Support for PowerPC e200 cores

[Thomas Doerfler]

Till,

Till Straumann schrieb:
> FWIW, I strongly recommend to abandon the current PPC multilib mess.
> The e500 core is at the UISA-level still powerpc compatible and
> that's all gcc cares about.
> 
> The mvme3100 BSP I recently added uses an 8540 CPU and the compiler
> flags
> 
> -mcpu=powerpc -mno-float
> 
> which work just fine. I am convinced that we could drastically reduce
> the PPC multilib variants.
> 
> A new multilib-variant would only be needed if the 8450's 
> single-precision float APU were to be used.
> 

This may be ok for pure integer control stuff, but our automotive
customer who initiates the MPC55xx support will require single precision
float support by hardware. This is one of the key features of this chip
family and it will not be acceptable on the long run to disable GCC
created float instructions :-(

> Currently, the main reason for the many variants is that rtems' gcc
> specs map the -mcpu=xxx flags to a corresponding -DXXX which is
> tested in rtems libcpu and libbsp code.

Hm, I agree that we may have too many multilib variants, but on the 
other hand I assume that GCC will create different optimized code for a 
ppc405, a ppc750 and a ppc74xx. I agree that there is no reason to 
create a different multilib just to distinguish between microcontroller 
variants, but I would like to waste the potential performance gain 
resulting from a better optimization.

Which set of multilibs would you regard as "basic"?

Currently (in a 4.8 toolset) I see

m403/ m7400/ m505/ m860/ m601/ m603e/ m604/

which seems reasonable for me (601/603e/604 might be discussed).

> -- Till

wkr,
Thomas.

-- 
--------------------------------------------
embedded brains GmbH
Thomas Doerfler           Obere Lagerstr. 30
D-82178 Puchheim          Germany
Tel. : +49-89-18 90 80 79-2
Fax  : +49-89-18 90 80 79-9
email: Thomas.Doerfler at embedded-brains.de
PGP public key available on request

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.