Re: [patch V8 09/15] SSB 9

[Thomas Gleixner <tglx@linutronix.de>]

On Tue, 1 May 2018, speck for Thomas Gleixner wrote:
> On Tue, 1 May 2018, speck for Thomas Gleixner wrote:
> > >  
> > >  	init_intel_misc_features(c);
> > > +
> > > +	if (cpu_has(c, X86_FEATURE_SPEC_STORE_BYPASS_DISABLE))
> > > +		x86_set_spec_ctrl(SPEC_CTRL_RDS);
> > 
> > Groan no. x86_setup_ap_spec_ctrl() needs to be investigated not random crap
> > added to some random place again.
> > 
> > x86_setup_ap_spec_ctrl()
> > {
> >        if (boot_cpu_has(X86_FEATURE_IBRS))
> >                 x86_set_spec_ctrl(x86_spec_ctrl_base & ~x86_spec_ctrl_mask);
> > 
> > So that's the issue.
> 
> And if that does not work on your machine then
> 
>        if (boot_cpu_has(X86_FEATURE_IBRS))
>                 rdmsrl(MSR_IA32_SPEC_CTRL, x86_spec_ctrl_base);
> 
> in check_bugs() does not work either. So how is your crappy hack solving
> anything?

And if your machine has the SPEC CTL MSR then it should have set
FEATURE_IBRS as well because:

       if (cpu_has(c, X86_FEATURE_SPEC_CTRL)) {
	        set_cpu_cap(c, X86_FEATURE_IBRS);
                set_cpu_cap(c, X86_FEATURE_IBPB);
        }

And if FEATURE_RDS is set and FEATURE_SPEC_CTRL is not then something is
even more wrong.

Papering over symptoms without trying to understand the root cause is the
worst engineering approach.

Thanks,

	tglx