9:36 p.m.
Today, hotplug of a SCSI hostdev relies on the presence of an existing
SCSI controller, which can be defined in the initial domain XML. Hotplug
of a SCSI controller relies on the presence of a PCI bus, which does not
work in all environments.
These patches correct this behavior, in order to allow hotplug (and unplug)
work without incident.
Changes from v2->v3:
- Patch 2: Combine attach code within QEMU_CAPS_DEVICE check
- Patch 2: Restructure detach to cover ..._TYPE_NONE controllers
Changes from v1->v2:
- Patch 2: Check for VIR_..._TYPE_NONE on controller->info.type
- Patch 2: Check device address for TYPE_CCW controllers during detach
- Patch 1: Fixed an indentation change that was missed
Eric Farman (3):
qemu: Rename controller hotplug functions to not be PCI-specific
qemu: Separate calls based on controller bus type
qemu: Auto-generate controller for hotplugged hostdev
src/qemu/qemu_driver.c | 4 +--
src/qemu/qemu_hotplug.c | 62 +++++++++++++++++++++++++++++++++++------------
src/qemu/qemu_hotplug.h | 12 ++++-----
3 files changed, 54 insertions(+), 24 deletions(-)
--
1.7.9.5
9:36 p.m.
New subject: [libvirt] [PATCH 1/3] qemu: Rename controller hotplug functions to not be PCI-specific
For attach/detach of controller devices, we rename the functions to
remove 'PCI' from their title. The actual separation of PCI-specific
operations will be handled in the next patch.
Signed-off-by: Eric Farman <farman(a)linux.vnet.ibm.com>
---
src/qemu/qemu_driver.c | 4 ++--
src/qemu/qemu_hotplug.c | 16 ++++++++--------
src/qemu/qemu_hotplug.h | 12 ++++++------
3 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 378b542..065e104 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -6328,7 +6328,7 @@ qemuDomainAttachDeviceControllerLive(virQEMUDriverPtr driver,
switch (cont->type) {
case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
- ret = qemuDomainAttachPciControllerDevice(driver, vm, cont);
+ ret = qemuDomainAttachControllerDevice(driver, vm, cont);
break;
default:
virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
@@ -6420,7 +6420,7 @@ qemuDomainDetachDeviceControllerLive(virQEMUDriverPtr driver,
switch (cont->type) {
case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
- ret = qemuDomainDetachPciControllerDevice(driver, vm, dev);
+ ret = qemuDomainDetachControllerDevice(driver, vm, dev);
break;
default :
virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index 0d9a3aa..c81d0f3 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -348,9 +348,9 @@ error:
}
-int qemuDomainAttachPciControllerDevice(virQEMUDriverPtr driver,
- virDomainObjPtr vm,
- virDomainControllerDefPtr controller)
+int qemuDomainAttachControllerDevice(virQEMUDriverPtr driver,
+ virDomainObjPtr vm,
+ virDomainControllerDefPtr controller)
{
int ret = -1;
const char* type = virDomainControllerTypeToString(controller->type);
@@ -438,8 +438,8 @@ qemuDomainFindOrCreateSCSIDiskController(virQEMUDriverPtr driver,
cont->model = -1;
VIR_INFO("No SCSI controller present, hotplugging one");
- if (qemuDomainAttachPciControllerDevice(driver,
- vm, cont) < 0) {
+ if (qemuDomainAttachControllerDevice(driver,
+ vm, cont) < 0) {
VIR_FREE(cont);
return NULL;
}
@@ -3029,9 +3029,9 @@ static bool qemuDomainControllerIsBusy(virDomainObjPtr vm,
}
}
-int qemuDomainDetachPciControllerDevice(virQEMUDriverPtr driver,
- virDomainObjPtr vm,
- virDomainDeviceDefPtr dev)
+int qemuDomainDetachControllerDevice(virQEMUDriverPtr driver,
+ virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
{
int idx, ret = -1;
virDomainControllerDefPtr detach = NULL;
diff --git a/src/qemu/qemu_hotplug.h b/src/qemu/qemu_hotplug.h
index 355d809..75789d6 100644
--- a/src/qemu/qemu_hotplug.h
+++ b/src/qemu/qemu_hotplug.h
@@ -36,9 +36,9 @@ int qemuDomainChangeEjectableMedia(virQEMUDriverPtr driver,
int qemuDomainCheckEjectableMedia(virQEMUDriverPtr driver,
virDomainObjPtr vm,
enum qemuDomainAsyncJob asyncJob);
-int qemuDomainAttachPciControllerDevice(virQEMUDriverPtr driver,
- virDomainObjPtr vm,
- virDomainControllerDefPtr controller);
+int qemuDomainAttachControllerDevice(virQEMUDriverPtr driver,
+ virDomainObjPtr vm,
+ virDomainControllerDefPtr controller);
int qemuDomainAttachDeviceDiskLive(virConnectPtr conn,
virQEMUDriverPtr driver,
virDomainObjPtr vm,
@@ -78,9 +78,9 @@ int qemuDomainChangeNetLinkState(virQEMUDriverPtr driver,
int qemuDomainDetachDeviceDiskLive(virQEMUDriverPtr driver,
virDomainObjPtr vm,
virDomainDeviceDefPtr dev);
-int qemuDomainDetachPciControllerDevice(virQEMUDriverPtr driver,
- virDomainObjPtr vm,
- virDomainDeviceDefPtr dev);
+int qemuDomainDetachControllerDevice(virQEMUDriverPtr driver,
+ virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev);
int qemuDomainDetachNetDevice(virQEMUDriverPtr driver,
virDomainObjPtr vm,
virDomainDeviceDefPtr dev);
--
1.7.9.5
9:36 p.m.
New subject: [libvirt] [PATCH 2/3] qemu: Separate calls based on controller bus type
For systems without a PCI bus, attaching a SCSI controller fails:
[root@xxxxxxxx ~]# cat controller.xml
<controller type='scsi' model='virtio-scsi' index='0' />
[root@xxxxxxxx ~]# virsh attach-device guest01 controller.xml
error: Failed to attach device from controller.xml
error: XML error: No PCI buses available
A similar problem occurs with the detach of a controller:
[root@xxxxxxxx ~]# virsh detach-device guest01 controller.xml
error: Failed to detach device from controller.xml
error: operation failed: controller scsi:0 not found
The qemuDomainXXtachPciControllerDevice routines made assumptions
that any caller had a PCI bus. These routines now selectively calls
PCI functions where necessary, and assigns the device information
type to one appropriate for the bus in use.
Signed-off-by: Eric Farman <farman(a)linux.vnet.ibm.com>
---
src/qemu/qemu_hotplug.c | 41 +++++++++++++++++++++++++++++++++--------
1 file changed, 33 insertions(+), 8 deletions(-)
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index c81d0f3..99496e5 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -366,8 +366,22 @@ int qemuDomainAttachControllerDevice(virQEMUDriverPtr driver,
}
if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
- if (qemuDomainPCIAddressEnsureAddr(priv->pciaddrs, &controller->info)
< 0)
- goto cleanup;
+ if (controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
+ if (STRPREFIX(vm->def->os.machine, "s390-ccw") &&
+ virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_VIRTIO_CCW))
+ controller->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW;
+ else if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_VIRTIO_S390))
+ controller->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_S390;
+ }
+
+ if (controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE ||
+ controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) {
+ if (qemuDomainPCIAddressEnsureAddr(priv->pciaddrs,
&controller->info) < 0)
+ goto cleanup;
+ } else if (controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW) {
+ if (qemuDomainCCWAddressAssign(&controller->info, priv->ccwaddrs,
true) < 0)
+ goto cleanup;
+ }
releaseaddr = true;
if (qemuAssignDeviceControllerAlias(controller) < 0)
goto cleanup;
@@ -399,7 +413,8 @@ int qemuDomainAttachControllerDevice(virQEMUDriverPtr driver,
qemuDomainObjExitMonitor(driver, vm);
if (ret == 0) {
- controller->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
+ if (controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
+ controller->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
virDomainControllerInsertPreAlloced(vm->def, controller);
}
@@ -3049,14 +3064,24 @@ int qemuDomainDetachControllerDevice(virQEMUDriverPtr driver,
detach = vm->def->controllers[idx];
- if (!virDomainDeviceAddressIsValid(&detach->info,
- VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI)) {
- virReportError(VIR_ERR_OPERATION_FAILED, "%s",
- _("device cannot be detached without a PCI address"));
+ if (detach->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI &&
+ detach->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW &&
+ detach->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_S390) {
+ virReportError(VIR_ERR_OPERATION_FAILED,
+ _("device with '%s' address cannot be
detached"),
+ virDomainDeviceAddressTypeToString(detach->info.type));
goto cleanup;
}
- if (qemuIsMultiFunctionDevice(vm->def, &detach->info)) {
+ if (!virDomainDeviceAddressIsValid(&detach->info, detach->info.type)) {
+ virReportError(VIR_ERR_OPERATION_FAILED,
+ _("device with invalid '%s' address cannot be
detached"),
+ virDomainDeviceAddressTypeToString(detach->info.type));
+ goto cleanup;
+ }
+
+ if (detach->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI &&
+ qemuIsMultiFunctionDevice(vm->def, &detach->info)) {
virReportError(VIR_ERR_OPERATION_FAILED,
_("cannot hot unplug multifunction PCI device: %s"),
dev->data.disk->dst);
--
1.7.9.5
3:41 a.m.
New subject: [libvirt] [PATCH 2/3] qemu: Separate calls based on controller bus type
On 11/21/2013 04:36 AM, Eric Farman wrote:
For systems without a PCI bus, attaching a SCSI controller fails:
[root@xxxxxxxx ~]# cat controller.xml
<controller type='scsi' model='virtio-scsi' index='0' />
[root@xxxxxxxx ~]# virsh attach-device guest01 controller.xml
error: Failed to attach device from controller.xml
error: XML error: No PCI buses available
A similar problem occurs with the detach of a controller:
[root@xxxxxxxx ~]# virsh detach-device guest01 controller.xml
error: Failed to detach device from controller.xml
error: operation failed: controller scsi:0 not found
The qemuDomainXXtachPciControllerDevice routines made assumptions
that any caller had a PCI bus. These routines now selectively calls
PCI functions where necessary, and assigns the device information
type to one appropriate for the bus in use.
Signed-off-by: Eric Farman <farman(a)linux.vnet.ibm.com>
---
src/qemu/qemu_hotplug.c | 41 +++++++++++++++++++++++++++++++++--------
1 file changed, 33 insertions(+), 8 deletions(-)
ACK and I've pushed the series with the following changes:
+
+ if (controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE ||
+ controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) {
+ if (qemuDomainPCIAddressEnsureAddr(priv->pciaddrs,
&controller->info) < 0)
+ goto cleanup;
+ } else if (controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW) {
+ if (qemuDomainCCWAddressAssign(&controller->info,
priv->ccwaddrs, true) < 0)
I have changed this to:
if (qemuDomainCCWAddressAssign(&controller->info, priv->ccwaddrs,
!controller->info.addr.ccw.assigned) < 0)
to prevent overwriting the user-requested address and match the other
CCWAddressAssign calls
+ goto cleanup;
+ }
releaseaddr = true;
if (qemuAssignDeviceControllerAlias(controller) < 0)
goto cleanup;
+ }
+
+ if (detach->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI &&
s/!=/==/
+ qemuIsMultiFunctionDevice(vm->def, &detach->info))
{
virReportError(VIR_ERR_OPERATION_FAILED,
_("cannot hot unplug multifunction PCI device: %s"),
dev->data.disk->dst);
Jan
9:36 p.m.
New subject: [libvirt] [PATCH 3/3] qemu: Auto-generate controller for hotplugged hostdev
If a SCSI hostdev is included in an initial domain XML, without a
corresponding controller statement, one is created silently when the
guest is booted.
When hotplugging a SCSI hostdev, a presumption is that the controller
is already present in the domain either from the original XML, or via
an earlier hotplug.
[root@xxxxxxxx ~]# cat disk.xml
<hostdev mode='subsystem' type='scsi'>
<source>
<adapter name='scsi_host0'/>
<address bus='0' target='3' unit='1088438288'/>
</source>
</hostdev>
[root@xxxxxxxx ~]# virsh attach-device guest01 disk.xml
error: Failed to attach device from disk.xml
error: internal error: unable to execute QEMU command 'device_add': Bus
'scsi0.0' not found
Since the infrastructure is in place, we can also create a controller
silently for use by the hotplugged hostdev device.
Signed-off-by: Eric Farman <farman(a)linux.vnet.ibm.com>
---
src/qemu/qemu_hotplug.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index 99496e5..5f54c44 100644
--- a/src/qemu/qemu_hotplug.c
+++ b/src/qemu/qemu_hotplug.c
@@ -1508,6 +1508,7 @@ qemuDomainAttachHostScsiDevice(virQEMUDriverPtr driver,
{
int ret = -1;
qemuDomainObjPrivatePtr priv = vm->privateData;
+ virDomainControllerDefPtr cont = NULL;
char *devstr = NULL;
char *drvstr = NULL;
bool teardowncgroup = false;
@@ -1520,6 +1521,10 @@ qemuDomainAttachHostScsiDevice(virQEMUDriverPtr driver,
return -1;
}
+ cont = qemuDomainFindOrCreateSCSIDiskController(driver, vm,
hostdev->info->addr.drive.controller);
+ if (!cont)
+ return -1;
+
if (qemuPrepareHostdevSCSIDevices(driver, vm->def->name,
&hostdev, 1)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
--
1.7.9.5
4049
days inactive
4049
days old
4 comments
2 participants
participants (2)
-
Eric Farman -
Ján Tomko