Re: [PATCH v11 09/12] x86: Enable User-Mode Instruction Prevention at runtime

[Ricardo Neri]

On Wed, Nov 08, 2017 at 10:52:59AM +0100, Ingo Molnar wrote:
> 
> * Ricardo Neri <ricardo.neri-calderon@xxxxxxxxxxxxxxx> wrote:
> 
> > User-Mode Instruction Prevention (UMIP) is enabled by setting/clearing a
> > bit in %cr4.
> > 
> > It makes sense to enable UMIP at some point while booting, before user
> > spaces come up. Like SMAP and SMEP, is not critical to have it enabled
> > very early during boot. This is because UMIP is relevant only when there is
> > a user space to be protected from. Given these similarities, UMIP can be
> > enabled along with SMAP and SMEP.
> > 
> > At the moment, UMIP is disabled by default at build time. It can be enabled
> > at build time by selecting CONFIG_X86_INTEL_UMIP. If enabled at build time,
> > it can be disabled at run time by adding clearcpuid=514 to the kernel
> > parameters.
> 
> > +config X86_INTEL_UMIP
> > +	def_bool n
> > +	depends on CPU_SUP_INTEL
> > +	prompt "Intel User Mode Instruction Prevention" if EXPERT
> > +	---help---
> > +	  The User Mode Instruction Prevention (UMIP) is a security
> > +	  feature in newer Intel processors. If enabled, a general
> > +	  protection fault is issued if the instructions SGDT, SLDT,
> > +	  SIDT, SMSW and STR are executed in user mode.
> 
> Ok, I really like this latest series.

Thanks!

> 
> One small request: could you please make the feature default-y, which is what we 
> do for new hardware features that don't have a significant runtime cost. There's 
> no ABI breakage expected, right?

No, no ABI breakage should happen. The only visible change is that processes running
in user long mode will see a SIGSEGV signal if they try to use any of the instructions
that UMIP protects. Also, any process using SLDT or or STR will see the SIGSEGV signal.

I will submit an extra patch to do this default-y.
> 
> Another request: could you please extend the Kconfig description to _explain_ to 
> the user why this feature is useful - in a short sentence or so. I.e. point out 
> that these instructions expose information about hardware state that is not really 
> necessary for the vast majority of applications.

Sure, I will submit an extra with this explanation.
> 
> Plus:
> 
> > +       cr4_set_bits(X86_CR4_UMIP);
> 
> Please also inform admins that it's enabled, via something like:
> 
> 	pr_info("x86/cpu: Activated the Intel User Mode Instruction Prevention (UMIP) CPU feature\n");

Will do.
> 
> ... or so.
> 
> Please do these changes in a separate add-on patch #13 on top of this series, as 
> I've already started testing these bits.

Sure.

Thanks and BR,
Ricardo