Re: heads up new sub-port e500 v2

[Aldy Hernandez <aldyh at redhat dot com>]

On Wed, Oct 20, 2004 at 11:03:30PM -0700, Geoff Keating wrote:
> 
> On 19/10/2004, at 3:13 PM, Aldy Hernandez wrote:
> 
> >Stage 3 is open to new ports that do not add extra functionality.  As
> >such, I will be adding support for a new variant of the e500.
> 
> Although it's open for new entire chip ports, ports that affect an 
> existing machine description (like, say, the rs6000 one), are not 
> automatically acceptable.  It would depend on how likely the changes 
> are to break the current functionality.
> 
> (For example, new patterns that require TARGET_E500V2 are fine.  Adding 
> "&& ! TARGET_E500V2" to existing patterns is fine.  A complete new 
> scheduling description is probably fine.  Changing the rs6000 move 
> expander is not so fine.)

I believe the sub-port falls into these categories.

The most intrusive changes I see are changing a few define_insn's to
define_expands so as to support both an e500v2 version, and a classic
ppc version.

All other changes seem to be !TARGET_E500V2 and !TARGET_E500V2 related.

Is this acceptable?

Aldy