From patchwork Sat Dec 17 10:43:44 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Wang, Wei W" X-Patchwork-Id: 9478999 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 59D9060827 for ; Sat, 17 Dec 2016 11:27:30 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 4798128581 for ; Sat, 17 Dec 2016 11:27:30 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 39A6D28553; Sat, 17 Dec 2016 11:27:30 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 9CDA128553 for ; Sat, 17 Dec 2016 11:27:29 +0000 (UTC) Received: from localhost ([::1]:36178 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cID9E-0001HP-Nt for patchwork-qemu-devel@patchwork.kernel.org; Sat, 17 Dec 2016 06:27:28 -0500 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53607) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1cICV7-0000GP-Aq for qemu-devel@nongnu.org; Sat, 17 Dec 2016 05:46:02 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1cICV4-0008UU-8t for qemu-devel@nongnu.org; Sat, 17 Dec 2016 05:46:01 -0500 Received: from mga05.intel.com ([192.55.52.43]:21540) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1cICV3-0007ss-Uu for qemu-devel@nongnu.org; Sat, 17 Dec 2016 05:45:58 -0500 Received: from orsmga001.jf.intel.com ([10.7.209.18]) by fmsmga105.fm.intel.com with ESMTP; 17 Dec 2016 02:45:57 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos; i="5.33,363,1477983600"; d="scan'208"; a="1073174654" Received: from devel-ww.sh.intel.com ([10.239.48.105]) by orsmga001.jf.intel.com with ESMTP; 17 Dec 2016 02:45:55 -0800 From: Wei Wang To: marcandre.lureau@gmail.com, mst@redhat.com, stefanha@redhat.com, pbonzini@redhat.com, qemu-devel@nongnu.org, virtio-dev@lists.oasis-open.org Date: Sat, 17 Dec 2016 18:43:44 +0800 Message-Id: <1481971427-11094-35-git-send-email-wei.w.wang@intel.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1481971427-11094-1-git-send-email-wei.w.wang@intel.com> References: <1481971427-11094-1-git-send-email-wei.w.wang@intel.com> X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 192.55.52.43 Subject: [Qemu-devel] [PATCH v1 34/37] vhost-pci-slave: add "peer_reset" X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Wei Wang Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" X-Virus-Scanned: ClamAV using ClamSMTP Setting "peer_reset" flags that the peer device is about to be reset. This happens when the slave side device driver only accepts a subset of the feature bits offered by the peer device. The peer device is expected to have a second run of the vhost-user protocol for the re-setup. Signed-off-by: Wei Wang --- hw/net/vhost-pci-net.c | 4 ++++ hw/virtio/vhost-pci-slave.c | 17 +++++++++++++++-- include/hw/virtio/vhost-pci-slave.h | 1 + 3 files changed, 20 insertions(+), 2 deletions(-) diff --git a/hw/net/vhost-pci-net.c b/hw/net/vhost-pci-net.c index 567fb86..6d4db6c 100644 --- a/hw/net/vhost-pci-net.c +++ b/hw/net/vhost-pci-net.c @@ -141,6 +141,10 @@ static void vpnet_set_features(VirtIODevice *vdev, uint64_t features) int ret; if (need_send) { + if (vp_slave->feature_bits != features) { + vp_slave->feature_bits = features; + vp_slave->peer_reset = 1; + } need_send = 0; ret = vp_slave_send_feature_bits(features); if (ret < 0) diff --git a/hw/virtio/vhost-pci-slave.c b/hw/virtio/vhost-pci-slave.c index d3c3443..5cfa335 100644 --- a/hw/virtio/vhost-pci-slave.c +++ b/hw/virtio/vhost-pci-slave.c @@ -37,6 +37,14 @@ static void vp_slave_cleanup(void) if (!vp_slave->vdev) return; + /* + * if this cleanup is not invoked due to peer reset, re-initialize + * the feature bits, so that the slave can work with a new master + */ + if (!vp_slave->peer_reset) + vp_slave->feature_bits = + 1ULL << VHOST_USER_F_PROTOCOL_FEATURES; + nregions = vp_slave->pmem_msg.nregions; for (i = 0; i < nregions; i++) { ret = munmap(vp_slave->mr_map_base[i], vp_slave->mr_map_size[i]); @@ -163,8 +171,10 @@ static void vp_slave_set_device_type(VhostUserMsg *msg) switch (vp_slave->dev_type) { case VIRTIO_ID_NET: - vp_slave->feature_bits |= (VHOST_PCI_FEATURE_BITS - | VHOST_PCI_NET_FEATURE_BITS); + /* Don't initialize the feature bits if they have been negotiated */ + if (!vp_slave->peer_reset) + vp_slave->feature_bits |= (VHOST_PCI_FEATURE_BITS + | VHOST_PCI_NET_FEATURE_BITS); break; default: error_report("device type %d is not supported", vp_slave->dev_type); @@ -351,6 +361,8 @@ static int vp_slave_set_vhost_pci(CharBackend *chr_be, VhostUserMsg *msg) switch (cmd) { case VHOST_USER_SET_VHOST_PCI_start: + if (vp_slave->peer_reset) + vp_slave->peer_reset = 0; ret = vp_slave_device_create(vp_slave->dev_type); if (ret < 0) return ret; @@ -519,6 +531,7 @@ int vhost_pci_slave_init(QemuOpts *opts) vp_slave->sub_mr = NULL; QLIST_INIT(&vp_slave->pvq_list); vp_slave->pvq_num = 0; + vp_slave->peer_reset = 0; qemu_chr_fe_init(&vp_slave->chr_be, chr, &error_abort); qemu_chr_fe_set_handlers(&vp_slave->chr_be, vp_slave_can_read, vp_slave_read, vp_slave_event, diff --git a/include/hw/virtio/vhost-pci-slave.h b/include/hw/virtio/vhost-pci-slave.h index e60580a..dd78ed4 100644 --- a/include/hw/virtio/vhost-pci-slave.h +++ b/include/hw/virtio/vhost-pci-slave.h @@ -21,6 +21,7 @@ typedef struct VhostPCISlave { CharBackend chr_be; /* Ponnter to the slave device */ VirtIODevice *vdev; + bool peer_reset; uint16_t dev_type; uint64_t feature_bits; /* hotplugged memory should be mapped following the offset */