From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:43869) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1c7YBy-0004rU-DX for qemu-devel@nongnu.org; Thu, 17 Nov 2016 20:42:15 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1c7YBv-0006GU-7m for qemu-devel@nongnu.org; Thu, 17 Nov 2016 20:42:14 -0500 Received: from mx0a-001b2d01.pphosted.com ([148.163.156.1]:57639) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1c7YBu-0006FP-VL for qemu-devel@nongnu.org; Thu, 17 Nov 2016 20:42:11 -0500 Received: from pps.filterd (m0098399.ppops.net [127.0.0.1]) by mx0a-001b2d01.pphosted.com (8.16.0.17/8.16.0.17) with SMTP id uAI1cZYG076917 for ; Thu, 17 Nov 2016 20:42:09 -0500 Received: from e33.co.us.ibm.com (e33.co.us.ibm.com [32.97.110.151]) by mx0a-001b2d01.pphosted.com with ESMTP id 26sp8tvahd-1 (version=TLSv1.2 cipher=AES256-SHA bits=256 verify=NOT) for ; Thu, 17 Nov 2016 20:42:09 -0500 Received: from localhost by e33.co.us.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Thu, 17 Nov 2016 18:42:08 -0700 From: Michael Roth Date: Thu, 17 Nov 2016 19:40:25 -0600 In-Reply-To: <1479433227-29238-1-git-send-email-mdroth@linux.vnet.ibm.com> References: <1479433227-29238-1-git-send-email-mdroth@linux.vnet.ibm.com> Message-Id: <1479433227-29238-2-git-send-email-mdroth@linux.vnet.ibm.com> Subject: [Qemu-devel] [PATCH for-2.8 1/3] migration: alternative way to set instance_id in SaveStateEntry List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: qemu-ppc@nongnu.org, david@gibson.dropbear.id.au, duanj@linux.vnet.ibm.com, bharata@linux.vnet.ibm.com, dgilbert@redhat.com, quintela@redhat.com, amit.shah@redhat.com From: Jianjun Duan Currently migrated Devices are identified with an idstr which is calculated automatically using their path in the QOM composition tree. In some cases these Devices may have a more reliable identifier that may be preferable in situations where there's a chance their path in the composition tree might change in the future as a resulting of changes in how the device is modeled in QEMU. One such Device is the "spapr-dr-connector" used to handle hotplug for various resources on pseries machine types, where the PAPR specification mandates that each DRC have a 32-bit globally unique identifier associated with it. As such, this identifier is also ideal as a reliable way to identify a particular DRC in the migration stream, so we introduce support here for using a caller-side supplied instance_id for Devices in preparation for that. register_savevm_live() and vmstate_register_with_alias_id() already provide a means to let the caller supply an instance_id instead of relying on the migration subsystem to generate one automatically, but in cases where we're registering SaveVMHandlers or VMSDs (respectively) that are associated with a Device, the instance_id is ignored in favor of a string identifier based solely on the QOM path. This patch generalizes this so that an instance_id can also be supplied by the caller for Devices. Since VMSD registration for Devices is generally handled automatically by qdev, we also introduce a DeviceClass->dev_get_instance_id() method that, when set, is called by qdev to obtain the corresponding instance_id that should be used for a particular Device. Otherwise we maintain the original behavior of passing instance_id == -1 and thus falling back to the previous logic of using the QOM path. Signed-off-by: Jianjun Duan * moved usage of DeviceClass->dev_get_instance_id() out of savevm.c and into caller (qdev) instead * clarified usage/intent in comments and commit msg Signed-off-by: Michael Roth --- hw/core/qdev.c | 6 +++++- include/hw/qdev-core.h | 9 +++++++++ migration/savevm.c | 4 ++-- 3 files changed, 16 insertions(+), 3 deletions(-) diff --git a/hw/core/qdev.c b/hw/core/qdev.c index 5783442..c8c0c44 100644 --- a/hw/core/qdev.c +++ b/hw/core/qdev.c @@ -933,7 +933,11 @@ static void device_set_realized(Object *obj, bool value, Error **errp) } if (qdev_get_vmsd(dev)) { - vmstate_register_with_alias_id(dev, -1, qdev_get_vmsd(dev), dev, + int instance_id = DEVICE_GET_CLASS(dev)->dev_get_instance_id + ? DEVICE_GET_CLASS(dev)->dev_get_instance_id(dev) + : -1; + + vmstate_register_with_alias_id(dev, instance_id, qdev_get_vmsd(dev), dev, dev->instance_id_alias, dev->alias_required_for_version); } diff --git a/include/hw/qdev-core.h b/include/hw/qdev-core.h index 2c97347..8ba82af 100644 --- a/include/hw/qdev-core.h +++ b/include/hw/qdev-core.h @@ -139,6 +139,15 @@ typedef struct DeviceClass { qdev_initfn init; /* TODO remove, once users are converted to realize */ qdev_event exit; /* TODO remove, once users are converted to unrealize */ const char *bus_type; + + /* When this field is set, instead of using the device's QOM path, + * SaveStateEntry's for devices will be identified using a combination + * of the corresponding VMSD name and an instance_id returned by this + * function. This should only be necessary for situations where the + * QOM path is anticipated to change and a more stable identifier is + * desired to identify a device in the migration stream. + */ + int (*dev_get_instance_id)(DeviceState *dev); } DeviceClass; typedef struct NamedGPIOList NamedGPIOList; diff --git a/migration/savevm.c b/migration/savevm.c index 0363372..a95fff9 100644 --- a/migration/savevm.c +++ b/migration/savevm.c @@ -556,7 +556,7 @@ int register_savevm_live(DeviceState *dev, se->is_ram = 1; } - if (dev) { + if (dev && instance_id == -1) { char *id = qdev_get_dev_path(dev); if (id) { pstrcpy(se->idstr, sizeof(se->idstr), id); @@ -640,7 +640,7 @@ int vmstate_register_with_alias_id(DeviceState *dev, int instance_id, se->vmsd = vmsd; se->alias_id = alias_id; - if (dev) { + if (dev && instance_id == -1) { char *id = qdev_get_dev_path(dev); if (id) { pstrcpy(se->idstr, sizeof(se->idstr), id); -- 1.9.1