linux-kernel - Re: [PATCH 1/4] KVM: guest_memfd: add generic post

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <589ccb59-ae79-49db-8017-f6d28d7f6982@amd.com>
Date: Fri, 22 Nov 2024 12:40:08 -0600
From: Mike Day <michael.day@....com>
To: Nikita Kalyazin <kalyazin@...zon.com>, pbonzini@...hat.com,
 corbet@....net, kvm@...r.kernel.org, linux-doc@...r.kernel.org,
 linux-kernel@...r.kernel.org
Cc: jthoughton@...gle.com, brijesh.singh@....com, michael.roth@....com,
 graf@...zon.de, jgowans@...zon.com, roypat@...zon.co.uk, derekmn@...zon.com,
 nsaenz@...zon.es, xmarcalx@...zon.com
Subject: Re: [PATCH 1/4] KVM: guest_memfd: add generic post_populate callback



On 10/24/24 04:54, Nikita Kalyazin wrote:
> This adds a generic implementation of the `post_populate` callback for
> the `kvm_gmem_populate`.  The only thing it does is populates the pages
> with data provided by userspace if the user pointer is not NULL,
> otherwise it clears the pages.
> This is supposed to be used by KVM_X86_SW_PROTECTED_VM VMs.
> 
> Signed-off-by: Nikita Kalyazin <kalyazin@...zon.com>
> ---
>   virt/kvm/guest_memfd.c | 21 +++++++++++++++++++++
>   1 file changed, 21 insertions(+)
> 
> diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c
> index 8f079a61a56d..954312fac462 100644
> --- a/virt/kvm/guest_memfd.c
> +++ b/virt/kvm/guest_memfd.c
> @@ -620,6 +620,27 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot,
>   EXPORT_SYMBOL_GPL(kvm_gmem_get_pfn);
>   
>   #ifdef CONFIG_KVM_GENERIC_PRIVATE_MEM

KVM_AMD_SEV can select KVM_GENERIC_PRIVATE_MEM, so to guarantee this is only for
software protection it might be good to use:

#if defined CONFIG_KVM_GENERIC_PRIVATE_MEM && !defined CONFIG_KVM_AMD_SEV

That could end up too verbose so there should probably be some more concise mechanism
to guarantee this generic callback isn't used for a hardware-protected guest.

Mike

> +static int kvm_gmem_post_populate_generic(struct kvm *kvm, gfn_t gfn_start, kvm_pfn_t pfn,
> +				  void __user *src, int order, void *opaque)
> +{
> +	int ret = 0, i;
> +	int npages = (1 << order);
> +	gfn_t gfn;
> +
> +	if (src) {
> +		void *vaddr = kmap_local_pfn(pfn);
> +
> +		ret = copy_from_user(vaddr, src, npages * PAGE_SIZE);
> +		if (ret)
> +			ret = -EINVAL;
> +		kunmap_local(vaddr);
> +	} else
> +		for (gfn = gfn_start, i = 0; gfn < gfn_start + npages; gfn++, i++)
> +			clear_highpage(pfn_to_page(pfn + i));
> +
> +	return ret;
> +}
> +
>   long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long npages,
>   		       kvm_gmem_populate_cb post_populate, void *opaque)
>   {