Re: [patch 1/2] x86_64 page fault NMI-safe

[Linus Torvalds]

On Thu, Jul 15, 2010 at 11:43 AM, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> But maybe I'm missing something.

Hmm. Of course - one way of solving this might be to just make the
32-bit case switch stacks in software. That might be a good idea
regardless, and would not be complicated. We already do that for
sysenter, but the NMI case would be simpler because we don't need to
worry about being re-entered by NMI/DEBUG during the stack switch.

And since we have to play some games with moving the data on the stack
around _anyway_, doing the whole "switch stacks entirely rather than
just subtract a bit from the old stack" would be fairly logical.

So I think you may end up being right: we don't need to save the
original NMI stack pointer, because we can make sure that the
replacement stack (that we need anyway) is always deterministic.

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