Re: Bug#440720: [SPARC]: non-SMP kernel fail on SunFIres with >= 2 CPUs
{I'm adding the former CCs to the mail again, this info should be
included in the bug report and on -boot}
Hi,
>> Also I'm pretty sure that non-smp kernels just don't work on the
>> machine. As far as I understand the way those machines work is that at
>> least two CPUs have to be in an operating state as they share one CPU
>> Data switch (if you have a look into such a machine you see that always
>> 2 CPUs + their memory are sitting in a CPU bay, and as far as I know you
>> can't run the system with an odd number of CPUs.
>
> I have 3 identical Sun Blade with 16 used CPU-Slots (each) where each can
> has two CPU's and shared memory between them and this machine machine
> does not boot with a NON-SMP Kernel, even if I remove 15 CPU-Cards.
Out of curiosity: Are those machines using UltraSparc III CPUs?
This sounds like another reason to have a SMP sparc installer.
What I'm still wondering about is if Sparc SMP Kernels are supposed to
run on all single CPU machines. I know one single-CPU machine where
running a SMP kernel results in an OOPS.
>> All processors share the same pysical memory address space and use -
>> depending on the number of CPUs - different cache coherence protocols,
>> so as far as I understand it such a system can't work at all without
>> having all CPUs properly initialized. But probably somebody with a
>> better knowledge about this architecture can give us some insight on
>> that, therefore I'm forwarding the message to debian-sparc, too.
>
> Since I do not know, how many CPU-Slots you have, but do you have tried
> to run the machine with ONLY ONE CPU-Card?
One CPU card contains 2 CPUs, and as far as I know you're not supposed
to run the machine with an odd number of CPUs. But I didn't find any
real proper information about that - I didn't spend much time on reading
books, though. The machine has 4 slots, 2 CPUs per slot :)
But removing CPUs to install a machine is nothing one can suggest people
to install a machine, especially when you can't be sure that it'll work
at all.
Best regards,
Bernd
--
Bernd Zeimetz
<bernd@bzed.de> <http://bzed.de/>
Reply to: