svn commit: r222183 - head/lib/clang

[John Baldwin]

On Monday, May 23, 2011 8:05:12 am Bjoern A. Zeeb wrote:
> On May 22, 2011, at 10:01 PM, Warner Losh wrote:
> 
> > 
> > On May 22, 2011, at 3:30 PM, Mark Linimon wrote:
> > 
> >> On Sun, May 22, 2011 at 10:22:56PM +0200, Roman Divacky wrote:
> >>> The problem here is deeper in my opinion. What FreeBSD calls
> >>> amd64 the rest of the world (ie. linux) calls x86_64, I think
> >>> that instead of this we should teach llvm/clang about "amd64".
> >>> Maybe as a FreeBSD-only diff.
> >> 
> >> If we move away from "amd64", we are going to need a _substantial_ amount
> >> of work on ports.
> > 
> > Yea.  That's why I think, although I'd like to move away from it, we're stuck with amd64 for both MACHINE and MACHINE_ARCH for quite some time.  
In that case, we'll just have to configure clang the same way we configure gcc with the x86_64 monicker.
> 
> Is there an authoritative source for these names?  I am not quite sure
> what led to the confusion in first place but I guess it's the "oh Intel
> is doing EMT64 let's not call this amd" kind of thing.
> 
> I seem to remember that apart from the linux kernel and OSX most others
> incl. windows, java and sun used to call it amd64 as well when it came to
> technical things.  So I am not exactly sure why this regularly comes up
> but I'd assume it's because of "what linux calls it must be right"?
> 
> I admit that it's confusing to got to x86-64.org and see it called the
> amd64 architecture.  I am sure Peter and David know a lot more about this.
> 
> Despite remembering Dragonfly having done the change I cannot see FreeBSD
> doing the same in a short timeframe, and yet I have no idea why we wanted to?

AMD changed the name.  It used to be 'amd64' and we used that name as result.
Later they changed it to x86-64, possibly because Linux and the GNU folks had
adopted that instead of 'amd64'.

Renaming it now would be a PITA.

-- 
John Baldwin