[U-Boot-Users] Re: Patch: Support for PQ27e (8247/48/71/72) chips and MPC8272ADS board

Wolfgang Denk wd at denx.de
Mon Mar 15 11:07:15 CET 2004


In message <16469.31569.535909.246506 at gargle.gargle.HOWL> you wrote:
>
>     Wolfgang> -#undef CPM_DATAONLY_SIZE
>     Wolfgang> -#define CPM_DATAONLY_SIZE ((uint)(8 * 1024) - CPM_DATAONLY_BASE)
> 
>     Wolfgang> Are you 100% sure this does not break any existing boards?
> 
> I checked it on four different boards with different PQ2 chips. In fact,
> for older (pre-PQ27e) chips this patch does not change memory map so
> they aren't affected in any way. Am I missing anything?

I don't know. I just want to understand the consequences.

>     Wolfgang> + gd->CPM_DATAONLY_SIZE;
>     Wolfgang> + if (is_pq27e())
>     Wolfgang> + gd->dp_alloc_top = gd->dp_alloc_base
>     Wolfgang> + gd->PQ27E_DATAONLY_SIZE;
>     Wolfgang> + else
>     Wolfgang> + gd->dp_alloc_top = gd->dp_alloc_base
>     Wolfgang> + gd->CPM_DATAONLY_SIZE;

First, ther is at least one redundand "gd->CPM_DATAONLY_SIZE;" here.

>     Wolfgang> Please do not add a board-specific #define
>     Wolfgang> PQ27E_DATAONLY_SIZE when we already have a
>     Wolfgang> CPM_DATAONLY_SIZE which could be used.
> 
> There should be some misunderstanding here. PQ27E_DATAONLY_SIZE is NOT
> board-specific (I won't add board-specific things into common
> files). PQ27e is the common name for the 8247/8248/8271/8272 (the `e'

Sorry for chosing a poor description.

> stands for `encryption'). These chips have got much less internal memory
> (DPRAM) and at different addresses so they must have separate
> #defines. CPM_DATAONLY_SIZE is just incorrect for PQ27e. PQ27e have got
> only 4K of "data only" RAM and not 8K (and at different base, BTW).

Why do you need a separate  #define  then?  Isn't  it  sufficient  to
#define a correct value for CPM_DATAONLY_SIZE then?

>     Wolfgang> 	PQ27E_FCC_SPECIAL_BASE vs. CPM_FCC_SPECIAL_BASE
> 
> It's for the same reason. PQ27e have not got RAM at 0xB000, other PQ2s

So why not just #define a correct value?

> After the explanations, what changes would you suggest?

Use the existing variables and just assign correct values?

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd at denx.de
When the bosses talk about improving  productivity,  they  are  never
talking about themselves.




More information about the U-Boot mailing list