Re: 2.1.120 broken on P5 SMP machines + fix
Linus Torvalds (torvalds@transmeta.com)
Sun, 6 Sep 1998 11:08:29 -0700 (PDT)
On Sun, 6 Sep 1998, Thomas Bogendoerfer wrote:
>
> I'm not sure, but it looks like the F00F workaround gets only enabled for
> one CPU, when we do smp_begin() before check_bugs(). So it makes sense,
> when people see that the first start of a f00f trigger program gives
> illegal instruction while the second one hangs the machine. When we do
> smp_begin() all of the processors get started, but check_bugs() only
> reload the idt for the boot processor in trap_init_f00f_bug(), right ?
>
> > I'll just switch the check_bugs things around again and be red in the
> > face. Sorry,
>
> no problem. But it might be better to leave it the way it is in 2.1.120
> and stick in my patch.
I think the real problem is that "set_intr_gate()" really should just know
about the F0 0F bug, and be able to insert stuff into the IDT regardless
of whether the workaround is active or not. I'll write at least the basics
for that to be possible, and hope that somebody with a dual Pentium can do
the final details to make it all happy..
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html