On Wed, 12 Jun 2002, Andre Hedrick wrote:
> The hardware changed and the interrupt parser feature that stablized the
> old chipsets under SMP is gone. The new chipsets (20268 and above) do not
> have a location with sticky bits. So in some cases I expect things to go
> south, but in general they should not. Otherwise promise would be all
> over the issue.
Okay, thanks, that clarifies it, and if I read it right this may be a
permanent restriction. If the system is running noapic so the ints all go
to CPU0 does that help the situation? In a sane system I agree with Alan
that this is not a performance hit, or at least not significant.
> > On Wed, 12 Jun 2002, Alan Cox wrote:
> > > Then I suggest you give up computing, because PC hardware doesnt make
> > > your grade. BTW the general open promise bugs *dont* include data
> > > corruption so I suspect it may be your h/w thats hosed.
> Andre Hedrick
> LAD Storage Consulting Group
Guys, I wasn't questioning anyone's competence, just agreeing with the
original poster that if this is going to be an ongoing issue of stability
it might be held back for a version until there is time to either program
around it or characterize the conditions under which it causes problems.
And if that means that I can't trust it SMP, I don't think I'm a bad guy
to suggest the config drop the driver if SMP support is selected.
I don't mind moving a card to another system, and even if I didn't have
another system I would rather not take that particular risk. And if it
won't work reliably SMP then perhaps Promise should be taking some action.
There are SMP W2k machines out there, too.
-- bill davidsen <firstname.lastname@example.org> CTO, TMR Associates, Inc Doing interesting things with little computers since 1979.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to email@example.com More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Jun 15 2002 - 22:00:26 EST