[U-Boot] [PATCH 4/6] gpt: Support for GPT (GUID Partition Table) restoration

[Stephen Warren <swarren@wwwdotorg.org>]

On 09/06/2012 05:19 AM, Lukasz Majewski wrote:
> Hi Stephen,
> 
>> On 08/24/2012 02:13 AM, Lukasz Majewski wrote:
>>> The restoration of GPT table (both primary and secondary) is now
>>> possible. Simple GUID generation is supported.
>>
>>> diff --git a/include/part.h b/include/part.h
>>
>>> +int set_gpt_table(block_dev_desc_t *dev_desc,
>>> +		   int parts, unsigned int *blocks, char *name[]);
>>
>> That interface seems very limiting; what if you want to select
>> specific type UUIDs, set the partition attributes, leave gaps between
>> the partitions, have the physical order of partitions (their block
>> numbers) be different to the partition table ordering, etc.
>>
>> I wonder if it wouldn't be better to take an array of structures here,
>> and have each entry define all the properties of the partition. That
>> would allow fields to be added to the structure to describe more
>> information as we add more functionality.
> 
> I agree.
> 
> I'd propose following approach:
> 
> - set_gpt_table will receive a table with pointers to partition entries
>   (up to 128 entries in this table). In this way one can easily define
>   partition properties (which can be different for different
>   commands/boards).
> 
> - I think, that Legacy MBR, and Primary/Secondary GPT shall be
>   generated in the set_gpt_table function. However I expect a problem
>   with one Primary/Secondary GPT field
>   - the Disk GUID. We can read it from environment, generate it or
>   hardcode it at ./include/boards/{target}.h
> 
> 
> What is your opinion?

Perhaps rather than /just/ taking a pointer to a table of partitions,
take a pointer to a struct, which can hold both global properties and a
pointer to the list of partitions?