From mboxrd@z Thu Jan 1 00:00:00 1970 From: Casey Leedom Subject: Re: KVM vs. PCI-E Function Level Reset (FLR) ... Date: Wed, 14 Jul 2010 17:55:26 -0700 Message-ID: <201007141755.26185.leedom@chelsio.com> References: <201007150033.o6F0XUBj024880@stargate.chelsio.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit To: kvm@vger.kernel.org Return-path: Received: from stargate.chelsio.com ([67.207.112.58]:10586 "EHLO stargate.chelsio.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757942Ab0GOBCd (ORCPT ); Wed, 14 Jul 2010 21:02:33 -0400 Received: from maui.asicdesigners.com (maui.asicdesigners.com [10.192.180.15]) by stargate.chelsio.com (8.13.1/8.13.1) with SMTP id o6F12WJW025182 for ; Wed, 14 Jul 2010 18:02:32 -0700 In-Reply-To: <201007150033.o6F0XUBj024880@stargate.chelsio.com> Sender: kvm-owner@vger.kernel.org List-ID: We've done a bit of testing and determined that the KVM support code does call pci_reset_function() from kvm_free_assigned_device() when a Virtual Machine is powered down. But pci_reset_function() is not called when a Virtual Machine is rebooted. It really should be reset for a reboot to track what happens to a Physical Host when a reboot is performed ... We've also determined that writing the FLR bit in the PCI-E Express Capability in an Assigned Device from the Guest OS/Driver does not result in a call to pci_reset_function() in the Hypervisor. This seems to be a problem ... Casey