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

Re: architecture naming discussion

On 4/27/06, Pjotr Kourzanov <peter.kourzanov@xs4all.nl> wrote:
> FPU is a CPU-related thing, so it should be part of $cpu I think.
But software fpu emulation is a kernel feature, so arguable. Alghough
I must agree at this point.

> For ARM OABI this means having separate $cpu architectures
> for: arm (hard-float), arm-softfloat (soft-float), arm-vfp (softfp),
> arm-iwmmxt and whatnot.
Now let's see how many architectures are there for arm cpus. :) A
better solution won't involve separate architectures for each case, I

> Isn't ARM EABI a solution to this? The discussion on a proper name
> for "arm-eabi" (which is my preference) is still ongoing, but it seems
> that most likely armel/armeb will be chosen (where "e" is for EABI
> and "l"/"b" is for little/big endian).
I'd rather arm eabi become mainstream 'arm' and old abi forgotten and buried.

But then, soft/hard fpu problem affects many architectures out there,
not only arms.

> I didn't forget about kfreebsd-i386-uclibc which in your case would
> be kfreebsd-uclibc-i386.
. o O ( darwin-uclibc-sparc )

I am free of all prejudices. I hate every one equally.

Reply to: