Re: [PATCH 3/4] KVM: Count the number of dirty pages for dirty logging

[Avi Kivity <avi@redhat.com>]

On 11/14/2011 11:23 AM, Takuya Yoshikawa wrote:
> Needed for the next patch which uses this number to decide how to write
> protect a slot.
>
>  	/* If nothing is dirty, don't bother messing with page tables. */
> -	if (is_dirty) {
> +	if (memslot->nr_dirty_pages) {
>  		struct kvm_memslots *slots, *old_slots;
>  		unsigned long *dirty_bitmap;
>  
> @@ -3504,6 +3500,7 @@ int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
>  			goto out;
>  		memcpy(slots, kvm->memslots, sizeof(struct kvm_memslots));
>  		slots->memslots[log->slot].dirty_bitmap = dirty_bitmap;
> +		slots->memslots[log->slot].nr_dirty_pages = 0;
>  		slots->generation++;
>  
>  #endif /* !CONFIG_S390 */
> @@ -1491,7 +1492,8 @@ void mark_page_dirty_in_slot(struct kvm *kvm, struct kvm_memory_slot *memslot,
>  	if (memslot && memslot->dirty_bitmap) {
>  		unsigned long rel_gfn = gfn - memslot->base_gfn;
>  
> -		__set_bit_le(rel_gfn, memslot->dirty_bitmap);
> +		if (!__test_and_set_bit_le(rel_gfn, memslot->dirty_bitmap))
> +			memslot->nr_dirty_pages++;
>  	}
>  }
>  

The two assignments to ->nr_dirty_pages can race, no?  Nothing protects it.

btw mark_page_dirty() itself seems to assume mmu_lock protection that
doesn't exist.  Marcelo?

-- 
error compiling committee.c: too many arguments to function