Re: [PATCH v3 3/5] arch/x86/mm: Refactor cond_ibpb() to support other use cases

[Singh, Balbir]

On Sat, 2020-04-18 at 11:59 +0200, Thomas Gleixner wrote:
> 
> "Singh, Balbir" <sblbir@xxxxxxxxxx> writes:
> > On Fri, 2020-04-17 at 15:07 +0200, Thomas Gleixner wrote:
> > > 
> > > Balbir Singh <sblbir@xxxxxxxxxx> writes:
> > > > 
> > > >  /*
> > > > - * Use bit 0 to mangle the TIF_SPEC_IB state into the mm pointer
> > > > which is
> > > > - * stored in cpu_tlb_state.last_user_mm_ibpb.
> > > > + * Bits to mangle the TIF_SPEC_IB state into the mm pointer which is
> > > > + * stored in cpu_tlb_state.last_user_mm_spec.
> > > >   */
> > > >  #define LAST_USER_MM_IBPB    0x1UL
> > > > +#define LAST_USER_MM_SPEC_MASK       (LAST_USER_MM_IBPB)
> > > > 
> > > >       /* Reinitialize tlbstate. */
> > > > -     this_cpu_write(cpu_tlbstate.last_user_mm_ibpb,
> > > > LAST_USER_MM_IBPB);
> > > > +     this_cpu_write(cpu_tlbstate.last_user_mm_spec,
> > > > LAST_USER_MM_IBPB);
> > > 
> > > Shouldn't that be LAST_USER_MM_MASK?
> > > 
> > 
> > No, that crashes the system for SW flushes, because it tries to flush the
> > L1D
> > via the software loop and early enough we don't have the l1d_flush_pages
> > allocated. LAST_USER_MM_MASK has LAST_USER_MM_FLUSH_L1D bit set.
> 
> You can trivially prevent this by checking l1d_flush_pages != NULL.
> 

But why would we want to flush on reinit? It is either coming back from a low
power state or initialising, is it worth adding a check for != NULL everytime
we flush to handle this case?

Thanks,
Balbir