From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paolo Bonzini Subject: Re: [PATCH uq/master 0/3] Fix MSI injection at load time Date: Tue, 30 Oct 2012 17:47:41 +0100 Message-ID: <509004AD.3010500@redhat.com> References: <1351599394-24876-1-git-send-email-pbonzini@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: qemu-devel@nongnu.org, avi@redhat.com, mtosatti@redhat.com, jan.kiszka@siemens.com, kvm@vger.kernel.org, Amit Shah To: unlisted-recipients:; (no To-header on input) Return-path: Received: from mail-we0-f174.google.com ([74.125.82.174]:33262 "EHLO mail-we0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S964874Ab2J3Qru (ORCPT ); Tue, 30 Oct 2012 12:47:50 -0400 Received: by mail-we0-f174.google.com with SMTP id t9so221262wey.19 for ; Tue, 30 Oct 2012 09:47:49 -0700 (PDT) In-Reply-To: <1351599394-24876-1-git-send-email-pbonzini@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: Il 30/10/2012 13:16, Paolo Bonzini ha scritto: > A simplified reproducer (that doesn't hang Linux, > but shows the message) is to start the VM without a backend for the > virtserialport, and to resume it with a backend, for example > > $ qemu-system-x86_64 -device virtio-serial-pci -device virtserialport test.img --enable-kvm -m 512 > $ qemu-system-x86_64 -device virtio-serial-pci -chardev stdio,id=vs0 -device virtserialport,chardev=vs0 test.img --enable-kvm -m 512 -incoming 'exec:cat foo.ckp' Jan, Amit, the same bug is also happening without MSI. The reproducer is the same as above, but with pci=nomsi for Linux guests. After migration, "cat /dev/vport0p1" will not block, and if you have a NIC on the same line as virtio-serial it will also not work. Do any of you have some time to look at it? Paolo