[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Any objections to armeb port posts on this list?

+++ Lennart Sorensen [05-09-22 09:10 -0400]:
> On Thu, Sep 22, 2005 at 04:30:35AM -0400, smcdougall27@aol.com wrote:
> So since we have mips and mipsel (endian little) it seems reasonable to
> call it arm and armeb (endian big).

Yep, use this list - anything else would be daft.

> I really hope they avoid the legacy arm (el) bad decisions while they
> are at it (like avoid floating point emulation rather than softfloat and
> such).  Unless there really WAS a good reason for doing that on the
> netwinder.

It made sense back then as there were chips with hardware FPU used in the
original kernel port (ARM7500), and that was the way it was 'always done'.
There were enough things to worry about getting the port done without doing
anything as radical as defaulting to soft-float. (I am speculating a bit
here as I missed the beginnings of the arm port).

But clearly it makes sense to default to soft-float or VFP now, and ARM ltd
is pushing for ABI changes to makes thing more compatible between gcc and
their compilers. I think now is a good time to look at starting the armeb
port with a new ABI and changing the old port to match. This will be painful
for a while and needs some care, but needs doing.

I will be discussing this with Phil Blundell next week in Cambridge (UK),
and we'll see if we can work out a plan.

Anyone else with ideas about how to manage the transition please speak up.

Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK  Tel +44 (0) 1223 811679
work: http://www.aleph1.co.uk/     play: http://www.chaos.org.uk/~wookey/

Reply to: