Hi,
On 4/15/2017 4:03 AM, Adam Ford wrote:
> On Wed, Apr 12, 2017 at 11:24 PM, Vignesh R <vigne...@ti.com> wrote:
>>
>>
>> On Wednesday 12 April 2017 06:08 PM, Adam Ford wrote:
>>> This patch fixes and issue where DM_MMC adds a 0x100 byte offset to the
>>> base register. This is necessary for AM33xx, OMAP4+ and newer devices, but
>>> it is not necessary for OMAP34XX boards.
>>>
>>> This patch will now only apply the x100 byte offset correction if DM_MMC
>>> is enabled and the device is not OMAP34XX.
>>>
>>
>> Although this is an easy solution, I think its better to handle this w/o
>> ifdefs and instead using compatible string. See how omap3 base address
>> is handled in Linux omap_hsmmc driver (drivers/mmc/host/omap_hsmmc.c).
>>
>
> I have the basic ideal, but I am not sure how extract the
> udevice_id->data from the corresponding omap_hsmmc_ids
>
> Like Linux, I was going to create a structure
>
> static const struct omap_mmc_of_data omap4_mmc_of_data = {
> .reg_offset = 0x100,
> };
>
> and attach it to the necessary .compatible parts.
> {
> .compatible = "ti,omap4-hsmmc",
> .data = &omap4_mmc_of_data
> }
>
> When ."compatible" matches, I need to extract the corresponding
> ->data, but I don't what what the best driver example to follow would
> be.
>
> Can you point me to to good example or give me some suggestions on how
> to extract it? This device tree decoding stuff is new to me, but I
> want to be helpful too.
>
You could look at drivers/spi/omap3_spi.c in U-Boot tree, that handles
similar problem.
>
>> Also, please send rest of the series in reply to Patch 0/7,
>> git send-email should automatically do that for you.
>
> I'll read the docs again on submitting patch series, and try to figure
> it out once I address the MMC offset thing. This is my first patch
> series.
>
Thanks for the patch!
Maybe, this link could help: https://kernelnewbies.org/FirstKernelPatch
Regards
Vignesh
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
https://lists.denx.de/listinfo/u-boot
