From mboxrd@z Thu Jan  1 00:00:00 1970
From: Peter Teoh <htmldeveloper@gmail.com>
Subject: tlb flushing after spin_unlock of mmu lock
Date: Sat, 11 Apr 2009 00:45:21 -0400
Message-ID: <804dabb00904102145r446de3dfw6e2a287736fb170c@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
To: kvm@vger.kernel.org
Return-path: <kvm-owner@vger.kernel.org>
Received: from mail-gx0-f160.google.com ([209.85.217.160]:47046 "EHLO
	mail-gx0-f160.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752290AbZDKEpX (ORCPT <rfc822;kvm@vger.kernel.org>);
	Sat, 11 Apr 2009 00:45:23 -0400
Received: by gxk4 with SMTP id 4so3071437gxk.13
        for <kvm@vger.kernel.org>; Fri, 10 Apr 2009 21:45:21 -0700 (PDT)
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

In this function, the TLB flushing comes before spin unlock,

void kvm_mmu_slot_remove_write_access(struct kvm *kvm, int slot)
{
       struct kvm_mmu_page *sp;

       spin_lock(&kvm->mmu_lock);

       kvm_flush_remote_tlbs(kvm);
       spin_unlock(&kvm->mmu_lock);
}


but in kvm_vm_ioctl_set_memory_alias():

       spin_unlock(&kvm->mmu_lock);
       kvm_mmu_zap_all(kvm);

it comes after inside kvm_mmu_zap_all().   Does it sound logical?


-- 
Regards,
Peter Teoh