linux-kernel - Re: [PATCH v2 03/13] KVM: Allow late setting of KVM_MEM_USERFAULT on guest

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: <aBqjV40pb-s9gvsz@google.com>
Date: Tue, 6 May 2025 17:03:35 -0700
From: Sean Christopherson <seanjc@...gle.com>
To: James Houghton <jthoughton@...gle.com>
Cc: Paolo Bonzini <pbonzini@...hat.com>, Jonathan Corbet <corbet@....net>, Marc Zyngier <maz@...nel.org>, 
	Oliver Upton <oliver.upton@...ux.dev>, Yan Zhao <yan.y.zhao@...el.com>, 
	Nikita Kalyazin <kalyazin@...zon.com>, Anish Moorthy <amoorthy@...gle.com>, 
	Peter Gonda <pgonda@...gle.com>, Peter Xu <peterx@...hat.com>, 
	David Matlack <dmatlack@...gle.com>, wei.w.wang@...el.com, kvm@...r.kernel.org, 
	linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org, 
	linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.linux.dev
Subject: Re: [PATCH v2 03/13] KVM: Allow late setting of KVM_MEM_USERFAULT on
 guest_memfd memslot

On Thu, Jan 09, 2025, James Houghton wrote:
> Currently guest_memfd memslots can only be deleted. Slightly change the
> logic to allow KVM_MR_FLAGS_ONLY changes when the only flag being
> changed is KVM_MEM_USERFAULT.
> 
> Signed-off-by: James Houghton <jthoughton@...gle.com>
> ---
>  virt/kvm/kvm_main.c | 15 +++++++++++----
>  1 file changed, 11 insertions(+), 4 deletions(-)
> 
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> index 4bceae6a6401..882c1f7b4aa8 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -2015,9 +2015,6 @@ int __kvm_set_memory_region(struct kvm *kvm,
>  		if ((kvm->nr_memslot_pages + npages) < kvm->nr_memslot_pages)
>  			return -EINVAL;
>  	} else { /* Modify an existing slot. */
> -		/* Private memslots are immutable, they can only be deleted. */
> -		if (mem->flags & KVM_MEM_GUEST_MEMFD)
> -			return -EINVAL;
>  		if ((mem->userspace_addr != old->userspace_addr) ||
>  		    (npages != old->npages) ||
>  		    ((mem->flags ^ old->flags) & KVM_MEM_READONLY))
> @@ -2031,6 +2028,16 @@ int __kvm_set_memory_region(struct kvm *kvm,
>  			return 0;
>  	}
>  
> +	/*
> +	 * Except for being able to set KVM_MEM_USERFAULT, private memslots are
> +	 * immutable, they can only be deleted.
> +	 */
> +	if (mem->flags & KVM_MEM_GUEST_MEMFD &&
> +	    !(change == KVM_MR_CREATE ||
> +	      (change == KVM_MR_FLAGS_ONLY &&
> +	       (mem->flags ^ old->flags) == KVM_MEM_USERFAULT)))
> +		return -EINVAL;

Oof.  I don't even want to decipher this.  Let's just drop the blankent immutable
restriction, and simply say guest_memfd slots can't be MOVED.  guest_memfd doesn't
support RO memslots (and never will), and doesn't support dirty logging, so this
is effectively dead code anyways.  Then the diff is much nicer:

@@ -2012,14 +2027,15 @@ static int kvm_set_memory_region(struct kvm *kvm,
                if ((kvm->nr_memslot_pages + npages) < kvm->nr_memslot_pages)
                        return -EINVAL;
        } else { /* Modify an existing slot. */
-               /* Private memslots are immutable, they can only be deleted. */
-               if (mem->flags & KVM_MEM_GUEST_MEMFD)
-                       return -EINVAL;
                if ((mem->userspace_addr != old->userspace_addr) ||
                    (npages != old->npages) ||
                    ((mem->flags ^ old->flags) & KVM_MEM_READONLY))
                        return -EINVAL;
 
+               /* Moving a guest_memfd memslot isn't supported. */
+               if (base_gfn != old->base_gfn && mem->flags & KVM_MEM_GUEST_MEMFD)
+                       return -EINVAL;
+
                if (base_gfn != old->base_gfn)
                        change = KVM_MR_MOVE;
                else if (mem->flags != old->flags)

> +
>  	if ((change == KVM_MR_CREATE || change == KVM_MR_MOVE) &&
>  	    kvm_check_memslot_overlap(slots, id, base_gfn, base_gfn + npages))
>  		return -EEXIST;
> @@ -2046,7 +2053,7 @@ int __kvm_set_memory_region(struct kvm *kvm,
>  	new->npages = npages;
>  	new->flags = mem->flags;
>  	new->userspace_addr = mem->userspace_addr;
> -	if (mem->flags & KVM_MEM_GUEST_MEMFD) {
> +	if (mem->flags & KVM_MEM_GUEST_MEMFD && change == KVM_MR_CREATE) {
>  		r = kvm_gmem_bind(kvm, new, mem->guest_memfd, mem->guest_memfd_offset);
>  		if (r)
>  			goto out;
> -- 
> 2.47.1.613.gc27f4b7a9f-goog
>