From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lucas Meneghel Rodrigues Subject: Problems with e1000 network card on qemu.git Date: Thu, 23 Sep 2010 08:58:43 -0300 Message-ID: <1285243123.2619.28.camel@freedom> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , Avi Kivity , aliguori@us.ibm.com To: qemu-devel@nongnu.org, KVM mailing list Return-path: Received: from mx1.redhat.com ([209.132.183.28]:50537 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752561Ab0IWL6w (ORCPT ); Thu, 23 Sep 2010 07:58:52 -0400 Sender: kvm-owner@vger.kernel.org List-ID: Hi folks: As most of you might know, we run some daily sanity and functional tests with both qemu-kvm.git and qemu.git. I decided to write asking for help with regards to what it appears to be a problem with the e1000 nw card (the default). Here is a list of what it fails pretty much every day for qemu.git: 1) Problems logging into guest with e1000 card kvm.qemu-git.RHEL.5.5.i386.e1000.boot FAIL Could not log into guest 'vm1' kvm.qemu-git.RHEL.5.5.i386.e1000.reboot FAIL Could not log into guest 'vm1' kvm.qemu-git.RHEL.5.5.i386.e1000.shutdown FAIL Could not log into guest 'vm1' ^ At first I thought this was some sort of intermittent, harmless problem, but it reproduces fairly frequently, thing that pretty much doesn't happen with virtio. kvm.qemu-git.RHEL.5.5.i386.virtio_net.boot GOOD completed successfully kvm.qemu-git.RHEL.5.5.i386.virtio_net.reboot GOOD completed successfully kvm.qemu-git.RHEL.5.5.i386.virtio_net.shutdown GOOD completed successfully Looking at test logs and screenshots the only thing that I found abnormal was: 2010-09-23 05:51:02: EXT3-fs: recovery complete. 2010-09-23 05:51:02: EXT3-fs: mounted filesystem with ordered data mode. 2010-09-23 05:51:05: type=1404 audit(1285235453.446:2): enforcing=1 old_enforcing=0 auid=4294967295 ses=4294967295 2010-09-23 05:51:06: type=1403 audit(1285235454.059:3): policy loaded auid=4294967295 ses=4294967295 2010-09-23 05:53:16: hdc: drive_cmd: status=0x41 { DriveReady Error } 2010-09-23 05:53:16: hdc: drive_cmd: error=0x04 { AbortedCommand } 2010-09-23 05:53:16: ide: failed opcode was: 0xec ^ Serial logs for the RHEL5.5 guest. If someone wants the full serial log, please let me know. 2) Qemu dies during unattended install kvm.qemu-git.RHEL.5.5.x86_64.e1000.unattended_install.cdrom ERROR Guest died before end of OS install ^ This is the bug https://bugs.launchpad.net/qemu/+bug/588955 Which is, according to preliminary investigation done, a bug on the slirp code. IMHO, we really need to start looking into those issues, even because e1000 is the default network card. Cheers, Lucas