Re: No asm/smp.h == no compile

Nick STEPHEN (stephen@gr.osf.org)
Tue, 19 Mar 96 14:51:04 +0100


>> Another point of interest. I heard a rumor that Linus was going to
>> do the UltraSparc port. If this is true, will he be the only person
>> who has ported a 64 bit OS to more than 1 architecture?
>
>Again, it depends on what you consider an "operating system". OSF/1 MACH
>has been ported to both Alpha and the PowerPC, although I am not sure if the
>PowerPC Linux environment is "64-bit". And I am not sure what CMU itself
>has done along the 64-bit lines. And since Linus doing the UltraSparc port
>is still "future", only time will tell if he is "first" or not.

Linux for the PowerPC has currently only been ported to 32-bit
processors, and is not 64-bit clean. This is principally due to the
machines available, but work was also started before the Linux kernel
was 64-bit clean.

The OSF Mach3 microkernel has been ported to the Alpha and this port
is 64-bit clean. It has also been ported to the PowerPC, but currently
only to some of the 32-bit processors.

The OSF/1 server has been ported to Alpha (64 bit) but not at all to
the PowerPC processors (although this would be simple). On the PowerPC
line we use the MkLinux server, a linux kernel with the machine
dependant layer replaced by calls to the Mach microkernel.

Linux on PowerPc : http://www.linuxppc.org

MkLinux on PowerMac and Intel : http://www.gr.osf.org/mklinux

Obviously, once the generic code in the kernel or microkernel is
64-bit clean, it becomes much easier to port from a 32-bit to 64-bit
processor in the same family.

>Offhand, I would say that Linux has a "bleeding edge".

Of course it has - as have all other operating systems undergoing
maintenance/improvement. The major difference with Linux is that the
bleeding edge is available for the public to cut themselves on, rather
than being hidden in some company's development group. Most users can
isolate themselves from this bleeding edge by running an older, stable
kernel.

[ Nick ]

PS. I'm not on the alpha mailing list, please cc me in any
followups. Thanks.

[ Nick ]
__
Nick Stephen Web: http://www.gr.osf.org/~stephen
OSF Research Institute Email: stephen@gr.osf.org
2, Avenue de Vignate Phone: +33 76 63 48 72
38610 Gieres - France Fax: +33 76 51 05 32