Re: [PATCH][1/3] Move kvm_vcpu_iotcl_get_dirty_log to arch

[Avi Kivity <avi-atKUWr5tajBWk0Htik3J/w@public.gmane.org>]

Carsten Otte wrote:
> Zhang, Xiantao wrote:
>> Move kvm_vcpu_ioctl_get_dirty_log to arch, and still keep the interface
>> in common.
>
> diff --git a/drivers/kvm/kvm_main.c b/drivers/kvm/kvm_main.c
> index 45b18e1..3400265 100644
> --- a/drivers/kvm/kvm_main.c
> +++ b/drivers/kvm/kvm_main.c
> @@ -419,19 +419,14 @@ int kvm_vm_ioctl_set_memory_region(struct kvm *kvm,
>      return kvm_set_memory_region(kvm, mem, user_alloc);
>  }
>
> -/*
> - * Get (and clear) the dirty memory log for a memory slot.
> - */
> -static int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
> -                      struct kvm_dirty_log *log)
> +int kvm_get_dirty_log(struct kvm *kvm,
> +            struct kvm_dirty_log *log, int *is_dirty)
>
> Any reason to remove that comment? It looks helpful to me.
>
>
>  {
>      struct kvm_memory_slot *memslot;
>      int r, i;
>      int n;
>      unsigned long any = 0;
>
> -    mutex_lock(&kvm->lock);
> -
>      r = -EINVAL;
>      if (log->slot >= KVM_MEMORY_SLOTS)
>          goto out;
> @@ -450,17 +445,11 @@ static int kvm_vm_ioctl_get_dirty_log(struct kvm 
> *kvm,
>      if (copy_to_user(log->dirty_bitmap, memslot->dirty_bitmap, n))
>          goto out;
>
> -    /* If nothing is dirty, don't bother messing with page tables. */
> -    if (any) {
> -        kvm_mmu_slot_remove_write_access(kvm, log->slot);
> -        kvm_flush_remote_tlbs(kvm);
> -        memset(memslot->dirty_bitmap, 0, n);
> -    }
> +    if (any)
> +        *is_dirty = 1;
>
>      r = 0;
> -
>  out:
> -    mutex_unlock(&kvm->lock);
>      return r;
>  }
>
> This split won't work on s390. I think kvm_get_dirty_log should be 
> arch dependent alltogether: we're not going to have a dirty bitmap on 
> s390, we rather rely on our hardware support to update our storage 
> keys accordingly without guest/host intervention required. We'd want 
> to use that, and thus this implementation of kvm_get_dirty_log makes 
> no sense for us. On the other hand, I'd really want to keep the ioctl 
> common, so that guest migration code in userland can be common for all 
> archs.

On the other hand, it will work for all others IIUC.  Duplicating it in 
all other archs is not a good idea.

We can special case it using #ifdef ARCH_HAS_KVM_GUEST_DIRTY_BITMAP or 
something.

(hides from the #ifdef police)

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


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/