[U-Boot] [PATCH 0/2] New abstraction layer for handling MMC and FAT commands

[Marek Vasut]

Dear Wolfgang Denk,

> Dear Marek,
> 
> In message <201207271330.46967.marex at denx.de> you wrote:
> > > This makes no sense to me.  MMC is just one of the storagte devices we
> > > support, and especially with the upcoming support of a clean device
> > > model it makes no sense to handle it special in any way.
> > 
> > Sure, but that might still be a release or so away. So let's not hinder
> > people from doing so, rather let's coordinate. Also, I consider there
> > patches helpful, as we can use that abstraction during the DM work.
> 
> Fine.
> 
> But why adding special code to only one storage media, and one file
> system type?

Well, that's other thing. But I hope Lukasz will eventually extend this work 
over other media as the DFU gains support for more stuff. Right, Lukasz?

> > > The same holds for FAT - it is just one out of a number of file
> > > systems we support, and it makes no sense to add any code to FAT
> > > support that makes it incompatible with other file systems.
> > 
> > It's not incompatible, how?
> 
> I cannot use - say - a USB storage device instead of MMC.

Not _yet_ .

> Why is this so?

Because it's not implemented yet.

> I will reply to that in the context of the DFU thread, it makes more
> sense there.
> 
> > > Any such addition should be implemented in a generic way, that is
> > > agnostic of the underlying storage device and the file system used on
> > > top of it.
> > 
> > Should, that's the proper word here. It isn't in most of FSs in uboot
> > though and I think this is better than nothing.
> 
> Wrong.  Adding special code to one file system which makes it
> "special" compared to other file systems is actually really, really
> bad.

I hope it'll not only be fat over the time.

> Best regards,
> 
> Wolfgang Denk

Best regards,
Marek Vasut