warren <swarren@wwwdotorg.org> wrote:
>> On 01/16/2013 02:14 PM, Tom Warren wrote:
>>> Common Tegra files are in arch-tegra, shared between T20/T30/T114.
>>> Tegra114-specific headers are in arch-tegra114. Note that some of
>>> these will be filled in as more T114 support is added (drivers,
>>> WB/LP0 support, etc.).
>>
>>> diff --git a/arch/arm/include/asm/arch-tegra114/gpio.h b/arch/arm/include/asm/arch-tegra114/gpio.h
>>
>> I think the Tegra30 and Tegra114 GPIO controllers are basically
>> identical. Can this file be shared at all?
>>
> 
> gpio.h is identical for T30 and T114, but not for T20.  I've already
> commonized what I could in arch-tegra/gpio.h, but to do the same for
> the gpio_ctlr struct and gpio_pin enums would involve #ifdefs, which
> is moving in the wrong direction, IMO.  If you have a way to share one
> or two files between T20/T30/T114/future SoCs, I'll give it a try.

This may not work well for U-Boot since it uses structs to define the
register layouts rather than simply #defining the various register
addresses, but:

We use the exact same driver code for Tegra20/30/114. The difference in
register layout between Tegra20/30 can be described algorithmically with
the following data:

> struct tegra_gpio_soc_config {
>         u32 bank_stride;
>         u32 upper_offset;
> };
> 
> static struct tegra_gpio_soc_config tegra20_gpio_config = {
>         .bank_stride = 0x80,
>         .upper_offset = 0x800,
> };
> 
> static struct tegra_gpio_soc_config tegra30_gpio_config = {
>         .bank_stride = 0x100,
>         .upper_offset = 0x80,
> };
> 
> static struct of_device_id tegra_gpio_of_match[] = {
>         { .compatible = "nvidia,tegra30-gpio", .data = &tegra30_gpio_config },
>         { .compatible = "nvidia,tegra20-gpio", .data = &tegra20_gpio_config },
>         { },
> };

which feeds into a few register offset #defines:

> #define GPIO_REG(x)             (GPIO_BANK(x) * tegra_gpio_bank_stride + \
...
> #define GPIO_CNF(x)             (GPIO_REG(x) + 0x00)
...
> #define GPIO_MSK_CNF(x)         (GPIO_REG(x) + tegra_gpio_upper_offset + 0x00)

Alternatively, can the Tegra114 GPIO header simply #include the Tegra30
GPIO header and do nothing else, via some relative include or similar?
I'm not sure how gross that'd turn out to be.