Re: Pending IDE update for SMP, bugfixes, ...

Dominik Kubla (kubla@sundiver.zdv.uni-mainz.de)
Fri, 17 Apr 1998 00:47:19 +0200


On Thu, Apr 16, 1998 at 11:03:47AM -0400, mlord wrote:
> Andrzej Krzysztofowicz wrote:
> ...
> > I don't use a SMP machine, but have some experience with ali chipset,
> > so my $0.02:
> >
> > 1. I don't think it is safe to enable other CPU's interrupts while
> > probing ports for chipset detection. Some of ports beeing probed
> > may be used by other hardware (eg. for some MB's 0x34 is same as 0x20
> > and probing it causes interrupt controller reprogramming)
>
> This does not worry me, because:
>
> (1) nobody with an SMP machine will have an ALI controller,

But it should: there are already Pentium-II boards with the ALi Aladdin Pro
chipset, eg the DataExpert MAL8621. And since this is a 100MHz bus clock
chipset, you can expext to see it on SMP systems as well.

And what about distributions providing a SMP kernel by default? And this is
not without precedent. Just look at the boot message of WNT4 and you will
see... right: MP kernel. But that's going to change with 5.0, then you will
have to pay extra to use your second CPU. So we have a good opportunity to
beat WNT5 in the "power-user" segment, if Linux get's the SMP support right...

> and (2) the probe only happens if the user asks for it with
> a kernel boot parameter.

Security by obscurity. If you were in system support then you would not think
this to be safe enough: Joe User will try any combination of boot paramaters
just to get another 1% of "speed". And if this breaks his system he will
surely fault the software, not his own stupidity.

Yours,
Dominik Kubla

-- 
               "Waga hoshi wa doko ni tabine ya ama-no-gawa."
      In the Summer Evening, Where is My Star Hidden in the Milky Way?
                         Kobayashi Issa (1763-1828)

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu