Date:   Mon, 13 Mar 2023 10:08:37 -0700
In-Reply-To: <ZA7eTpG5tpo5yIo3@gao-cwp>
Mime-Version: 1.0
References: <20230310214232.806108-1-seanjc@google.com> <20230310214232.806108-4-seanjc@google.com>
 <ZA7eTpG5tpo5yIo3@gao-cwp>
Message-ID: <ZA9YlTknLKRKcCy3@google.com>
Subject: Re: [PATCH v2 03/18] x86/reboot: Harden virtualization hooks for
 emergency reboot
From:   Sean Christopherson <seanjc@google.com>
To:     Chao Gao <chao.gao@intel.com>
Cc:     Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org,
        Paolo Bonzini <pbonzini@redhat.com>,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        Andrew Cooper <Andrew.Cooper3@citrix.com>,
        Kai Huang <kai.huang@intel.com>
Content-Type: text/plain; charset="us-ascii"
Precedence: bulk

On Mon, Mar 13, 2023, Chao Gao wrote:
> On Fri, Mar 10, 2023 at 01:42:17PM -0800, Sean Christopherson wrote:
> >+void cpu_emergency_register_virt_callback(cpu_emergency_virt_cb *callback)
> >+{
> >+	if (WARN_ON_ONCE(rcu_access_pointer(cpu_emergency_virt_callback)))
> >+		return;
> >+
> >+	rcu_assign_pointer(cpu_emergency_virt_callback, callback);
> 
> Was it intentional to not call synchronize_rcu() (in the original
> code), different from the un-registration path?

Yes, synchronize_rcu() is needed when removing a callback to ensure any in-flight
invocations of the callback complete before KVM is unloaded, i.e. to prevent
use-after-free of the KVM module code.  Registering a callback doesn't have the
same concerns, and adding a synchronize_rcu() wouldn't do anything in terms of
ensuring virtualization isn't enabled after an emergency restart/shutdown is
initiated.