KVM: Disallow user memslot with size that exceeds "unsigned long"
authorSean Christopherson <seanjc@google.com>
Thu, 4 Nov 2021 00:25:03 +0000 (00:25 +0000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 8 Dec 2021 08:03:21 +0000 (09:03 +0100)
commit6a44f200f1f15d041dc381dd13dd840ac36af863
tree163fb1dda34238b7b2fd8f0fe1a97128e9850b41
parent775191dd4c1bcf546ab7ce87f3476537615437b7
KVM: Disallow user memslot with size that exceeds "unsigned long"

commit 6b285a5587506bae084cf9a3ed5aa491d623b91b upstream.

Reject userspace memslots whose size exceeds the storage capacity of an
"unsigned long".  KVM's uAPI takes the size as u64 to support large slots
on 64-bit hosts, but does not account for the size being truncated on
32-bit hosts in various flows.  The access_ok() check on the userspace
virtual address in particular casts the size to "unsigned long" and will
check the wrong number of bytes.

KVM doesn't actually support slots whose size doesn't fit in an "unsigned
long", e.g. KVM's internal kvm_memory_slot.npages is an "unsigned long",
not a "u64", and misc arch specific code follows that behavior.

Fixes: fa3d315a4ce2 ("KVM: Validate userspace_addr of memslot when registered")
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Message-Id: <20211104002531.1176691-3-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
virt/kvm/kvm_main.c