Re: [PATCH v6 06/14] KVM: Add memslot flag to let userspace force an exit on missing hva mappings

[Sean Christopherson <seanjc@google.com>]

On Thu, Feb 01, 2024, Anish Moorthy wrote:
> On Thu, Feb 1, 2024 at 8:09 AM Sean Christopherson <seanjc@google.com> wrote:
> >
> > On Wed, Jan 31, 2024, Anish Moorthy wrote:
> > > On Tue, Jan 30, 2024 at 4:26 PM James Houghton <jthoughton@google.com> wrote:
> > > >
> > > > Feel free to add:
> > > >
> > > > Reviewed-by: James Houghton <jthoughton@google.com>
> > >
> > > > If we include KVM_MEM_GUEST_MEMFD here, we should point the reader to
> > > > KVM_SET_USER_MEMORY_REGION2 and explain that using
> > > > KVM_SET_USER_MEMORY_REGION with this flag will always fail.
> > >
> > > Done and done (I've split the guest memfd doc update off into its own
> > > commit too).
> > >
> > > > > @@ -3070,6 +3074,15 @@ kvm_pfn_t __gfn_to_pfn_memslot(const struct kvm_memory_slot *slot, gfn_t gfn,
> > > > >                 writable = NULL;
> > > > >         }
> > > > >
> > > > > +       if (!atomic && can_exit_on_missing
> > > > > +           && kvm_is_slot_exit_on_missing(slot)) {
> >
> > Operators go on the preceding line:
> 
> Thanks. On a side note, is this actually documented anywhere? I
> searched coding-style.rst but couldn't find it.

Maybe?  But the fact there are very few, if any, patterns like this in KVM should
be a big clue that it's not the One True Way.  The formal docs will never be 100%
complete, and preferences do evolve and change, but if your code sticks out like
a sore thumb, odds are good you're doing something wrong.