Re: Correct way of tracking reads on given gfn ?

[Paolo Bonzini <pbonzini@redhat.com>]

Il 09/09/2013 11:45, Sunil ha scritto:
> Hello List,
> 
> (1) What is the correct way of tracking reads on given gfn ?
> 
> Looks like functions rmap_write_protect(kvm, gfn) and
> kvm_mmu_rmap_write_protect(kvm, gfn, slot) removes the write
> permission for the given page, by flipping write bit using
> PT_WRITABLE_MASK. Thus trapping writes.
> 
> For trapping reads, are there similar routines ? If not, I see
> equivalent flag PM_PRESENT_MASK. Thus one way probably is writing
> wrapper routines similar to above to flip both read(present) and write
> bits. Or would it be just enough to drop the spte instead using
> function drop_spte() ?

If you are using EPT and your processor has the accessed and dirty bits
for EPT page tables (cat /sys/module/kvm_intel/parameters/eptad), you
could use the accessed bit to track reads.  They wouldn't trap, but you
would be able to poll the EPT page tables.

Paolo

> (2) Is kvm_flush_remote_tlbs() required after either of above approaches ?
> Is kvm_flush_remote_tlbs() correct function to flush tlbs ?
> 
> --
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>