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

[Mathieu Desnoyers]

* Linus Torvalds (torvalds@xxxxxxxxxxxxxxxxxxxx) wrote:
> On Thu, Jul 15, 2010 at 9:44 AM, Mathieu Desnoyers
> <mathieu.desnoyers@xxxxxxxxxxxx> wrote:
> >
> > The first thing that strikes me is that we could be interrupted by a standard
> > interrupt on top of the iret instruction below.
> 
> No, that can never happen.
> 
> Why? Simple: regular interrupts aren't ever enabled in eflags. So the
> only kinds of traps we can get are NMI's (that don't follow the normal
> rules), and exceptions.

Ah, you're right, since NMIs are an intr gate, IF is disabled in the EFLAGS all
along.

> 
> Of course, if there is some trap that re-enables interrupts even if
> the trap happened in an interrupt-disabled region, then that would
> change things, but that would be a bad bug regardless (and totally
> independently of any NMI issues). So in that sense it's a "could
> happen", but it's something that would be a totally separate bug.

Yep, this kind of scenario would therefore be a bug that does not belong to the
specific realm of nmis.

Thanks,

Mathieu


-- 
Mathieu Desnoyers
Operating System Efficiency R&D Consultant
EfficiOS Inc.
http://www.efficios.com
--
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/