[U-Boot] [PATCH v2 5/6] apalis/colibri_t30: move environment location

[Stephen Warren]

On 09/30/2016 05:44 AM, Marcel Ziswiler wrote:
> On Wed, 2016-09-28 at 12:11 -0600, Stephen Warren wrote:
>> On 09/28/2016 03:35 AM, Marcel Ziswiler wrote:
>>>
>>> Now with the config block handling in place move the U-Boot
>>> environment
>>> location before the config block at the end of 1st "boot sector" as
>>> deployed during production using our downstream BSP.
>>>
>>> diff --git a/include/configs/apalis_t30.h b/include/configs/apalis_t30.h
>>>
>>> -/* Environment in eMMC, at the end of 2nd "boot sector" */
>>> +/* Environment in eMMC, before config block at the end of 1st "boot sector" */
>>>  #define CONFIG_ENV_IS_IN_MMC
>>> -#define CONFIG_ENV_OFFSET		(-CONFIG_ENV_SIZE)
>>> +#define CONFIG_ENV_OFFSET		(-CONFIG_ENV_SIZE + \
>>> +					 CONFIG_TDX_CFG_BLOCK_OFFSET)
>> I'm not convinced that ENV_OFFSET actually points where the description
>> says it does. (-CONFIG_ENV_SIZE + CONFIG_TDX_CFG_BLOCK_OFFSET) is (-8192
>> + -512). The original comment states that an ENV_OFFSET of -8192 locates
>> the environment at the end of a sector, so shifting it 512 bytes earlier
>> in the flash can't possible now align it with a different sector?
>
> Why not?

If -8192 is aligned to the last sector boundary, the sectors must be 
8192 bytes in size (or the last sector isn't a whole sector, which is 
highly unlikely). If so, then 512 bytes away from that isn't also at a 
sector boundary.

>> Should  ENV_OFFSET be shifted down by a whole sector size by this patch, i.e. should
>> +CONFIG_TXT_CFG_BLOCK_OFFSET be rounded up to a sector size to  match the comment?
>
> Yes, a block, sector or whatever you wana call it is actually 512 bytes
> in size on all eMMCs or even MMC/SD cards that I have seen so far.
>
> I guess while it would be kind of nice having this enforced to some
> intrinsic size in software ultimately this is a board configuration
> option concerning our Toradex factory configuration block as programmed
> during production. We just want to make sure that one is preserved and
> the environment is stored right before.

Ah, I see. the existing comment uses the term "sector" incorrectly, 
which is where the confusion arises. eMMC do indeed have 512-byte 
sectors. I suppose the old (and new) comments should refer to HW boot 
partitions, not "sectors"?