On Thu, Aug 19, 2021, Marc Orr wrote:

On Thu, Aug 19, 2021 at 3:58 PM Sean Christopherson [off-list ref] wrote:

quoted

On Thu, Aug 19, 2021, Peter Gonda wrote:

quoted

Marc I think that only having the spin lock could result in
deadlocking. If userspace double migrated 2 VMs, A and B for
discussion, A could grab VM_A.spin_lock then VM_A.kvm_mutex. Meanwhile
B could grab VM_B.spin_lock and VM_B.kvm_mutex. Then A attempts to
grab VM_B.spin_lock and we have a deadlock. If the same happens with
the proposed scheme when A attempts to lock B, VM_B.spin_lock will be
open but the bool will mark the VM under migration so A will unlock
and bail. Sean originally proposed a global spin lock but I thought a
per kvm_sev_info struct would also be safe.

Close.  The issue is taking kvm->lock from both VM_A and VM_B.  If userspace
double migrates we'll end up with lock ordering A->B and B-A, so we need a way
to guarantee one of those wins.  My proposed solution is to use a flag as a sort
of one-off "try lock" to detect a mean userspace.

Got it now. Thanks to you both, for the explanation. By the way, just
to make sure I completely follow, I assume that if a "double
migration" occurs, then user space is mis-behaving -- correct?

Yep.

But presumably, we need to reason about how to respond to such mis-behavior
so that buggy or malicious user-space code cannot stumble over/exploit this
scenario?

That's what the anti-deadlock flag is for. :-)  With that in place, there's no
meaningful difference between say a bad userspace doing double migrate and a bad
userspace migrating from garbage, e.g. passing in a bogus fd.