Re: [PATCH]: Fix MTRR handling on HT CPUs (improved)

arjan@fenrus.demon.nl
Fri, 25 Jan 2002 18:51:24 +0000 (GMT)


In article <E16UBUl-0003J9-00@the-village.bc.nu> you wrote:
> As Dave pointed out I was mixing them

>> just not do it on the right CPU (you're _not_ supposed to read to see if
>> you are writing the same value: MTRR's can at least in theory have
>> side-effects, so it's not the same check as for the microcode update).

> So why not just set it twice - surely that is harmless ? Why add complex
> code ?

mtrr code does

"read value" -> store in variable
"set uncached"
<do lots of fiddling>
"write stored value"

on all cpus precisely in parallel.

Now... on HT the second half will store the value the first half just set to
uncached. and worse: it will RESTORE that into the final mtrr..

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/