Re: [kvm-devel] [PATCH 27/50] KVM: Support assigning userspace memory to the guest

[Avi Kivity <avi@qumranet.com>]

Avi Kivity wrote:
> From: Izik Eidus <izike@qumranet.com>
>
> Instead of having the kernel allocate memory to the guest, let userspace
> allocate it and pass the address to the kernel.
>
> This is required for s390 support, but also enables features like memory
> sharing and using hugetlbfs backed memory.
>
>   

[...]

> @@ -728,11 +752,27 @@ static int kvm_vm_ioctl_set_memory_region(struct kvm *kvm,
>  
>  		memset(new.phys_mem, 0, npages * sizeof(struct page *));
>  		memset(new.rmap, 0, npages * sizeof(*new.rmap));
> -		for (i = 0; i < npages; ++i) {
> -			new.phys_mem[i] = alloc_page(GFP_HIGHUSER
> -						     | __GFP_ZERO);
> -			if (!new.phys_mem[i])
> +		if (user_alloc) {
> +			unsigned long pages_num;
> +
> +			new.user_alloc = 1;
> +			down_read(&current->mm->mmap_sem);
> +
> +			pages_num = get_user_pages(current, current->mm,
> +						   mem->userspace_addr,
> +						   npages, 1, 0, new.phys_mem,
> +						   NULL);
> +
>   

I just combined a patch that changes the 'force' parameter to 
get_user_pages from 0 to 1, into this patch, to avoid introducing a bug 
and its fix in the same patchset.  I won't be resending this patch since 
the change is too trivial.  Same change applies to patch 48, "KVM: MMU: 
Partial swapping of guest memory".

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