Subject: Re: Possible bug in arm32 strongarm optimisations.
To: Richard Earnshaw <rearnsha@arm.com>
From: David Brownlee <abs@netbsd.org>
List: port-arm32
Date: 10/31/2000 13:25:45
On Mon, 30 Oct 2000, Richard Earnshaw wrote:

> NetBSD always uses soft-float, so the co-proc flag won't affect anything.
>
	Thanks.

> >
> > 	Are there any other interesting armv4 instructions?
>
> The only thing that you lose between -march=armv3m and -march=armv4 that
> would be of value on a SA+RPC is ldrsb.  Given that adding this would only
> have minimal affect on normal user code, there seems little point in
> adding a special option just to work around the RPCs obsolete bus
> architecture.

	You mention user-code - would it have any benefit in kernel-code?

                David/absolute
			       -- www.netbsd.org: A pmap for every occasion --