[Info-vax] Intel x86-64 Processor Design Security Vulnerability?

[already5chosen at yahoo.com]

On Wednesday, January 3, 2018 at 8:45:17 PM UTC+2, Simon Clubley wrote:
> On 2018-01-02, Stephen Hoffman <seaohveh at hoffmanlabs.invalid> wrote:
> > On 2018-01-02 18:24:24 +0000, Simon Clubley said:
> >> 
> >> If this is what it sounds like, I wonder if a microkernel based 
> >> operating system would be more protected against this than a monolithic 
> >> operating system would be ?
> >
> > For those inclined, I'd expect that OpenVMS can be targeted by 
> > rowhammer on both Itanium and Alpha.  Rowhammer is likely too "new" to 
> > particularly have hardware mitigations available in either platform, 
> > too.
> >
> > VSI hasn't specifically indicated if they're porting OpenVMS directly 
> > or if they're going to repeat the earlier experimental Mach port.  But 
> > I'd expect that they're porting without using Mach, given there were no 
> > discussions of Mach at the boot camp.
> >
> 
> No, I wasn't expecting VSI to start using a microkernel (although I was
> indeed aware of the older VMS on Mach experiment).
> 
> I was thinking more that the microkernel internals could be more
> shielded than in a monolithic kernel, although until we know the
> details, that's just a guess.
> 
> What is interesting is that people put down microkernels because of
> the additional overheads they incur, but now monolithic kernels are
> likely to have most, if not all, of those overheads on Intel 64-bit
> processors without getting the benefits that come with microkernels.
> 
> In fact, some microkernel designs mitigate a part of this overhead,
> so I wonder if the monolithic kernels are actually going to become 
> slightly _slower_ than a microkernel for some things ?
> 
> Simon.
> 
> -- 
> Simon Clubley, clubley at remove_me.eisner.decus.org-Earth.UFP
> Microsoft: Bringing you 1980s technology to a 21st century world

Details of exploit are still under embargo, but if it is what most people guess it is then microkernels that are built around SAS ideas (which are most of them, right ?) are at least as vulnerable as monolithic kernels. The difference is that mitigation, currently applied to Linux and supposedly Windows, would have significantly bigger performance impact if applied to SAS microkernel.