[parisc-linux] parisc64 vs kernel64

Matthew Wilcox matthew@wil.cx
Sat, 24 Feb 2001 20:42:37 +0000

On Fri, Feb 23, 2001 at 02:12:22PM -0800, Grant Grundler wrote:
> How much is driven by how gcc/binutils/config.guess type stuff works?


> We've made an effort to keep them the same. That's why most of the
> files are common. The fact that PA2.0 cpus can run PA1.1 kernel bits
> makes it easy for us. I'm *know* the R4000 MIPs processor had
> good backwards compatibility between 32/64-bit - don't know about
> later ones or Sparc/Sparc64.

We had a discussion of this in the meeting, but it's worth mentioning for
the record.  Ralf had two teams of people pulling at him -- one was the embedded
MIPS people who wanted to keep everything simple and small, and the other
was SGI who wanted 64-bit, SMP, scalability.  So MIPS vs MIPS64 is also a
tradeoff between make-it-small and make-it-fast.

For both Sparc & MIPS there was also the opportunity to discard
backwards-compatibility stuff.  Compare the Sparc64 and Sparc32 atomic_t
operations.  The changes between PA-RISC and PA-RISC64 really aren't
as major.  There aren't many places we could remove cruft by guaranteeing
we're omitting certain types of processor.

It would almost certainly make sense to have a `parisc1.0' architecture
if someone were crazy enough to do that :-)

> The Processor family selection is a good idea regardless of
> how arch/parisc* source is organized.


Revolutions do not require corporate support.