Re: [ECOS] Supporting Paged Flash in eCos/RedBoot

[Jonathan Larmour <jifl at eCosCentric dot com>]

Jonathan Larmour wrote:
> Ian Campbell wrote:
>
> > Hi,
> >
> > Attached is a rough first draft of a patch to implement an improved
> > flash interface -- It's a first pass and I certainly wouldn't expect you
> > to apply it as is...
> >
> > The main bulk of the patch modifies the API of cyg/io/flash.h such that
> > each function takes a struct flash_info * as it's first parameter and
> > all addresses are described as unsigned long offsets
>
>
> NB The flash drivers were written when Gary thought he could get away 
> with using things like ints and longs instead of the normal eCos-y types 
> like cyg_uint32 etc. 

Just to be clear, at the time the intention was only to get flash drivers 
for RedBoot, so things like namespacing were irrelevant as it was just 
RedBoot, right? :-); and (I've always assumed) ints/longs were sufficient 
rather than the eCos-y types because it wasn't at that point intended for 
RedBoot to work on a target where sizeof(int) != 4, or sizeof(long) isn't 
the same size as a function or data address[1]. Besides, as always, there 
were deadlines to meet...

However eCos has always been intended to be more portable than that and we 
need to address that if we can, even if not all at once, at least if we 
change the API, we should do the right thing in that API, coz APIs tend to 
stick around :-).

Jifl
[1] Some obscure targets do have differently sized function and data 
pointers, primarily harvard architectures, scarily enough.
--
eCosCentric    http://www.eCosCentric.com/    The eCos and RedBoot experts
--[ "You can complain because roses have thorns, or you ]--
--[  can rejoice because thorns have roses." -Lincoln   ]-- Opinions==mine


--
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss