[U-Boot] [PATCH V2 3/3] fs: add partition switch libary, implement ls and fsload commands

[Stephen Warren]

On 10/19/2012 01:18 PM, Rob Herring wrote:
> On 10/19/2012 11:56 AM, Stephen Warren wrote:
>> On 10/18/2012 05:23 PM, Tom Rini wrote:
>>> On 10/18/12 16:12, Rob Herring wrote:
>>>> On 10/18/2012 06:01 PM, Tom Rini wrote:
>> ...
>>>>>>> On 10/11/2012 01:59 PM, Stephen Warren wrote:
>>>>>>>> Implement "ls" and "fsload" commands that act like 
>>>>>>>> {fat,ext2}{ls,load}, and transparently handle either 
>>>>>>>> file-system. This scheme could easily be extended to
>>>>>>>> other filesystem types; I only didn't do it for zfs
>>>>>>>> because I don't have any filesystems of that type to test
>>>>>>>> with.
>> ...
>>>>> Baring further discussion, I intend to grab this really soon,
>>>>> as it sounds like it's a functional starting point, however we
>>>>> wish to make this happen.  Or am I not following?  Thanks!
>>>
>>>> It's your call. I'd rather see clean-up first and features
>>>> second, but that's just me. Either way works. The amount of
>>>> duplication in u-boot just annoys me. Hopefully the DM work will
>>>> fix some of it.
>>>
>>> I too would like to see more clean-up,
>>
>> Which clean-up exactly?
>>
>> The only duplication I see here is that ext2load/fatload could be
>> modified to simply call into do_fsload. That'd be pretty simple, I
>> think, assuming the behaviour change was OK (e.g. fatload would
>> suddenly support either FAT or ext2*), and that cmd_fs.c and fs.c
>> would both always be pulled in.
> 
> Can't you make do_fsload support either specifying the fs for legacy use
> or detecting it on the new commands?

Yes, I suppose I could:

* Add a bit-mask of legal filesystems as a parameter to fs_set_blk_dev().

* Move the body of do_fsload() into some common called by do_fsload(),
do_ext2load(), do_fatload(), each passing in the appropriate bit-mask,
which gets passed down to fs_set_blk_dev().

That'd be easy, and probably entail only extremely minimal code-size
increases for an ext2-only or FAT-only build; just a few bytes for a few
more function calls. Sound like a plan?