> On Sun, 8 Mar 1998, Phil Brutsche wrote:
>
> > Besides, what if someone wanted to CHANGE those options? They would have
> > to re-compile their kernel every time. It's so much more convenient to
> > have the initializations in a user-space program.
>
> I would agree with this except for one thing ... the halt-on-idle should
> be set if a 6x86 is detected should be the default behavior. The reson
Seconded. IMO, two features that are not present in the current 2.1 kernel
that should be for Cyrix, are: 1.) HLT-on-idle, and 2.) NO_LOCK set by
default. The justification is this: If these options go into the kernel
by default, users and vendors will have a Cyrix system that boots up with
the power-saving/heat-reduction feature on by default (a big win), and the
Cyrix-coma workaround enabled by default (a huge win for ISP-targetted
distributions, like RedHat, Debian, et al). In fact, I'd go a step further
and say "don't even make them options". Turn them on by default and let
the user *disable* them from user-space if they want. Unless someone has
conclusive evidence that either of these options causes kernel hangs, why
not? In the Cyrix research that I've done, these are both rock-solid and
safe (and recommended) register settings. I've already hand-pecked them
into my 2.1.8x kernel with great satisfaction. I'm using a step 1 rev 7
6x86 120Mhz CPU.
Recall that we already safeguard the kernel against the f00f Pentium bug;
why leave the Cyrix one wide open when it's so simple to add to the kernel?
Certainly it's a bigger chalenge for distributions to figure out what type
of CPU the novice has and enable some rc script at boot time that deals
with it (and can deal with a CPU change apres install).
My $0.02. :)
cheers,
-bp
-- B. James Phillippe <bryan@terran.org> Linux Software Engineer, WGT Inc. http://w3.terran.org/~bryan
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu