9:44 p.m.
This are some additional features to libxl driver. Some of them require change
in domain config structures/syntax. Details described with each patch.
The last two patches are bugfix for deadlock during daemon startup.
Marek Marczykowski (12):
libxl: allow script for any network interface, not only bridge
libxl: PCI passthrough support
libxl: nodeDevice* support for PCI devices
libxl: populate xenstore memory entries at startup
conf: add 'script' attribute to disk specification
libxl: use disk 'script' attribute
conf: support backend domain name in disk and network devices
libxl: support backend domain setting for disk and net devices
libxl: fill HVM SDL and VNC settings based on <graphics/> entries
RFC: libxl: special 'stubdom-dm' emulator to use qemu in stub domain
conf: virDomainObjListRemoveLocked function
libxl: fix deadlock in libxlReconnectDomain
docs/schemas/domaincommon.rng | 14 ++
src/conf/domain_conf.c | 54 ++++++++
src/conf/domain_conf.h | 5 +
src/libvirt_private.syms | 1 +
src/libxl/libxl_conf.c | 297 +++++++++++++++++++++++++++++++++---------
src/libxl/libxl_conf.h | 6 +-
src/libxl/libxl_driver.c | 253 +++++++++++++++++++++++++++++++----
7 files changed, 545 insertions(+), 85 deletions(-)
--
1.8.1.4
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 12/12] libxl: fix deadlock in libxlReconnectDomain
Use virDomainObjListRemoveLocked instead of virDomainObjListRemove, as
driver->domains is already taken by virDomainObjListForEach.
Above deadlock can be triggered when libvirtd is started after some
domain have been started by hand (in which case driver will not find
libvirt-xml domain config).
---
src/libxl/libxl_driver.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index b0f0c6a..687190c 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -1060,7 +1060,7 @@ libxlReconnectDomain(virDomainObjPtr vm,
out:
libxlVmCleanup(driver, vm, VIR_DOMAIN_SHUTOFF_UNKNOWN);
if (!vm->persistent)
- virDomainObjListRemove(driver->domains, vm);
+ virDomainObjListRemoveLocked(driver->domains, vm);
else
virObjectUnlock(vm);
--
1.8.1.4
12:48 p.m.
New subject: [libvirt] [PATCH RESENT 12/12] libxl: fix deadlock in libxlReconnectDomain
Marek Marczykowski wrote:
Use virDomainObjListRemoveLocked instead of virDomainObjListRemove,
as
driver->domains is already taken by virDomainObjListForEach.
Above deadlock can be triggered when libvirtd is started after some
domain have been started by hand (in which case driver will not find
libvirt-xml domain config).
Looks good, but let's see what others have to say about the
virDomainObjListRemoveLocked() addition.
Regards,
Jim
---
src/libxl/libxl_driver.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index b0f0c6a..687190c 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -1060,7 +1060,7 @@ libxlReconnectDomain(virDomainObjPtr vm,
out:
libxlVmCleanup(driver, vm, VIR_DOMAIN_SHUTOFF_UNKNOWN);
if (!vm->persistent)
- virDomainObjListRemove(driver->domains, vm);
+ virDomainObjListRemoveLocked(driver->domains, vm);
else
virObjectUnlock(vm);
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 01/12] libxl: allow script for any network interface, not only bridge
This can be useful for route or NAT networks, or any other custom
network setup. Especially configuration example in documentation uses
<script/> tag with type 'ethernet'.
---
src/libxl/libxl_conf.c | 17 +++++------------
1 file changed, 5 insertions(+), 12 deletions(-)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index b208dd8..ffc7bbb 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -608,18 +608,11 @@ libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic)
virReportOOMError();
return -1;
}
- if (l_nic->script &&
- (x_nic->script = strdup(l_nic->script)) == NULL) {
- virReportOOMError();
- return -1;
- }
- } else {
- if (l_nic->script) {
- virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
- _("scripts are not supported on interfaces of type
%s"),
- virDomainNetTypeToString(l_nic->type));
- return -1;
- }
+ }
+ if (l_nic->script &&
+ (x_nic->script = strdup(l_nic->script)) == NULL) {
+ virReportOOMError();
+ return -1;
}
return 0;
--
1.8.1.4
4:10 a.m.
New subject: [libvirt] [PATCH RESENT 01/12] libxl: allow script for any network interface, not only bridge
On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
This can be useful for route or NAT networks, or any other custom
network setup. Especially configuration example in documentation uses
<script/> tag with type 'ethernet'.
---
src/libxl/libxl_conf.c | 17 +++++------------
1 file changed, 5 insertions(+), 12 deletions(-)
The script should not not have been allowed for type='bridge' in
the first place, it is only intended for type='ethernet' usage
and nothing else.
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
11:59 a.m.
New subject: [libvirt] [PATCH RESENT 01/12] libxl: allow script for any network interface, not only bridge
On 04/10/2013 05:10 AM, Daniel P. Berrange wrote:
On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
> This can be useful for route or NAT networks, or any other custom
> network setup. Especially configuration example in documentation uses
> <script/> tag with type 'ethernet'.
> ---
> src/libxl/libxl_conf.c | 17 +++++------------
> 1 file changed, 5 insertions(+), 12 deletions(-)
The script should not not have been allowed for type='bridge' in
the first place, it is only intended for type='ethernet' usage
and nothing else.
I thought that it was also allowed/necessary for type='bridge' in xen
domains. Is this incorrect?
(I agree that it shouldn't be allowed for anything new, though)
10:30 p.m.
New subject: [libvirt] [PATCH RESENT 01/12] libxl: allow script for any network interface, not only bridge
Laine Stump wrote:
On 04/10/2013 05:10 AM, Daniel P. Berrange wrote:
> On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
>
>> This can be useful for route or NAT networks, or any other custom
>> network setup. Especially configuration example in documentation uses
>> <script/> tag with type 'ethernet'.
>> ---
>> src/libxl/libxl_conf.c | 17 +++++------------
>> 1 file changed, 5 insertions(+), 12 deletions(-)
>>
> The script should not not have been allowed for type='bridge' in
> the first place, it is only intended for type='ethernet' usage
> and nothing else.
>
I thought that it was also allowed/necessary for type='bridge' in xen
domains. Is this incorrect?
I think it is only necessary if something other than the default
(/etc/xen/scripts/vif-bridge) is desired. It has been allowed in both
xen drivers for as long as I remember.
(I agree that it shouldn't be allowed for anything new, though)
Agreed, ethernet and bridge only.
Regards,
Jim
4:05 p.m.
New subject: [libvirt] [PATCH RESENT 01/12] libxl: allow script for any network interface, not only bridge
Jim Fehlig wrote:
Laine Stump wrote:
> On 04/10/2013 05:10 AM, Daniel P. Berrange wrote:
>
>
>> On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
>>
>>
>>> This can be useful for route or NAT networks, or any other custom
>>> network setup. Especially configuration example in documentation uses
>>> <script/> tag with type 'ethernet'.
>>> ---
>>> src/libxl/libxl_conf.c | 17 +++++------------
>>> 1 file changed, 5 insertions(+), 12 deletions(-)
>>>
>>>
>> The script should not not have been allowed for type='bridge' in
>> the first place, it is only intended for type='ethernet' usage
>> and nothing else.
>>
>>
> I thought that it was also allowed/necessary for type='bridge' in xen
> domains. Is this incorrect?
>
>
I think it is only necessary if something other than the default
(/etc/xen/scripts/vif-bridge) is desired. It has been allowed in both
xen drivers for as long as I remember.
Any consensus here? I double-checked the legacy xen driver and it does
in fact support <script> for <interface type='bridge'>. I'm not
sure how
many users specify something other than the default, but I'm loath to
break them when moving from the old xen toolstack to libxl.
Marek,
Assuming we continue to allow <script> for type='bridge' interfaces in
the libxl driver, you'll have to change your patch to only allow it for
type 'bridge' and 'ethernet'.
BTW, sorry for the delay in getting to your patches. I have a bit of
time to work on libxl stuff and will get to reviewing them.
Regards,
Jim
9:23 a.m.
New subject: [libvirt] [PATCH RESENT 01/12] libxl: allow script for any network interface, not only bridge
Jim Fehlig wrote:
Jim Fehlig wrote:
> Laine Stump wrote:
>
>
>> On 04/10/2013 05:10 AM, Daniel P. Berrange wrote:
>>
>>
>>
>>> On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
>>>
>>>
>>>
>>>> This can be useful for route or NAT networks, or any other custom
>>>> network setup. Especially configuration example in documentation uses
>>>> <script/> tag with type 'ethernet'.
>>>> ---
>>>> src/libxl/libxl_conf.c | 17 +++++------------
>>>> 1 file changed, 5 insertions(+), 12 deletions(-)
>>>>
>>>>
>>>>
>>> The script should not not have been allowed for type='bridge' in
>>> the first place, it is only intended for type='ethernet' usage
>>> and nothing else.
>>>
>>>
>>>
>> I thought that it was also allowed/necessary for type='bridge' in xen
>> domains. Is this incorrect?
>>
>>
>>
> I think it is only necessary if something other than the default
> (/etc/xen/scripts/vif-bridge) is desired. It has been allowed in both
> xen drivers for as long as I remember.
>
>
Any consensus here? I double-checked the legacy xen driver and it does
in fact support <script> for <interface type='bridge'>. I'm not
sure how
many users specify something other than the default, but I'm loath to
break them when moving from the old xen toolstack to libxl.
Marek,
Assuming we continue to allow <script> for type='bridge' interfaces in
the libxl driver, you'll have to change your patch to only allow it for
type 'bridge' and 'ethernet'.
Please send a V2 with this change.
Regards,
Jim
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
---
src/libxl/libxl_conf.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++++
src/libxl/libxl_conf.h | 2 ++
2 files changed, 74 insertions(+)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index ffc7bbb..7668305 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -759,6 +759,74 @@ error:
return -1;
}
+int
+libxlMakePci(virDomainHostdevDefPtr l_hostdev,
+ libxl_device_pci *x_pci)
+{
+ if (l_hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS) {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("This driver
supports only subsystem host device mode (PCI devices only)"));
+ return -1;
+ }
+ if (l_hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI) {
+ virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("This driver
supports only PCI host devices"));
+ return -1;
+ }
+
+ x_pci->bus = l_hostdev->source.subsys.u.pci.bus;
+ x_pci->dev = l_hostdev->source.subsys.u.pci.slot;
+ x_pci->func = l_hostdev->source.subsys.u.pci.function;
+ return 0;
+}
+
+static int
+libxlMakePciList(virDomainDefPtr def,
+ libxl_domain_config *d_config)
+{
+ virDomainHostdevDefPtr *l_hostdevs = def->hostdevs;
+ int nhostdevs = def->nhostdevs;
+ int npcidevs = 0;
+ libxl_device_pci *x_pcidevs;
+ int i, j;
+
+ if (nhostdevs == 0)
+ return 0;
+
+ /* count PCI devices, ignore others (USB) */
+ for (i = 0; i < nhostdevs; i++) {
+ if (l_hostdevs[i]->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
+ l_hostdevs[i]->source.subsys.type ==
VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI)
+ npcidevs++;
+ }
+ if (VIR_ALLOC_N(x_pcidevs, npcidevs) < 0) {
+ virReportOOMError();
+ return -1;
+ }
+
+ for (i = 0, j = 0; i < nhostdevs; i++) {
+ if (l_hostdevs[i]->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS ||
+ l_hostdevs[i]->source.subsys.type !=
VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI)
+ continue;
+
+ libxl_device_pci_init(&x_pcidevs[j]);
+
+ if (libxlMakePci(l_hostdevs[i], &x_pcidevs[j]) < 0)
+ goto error;
+ j++;
+ }
+
+ d_config->pcidevs = x_pcidevs;
+ d_config->num_pcidevs = npcidevs;
+
+ return 0;
+
+error:
+ for (i = 0; i < npcidevs; i++) {
+ libxl_device_pci_dispose(&x_pcidevs[i]);
+ }
+ VIR_FREE(x_pcidevs);
+ return -1;
+}
+
virCapsPtr
libxlMakeCapabilities(libxl_ctx *ctx)
{
@@ -817,6 +885,10 @@ libxlBuildDomainConfig(libxlDriverPrivatePtr driver,
goto error;
}
+ if (libxlMakePciList(def, d_config) < 0) {
+ goto error;
+ }
+
d_config->on_reboot = def->onReboot;
d_config->on_poweroff = def->onPoweroff;
d_config->on_crash = def->onCrash;
diff --git a/src/libxl/libxl_conf.h b/src/libxl/libxl_conf.h
index f8602b4..b3ab3bf 100644
--- a/src/libxl/libxl_conf.h
+++ b/src/libxl/libxl_conf.h
@@ -119,6 +119,8 @@ libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic);
int
libxlMakeVfb(libxlDriverPrivatePtr driver,
virDomainGraphicsDefPtr l_vfb, libxl_device_vfb *x_vfb);
+int
+libxlMakePci(virDomainHostdevDefPtr l_hostdev, libxl_device_pci *x_pci);
int
libxlBuildDomainConfig(libxlDriverPrivatePtr driver,
--
1.8.1.4
4:13 a.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
---
src/libxl/libxl_conf.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++++
src/libxl/libxl_conf.h | 2 ++
2 files changed, 74 insertions(+)
This needs todo more than just create the config. You need to use
the libvirt PCI APIs to validate that it is safe to assign the
devices requested, and track which devices are given to which
guest, to avoid concurrent assignemnt to 2 guests. Basically I'd
expect this to borrow all the code currently in the QMEU driver
related to the 'activePciHostdevs' and 'inactivePciHostdevs'
data structures.
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
10:50 p.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
Daniel P. Berrange wrote:
On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
> ---
> src/libxl/libxl_conf.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++++
> src/libxl/libxl_conf.h | 2 ++
> 2 files changed, 74 insertions(+)
>
This needs todo more than just create the config. You need to use
the libvirt PCI APIs to validate that it is safe to assign the
devices requested, and track which devices are given to which
guest, to avoid concurrent assignemnt to 2 guests. Basically I'd
expect this to borrow all the code currently in the QMEU driver
related to the 'activePciHostdevs' and 'inactivePciHostdevs'
data structures.
Right. Chunyan already posted such a patch series [1], but Laine noted
that it would be useful to maintain the state of PCI device assignment
for coordination among multiple drivers. Chunyan posted a design [2]
for a "hostdev passthrough driver", and IIRC she has started coding
after receiving feedback from the list.
Regards,
Jim
[1] https://www.redhat.com/archives/libvir-list/2013-January/msg00697.html
[2] https://www.redhat.com/archives/libvir-list/2013-March/msg01331.html
7:43 p.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
On 11.04.2013 05:50, Jim Fehlig wrote:
Right. Chunyan already posted such a patch series [1], but Laine
noted
that it would be useful to maintain the state of PCI device assignment
for coordination among multiple drivers. Chunyan posted a design [2]
for a "hostdev passthrough driver", and IIRC she has started coding
after receiving feedback from the list.
Regards,
Jim
[1] https://www.redhat.com/archives/libvir-list/2013-January/msg00697.html
[2] https://www.redhat.com/archives/libvir-list/2013-March/msg01331.html
Chunyan, any progress on this? Perhaps I can help you somehow?
--
Best Regards / Pozdrawiam,
Marek Marczykowski
Invisible Things Lab
10:22 p.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
In testing, could send out in a couple of days I think.
2013/5/3 Marek Marczykowski <marmarek(a)invisiblethingslab.com>
On 11.04.2013 05:50, Jim Fehlig wrote:
> Right. Chunyan already posted such a patch series [1], but Laine noted
> that it would be useful to maintain the state of PCI device assignment
> for coordination among multiple drivers. Chunyan posted a design [2]
> for a "hostdev passthrough driver", and IIRC she has started coding
> after receiving feedback from the list.
>
> Regards,
> Jim
>
> [1]
https://www.redhat.com/archives/libvir-list/2013-January/msg00697.html
> [2] https://www.redhat.com/archives/libvir-list/2013-March/msg01331.html
Chunyan, any progress on this? Perhaps I can help you somehow?
--
Best Regards / Pozdrawiam,
Marek Marczykowski
Invisible Things Lab
9:33 a.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
Jim Fehlig wrote:
Daniel P. Berrange wrote:
> On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
>
>
>> ---
>> src/libxl/libxl_conf.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++++
>> src/libxl/libxl_conf.h | 2 ++
>> 2 files changed, 74 insertions(+)
>>
>>
> This needs todo more than just create the config. You need to use
> the libvirt PCI APIs to validate that it is safe to assign the
> devices requested, and track which devices are given to which
> guest, to avoid concurrent assignemnt to 2 guests. Basically I'd
> expect this to borrow all the code currently in the QMEU driver
> related to the 'activePciHostdevs' and 'inactivePciHostdevs'
> data structures.
>
>
Right. Chunyan already posted such a patch series [1], but Laine noted
that it would be useful to maintain the state of PCI device assignment
for coordination among multiple drivers. Chunyan posted a design [2]
for a "hostdev passthrough driver", and IIRC she has started coding
after receiving feedback from the list.
I see that Chunyan has now posted patches for this work
https://www.redhat.com/archives/libvir-list/2013-May/msg01162.html
Could you can help review and test those?
Regards,
Jim
6:38 p.m.
New subject: [libvirt] [PATCH RESENT 02/12] libxl: PCI passthrough support
On 22.05.2013 16:33, Jim Fehlig wrote:
Jim Fehlig wrote:
> Daniel P. Berrange wrote:
>
>> On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
>>
>>
>>> ---
>>> src/libxl/libxl_conf.c | 72
++++++++++++++++++++++++++++++++++++++++++++++++++
>>> src/libxl/libxl_conf.h | 2 ++
>>> 2 files changed, 74 insertions(+)
>>>
>>>
>> This needs todo more than just create the config. You need to use
>> the libvirt PCI APIs to validate that it is safe to assign the
>> devices requested, and track which devices are given to which
>> guest, to avoid concurrent assignemnt to 2 guests. Basically I'd
>> expect this to borrow all the code currently in the QMEU driver
>> related to the 'activePciHostdevs' and 'inactivePciHostdevs'
>> data structures.
>>
>>
>
> Right. Chunyan already posted such a patch series [1], but Laine noted
> that it would be useful to maintain the state of PCI device assignment
> for coordination among multiple drivers. Chunyan posted a design [2]
> for a "hostdev passthrough driver", and IIRC she has started coding
> after receiving feedback from the list.
>
I see that Chunyan has now posted patches for this work
https://www.redhat.com/archives/libvir-list/2013-May/msg01162.html
Could you can help review and test those?
Sure, will look at it tomorrow.
--
Best Regards,
Marek Marczykowski
Invisible Things Lab
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 03/12] libxl: nodeDevice* support for PCI devices
For now only for PCI devices. Mostly copy-paste from old xen driver.
---
src/libxl/libxl_driver.c | 193 +++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 193 insertions(+)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 40a7a6b..011edf8 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -42,6 +42,7 @@
#include "libxl.h"
#include "libxl_driver.h"
#include "libxl_conf.h"
+#include "node_device_conf.h"
#include "xen_xm.h"
#include "virtypedparam.h"
#include "viruri.h"
@@ -3666,6 +3667,195 @@ libxlDomainUpdateDeviceFlags(virDomainPtr dom, const char *xml,
return libxlDomainModifyDeviceFlags(dom, xml, flags, LIBXL_DEVICE_UPDATE);
}
+static int
+libxlNodeDeviceGetPciInfo(virNodeDevicePtr dev,
+ unsigned *domain,
+ unsigned *bus,
+ unsigned *slot,
+ unsigned *function)
+{
+ virNodeDeviceDefPtr def = NULL;
+ virNodeDevCapsDefPtr cap;
+ char *xml = NULL;
+ int ret = -1;
+
+ xml = virNodeDeviceGetXMLDesc(dev, 0);
+ if (!xml)
+ goto out;
+
+ def = virNodeDeviceDefParseString(xml, EXISTING_DEVICE, NULL);
+ if (!def)
+ goto out;
+
+ cap = def->caps;
+ while (cap) {
+ if (cap->type == VIR_NODE_DEV_CAP_PCI_DEV) {
+ *domain = cap->data.pci_dev.domain;
+ *bus = cap->data.pci_dev.bus;
+ *slot = cap->data.pci_dev.slot;
+ *function = cap->data.pci_dev.function;
+ break;
+ }
+
+ cap = cap->next;
+ }
+
+ if (!cap) {
+ virReportError(VIR_ERR_INVALID_ARG,
+ _("device %s is not a PCI device"), dev->name);
+ goto out;
+ }
+
+ ret = 0;
+out:
+ virNodeDeviceDefFree(def);
+ VIR_FREE(xml);
+ return ret;
+}
+
+static int
+libxlNodeDeviceDettach(virNodeDevicePtr dev)
+{
+ virPCIDevicePtr pci;
+ unsigned domain, bus, slot, function;
+ int ret = -1;
+
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ return -1;
+
+ pci = virPCIDeviceNew(domain, bus, slot, function);
+ if (!pci)
+ return -1;
+
+ if (virPCIDeviceDetach(pci, NULL, NULL, "pciback") < 0)
+ goto out;
+
+ ret = 0;
+out:
+ virPCIDeviceFree(pci);
+ return ret;
+}
+
+static int
+libxlNodeDeviceAssignedDomainId(virNodeDevicePtr dev)
+{
+ int numdomains;
+ int numpcidevs;
+ int ret = -1, i, j;
+ int *ids = NULL;
+ char *bdf = NULL;
+ char *xref = NULL;
+ unsigned int domain, bus, slot, function;
+ libxl_device_pci *pcidevs = NULL;
+ virConnectPtr conn = dev->conn;
+ libxlDriverPrivatePtr driver = conn->privateData;
+
+ /* Get active domains */
+ numdomains = libxlNumDomains(conn);
+ if (numdomains < 0) {
+ return ret;
+ }
+ if (numdomains > 0){
+ if (VIR_ALLOC_N(ids, numdomains) < 0) {
+ virReportOOMError();
+ goto out;
+ }
+ if ((numdomains = libxlListDomains(conn, &ids[0], numdomains)) < 0) {
+ goto out;
+ }
+ }
+
+ /* Get pci bdf */
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ goto out;
+
+ if (virAsprintf(&bdf, "%04x:%02x:%02x.%0x",
+ domain, bus, slot, function) < 0) {
+ virReportOOMError();
+ goto out;
+ }
+
+ libxlDriverLock(driver);
+ /* Check if bdf is assigned to one of active domains */
+ for (i = 0; i < numdomains; i++) {
+ pcidevs = libxl_device_pci_list(driver->ctx, ids[i], &numpcidevs);
+ if (pcidevs == NULL) {
+ continue;
+ } else {
+ for (j = 0; j < numpcidevs; j++) {
+ if (pcidevs[j].bus == bus && pcidevs[j].dev == slot &&
pcidevs[j].func == function) {
+ ret = ids[i];
+ break;
+ }
+ }
+ VIR_FREE(pcidevs);
+ }
+ }
+ libxlDriverUnlock(driver);
+
+ VIR_FREE(xref);
+ VIR_FREE(bdf);
+out:
+ VIR_FREE(ids);
+
+ return ret;
+}
+
+static int
+libxlNodeDeviceReAttach(virNodeDevicePtr dev)
+{
+ virPCIDevicePtr pci;
+ unsigned domain, bus, slot, function;
+ int ret = -1;
+ int domid;
+
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ return -1;
+
+ pci = virPCIDeviceNew(domain, bus, slot, function);
+ if (!pci)
+ return -1;
+
+ /* Check if device is assigned to an active guest */
+ if ((domid = libxlNodeDeviceAssignedDomainId(dev)) >= 0) {
+ virReportError(VIR_ERR_INTERNAL_ERROR,
+ _("Device %s has been assigned to guest %d"),
+ dev->name, domid);
+ goto out;
+ }
+
+ if (virPCIDeviceReattach(pci, NULL, NULL, "pciback") < 0)
+ goto out;
+
+ ret = 0;
+out:
+ virPCIDeviceFree(pci);
+ return ret;
+}
+
+static int
+libxlNodeDeviceReset(virNodeDevicePtr dev)
+{
+ virPCIDevicePtr pci;
+ unsigned domain, bus, slot, function;
+ int ret = -1;
+
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ return -1;
+
+ pci = virPCIDeviceNew(domain, bus, slot, function);
+ if (!pci)
+ return -1;
+
+ if (virPCIDeviceReset(pci, NULL, NULL) < 0)
+ goto out;
+
+ ret = 0;
+out:
+ virPCIDeviceFree(pci);
+ return ret;
+}
+
static unsigned long long
libxlNodeGetFreeMemory(virConnectPtr conn)
{
@@ -4233,6 +4423,9 @@ static virDriver libxlDriver = {
.nodeGetFreeMemory = libxlNodeGetFreeMemory, /* 0.9.0 */
.domainEventRegister = libxlDomainEventRegister, /* 0.9.0 */
.domainEventDeregister = libxlDomainEventDeregister, /* 0.9.0 */
+ .nodeDeviceDettach = libxlNodeDeviceDettach, /* 1.0.4 */
+ .nodeDeviceReAttach = libxlNodeDeviceReAttach, /* 1.0.4 */
+ .nodeDeviceReset = libxlNodeDeviceReset, /* 1.0.4 */
.domainManagedSave = libxlDomainManagedSave, /* 0.9.2 */
.domainHasManagedSaveImage = libxlDomainHasManagedSaveImage, /* 0.9.2 */
.domainManagedSaveRemove = libxlDomainManagedSaveRemove, /* 0.9.2 */
--
1.8.1.4
9:36 a.m.
New subject: [libvirt] [PATCH RESENT 03/12] libxl: nodeDevice* support for PCI devices
Marek Marczykowski wrote:
For now only for PCI devices. Mostly copy-paste from old xen driver.
This one is (or will be) covered by Chanyan's work as well right?
Regards,
Jim
---
src/libxl/libxl_driver.c | 193 +++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 193 insertions(+)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 40a7a6b..011edf8 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -42,6 +42,7 @@
#include "libxl.h"
#include "libxl_driver.h"
#include "libxl_conf.h"
+#include "node_device_conf.h"
#include "xen_xm.h"
#include "virtypedparam.h"
#include "viruri.h"
@@ -3666,6 +3667,195 @@ libxlDomainUpdateDeviceFlags(virDomainPtr dom, const char *xml,
return libxlDomainModifyDeviceFlags(dom, xml, flags, LIBXL_DEVICE_UPDATE);
}
+static int
+libxlNodeDeviceGetPciInfo(virNodeDevicePtr dev,
+ unsigned *domain,
+ unsigned *bus,
+ unsigned *slot,
+ unsigned *function)
+{
+ virNodeDeviceDefPtr def = NULL;
+ virNodeDevCapsDefPtr cap;
+ char *xml = NULL;
+ int ret = -1;
+
+ xml = virNodeDeviceGetXMLDesc(dev, 0);
+ if (!xml)
+ goto out;
+
+ def = virNodeDeviceDefParseString(xml, EXISTING_DEVICE, NULL);
+ if (!def)
+ goto out;
+
+ cap = def->caps;
+ while (cap) {
+ if (cap->type == VIR_NODE_DEV_CAP_PCI_DEV) {
+ *domain = cap->data.pci_dev.domain;
+ *bus = cap->data.pci_dev.bus;
+ *slot = cap->data.pci_dev.slot;
+ *function = cap->data.pci_dev.function;
+ break;
+ }
+
+ cap = cap->next;
+ }
+
+ if (!cap) {
+ virReportError(VIR_ERR_INVALID_ARG,
+ _("device %s is not a PCI device"), dev->name);
+ goto out;
+ }
+
+ ret = 0;
+out:
+ virNodeDeviceDefFree(def);
+ VIR_FREE(xml);
+ return ret;
+}
+
+static int
+libxlNodeDeviceDettach(virNodeDevicePtr dev)
+{
+ virPCIDevicePtr pci;
+ unsigned domain, bus, slot, function;
+ int ret = -1;
+
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ return -1;
+
+ pci = virPCIDeviceNew(domain, bus, slot, function);
+ if (!pci)
+ return -1;
+
+ if (virPCIDeviceDetach(pci, NULL, NULL, "pciback") < 0)
+ goto out;
+
+ ret = 0;
+out:
+ virPCIDeviceFree(pci);
+ return ret;
+}
+
+static int
+libxlNodeDeviceAssignedDomainId(virNodeDevicePtr dev)
+{
+ int numdomains;
+ int numpcidevs;
+ int ret = -1, i, j;
+ int *ids = NULL;
+ char *bdf = NULL;
+ char *xref = NULL;
+ unsigned int domain, bus, slot, function;
+ libxl_device_pci *pcidevs = NULL;
+ virConnectPtr conn = dev->conn;
+ libxlDriverPrivatePtr driver = conn->privateData;
+
+ /* Get active domains */
+ numdomains = libxlNumDomains(conn);
+ if (numdomains < 0) {
+ return ret;
+ }
+ if (numdomains > 0){
+ if (VIR_ALLOC_N(ids, numdomains) < 0) {
+ virReportOOMError();
+ goto out;
+ }
+ if ((numdomains = libxlListDomains(conn, &ids[0], numdomains)) < 0) {
+ goto out;
+ }
+ }
+
+ /* Get pci bdf */
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ goto out;
+
+ if (virAsprintf(&bdf, "%04x:%02x:%02x.%0x",
+ domain, bus, slot, function) < 0) {
+ virReportOOMError();
+ goto out;
+ }
+
+ libxlDriverLock(driver);
+ /* Check if bdf is assigned to one of active domains */
+ for (i = 0; i < numdomains; i++) {
+ pcidevs = libxl_device_pci_list(driver->ctx, ids[i], &numpcidevs);
+ if (pcidevs == NULL) {
+ continue;
+ } else {
+ for (j = 0; j < numpcidevs; j++) {
+ if (pcidevs[j].bus == bus && pcidevs[j].dev == slot &&
pcidevs[j].func == function) {
+ ret = ids[i];
+ break;
+ }
+ }
+ VIR_FREE(pcidevs);
+ }
+ }
+ libxlDriverUnlock(driver);
+
+ VIR_FREE(xref);
+ VIR_FREE(bdf);
+out:
+ VIR_FREE(ids);
+
+ return ret;
+}
+
+static int
+libxlNodeDeviceReAttach(virNodeDevicePtr dev)
+{
+ virPCIDevicePtr pci;
+ unsigned domain, bus, slot, function;
+ int ret = -1;
+ int domid;
+
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ return -1;
+
+ pci = virPCIDeviceNew(domain, bus, slot, function);
+ if (!pci)
+ return -1;
+
+ /* Check if device is assigned to an active guest */
+ if ((domid = libxlNodeDeviceAssignedDomainId(dev)) >= 0) {
+ virReportError(VIR_ERR_INTERNAL_ERROR,
+ _("Device %s has been assigned to guest %d"),
+ dev->name, domid);
+ goto out;
+ }
+
+ if (virPCIDeviceReattach(pci, NULL, NULL, "pciback") < 0)
+ goto out;
+
+ ret = 0;
+out:
+ virPCIDeviceFree(pci);
+ return ret;
+}
+
+static int
+libxlNodeDeviceReset(virNodeDevicePtr dev)
+{
+ virPCIDevicePtr pci;
+ unsigned domain, bus, slot, function;
+ int ret = -1;
+
+ if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function)
< 0)
+ return -1;
+
+ pci = virPCIDeviceNew(domain, bus, slot, function);
+ if (!pci)
+ return -1;
+
+ if (virPCIDeviceReset(pci, NULL, NULL) < 0)
+ goto out;
+
+ ret = 0;
+out:
+ virPCIDeviceFree(pci);
+ return ret;
+}
+
static unsigned long long
libxlNodeGetFreeMemory(virConnectPtr conn)
{
@@ -4233,6 +4423,9 @@ static virDriver libxlDriver = {
.nodeGetFreeMemory = libxlNodeGetFreeMemory, /* 0.9.0 */
.domainEventRegister = libxlDomainEventRegister, /* 0.9.0 */
.domainEventDeregister = libxlDomainEventDeregister, /* 0.9.0 */
+ .nodeDeviceDettach = libxlNodeDeviceDettach, /* 1.0.4 */
+ .nodeDeviceReAttach = libxlNodeDeviceReAttach, /* 1.0.4 */
+ .nodeDeviceReset = libxlNodeDeviceReset, /* 1.0.4 */
.domainManagedSave = libxlDomainManagedSave, /* 0.9.2 */
.domainHasManagedSaveImage = libxlDomainHasManagedSaveImage, /* 0.9.2 */
.domainManagedSaveRemove = libxlDomainManagedSaveRemove, /* 0.9.2 */
6:32 p.m.
New subject: [libvirt] [PATCH RESENT 03/12] libxl: nodeDevice* support for PCI devices
On 22.05.2013 16:36, Jim Fehlig wrote:
Marek Marczykowski wrote:
> For now only for PCI devices. Mostly copy-paste from old xen driver.
>
This one is (or will be) covered by Chanyan's work as well right?
Right.
Regards,
Jim
> ---
> src/libxl/libxl_driver.c | 193 +++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 193 insertions(+)
>
> diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
> index 40a7a6b..011edf8 100644
> --- a/src/libxl/libxl_driver.c
> +++ b/src/libxl/libxl_driver.c
> @@ -42,6 +42,7 @@
> #include "libxl.h"
> #include "libxl_driver.h"
> #include "libxl_conf.h"
> +#include "node_device_conf.h"
> #include "xen_xm.h"
> #include "virtypedparam.h"
> #include "viruri.h"
> @@ -3666,6 +3667,195 @@ libxlDomainUpdateDeviceFlags(virDomainPtr dom, const char
*xml,
> return libxlDomainModifyDeviceFlags(dom, xml, flags, LIBXL_DEVICE_UPDATE);
> }
>
> +static int
> +libxlNodeDeviceGetPciInfo(virNodeDevicePtr dev,
> + unsigned *domain,
> + unsigned *bus,
> + unsigned *slot,
> + unsigned *function)
> +{
> + virNodeDeviceDefPtr def = NULL;
> + virNodeDevCapsDefPtr cap;
> + char *xml = NULL;
> + int ret = -1;
> +
> + xml = virNodeDeviceGetXMLDesc(dev, 0);
> + if (!xml)
> + goto out;
> +
> + def = virNodeDeviceDefParseString(xml, EXISTING_DEVICE, NULL);
> + if (!def)
> + goto out;
> +
> + cap = def->caps;
> + while (cap) {
> + if (cap->type == VIR_NODE_DEV_CAP_PCI_DEV) {
> + *domain = cap->data.pci_dev.domain;
> + *bus = cap->data.pci_dev.bus;
> + *slot = cap->data.pci_dev.slot;
> + *function = cap->data.pci_dev.function;
> + break;
> + }
> +
> + cap = cap->next;
> + }
> +
> + if (!cap) {
> + virReportError(VIR_ERR_INVALID_ARG,
> + _("device %s is not a PCI device"), dev->name);
> + goto out;
> + }
> +
> + ret = 0;
> +out:
> + virNodeDeviceDefFree(def);
> + VIR_FREE(xml);
> + return ret;
> +}
> +
> +static int
> +libxlNodeDeviceDettach(virNodeDevicePtr dev)
> +{
> + virPCIDevicePtr pci;
> + unsigned domain, bus, slot, function;
> + int ret = -1;
> +
> + if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot,
&function) < 0)
> + return -1;
> +
> + pci = virPCIDeviceNew(domain, bus, slot, function);
> + if (!pci)
> + return -1;
> +
> + if (virPCIDeviceDetach(pci, NULL, NULL, "pciback") < 0)
> + goto out;
> +
> + ret = 0;
> +out:
> + virPCIDeviceFree(pci);
> + return ret;
> +}
> +
> +static int
> +libxlNodeDeviceAssignedDomainId(virNodeDevicePtr dev)
> +{
> + int numdomains;
> + int numpcidevs;
> + int ret = -1, i, j;
> + int *ids = NULL;
> + char *bdf = NULL;
> + char *xref = NULL;
> + unsigned int domain, bus, slot, function;
> + libxl_device_pci *pcidevs = NULL;
> + virConnectPtr conn = dev->conn;
> + libxlDriverPrivatePtr driver = conn->privateData;
> +
> + /* Get active domains */
> + numdomains = libxlNumDomains(conn);
> + if (numdomains < 0) {
> + return ret;
> + }
> + if (numdomains > 0){
> + if (VIR_ALLOC_N(ids, numdomains) < 0) {
> + virReportOOMError();
> + goto out;
> + }
> + if ((numdomains = libxlListDomains(conn, &ids[0], numdomains)) < 0)
{
> + goto out;
> + }
> + }
> +
> + /* Get pci bdf */
> + if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot,
&function) < 0)
> + goto out;
> +
> + if (virAsprintf(&bdf, "%04x:%02x:%02x.%0x",
> + domain, bus, slot, function) < 0) {
> + virReportOOMError();
> + goto out;
> + }
> +
> + libxlDriverLock(driver);
> + /* Check if bdf is assigned to one of active domains */
> + for (i = 0; i < numdomains; i++) {
> + pcidevs = libxl_device_pci_list(driver->ctx, ids[i], &numpcidevs);
> + if (pcidevs == NULL) {
> + continue;
> + } else {
> + for (j = 0; j < numpcidevs; j++) {
> + if (pcidevs[j].bus == bus && pcidevs[j].dev == slot
&& pcidevs[j].func == function) {
> + ret = ids[i];
> + break;
> + }
> + }
> + VIR_FREE(pcidevs);
> + }
> + }
> + libxlDriverUnlock(driver);
> +
> + VIR_FREE(xref);
> + VIR_FREE(bdf);
> +out:
> + VIR_FREE(ids);
> +
> + return ret;
> +}
> +
> +static int
> +libxlNodeDeviceReAttach(virNodeDevicePtr dev)
> +{
> + virPCIDevicePtr pci;
> + unsigned domain, bus, slot, function;
> + int ret = -1;
> + int domid;
> +
> + if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot,
&function) < 0)
> + return -1;
> +
> + pci = virPCIDeviceNew(domain, bus, slot, function);
> + if (!pci)
> + return -1;
> +
> + /* Check if device is assigned to an active guest */
> + if ((domid = libxlNodeDeviceAssignedDomainId(dev)) >= 0) {
> + virReportError(VIR_ERR_INTERNAL_ERROR,
> + _("Device %s has been assigned to guest %d"),
> + dev->name, domid);
> + goto out;
> + }
> +
> + if (virPCIDeviceReattach(pci, NULL, NULL, "pciback") < 0)
> + goto out;
> +
> + ret = 0;
> +out:
> + virPCIDeviceFree(pci);
> + return ret;
> +}
> +
> +static int
> +libxlNodeDeviceReset(virNodeDevicePtr dev)
> +{
> + virPCIDevicePtr pci;
> + unsigned domain, bus, slot, function;
> + int ret = -1;
> +
> + if (libxlNodeDeviceGetPciInfo(dev, &domain, &bus, &slot,
&function) < 0)
> + return -1;
> +
> + pci = virPCIDeviceNew(domain, bus, slot, function);
> + if (!pci)
> + return -1;
> +
> + if (virPCIDeviceReset(pci, NULL, NULL) < 0)
> + goto out;
> +
> + ret = 0;
> +out:
> + virPCIDeviceFree(pci);
> + return ret;
> +}
> +
> static unsigned long long
> libxlNodeGetFreeMemory(virConnectPtr conn)
> {
> @@ -4233,6 +4423,9 @@ static virDriver libxlDriver = {
> .nodeGetFreeMemory = libxlNodeGetFreeMemory, /* 0.9.0 */
> .domainEventRegister = libxlDomainEventRegister, /* 0.9.0 */
> .domainEventDeregister = libxlDomainEventDeregister, /* 0.9.0 */
> + .nodeDeviceDettach = libxlNodeDeviceDettach, /* 1.0.4 */
> + .nodeDeviceReAttach = libxlNodeDeviceReAttach, /* 1.0.4 */
> + .nodeDeviceReset = libxlNodeDeviceReset, /* 1.0.4 */
> .domainManagedSave = libxlDomainManagedSave, /* 0.9.2 */
> .domainHasManagedSaveImage = libxlDomainHasManagedSaveImage, /* 0.9.2 */
> .domainManagedSaveRemove = libxlDomainManagedSaveRemove, /* 0.9.2 */
>
--
Best Regards,
Marek Marczykowski
Invisible Things Lab
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
libxl uses some xenstore entries for hints in memory management
(especially when starting new domain). This includes dom0 memory limit
and Xen free memory margin, based on current system state. Entries are
created at first usage, so force such usage at daemon startup, which most
likely will be before any domain startup.
---
src/libxl/libxl_driver.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 011edf8..89546a5 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -1115,6 +1115,7 @@ libxlStartup(bool privileged,
char *log_file = NULL;
virCommandPtr cmd;
int status, ret = 0;
+ unsigned int free_mem;
char ebuf[1024];
/* Disable libxl driver if non-root */
@@ -1240,6 +1241,13 @@ libxlStartup(bool privileged,
goto error;
}
+ /* This will fill xenstore info about free and dom0 memory - if missing,
+ * should be called before starting first domain */
+ if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
+ VIR_ERROR(_("cannot get free memory info"));
+ goto error;
+ }
+
if (!(libxl_driver->xmlconf =
virDomainXMLConfNew(&libxlDomainXMLPrivateDataCallbacks,
NULL)))
goto error;
--
1.8.1.4
11:09 p.m.
New subject: [libvirt] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
Marek Marczykowski wrote:
libxl uses some xenstore entries for hints in memory management
(especially when starting new domain). This includes dom0 memory limit
and Xen free memory margin, based on current system state. Entries are
created at first usage, so force such usage at daemon startup, which most
likely will be before any domain startup.
Hmm, I'd like to get the xen developer's opinion on this change.
Perhaps Ian C. or Ian J. have some comment...
---
src/libxl/libxl_driver.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 011edf8..89546a5 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -1115,6 +1115,7 @@ libxlStartup(bool privileged,
char *log_file = NULL;
virCommandPtr cmd;
int status, ret = 0;
+ unsigned int free_mem;
uint32_t to match libxl_get_free_memory() definition?
char ebuf[1024];
/* Disable libxl driver if non-root */
@@ -1240,6 +1241,13 @@ libxlStartup(bool privileged,
goto error;
}
+ /* This will fill xenstore info about free and dom0 memory - if missing,
+ * should be called before starting first domain */
+ if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
+ VIR_ERROR(_("cannot get free memory info"));
+ goto error;
+ }
Should failure of libxl_get_free_memory() really be fatal and prevent
the driver from loading?
Regards,
Jim
+
if (!(libxl_driver->xmlconf =
virDomainXMLConfNew(&libxlDomainXMLPrivateDataCallbacks,
NULL)))
goto error;
2:52 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
> + /* This will fill xenstore info about free and dom0 memory
- if missing,
> + * should be called before starting first domain */
> + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
> + VIR_ERROR(_("cannot get free memory info"));
> + goto error;
> + }
>
Should failure of libxl_get_free_memory() really be fatal and prevent
the driver from loading?
I'm not sure it is intended to be called like this...
I think it is intended to be called as part of starting every domain, to
check if there is enough free memory for that domain, rather than
calling it once at start of day.
In that context if it fails or returns less than the required amount of
memory then that would be fatal for starting that domain.
In xl we use this as part of the auto balloon of dom0, see
xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it require
dom0_mem? Perhaps this is handled at a higher level?
Ian.
Regards,
Jim
> +
> if (!(libxl_driver->xmlconf =
virDomainXMLConfNew(&libxlDomainXMLPrivateDataCallbacks,
> NULL)))
> goto error;
>
_______________________________________________
Xen-devel mailing list
Xen-devel(a)lists.xen.org
http://lists.xen.org/xen-devel
5:53 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On 11.04.2013 09:52, Ian Campbell wrote:
On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>> + /* This will fill xenstore info about free and dom0 memory - if missing,
>> + * should be called before starting first domain */
>> + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
>> + VIR_ERROR(_("cannot get free memory info"));
>> + goto error;
>> + }
>>
>
> Should failure of libxl_get_free_memory() really be fatal and prevent
> the driver from loading?
I'm not sure it is intended to be called like this...
I think it is intended to be called as part of starting every domain, to
check if there is enough free memory for that domain, rather than
calling it once at start of day.
In that context if it fails or returns less than the required amount of
memory then that would be fatal for starting that domain.
In xl we use this as part of the auto balloon of dom0, see
xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it require
dom0_mem? Perhaps this is handled at a higher level?
The problem is how libxl set initial value for freemem-slack. If, for any
reason, dom0 hasn't (almost) all memory assigned before creating first domain,
15% of host memory will no longer be used at all. This "any reason" can be
dom0_mem, which is covered by "auto" value for autoballoon in xen-unstable
(actually only for xl, not libxl in general). But this can also happen if
somebody calls xl set-mem 0 <some value>. The later case doesn't mean the user
want to disable autoballoon completely.
And to answer you question - libvirt rely on libxl autoballoon.
--
Best Regards / Pozdrawiam,
Marek Marczykowski
Invisible Things Lab
6 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On Thu, 2013-04-11 at 11:53 +0100, Marek Marczykowski wrote:
On 11.04.2013 09:52, Ian Campbell wrote:
> On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>>> + /* This will fill xenstore info about free and dom0 memory - if
missing,
>>> + * should be called before starting first domain */
>>> + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
>>> + VIR_ERROR(_("cannot get free memory info"));
>>> + goto error;
>>> + }
>>>
>>
>> Should failure of libxl_get_free_memory() really be fatal and prevent
>> the driver from loading?
>
> I'm not sure it is intended to be called like this...
>
> I think it is intended to be called as part of starting every domain, to
> check if there is enough free memory for that domain, rather than
> calling it once at start of day.
>
> In that context if it fails or returns less than the required amount of
> memory then that would be fatal for starting that domain.
>
> In xl we use this as part of the auto balloon of dom0, see
> xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it require
> dom0_mem? Perhaps this is handled at a higher level?
The problem is how libxl set initial value for freemem-slack. If, for any
reason, dom0 hasn't (almost) all memory assigned before creating first domain,
15% of host memory will no longer be used at all.
This is probably one for Stefano who understands this aspect best, I
think. (He's back next week, so this may need to wait). IanJ also knows
this stuff better than I, I've added both of them to the CC.
Note that whatever problem you are envisaging is also true if /creating
first domain/starting libvirtd/ (or whatever the relevant daemon is
called).
This "any reason" can be
dom0_mem, which is covered by "auto" value for autoballoon in xen-unstable
(actually only for xl, not libxl in general). But this can also happen if
somebody calls xl set-mem 0 <some value>. The later case doesn't mean the user
want to disable autoballoon completely.
And to answer you question - libvirt rely on libxl autoballoon.
AIUI autoballoon and dom0_mem (and I think mem-set at start of day too)
are mutually exclusive, which means that as it stands the libvirt
backend is incompatible with dom0_mem -- obviously this is something
which should be addressed. I don't think this patch is the right way to
do that but I'm not sure what is, hopefully Stefano or IanJ have some
ideas...
Ian.
6:10 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On Thu, 11 Apr 2013, Marek Marczykowski wrote:
On 11.04.2013 09:52, Ian Campbell wrote:
> On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>>> + /* This will fill xenstore info about free and dom0 memory - if
missing,
>>> + * should be called before starting first domain */
>>> + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
>>> + VIR_ERROR(_("cannot get free memory info"));
>>> + goto error;
>>> + }
>>>
>>
>> Should failure of libxl_get_free_memory() really be fatal and prevent
>> the driver from loading?
>
> I'm not sure it is intended to be called like this...
>
> I think it is intended to be called as part of starting every domain, to
> check if there is enough free memory for that domain, rather than
> calling it once at start of day.
>
> In that context if it fails or returns less than the required amount of
> memory then that would be fatal for starting that domain.
>
> In xl we use this as part of the auto balloon of dom0, see
> xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it require
> dom0_mem? Perhaps this is handled at a higher level?
The problem is how libxl set initial value for freemem-slack. If, for any
reason, dom0 hasn't (almost) all memory assigned before creating first domain,
15% of host memory will no longer be used at all. This "any reason" can be
dom0_mem, which is covered by "auto" value for autoballoon in xen-unstable
(actually only for xl, not libxl in general).
This is the (in)famous incompatibility between autoballoon and dom0_mem.
But this can also happen if
somebody calls xl set-mem 0 <some value>. The later case doesn't mean the user
want to disable autoballoon completely.
Calling "xl set-mem 0" manually should be compatible with
autoballoon. However it wouldn't work with dom0_mem.
And to answer you question - libvirt rely on libxl autoballoon.
Could we introduce something similar to autoballoon=auto to libvirt?
Maybe we should push down the autoballoon option to libxl: we should
probably rename autoballoon to libxl_memory_management, enable it
automatically during initialization (and call
libxl__fill_dom0_memory_info) unless dom0_mem has been passed as an
option, in which case we would disable it. If we do it at the libxl
level we would cover xl as well as libvirt and also fix the "xl set-mem
0" case.
6:23 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On 19.04.2013 13:10, Stefano Stabellini wrote:
On Thu, 11 Apr 2013, Marek Marczykowski wrote:
> On 11.04.2013 09:52, Ian Campbell wrote:
>> On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>>>> + /* This will fill xenstore info about free and dom0 memory - if
missing,
>>>> + * should be called before starting first domain */
>>>> + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
>>>> + VIR_ERROR(_("cannot get free memory info"));
>>>> + goto error;
>>>> + }
>>>>
>>>
>>> Should failure of libxl_get_free_memory() really be fatal and prevent
>>> the driver from loading?
>>
>> I'm not sure it is intended to be called like this...
>>
>> I think it is intended to be called as part of starting every domain, to
>> check if there is enough free memory for that domain, rather than
>> calling it once at start of day.
>>
>> In that context if it fails or returns less than the required amount of
>> memory then that would be fatal for starting that domain.
>>
>> In xl we use this as part of the auto balloon of dom0, see
>> xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it require
>> dom0_mem? Perhaps this is handled at a higher level?
>
> The problem is how libxl set initial value for freemem-slack. If, for any
> reason, dom0 hasn't (almost) all memory assigned before creating first domain,
> 15% of host memory will no longer be used at all. This "any reason" can be
> dom0_mem, which is covered by "auto" value for autoballoon in xen-unstable
> (actually only for xl, not libxl in general).
This is the (in)famous incompatibility between autoballoon and dom0_mem.
> But this can also happen if
> somebody calls xl set-mem 0 <some value>. The later case doesn't mean the
user
> want to disable autoballoon completely.
Calling "xl set-mem 0" manually should be compatible with
autoballoon.
Unless it is called before first domain start (which case is covered by my patch).
However it wouldn't work with dom0_mem.
> And to answer you question - libvirt rely on libxl autoballoon.
Could we introduce something similar to autoballoon=auto to libvirt?
Maybe we should push down the autoballoon option to libxl: we should
probably rename autoballoon to libxl_memory_management, enable it
automatically during initialization (and call
libxl__fill_dom0_memory_info) unless dom0_mem has been passed as an
option, in which case we would disable it. If we do it at the libxl
level we would cover xl as well as libvirt and also fix the "xl set-mem
0" case.
Looks good for me, but IIUC it's too late for such change in 4.3 and it
doesn't qualify for later backport, right? If so, some approach still is
needed in libvirt for xen 4.2 and 4.3 (like my simple patch). This still
doesn't make libvirt compatible with dom0_mem, but at least cover one
(independent) case. Perhaps additionally autoballoon=auto code should be
copied from xl to libvirt to cover dom0_mem case with xen 4.2 and 4.3?
--
Best Regards / Pozdrawiam,
Marek Marczykowski
Invisible Things Lab
6:33 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On Fri, 19 Apr 2013, Marek Marczykowski wrote:
>> And to answer you question - libvirt rely on libxl
autoballoon.
>
> Could we introduce something similar to autoballoon=auto to libvirt?
>
> Maybe we should push down the autoballoon option to libxl: we should
> probably rename autoballoon to libxl_memory_management, enable it
> automatically during initialization (and call
> libxl__fill_dom0_memory_info) unless dom0_mem has been passed as an
> option, in which case we would disable it. If we do it at the libxl
> level we would cover xl as well as libvirt and also fix the "xl set-mem
> 0" case.
Looks good for me, but IIUC it's too late for such change in 4.3 and it
doesn't qualify for later backport, right? If so, some approach still is
needed in libvirt for xen 4.2 and 4.3 (like my simple patch). This still
doesn't make libvirt compatible with dom0_mem, but at least cover one
(independent) case.
Perhaps additionally autoballoon=auto code should be
copied from xl to libvirt to cover dom0_mem case with xen 4.2 and 4.3?
Might be a decent solution for the moment.
10:05 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
Marek Marczykowski wrote:
On 19.04.2013 13:10, Stefano Stabellini wrote:
> On Thu, 11 Apr 2013, Marek Marczykowski wrote:
>
>> On 11.04.2013 09:52, Ian Campbell wrote:
>>
>>> On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>>>
>>>>> + /* This will fill xenstore info about free and dom0 memory - if
missing,
>>>>> + * should be called before starting first domain */
>>>>> + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem))
{
>>>>> + VIR_ERROR(_("cannot get free memory info"));
>>>>> + goto error;
>>>>> + }
>>>>>
>>>>>
>>>> Should failure of libxl_get_free_memory() really be fatal and prevent
>>>> the driver from loading?
>>>>
>>> I'm not sure it is intended to be called like this...
>>>
>>> I think it is intended to be called as part of starting every domain, to
>>> check if there is enough free memory for that domain, rather than
>>> calling it once at start of day.
>>>
>>> In that context if it fails or returns less than the required amount of
>>> memory then that would be fatal for starting that domain.
>>>
>>> In xl we use this as part of the auto balloon of dom0, see
>>> xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it require
>>> dom0_mem? Perhaps this is handled at a higher level?
>>>
>> The problem is how libxl set initial value for freemem-slack. If, for any
>> reason, dom0 hasn't (almost) all memory assigned before creating first
domain,
>> 15% of host memory will no longer be used at all. This "any reason" can
be
>> dom0_mem, which is covered by "auto" value for autoballoon in
xen-unstable
>> (actually only for xl, not libxl in general).
>>
> This is the (in)famous incompatibility between autoballoon and dom0_mem.
>
>
>
>> But this can also happen if
>> somebody calls xl set-mem 0 <some value>. The later case doesn't mean
the user
>> want to disable autoballoon completely.
>>
> Calling "xl set-mem 0" manually should be compatible with
> autoballoon.
>
Unless it is called before first domain start (which case is covered by my patch).
> However it wouldn't work with dom0_mem.
>
>
>
>> And to answer you question - libvirt rely on libxl autoballoon.
>>
> Could we introduce something similar to autoballoon=auto to libvirt?
>
> Maybe we should push down the autoballoon option to libxl: we should
> probably rename autoballoon to libxl_memory_management, enable it
> automatically during initialization (and call
> libxl__fill_dom0_memory_info) unless dom0_mem has been passed as an
> option, in which case we would disable it. If we do it at the libxl
> level we would cover xl as well as libvirt and also fix the "xl set-mem
> 0" case.
>
Looks good for me, but IIUC it's too late for such change in 4.3 and it
doesn't qualify for later backport, right? If so, some approach still is
needed in libvirt for xen 4.2 and 4.3 (like my simple patch). This still
doesn't make libvirt compatible with dom0_mem, but at least cover one
(independent) case. Perhaps additionally autoballoon=auto code should be
copied from xl to libvirt to cover dom0_mem case with xen 4.2 and 4.3?
After reading this thread again, it sounds like the best option is to
support the autoballoon option in libvirt, e.g. in
/etc/libvirt/libxl.conf. (We currently don't have a config file for the
libxl driver, but I think we'll need one anyhow for other knobs, similar
to qemu.conf.) As you note, even if autoballoon is pushed down to
libxl, libvirt would need to handle it for older libxl. And libvirt
needs to handle the dom0_mem case...
Regards,
Jim
11:58 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup\
On Wed, 22 May 2013, Jim Fehlig wrote:
Marek Marczykowski wrote:
> On 19.04.2013 13:10, Stefano Stabellini wrote:
>
>> On Thu, 11 Apr 2013, Marek Marczykowski wrote:
>>
>>> On 11.04.2013 09:52, Ian Campbell wrote:
>>>
>>>> On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>>>>
>>>>>> + /* This will fill xenstore info about free and dom0 memory
- if missing,
>>>>>> + * should be called before starting first domain */
>>>>>> + if (libxl_get_free_memory(libxl_driver->ctx,
&free_mem)) {
>>>>>> + VIR_ERROR(_("cannot get free memory info"));
>>>>>> + goto error;
>>>>>> + }
>>>>>>
>>>>>>
>>>>> Should failure of libxl_get_free_memory() really be fatal and
prevent
>>>>> the driver from loading?
>>>>>
>>>> I'm not sure it is intended to be called like this...
>>>>
>>>> I think it is intended to be called as part of starting every domain,
to
>>>> check if there is enough free memory for that domain, rather than
>>>> calling it once at start of day.
>>>>
>>>> In that context if it fails or returns less than the required amount of
>>>> memory then that would be fatal for starting that domain.
>>>>
>>>> In xl we use this as part of the auto balloon of dom0, see
>>>> xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it
require
>>>> dom0_mem? Perhaps this is handled at a higher level?
>>>>
>>> The problem is how libxl set initial value for freemem-slack. If, for any
>>> reason, dom0 hasn't (almost) all memory assigned before creating first
domain,
>>> 15% of host memory will no longer be used at all. This "any
reason" can be
>>> dom0_mem, which is covered by "auto" value for autoballoon in
xen-unstable
>>> (actually only for xl, not libxl in general).
>>>
>> This is the (in)famous incompatibility between autoballoon and dom0_mem.
>>
>>
>>
>>> But this can also happen if
>>> somebody calls xl set-mem 0 <some value>. The later case doesn't
mean the user
>>> want to disable autoballoon completely.
>>>
>> Calling "xl set-mem 0" manually should be compatible with
>> autoballoon.
>>
>
> Unless it is called before first domain start (which case is covered by my patch).
>
>
>> However it wouldn't work with dom0_mem.
>>
>>
>>
>>> And to answer you question - libvirt rely on libxl autoballoon.
>>>
>> Could we introduce something similar to autoballoon=auto to libvirt?
>>
>> Maybe we should push down the autoballoon option to libxl: we should
>> probably rename autoballoon to libxl_memory_management, enable it
>> automatically during initialization (and call
>> libxl__fill_dom0_memory_info) unless dom0_mem has been passed as an
>> option, in which case we would disable it. If we do it at the libxl
>> level we would cover xl as well as libvirt and also fix the "xl set-mem
>> 0" case.
>>
>
> Looks good for me, but IIUC it's too late for such change in 4.3 and it
> doesn't qualify for later backport, right? If so, some approach still is
> needed in libvirt for xen 4.2 and 4.3 (like my simple patch). This still
> doesn't make libvirt compatible with dom0_mem, but at least cover one
> (independent) case. Perhaps additionally autoballoon=auto code should be
> copied from xl to libvirt to cover dom0_mem case with xen 4.2 and 4.3?
>
After reading this thread again, it sounds like the best option is to
support the autoballoon option in libvirt, e.g. in
/etc/libvirt/libxl.conf. (We currently don't have a config file for the
libxl driver, but I think we'll need one anyhow for other knobs, similar
to qemu.conf.) As you note, even if autoballoon is pushed down to
libxl, libvirt would need to handle it for older libxl. And libvirt
needs to handle the dom0_mem case...
Given all the troubles that we had with the autoballoon option in
xl/libxl and how it clashes with dom0_mem, I would strongly advise to
consider implementing something like autoballoon=auto from the start.
6:44 p.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup\
On 22.05.2013 18:58, Stefano Stabellini wrote:
On Wed, 22 May 2013, Jim Fehlig wrote:
> Marek Marczykowski wrote:
>> On 19.04.2013 13:10, Stefano Stabellini wrote:
>>
>>> On Thu, 11 Apr 2013, Marek Marczykowski wrote:
>>>
>>>> On 11.04.2013 09:52, Ian Campbell wrote:
>>>>
>>>>> On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
>>>>>
>>>>>>> + /* This will fill xenstore info about free and dom0
memory - if missing,
>>>>>>> + * should be called before starting first domain */
>>>>>>> + if (libxl_get_free_memory(libxl_driver->ctx,
&free_mem)) {
>>>>>>> + VIR_ERROR(_("cannot get free memory
info"));
>>>>>>> + goto error;
>>>>>>> + }
>>>>>>>
>>>>>>>
>>>>>> Should failure of libxl_get_free_memory() really be fatal and
prevent
>>>>>> the driver from loading?
>>>>>>
>>>>> I'm not sure it is intended to be called like this...
>>>>>
>>>>> I think it is intended to be called as part of starting every domain,
to
>>>>> check if there is enough free memory for that domain, rather than
>>>>> calling it once at start of day.
>>>>>
>>>>> In that context if it fails or returns less than the required amount
of
>>>>> memory then that would be fatal for starting that domain.
>>>>>
>>>>> In xl we use this as part of the auto balloon of dom0, see
>>>>> xl_cmdimplg.c:freemem. Does libvirt do autoballooning or does it
require
>>>>> dom0_mem? Perhaps this is handled at a higher level?
>>>>>
>>>> The problem is how libxl set initial value for freemem-slack. If, for
any
>>>> reason, dom0 hasn't (almost) all memory assigned before creating
first domain,
>>>> 15% of host memory will no longer be used at all. This "any
reason" can be
>>>> dom0_mem, which is covered by "auto" value for autoballoon in
xen-unstable
>>>> (actually only for xl, not libxl in general).
>>>>
>>> This is the (in)famous incompatibility between autoballoon and dom0_mem.
>>>
>>>
>>>
>>>> But this can also happen if
>>>> somebody calls xl set-mem 0 <some value>. The later case
doesn't mean the user
>>>> want to disable autoballoon completely.
>>>>
>>> Calling "xl set-mem 0" manually should be compatible with
>>> autoballoon.
>>>
>>
>> Unless it is called before first domain start (which case is covered by my
patch).
>>
>>
>>> However it wouldn't work with dom0_mem.
>>>
>>>
>>>
>>>> And to answer you question - libvirt rely on libxl autoballoon.
>>>>
>>> Could we introduce something similar to autoballoon=auto to libvirt?
>>>
>>> Maybe we should push down the autoballoon option to libxl: we should
>>> probably rename autoballoon to libxl_memory_management, enable it
>>> automatically during initialization (and call
>>> libxl__fill_dom0_memory_info) unless dom0_mem has been passed as an
>>> option, in which case we would disable it. If we do it at the libxl
>>> level we would cover xl as well as libvirt and also fix the "xl set-mem
>>> 0" case.
>>>
>>
>> Looks good for me, but IIUC it's too late for such change in 4.3 and it
>> doesn't qualify for later backport, right? If so, some approach still is
>> needed in libvirt for xen 4.2 and 4.3 (like my simple patch). This still
>> doesn't make libvirt compatible with dom0_mem, but at least cover one
>> (independent) case. Perhaps additionally autoballoon=auto code should be
>> copied from xl to libvirt to cover dom0_mem case with xen 4.2 and 4.3?
>>
>
> After reading this thread again, it sounds like the best option is to
> support the autoballoon option in libvirt, e.g. in
> /etc/libvirt/libxl.conf. (We currently don't have a config file for the
> libxl driver, but I think we'll need one anyhow for other knobs, similar
> to qemu.conf.) As you note, even if autoballoon is pushed down to
> libxl, libvirt would need to handle it for older libxl. And libvirt
> needs to handle the dom0_mem case...
Given all the troubles that we had with the autoballoon option in
xl/libxl and how it clashes with dom0_mem, I would strongly advise to
consider implementing something like autoballoon=auto from the start.
Ok, will send such patch (almost ready). For now it will contain copy&paste
auto_autobaloon() from xl.c, so the same logic. Can be easily extended to
config option in the future, but probably I will not have time for this.
--
Best Regards,
Marek Marczykowski
Invisible Things Lab
6:01 a.m.
New subject: [libvirt] [Xen-devel] [PATCH RESENT 04/12] libxl: populate xenstore memory entries at startup
On Thu, 11 Apr 2013, Ian Campbell wrote:
On Thu, 2013-04-11 at 05:09 +0100, Jim Fehlig wrote:
> > + /* This will fill xenstore info about free and dom0 memory - if missing,
> > + * should be called before starting first domain */
> > + if (libxl_get_free_memory(libxl_driver->ctx, &free_mem)) {
> > + VIR_ERROR(_("cannot get free memory info"));
> > + goto error;
> > + }
> >
>
> Should failure of libxl_get_free_memory() really be fatal and prevent
> the driver from loading?
I'm not sure it is intended to be called like this...
I think it is intended to be called as part of starting every domain, to
check if there is enough free memory for that domain, rather than
calling it once at start of day.
Actually libxl_get_free_memory is just meant to return to total amount
of free memory in the system. Can be called at any time.
As a side effect of libxl_get_free_memory, freemem-slack will be
written to xenstore if it wasn't before.
I think that it is OK to call it as Marek did in this patch, even though
might be nicer if we had a proper "libxl_initialize_memory_accounting"
function.
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 05/12] conf: add 'script' attribute to disk specification
Script to be called to prepare custom device for domain. Done with Xen
in mind, it maps to libxl_device_disk.script.
XML configuration would be:
<disk type='block' device='disk'>
<source dev='/dev/mapper/custom-device'/>
<script path='/script/to/setup/custom-device'/>
<target dev='xvdc'/>
</disk>
---
src/conf/domain_conf.c | 10 ++++++++++
src/conf/domain_conf.h | 1 +
2 files changed, 11 insertions(+)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index f3fca7f..257a265 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -1102,6 +1102,7 @@ void virDomainDiskDefFree(virDomainDiskDefPtr def)
VIR_FREE(def->wwn);
VIR_FREE(def->vendor);
VIR_FREE(def->product);
+ VIR_FREE(def->script);
if (def->auth.secretType == VIR_DOMAIN_DISK_SECRET_TYPE_USAGE)
VIR_FREE(def->auth.secret.usage);
virStorageEncryptionFree(def->encryption);
@@ -3993,6 +3994,7 @@ virDomainDiskDefParseXML(virCapsPtr caps,
char *wwn = NULL;
char *vendor = NULL;
char *product = NULL;
+ char *script = NULL;
int expected_secret_usage = -1;
int auth_secret_usage = -1;
@@ -4148,6 +4150,9 @@ virDomainDiskDefParseXML(virCapsPtr caps,
if (target &&
STRPREFIX(target, "ioemu:"))
memmove(target, target+6, strlen(target)-5);
+ } else if (!script &&
+ xmlStrEqual(cur->name, BAD_CAST "script")) {
+ script = virXMLPropString(cur, "path");
} else if (xmlStrEqual(cur->name, BAD_CAST "geometry")) {
if (virXPathUInt("string(./geometry/@cyls)",
ctxt, &def->geometry.cylinders) < 0) {
@@ -4690,6 +4695,8 @@ virDomainDiskDefParseXML(virCapsPtr caps,
source = NULL;
def->dst = target;
target = NULL;
+ def->script = script;
+ script = NULL;
def->hosts = hosts;
hosts = NULL;
def->nhosts = nhosts;
@@ -4788,6 +4795,7 @@ cleanup:
VIR_FREE(wwn);
VIR_FREE(vendor);
VIR_FREE(product);
+ VIR_FREE(script);
ctxt->node = save_ctxt;
return def;
@@ -12899,6 +12907,8 @@ virDomainDiskDefFormat(virBufferPtr buf,
}
}
+ virBufferEscapeString(buf, " <script path='%s'/>\n",
def->script);
+
virDomainDiskGeometryDefFormat(buf, def);
virDomainDiskBlockIoDefFormat(buf, def);
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index edddf25..d55d209 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -668,6 +668,7 @@ struct _virDomainDiskDef {
bool rawio_specified;
int rawio; /* no = 0, yes = 1 */
int sgio; /* enum virDomainDiskSGIO */
+ char *script;
size_t nseclabels;
virSecurityDeviceLabelDefPtr *seclabels;
--
1.8.1.4
4:14 a.m.
New subject: [libvirt] [PATCH RESENT 05/12] conf: add 'script' attribute to disk specification
On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
Script to be called to prepare custom device for domain. Done with
Xen
in mind, it maps to libxl_device_disk.script.
XML configuration would be:
<disk type='block' device='disk'>
<source dev='/dev/mapper/custom-device'/>
<script path='/script/to/setup/custom-device'/>
<target dev='xvdc'/>
</disk>
NACK, all configuration must be explicitly represented in
the config, not delegated to arbitrary opaque shell scripts.
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
5:13 a.m.
New subject: [libvirt] [PATCH RESENT 05/12] conf: add 'script' attribute to disk specification
On 10.04.2013 11:14, Daniel P. Berrange wrote:
On Wed, Apr 10, 2013 at 04:44:43AM +0200, Marek Marczykowski wrote:
> Script to be called to prepare custom device for domain. Done with Xen
> in mind, it maps to libxl_device_disk.script.
>
> XML configuration would be:
> <disk type='block' device='disk'>
> <source dev='/dev/mapper/custom-device'/>
> <script path='/script/to/setup/custom-device'/>
> <target dev='xvdc'/>
> </disk>
NACK, all configuration must be explicitly represented in
the config, not delegated to arbitrary opaque shell scripts.
Ok, but how to do that if backend is in other domain (not dom0), so you cannot
directly execute commands there?
--
Best Regards / Pozdrawiam,
Marek Marczykowski
Invisible Things Lab
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 06/12] libxl: use disk 'script' attribute
Implement handling of previously introduced <script/> element for disk
config.
This can be used for custom backend configuration like non-standard
device-mapper nodes, or to prepare device in other domain (see the next
patch).
---
src/libxl/libxl_conf.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 7668305..4bd62e9 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -484,6 +484,11 @@ libxlMakeDisk(virDomainDiskDefPtr l_disk, libxl_device_disk *x_disk)
return -1;
}
+ if (l_disk->script && (x_disk->script = strdup(l_disk->script)) ==
NULL) {
+ virReportOOMError();
+ return -1;
+ }
+
if (l_disk->driverName) {
if (STREQ(l_disk->driverName, "tap") ||
STREQ(l_disk->driverName, "tap2")) {
--
1.8.1.4
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 07/12] conf: support backend domain name in disk and network devices
At least Xen supports backend drivers in another domain (aka "driver
domain"). This patch introduces XML config option for such setting as
'domain' element with 'name' attribute. Verification its content is left
for the driver.
In the future some option will be needed for USB devices (hostdev
objects), but for now libxl doesn't have support for PVUSB.
---
docs/schemas/domaincommon.rng | 14 ++++++++++++++
src/conf/domain_conf.c | 27 +++++++++++++++++++++++++++
src/conf/domain_conf.h | 2 ++
3 files changed, 43 insertions(+)
diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
index 8d7e6db..1423187 100644
--- a/docs/schemas/domaincommon.rng
+++ b/docs/schemas/domaincommon.rng
@@ -869,6 +869,13 @@
</optional>
<ref name="target"/>
<optional>
+ <element name="domain">
+ <attribute name="name">
+ <ref name="domainName"/>
+ </attribute>
+ </element>
+ </optional>
+ <optional>
<ref name="deviceBoot"/>
</optional>
<optional>
@@ -1834,6 +1841,13 @@
</element>
</optional>
<optional>
+ <element name="domain">
+ <attribute name="name">
+ <ref name="domainName"/>
+ </attribute>
+ </element>
+ </optional>
+ <optional>
<element name="model">
<attribute name="type">
<data type="string">
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 257a265..bf1fec6 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -1103,6 +1103,7 @@ void virDomainDiskDefFree(virDomainDiskDefPtr def)
VIR_FREE(def->vendor);
VIR_FREE(def->product);
VIR_FREE(def->script);
+ VIR_FREE(def->domain_name);
if (def->auth.secretType == VIR_DOMAIN_DISK_SECRET_TYPE_USAGE)
VIR_FREE(def->auth.secret.usage);
virStorageEncryptionFree(def->encryption);
@@ -1228,6 +1229,7 @@ void virDomainNetDefFree(virDomainNetDefPtr def)
VIR_FREE(def->virtPortProfile);
VIR_FREE(def->script);
+ VIR_FREE(def->domain_name);
VIR_FREE(def->ifname);
virDomainDeviceInfoClear(&def->info);
@@ -3995,6 +3997,7 @@ virDomainDiskDefParseXML(virCapsPtr caps,
char *vendor = NULL;
char *product = NULL;
char *script = NULL;
+ char *domain_name = NULL;
int expected_secret_usage = -1;
int auth_secret_usage = -1;
@@ -4153,6 +4156,9 @@ virDomainDiskDefParseXML(virCapsPtr caps,
} else if (!script &&
xmlStrEqual(cur->name, BAD_CAST "script")) {
script = virXMLPropString(cur, "path");
+ } else if (!domain_name &&
+ xmlStrEqual(cur->name, BAD_CAST "domain")) {
+ domain_name = virXMLPropString(cur, "name");
} else if (xmlStrEqual(cur->name, BAD_CAST "geometry")) {
if (virXPathUInt("string(./geometry/@cyls)",
ctxt, &def->geometry.cylinders) < 0) {
@@ -4447,6 +4453,11 @@ virDomainDiskDefParseXML(virCapsPtr caps,
ctxt->node = saved_node;
}
+ if (domain_name != NULL) {
+ def->domain_name = domain_name;
+ domain_name = NULL;
+ }
+
if (target == NULL) {
virReportError(VIR_ERR_NO_TARGET,
source ? "%s" : NULL, source);
@@ -4796,6 +4807,7 @@ cleanup:
VIR_FREE(vendor);
VIR_FREE(product);
VIR_FREE(script);
+ VIR_FREE(domain_name);
ctxt->node = save_ctxt;
return def;
@@ -5353,6 +5365,7 @@ virDomainNetDefParseXML(virCapsPtr caps,
char *mode = NULL;
char *linkstate = NULL;
char *addrtype = NULL;
+ char *domain_name = NULL;
virNWFilterHashTablePtr filterparams = NULL;
virDomainActualNetDefPtr actual = NULL;
xmlNodePtr oldnode = ctxt->node;
@@ -5451,6 +5464,9 @@ virDomainNetDefParseXML(virCapsPtr caps,
} else if (!script &&
xmlStrEqual(cur->name, BAD_CAST "script")) {
script = virXMLPropString(cur, "path");
+ } else if (!domain_name &&
+ xmlStrEqual(cur->name, BAD_CAST "domain")) {
+ domain_name = virXMLPropString(cur, "name");
} else if (xmlStrEqual(cur->name, BAD_CAST "model")) {
model = virXMLPropString(cur, "type");
} else if (xmlStrEqual(cur->name, BAD_CAST "driver")) {
@@ -5682,6 +5698,10 @@ virDomainNetDefParseXML(virCapsPtr caps,
def->script = script;
script = NULL;
}
+ if (domain_name != NULL) {
+ def->domain_name = domain_name;
+ domain_name = NULL;
+ }
if (ifname != NULL) {
def->ifname = ifname;
ifname = NULL;
@@ -5808,6 +5828,7 @@ cleanup:
VIR_FREE(mode);
VIR_FREE(linkstate);
VIR_FREE(addrtype);
+ VIR_FREE(domain_name);
virNWFilterHashTableFree(filterparams);
return def;
@@ -12909,6 +12930,10 @@ virDomainDiskDefFormat(virBufferPtr buf,
virBufferEscapeString(buf, " <script path='%s'/>\n",
def->script);
+ if (def->domain_name) {
+ virBufferEscapeString(buf, " <domain name='%s'/>\n",
def->domain_name);
+ }
+
virDomainDiskGeometryDefFormat(buf, def);
virDomainDiskBlockIoDefFormat(buf, def);
@@ -13456,6 +13481,8 @@ virDomainNetDefFormat(virBufferPtr buf,
return -1;
virBufferEscapeString(buf, "<script path='%s'/>\n",
def->script);
+ if (def->domain_name)
+ virBufferEscapeString(buf, "<domain name='%s'/>\n",
def->domain_name);
if (def->ifname &&
!((flags & VIR_DOMAIN_XML_INACTIVE) &&
(STRPREFIX(def->ifname, VIR_NET_GENERATED_PREFIX)))) {
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index d55d209..db3002b 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -669,6 +669,7 @@ struct _virDomainDiskDef {
int rawio; /* no = 0, yes = 1 */
int sgio; /* enum virDomainDiskSGIO */
char *script;
+ char *domain_name; /* backend domain name */
size_t nseclabels;
virSecurityDeviceLabelDefPtr *seclabels;
@@ -919,6 +920,7 @@ struct _virDomainNetDef {
unsigned long sndbuf;
} tune;
char *script;
+ char *domain_name; /* backend domain name */
char *ifname;
virDomainDeviceInfo info;
char *filter;
--
1.8.1.4
10:22 a.m.
New subject: [libvirt] [PATCH RESENT 07/12] conf: support backend domain name in disk and network devices
Marek Marczykowski wrote:
At least Xen supports backend drivers in another domain (aka
"driver
domain"). This patch introduces XML config option for such setting as
'domain' element with 'name' attribute. Verification its content is left
for the driver.
In the future some option will be needed for USB devices (hostdev
objects), but for now libxl doesn't have support for PVUSB.
---
docs/schemas/domaincommon.rng | 14 ++++++++++++++
src/conf/domain_conf.c | 27 +++++++++++++++++++++++++++
src/conf/domain_conf.h | 2 ++
3 files changed, 43 insertions(+)
diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
index 8d7e6db..1423187 100644
--- a/docs/schemas/domaincommon.rng
+++ b/docs/schemas/domaincommon.rng
@@ -869,6 +869,13 @@
</optional>
<ref name="target"/>
<optional>
+ <element name="domain">
+ <attribute name="name">
+ <ref name="domainName"/>
+ </attribute>
+ </element>
+ </optional>
+ <optional>
<ref name="deviceBoot"/>
</optional>
<optional>
@@ -1834,6 +1841,13 @@
</element>
</optional>
<optional>
+ <element name="domain">
+ <attribute name="name">
+ <ref name="domainName"/>
+ </attribute>
+ </element>
+ </optional>
+ <optional>
I'm certainly no expert in RNG, but do these need to include <empty/>?
BTW, you'll also need to document this in docs/formatdomain.html.in.
<element name="model">
<attribute name="type">
<data type="string">
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index 257a265..bf1fec6 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -1103,6 +1103,7 @@ void virDomainDiskDefFree(virDomainDiskDefPtr def)
VIR_FREE(def->vendor);
VIR_FREE(def->product);
VIR_FREE(def->script);
+ VIR_FREE(def->domain_name);
if (def->auth.secretType == VIR_DOMAIN_DISK_SECRET_TYPE_USAGE)
VIR_FREE(def->auth.secret.usage);
virStorageEncryptionFree(def->encryption);
@@ -1228,6 +1229,7 @@ void virDomainNetDefFree(virDomainNetDefPtr def)
VIR_FREE(def->virtPortProfile);
VIR_FREE(def->script);
+ VIR_FREE(def->domain_name);
VIR_FREE(def->ifname);
virDomainDeviceInfoClear(&def->info);
@@ -3995,6 +3997,7 @@ virDomainDiskDefParseXML(virCapsPtr caps,
char *vendor = NULL;
char *product = NULL;
char *script = NULL;
+ char *domain_name = NULL;
int expected_secret_usage = -1;
int auth_secret_usage = -1;
@@ -4153,6 +4156,9 @@ virDomainDiskDefParseXML(virCapsPtr caps,
} else if (!script &&
xmlStrEqual(cur->name, BAD_CAST "script")) {
script = virXMLPropString(cur, "path");
+ } else if (!domain_name &&
+ xmlStrEqual(cur->name, BAD_CAST "domain")) {
+ domain_name = virXMLPropString(cur, "name");
} else if (xmlStrEqual(cur->name, BAD_CAST "geometry")) {
if (virXPathUInt("string(./geometry/@cyls)",
ctxt, &def->geometry.cylinders) < 0) {
@@ -4447,6 +4453,11 @@ virDomainDiskDefParseXML(virCapsPtr caps,
ctxt->node = saved_node;
}
+ if (domain_name != NULL) {
+ def->domain_name = domain_name;
+ domain_name = NULL;
+ }
+
Is the 'if' necessary here? It looks like the other fields of def are
unconditionally assigned, e.g.
def->src = source;
source = NULL;
if (target == NULL) {
virReportError(VIR_ERR_NO_TARGET,
source ? "%s" : NULL, source);
@@ -4796,6 +4807,7 @@ cleanup:
VIR_FREE(vendor);
VIR_FREE(product);
VIR_FREE(script);
+ VIR_FREE(domain_name);
ctxt->node = save_ctxt;
return def;
@@ -5353,6 +5365,7 @@ virDomainNetDefParseXML(virCapsPtr caps,
char *mode = NULL;
char *linkstate = NULL;
char *addrtype = NULL;
+ char *domain_name = NULL;
virNWFilterHashTablePtr filterparams = NULL;
virDomainActualNetDefPtr actual = NULL;
xmlNodePtr oldnode = ctxt->node;
@@ -5451,6 +5464,9 @@ virDomainNetDefParseXML(virCapsPtr caps,
} else if (!script &&
xmlStrEqual(cur->name, BAD_CAST "script")) {
script = virXMLPropString(cur, "path");
+ } else if (!domain_name &&
+ xmlStrEqual(cur->name, BAD_CAST "domain")) {
+ domain_name = virXMLPropString(cur, "name");
} else if (xmlStrEqual(cur->name, BAD_CAST "model")) {
model = virXMLPropString(cur, "type");
} else if (xmlStrEqual(cur->name, BAD_CAST "driver")) {
@@ -5682,6 +5698,10 @@ virDomainNetDefParseXML(virCapsPtr caps,
def->script = script;
script = NULL;
}
+ if (domain_name != NULL) {
+ def->domain_name = domain_name;
+ domain_name = NULL;
+ }
The pattern in this function is to conditionally assign def->foo, so I
guess my above comment is a nit.
if (ifname != NULL) {
def->ifname = ifname;
ifname = NULL;
@@ -5808,6 +5828,7 @@ cleanup:
VIR_FREE(mode);
VIR_FREE(linkstate);
VIR_FREE(addrtype);
+ VIR_FREE(domain_name);
virNWFilterHashTableFree(filterparams);
return def;
@@ -12909,6 +12930,10 @@ virDomainDiskDefFormat(virBufferPtr buf,
virBufferEscapeString(buf, " <script path='%s'/>\n",
def->script);
+ if (def->domain_name) {
+ virBufferEscapeString(buf, " <domain
name='%s'/>\n", def->domain_name);
+ }
+
virDomainDiskGeometryDefFormat(buf, def);
virDomainDiskBlockIoDefFormat(buf, def);
@@ -13456,6 +13481,8 @@ virDomainNetDefFormat(virBufferPtr buf,
return -1;
virBufferEscapeString(buf, "<script path='%s'/>\n",
def->script);
+ if (def->domain_name)
+ virBufferEscapeString(buf, "<domain name='%s'/>\n",
def->domain_name);
if (def->ifname &&
!((flags & VIR_DOMAIN_XML_INACTIVE) &&
(STRPREFIX(def->ifname, VIR_NET_GENERATED_PREFIX)))) {
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index d55d209..db3002b 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -669,6 +669,7 @@ struct _virDomainDiskDef {
int rawio; /* no = 0, yes = 1 */
int sgio; /* enum virDomainDiskSGIO */
char *script;
+ char *domain_name; /* backend domain name */
size_t nseclabels;
virSecurityDeviceLabelDefPtr *seclabels;
@@ -919,6 +920,7 @@ struct _virDomainNetDef {
unsigned long sndbuf;
} tune;
char *script;
+ char *domain_name; /* backend domain name */
char *ifname;
virDomainDeviceInfo info;
char *filter;
Also, with danpb's "improving unit test coverage" mail fresh in my head,
this needs to include a test? I think changes to domain conf generally
require a one anyhow. Looks good, but a V2 is needed with the missing
documentation.
Regards,
Jim
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 08/12] libxl: support backend domain setting for disk and net devices
This implement handling of <domain name=''/> parameter introduced in
previous patch.
Lookup on domain name (to get domain ID) requires libxlDriverPrivate
object, so it must be passed down to libxlMakeDisk and libxlMakeNet from
top level callers.
---
src/libxl/libxl_conf.c | 62 +++++++++++++++++++++++++++++++++++++++++-------
src/libxl/libxl_conf.h | 4 ++--
src/libxl/libxl_driver.c | 50 +++++++++++++++++++++-----------------
3 files changed, 84 insertions(+), 32 deletions(-)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 4bd62e9..477e46d 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -472,7 +472,9 @@ error:
}
int
-libxlMakeDisk(virDomainDiskDefPtr l_disk, libxl_device_disk *x_disk)
+libxlMakeDisk(libxlDriverPrivatePtr driver,
+ virDomainDiskDefPtr l_disk,
+ libxl_device_disk *x_disk)
{
if (l_disk->src && (x_disk->pdev_path = strdup(l_disk->src)) ==
NULL) {
virReportOOMError();
@@ -549,11 +551,32 @@ libxlMakeDisk(virDomainDiskDefPtr l_disk, libxl_device_disk
*x_disk)
return -1;
}
+ if (l_disk->domain_name) {
+ uint32_t domid;
+ /* Do not use virDomainObjListFindByName as it causes deadlock here -
+ * we already have lock on this domain object, but
+ * virDomainObjListFindByName will try to take it again.
+ */
+ switch (libxl_name_to_domid(driver->ctx, l_disk->domain_name, &domid))
{
+ case ERROR_INVAL:
+ virReportError(VIR_ERR_XML_DETAIL,
+ _("Disk backend domain '%s' does not exists"),
+ l_disk->domain_name);
+ return -1;
+ case ERROR_NOMEM:
+ virReportOOMError();
+ return -1;
+ }
+ x_disk->backend_domid = domid;
+ }
+
return 0;
}
static int
-libxlMakeDiskList(virDomainDefPtr def, libxl_domain_config *d_config)
+libxlMakeDiskList(libxlDriverPrivatePtr driver,
+ virDomainDefPtr def,
+ libxl_domain_config *d_config)
{
virDomainDiskDefPtr *l_disks = def->disks;
int ndisks = def->ndisks;
@@ -566,7 +589,7 @@ libxlMakeDiskList(virDomainDefPtr def, libxl_domain_config *d_config)
}
for (i = 0; i < ndisks; i++) {
- if (libxlMakeDisk(l_disks[i], &x_disks[i]) < 0)
+ if (libxlMakeDisk(driver, l_disks[i], &x_disks[i]) < 0)
goto error;
}
@@ -583,7 +606,9 @@ error:
}
int
-libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic)
+libxlMakeNic(libxlDriverPrivatePtr driver,
+ virDomainNetDefPtr l_nic,
+ libxl_device_nic *x_nic)
{
/* TODO: Where is mtu stored?
*
@@ -620,11 +645,32 @@ libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic)
return -1;
}
+ if (l_nic->domain_name) {
+ uint32_t domid;
+ /* Do not use virDomainObjListFindByName as it causes deadlock here -
+ * we already have lock on this domain object, but
+ * virDomainObjListFindByName will try to take it again.
+ */
+ switch (libxl_name_to_domid(driver->ctx, l_nic->domain_name, &domid))
{
+ case ERROR_INVAL:
+ virReportError(VIR_ERR_XML_DETAIL,
+ _("Network backend domain '%s' does not
exists"),
+ l_nic->domain_name);
+ return -1;
+ case ERROR_NOMEM:
+ virReportOOMError();
+ return -1;
+ }
+ x_nic->backend_domid = domid;
+ }
+
return 0;
}
static int
-libxlMakeNicList(virDomainDefPtr def, libxl_domain_config *d_config)
+libxlMakeNicList(libxlDriverPrivatePtr driver,
+ virDomainDefPtr def,
+ libxl_domain_config *d_config)
{
virDomainNetDefPtr *l_nics = def->nets;
int nnics = def->nnets;
@@ -639,7 +685,7 @@ libxlMakeNicList(virDomainDefPtr def, libxl_domain_config *d_config)
for (i = 0; i < nnics; i++) {
x_nics[i].devid = i;
- if (libxlMakeNic(l_nics[i], &x_nics[i]))
+ if (libxlMakeNic(driver, l_nics[i], &x_nics[i]))
goto error;
}
@@ -878,11 +924,11 @@ libxlBuildDomainConfig(libxlDriverPrivatePtr driver,
goto error;
}
- if (libxlMakeDiskList(def, d_config) < 0) {
+ if (libxlMakeDiskList(driver, def, d_config) < 0) {
goto error;
}
- if (libxlMakeNicList(def, d_config) < 0) {
+ if (libxlMakeNicList(driver, def, d_config) < 0) {
goto error;
}
diff --git a/src/libxl/libxl_conf.h b/src/libxl/libxl_conf.h
index b3ab3bf..99948b5 100644
--- a/src/libxl/libxl_conf.h
+++ b/src/libxl/libxl_conf.h
@@ -113,9 +113,9 @@ virCapsPtr
libxlMakeCapabilities(libxl_ctx *ctx);
int
-libxlMakeDisk(virDomainDiskDefPtr l_dev, libxl_device_disk *x_dev);
+libxlMakeDisk(libxlDriverPrivatePtr driver, virDomainDiskDefPtr l_dev, libxl_device_disk
*x_dev);
int
-libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic);
+libxlMakeNic(libxlDriverPrivatePtr driver, virDomainNetDefPtr l_nic, libxl_device_nic
*x_nic);
int
libxlMakeVfb(libxlDriverPrivatePtr driver,
virDomainGraphicsDefPtr l_vfb, libxl_device_vfb *x_vfb);
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 89546a5..b0f0c6a 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -3166,8 +3166,9 @@ libxlDomainUndefine(virDomainPtr dom)
}
static int
-libxlDomainChangeEjectableMedia(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDiskDefPtr disk)
+libxlDomainChangeEjectableMedia(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm, virDomainDiskDefPtr
+ disk)
{
virDomainDiskDefPtr origdisk = NULL;
libxl_device_disk x_disk;
@@ -3196,7 +3197,7 @@ libxlDomainChangeEjectableMedia(libxlDomainObjPrivatePtr priv,
return -1;
}
- if (libxlMakeDisk(disk, &x_disk) < 0)
+ if (libxlMakeDisk(driver, disk, &x_disk) < 0)
goto cleanup;
if ((ret = libxl_cdrom_insert(priv->ctx, vm->def->id, &x_disk, NULL))
< 0) {
@@ -3221,8 +3222,9 @@ cleanup:
}
static int
-libxlDomainAttachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDeviceDefPtr dev)
+libxlDomainAttachDeviceDiskLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
{
virDomainDiskDefPtr l_disk = dev->data.disk;
libxl_device_disk x_disk;
@@ -3230,7 +3232,7 @@ libxlDomainAttachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
switch (l_disk->device) {
case VIR_DOMAIN_DISK_DEVICE_CDROM:
- ret = libxlDomainChangeEjectableMedia(priv, vm, l_disk);
+ ret = libxlDomainChangeEjectableMedia(driver, priv, vm, l_disk);
break;
case VIR_DOMAIN_DISK_DEVICE_DISK:
if (l_disk->bus == VIR_DOMAIN_DISK_BUS_XEN) {
@@ -3251,7 +3253,7 @@ libxlDomainAttachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
goto cleanup;
}
- if (libxlMakeDisk(l_disk, &x_disk) < 0)
+ if (libxlMakeDisk(driver, l_disk, &x_disk) < 0)
goto cleanup;
if ((ret = libxl_device_disk_add(priv->ctx, vm->def->id,
@@ -3282,8 +3284,9 @@ cleanup:
}
static int
-libxlDomainDetachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDeviceDefPtr dev)
+libxlDomainDetachDeviceDiskLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
{
virDomainDiskDefPtr l_disk = NULL;
libxl_device_disk x_disk;
@@ -3304,7 +3307,7 @@ libxlDomainDetachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
l_disk = vm->def->disks[i];
- if (libxlMakeDisk(l_disk, &x_disk) < 0)
+ if (libxlMakeDisk(driver, l_disk, &x_disk) < 0)
goto cleanup;
if ((ret = libxl_device_disk_remove(priv->ctx, vm->def->id,
@@ -3336,14 +3339,15 @@ cleanup:
}
static int
-libxlDomainAttachDeviceLive(libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
- virDomainDeviceDefPtr dev)
+libxlDomainAttachDeviceLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
{
int ret = -1;
switch (dev->type) {
case VIR_DOMAIN_DEVICE_DISK:
- ret = libxlDomainAttachDeviceDiskLive(priv, vm, dev);
+ ret = libxlDomainAttachDeviceDiskLive(driver, priv, vm, dev);
if (!ret)
dev->data.disk = NULL;
break;
@@ -3388,14 +3392,15 @@ libxlDomainAttachDeviceConfig(virDomainDefPtr vmdef,
virDomainDeviceDefPtr dev)
}
static int
-libxlDomainDetachDeviceLive(libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
- virDomainDeviceDefPtr dev)
+libxlDomainDetachDeviceLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
{
int ret = -1;
switch (dev->type) {
case VIR_DOMAIN_DEVICE_DISK:
- ret = libxlDomainDetachDeviceDiskLive(priv, vm, dev);
+ ret = libxlDomainDetachDeviceDiskLive(driver, priv, vm, dev);
break;
default:
@@ -3435,8 +3440,9 @@ libxlDomainDetachDeviceConfig(virDomainDefPtr vmdef,
virDomainDeviceDefPtr dev)
}
static int
-libxlDomainUpdateDeviceLive(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDeviceDefPtr dev)
+libxlDomainUpdateDeviceLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
{
virDomainDiskDefPtr disk;
int ret = -1;
@@ -3446,7 +3452,7 @@ libxlDomainUpdateDeviceLive(libxlDomainObjPrivatePtr priv,
disk = dev->data.disk;
switch (disk->device) {
case VIR_DOMAIN_DISK_DEVICE_CDROM:
- ret = libxlDomainChangeEjectableMedia(priv, vm, disk);
+ ret = libxlDomainChangeEjectableMedia(driver, priv, vm, disk);
if (ret == 0)
dev->data.disk = NULL;
break;
@@ -3601,13 +3607,13 @@ libxlDomainModifyDeviceFlags(virDomainPtr dom, const char *xml,
switch (action) {
case LIBXL_DEVICE_ATTACH:
- ret = libxlDomainAttachDeviceLive(priv, vm, dev);
+ ret = libxlDomainAttachDeviceLive(driver, priv, vm, dev);
break;
case LIBXL_DEVICE_DETACH:
- ret = libxlDomainDetachDeviceLive(priv, vm, dev);
+ ret = libxlDomainDetachDeviceLive(driver, priv, vm, dev);
break;
case LIBXL_DEVICE_UPDATE:
- ret = libxlDomainUpdateDeviceLive(priv, vm, dev);
+ ret = libxlDomainUpdateDeviceLive(driver, priv, vm, dev);
break;
default:
virReportError(VIR_ERR_INTERNAL_ERROR,
--
1.8.1.4
11:37 a.m.
New subject: [libvirt] [PATCH RESENT 08/12] libxl: support backend domain setting for disk and net devices
Marek Marczykowski wrote:
This implement handling of <domain name=''/> parameter
introduced in
previous patch.
Lookup on domain name (to get domain ID) requires libxlDriverPrivate
object, so it must be passed down to libxlMakeDisk and libxlMakeNet from
top level callers.
---
src/libxl/libxl_conf.c | 62 +++++++++++++++++++++++++++++++++++++++++-------
src/libxl/libxl_conf.h | 4 ++--
src/libxl/libxl_driver.c | 50 +++++++++++++++++++++-----------------
3 files changed, 84 insertions(+), 32 deletions(-)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 4bd62e9..477e46d 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -472,7 +472,9 @@ error:
}
int
-libxlMakeDisk(virDomainDiskDefPtr l_disk, libxl_device_disk *x_disk)
+libxlMakeDisk(libxlDriverPrivatePtr driver,
+ virDomainDiskDefPtr l_disk,
+ libxl_device_disk *x_disk)
Indentation looks off here.
{
if (l_disk->src && (x_disk->pdev_path = strdup(l_disk->src)) ==
NULL) {
virReportOOMError();
@@ -549,11 +551,32 @@ libxlMakeDisk(virDomainDiskDefPtr l_disk, libxl_device_disk
*x_disk)
return -1;
}
+ if (l_disk->domain_name) {
+ uint32_t domid;
+ /* Do not use virDomainObjListFindByName as it causes deadlock here -
+ * we already have lock on this domain object, but
+ * virDomainObjListFindByName will try to take it again.
+ */
+ switch (libxl_name_to_domid(driver->ctx, l_disk->domain_name, &domid))
{
+ case ERROR_INVAL:
+ virReportError(VIR_ERR_XML_DETAIL,
+ _("Disk backend domain '%s' does not
exists"),
I think this error would be better worded with "Backend domain '%s' does
not exist for disk '%s'", using l_disk->dst for the disk name.
+ l_disk->domain_name);
+ return -1;
+ case ERROR_NOMEM:
+ virReportOOMError();
+ return -1;
+ }
+ x_disk->backend_domid = domid;
IMO this could be coded a little more defensively, e.g. only assign
backend_domid on "case 0:", and have a default case that will also fail
if libxl_name_to_domid() ever returns something besides 0, ERROR_INVAL,
or ERROR_NOMEM.
+ }
+
return 0;
}
static int
-libxlMakeDiskList(virDomainDefPtr def, libxl_domain_config *d_config)
+libxlMakeDiskList(libxlDriverPrivatePtr driver,
+ virDomainDefPtr def,
+ libxl_domain_config *d_config)
Indentation.
{
virDomainDiskDefPtr *l_disks = def->disks;
int ndisks = def->ndisks;
@@ -566,7 +589,7 @@ libxlMakeDiskList(virDomainDefPtr def, libxl_domain_config
*d_config)
}
for (i = 0; i < ndisks; i++) {
- if (libxlMakeDisk(l_disks[i], &x_disks[i]) < 0)
+ if (libxlMakeDisk(driver, l_disks[i], &x_disks[i]) < 0)
goto error;
}
@@ -583,7 +606,9 @@ error:
}
int
-libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic)
+libxlMakeNic(libxlDriverPrivatePtr driver,
+ virDomainNetDefPtr l_nic,
+ libxl_device_nic *x_nic)
Indentation.
{
/* TODO: Where is mtu stored?
*
@@ -620,11 +645,32 @@ libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic)
return -1;
}
+ if (l_nic->domain_name) {
+ uint32_t domid;
+ /* Do not use virDomainObjListFindByName as it causes deadlock here -
+ * we already have lock on this domain object, but
+ * virDomainObjListFindByName will try to take it again.
+ */
+ switch (libxl_name_to_domid(driver->ctx, l_nic->domain_name, &domid))
{
+ case ERROR_INVAL:
+ virReportError(VIR_ERR_XML_DETAIL,
+ _("Network backend domain '%s' does not
exists"),
Same comment here about the error message.
+ l_nic->domain_name);
+ return -1;
+ case ERROR_NOMEM:
+ virReportOOMError();
+ return -1;
+ }
+ x_nic->backend_domid = domid;
And same comment here about the switch statement.
+ }
+
return 0;
}
static int
-libxlMakeNicList(virDomainDefPtr def, libxl_domain_config *d_config)
+libxlMakeNicList(libxlDriverPrivatePtr driver,
+ virDomainDefPtr def,
+ libxl_domain_config *d_config)
Indentation.
{
virDomainNetDefPtr *l_nics = def->nets;
int nnics = def->nnets;
@@ -639,7 +685,7 @@ libxlMakeNicList(virDomainDefPtr def, libxl_domain_config
*d_config)
for (i = 0; i < nnics; i++) {
x_nics[i].devid = i;
- if (libxlMakeNic(l_nics[i], &x_nics[i]))
+ if (libxlMakeNic(driver, l_nics[i], &x_nics[i]))
goto error;
}
@@ -878,11 +924,11 @@ libxlBuildDomainConfig(libxlDriverPrivatePtr driver,
goto error;
}
- if (libxlMakeDiskList(def, d_config) < 0) {
+ if (libxlMakeDiskList(driver, def, d_config) < 0) {
goto error;
}
- if (libxlMakeNicList(def, d_config) < 0) {
+ if (libxlMakeNicList(driver, def, d_config) < 0) {
goto error;
}
diff --git a/src/libxl/libxl_conf.h b/src/libxl/libxl_conf.h
index b3ab3bf..99948b5 100644
--- a/src/libxl/libxl_conf.h
+++ b/src/libxl/libxl_conf.h
@@ -113,9 +113,9 @@ virCapsPtr
libxlMakeCapabilities(libxl_ctx *ctx);
int
-libxlMakeDisk(virDomainDiskDefPtr l_dev, libxl_device_disk *x_dev);
+libxlMakeDisk(libxlDriverPrivatePtr driver, virDomainDiskDefPtr l_dev, libxl_device_disk
*x_dev);
This is more than 80 columns now.
int
-libxlMakeNic(virDomainNetDefPtr l_nic, libxl_device_nic *x_nic);
+libxlMakeNic(libxlDriverPrivatePtr driver, virDomainNetDefPtr l_nic, libxl_device_nic
*x_nic);
Same here.
int
libxlMakeVfb(libxlDriverPrivatePtr driver,
virDomainGraphicsDefPtr l_vfb, libxl_device_vfb *x_vfb);
diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c
index 89546a5..b0f0c6a 100644
--- a/src/libxl/libxl_driver.c
+++ b/src/libxl/libxl_driver.c
@@ -3166,8 +3166,9 @@ libxlDomainUndefine(virDomainPtr dom)
}
static int
-libxlDomainChangeEjectableMedia(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDiskDefPtr disk)
+libxlDomainChangeEjectableMedia(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm, virDomainDiskDefPtr
+ disk)
Whitespace issues here. There are no strict guidelines for the
whitespace format of function definitions/declarations, but the dominate
pattern in libvirt seems to be return type on first line, then name and
params on next line. If name and params exceeds 80 columns, params 2..n
get their own line. I've tried to move to this pattern in the libxl
driver as I touch offending code.
{
virDomainDiskDefPtr origdisk = NULL;
libxl_device_disk x_disk;
@@ -3196,7 +3197,7 @@ libxlDomainChangeEjectableMedia(libxlDomainObjPrivatePtr priv,
return -1;
}
- if (libxlMakeDisk(disk, &x_disk) < 0)
+ if (libxlMakeDisk(driver, disk, &x_disk) < 0)
goto cleanup;
if ((ret = libxl_cdrom_insert(priv->ctx, vm->def->id, &x_disk, NULL))
< 0) {
@@ -3221,8 +3222,9 @@ cleanup:
}
static int
-libxlDomainAttachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDeviceDefPtr dev)
+libxlDomainAttachDeviceDiskLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
Indentation.
{
virDomainDiskDefPtr l_disk = dev->data.disk;
libxl_device_disk x_disk;
@@ -3230,7 +3232,7 @@ libxlDomainAttachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
switch (l_disk->device) {
case VIR_DOMAIN_DISK_DEVICE_CDROM:
- ret = libxlDomainChangeEjectableMedia(priv, vm, l_disk);
+ ret = libxlDomainChangeEjectableMedia(driver, priv, vm, l_disk);
break;
case VIR_DOMAIN_DISK_DEVICE_DISK:
if (l_disk->bus == VIR_DOMAIN_DISK_BUS_XEN) {
@@ -3251,7 +3253,7 @@ libxlDomainAttachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
goto cleanup;
}
- if (libxlMakeDisk(l_disk, &x_disk) < 0)
+ if (libxlMakeDisk(driver, l_disk, &x_disk) < 0)
goto cleanup;
if ((ret = libxl_device_disk_add(priv->ctx, vm->def->id,
@@ -3282,8 +3284,9 @@ cleanup:
}
static int
-libxlDomainDetachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDeviceDefPtr dev)
+libxlDomainDetachDeviceDiskLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
Indentation.
{
virDomainDiskDefPtr l_disk = NULL;
libxl_device_disk x_disk;
@@ -3304,7 +3307,7 @@ libxlDomainDetachDeviceDiskLive(libxlDomainObjPrivatePtr priv,
l_disk = vm->def->disks[i];
- if (libxlMakeDisk(l_disk, &x_disk) < 0)
+ if (libxlMakeDisk(driver, l_disk, &x_disk) < 0)
goto cleanup;
if ((ret = libxl_device_disk_remove(priv->ctx, vm->def->id,
@@ -3336,14 +3339,15 @@ cleanup:
}
static int
-libxlDomainAttachDeviceLive(libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
- virDomainDeviceDefPtr dev)
+libxlDomainAttachDeviceLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
Indentation.
{
int ret = -1;
switch (dev->type) {
case VIR_DOMAIN_DEVICE_DISK:
- ret = libxlDomainAttachDeviceDiskLive(priv, vm, dev);
+ ret = libxlDomainAttachDeviceDiskLive(driver, priv, vm, dev);
if (!ret)
dev->data.disk = NULL;
break;
@@ -3388,14 +3392,15 @@ libxlDomainAttachDeviceConfig(virDomainDefPtr vmdef,
virDomainDeviceDefPtr dev)
}
static int
-libxlDomainDetachDeviceLive(libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
- virDomainDeviceDefPtr dev)
+libxlDomainDetachDeviceLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
Indentation.
{
int ret = -1;
switch (dev->type) {
case VIR_DOMAIN_DEVICE_DISK:
- ret = libxlDomainDetachDeviceDiskLive(priv, vm, dev);
+ ret = libxlDomainDetachDeviceDiskLive(driver, priv, vm, dev);
break;
default:
@@ -3435,8 +3440,9 @@ libxlDomainDetachDeviceConfig(virDomainDefPtr vmdef,
virDomainDeviceDefPtr dev)
}
static int
-libxlDomainUpdateDeviceLive(libxlDomainObjPrivatePtr priv,
- virDomainObjPtr vm, virDomainDeviceDefPtr dev)
+libxlDomainUpdateDeviceLive(libxlDriverPrivatePtr driver,
+ libxlDomainObjPrivatePtr priv, virDomainObjPtr vm,
+ virDomainDeviceDefPtr dev)
Indentation.
Looks good with the exception of the nits. Depends on patch 7 of course...
Regards,
Jim
{
virDomainDiskDefPtr disk;
int ret = -1;
@@ -3446,7 +3452,7 @@ libxlDomainUpdateDeviceLive(libxlDomainObjPrivatePtr priv,
disk = dev->data.disk;
switch (disk->device) {
case VIR_DOMAIN_DISK_DEVICE_CDROM:
- ret = libxlDomainChangeEjectableMedia(priv, vm, disk);
+ ret = libxlDomainChangeEjectableMedia(driver, priv, vm, disk);
if (ret == 0)
dev->data.disk = NULL;
break;
@@ -3601,13 +3607,13 @@ libxlDomainModifyDeviceFlags(virDomainPtr dom, const char *xml,
switch (action) {
case LIBXL_DEVICE_ATTACH:
- ret = libxlDomainAttachDeviceLive(priv, vm, dev);
+ ret = libxlDomainAttachDeviceLive(driver, priv, vm, dev);
break;
case LIBXL_DEVICE_DETACH:
- ret = libxlDomainDetachDeviceLive(priv, vm, dev);
+ ret = libxlDomainDetachDeviceLive(driver, priv, vm, dev);
break;
case LIBXL_DEVICE_UPDATE:
- ret = libxlDomainUpdateDeviceLive(priv, vm, dev);
+ ret = libxlDomainUpdateDeviceLive(driver, priv, vm, dev);
break;
default:
virReportError(VIR_ERR_INTERNAL_ERROR,
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 10/12] RFC: libxl: special 'stubdom-dm' emulator to use qemu in stub domain
Xen have feature of having device model in separate domain (called stub
domain). It used to be enabled by special 'stubdom-dm' device model
path. Recent xl have separate config option for this feature
(device_model_stubdomain_override), but I'm not sure if it worth
introducing another xen-specific option in general domain XML syntax.
Also use os->cmdline as extra arguments for qemu (for HVM only). This
use of os->cmdline is rather dirty hack, but I haven't idea what better
attribute use for it. Perhaps worth new attribute (in <emulator/>
element?). Anyway os->cmdline was unused for HVM domains.
---
src/libxl/libxl_conf.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 068a97a..181d344 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -478,6 +478,18 @@ libxlMakeDomBuildInfo(libxlDriverPrivatePtr driver,
goto error;
}
+ if (def->emulator && strcmp(def->emulator,
"stubdom-dm")==0) {
+ libxl_defbool_set(&b_info->device_model_stubdomain, 1);
+ }
+
+ if (def->os.cmdline && def->os.cmdline[0]) {
+ b_info->extra_hvm = virStringSplit(def->os.cmdline, " ", 0);
+ if (b_info->extra_hvm == NULL) {
+ virReportOOMError();
+ goto error;
+ }
+ }
+
/* Disable VNC and SDL until explicitly enabled */
libxl_defbool_set(&b_info->u.hvm.vnc.enable, 0);
libxl_defbool_set(&b_info->u.hvm.sdl.enable, 0);
--
1.8.1.4
12:32 p.m.
New subject: [libvirt] [PATCH RESENT 10/12] RFC: libxl: special 'stubdom-dm' emulator to use qemu in stub domain
Marek Marczykowski wrote:
Xen have feature of having device model in separate domain (called
stub
domain). It used to be enabled by special 'stubdom-dm' device model
path. Recent xl have separate config option for this feature
(device_model_stubdomain_override), but I'm not sure if it worth
introducing another xen-specific option in general domain XML syntax.
Also use os->cmdline as extra arguments for qemu (for HVM only). This
use of os->cmdline is rather dirty hack, but I haven't idea what better
attribute use for it. Perhaps worth new attribute (in <emulator/>
element?). Anyway os->cmdline was unused for HVM domains.
---
src/libxl/libxl_conf.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 068a97a..181d344 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -478,6 +478,18 @@ libxlMakeDomBuildInfo(libxlDriverPrivatePtr driver,
goto error;
}
+ if (def->emulator && strcmp(def->emulator,
"stubdom-dm")==0) {
+ libxl_defbool_set(&b_info->device_model_stubdomain, 1);
+ }
This is another case of libxl exposing some internal implementation
detail IMO. Apps know what emulator they want to use, they don't know
what special xen bit to set, e.g. stubdom, traditional, new, newer, ...
+
+ if (def->os.cmdline && def->os.cmdline[0]) {
+ b_info->extra_hvm = virStringSplit(def->os.cmdline, " ",
0);
+ if (b_info->extra_hvm == NULL) {
+ virReportOOMError();
+ goto error;
+ }
+ }
+
NACK to this change. qemu supports kernel, initrd, and append args for
direct kernel boot, which is useful when installing a guest OS. libxl
should support this for HVM guests. I'm pretty sure this works in the
legacy stack.
Regards,
Jim
/* Disable VNC and SDL until explicitly enabled */
libxl_defbool_set(&b_info->u.hvm.vnc.enable, 0);
libxl_defbool_set(&b_info->u.hvm.sdl.enable, 0);
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 09/12] libxl: fill HVM SDL and VNC settings based on <graphics/> entries
Vfb entries in domain config are used only by PV drivers. Qemu
parameters are build based on b_info struct. So fill it with the same
data as vfb entries (actually the first one).
This will additionally allow graphic-less domain, when no <graphics/>
entries are present in domain XML (previously VNC was always enabled).
---
src/libxl/libxl_conf.c | 131 +++++++++++++++++++++++++++++++++----------------
1 file changed, 90 insertions(+), 41 deletions(-)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 477e46d..068a97a 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -347,7 +347,64 @@ error:
}
static int
-libxlMakeDomBuildInfo(virDomainDefPtr def, libxl_domain_config *d_config)
+libxlMakeVNCInfo(libxlDriverPrivatePtr driver,
+ virDomainGraphicsDefPtr l_vfb,
+ libxl_vnc_info *x_vnc)
+{
+ unsigned short port;
+ const char *listenAddr;
+
+ libxl_defbool_set(&x_vnc->enable, 1);
+ /* driver handles selection of free port */
+ libxl_defbool_set(&x_vnc->findunused, 0);
+ if (l_vfb->data.vnc.autoport) {
+
+ if (virPortAllocatorAcquire(driver->reservedVNCPorts, &port) < 0)
+ return -1;
+ if (port == 0) {
+ virReportError(VIR_ERR_INTERNAL_ERROR,
+ "%s", _("Unable to find an unused VNC port"));
+ return -1;
+ }
+ l_vfb->data.vnc.port = port;
+ }
+ x_vnc->display = l_vfb->data.vnc.port - LIBXL_VNC_PORT_MIN;
+
+ listenAddr = virDomainGraphicsListenGetAddress(l_vfb, 0);
+ if (listenAddr) {
+ /* libxl_device_vfb_init() does strdup("127.0.0.1") */
+ VIR_FREE(x_vnc->listen);
+ if ((x_vnc->listen = strdup(listenAddr)) == NULL) {
+ virReportOOMError();
+ return -1;
+ }
+ }
+ return 0;
+}
+
+static int
+libxlMakeSDLInfo(virDomainGraphicsDefPtr l_vfb,
+ libxl_sdl_info *x_sdl)
+{
+ libxl_defbool_set(&x_sdl->enable, 1);
+ if (l_vfb->data.sdl.display &&
+ (x_sdl->display = strdup(l_vfb->data.sdl.display)) == NULL) {
+ virReportOOMError();
+ return -1;
+ }
+ if (l_vfb->data.sdl.xauth &&
+ (x_sdl->xauthority =
+ strdup(l_vfb->data.sdl.xauth)) == NULL) {
+ virReportOOMError();
+ return -1;
+ }
+ return 0;
+}
+
+static int
+libxlMakeDomBuildInfo(libxlDriverPrivatePtr driver,
+ virDomainDefPtr def,
+ libxl_domain_config *d_config)
{
libxl_domain_build_info *b_info = &d_config->b_info;
int hvm = STREQ(def->os.type, "hvm");
@@ -421,6 +478,34 @@ libxlMakeDomBuildInfo(virDomainDefPtr def, libxl_domain_config
*d_config)
goto error;
}
+ /* Disable VNC and SDL until explicitly enabled */
+ libxl_defbool_set(&b_info->u.hvm.vnc.enable, 0);
+ libxl_defbool_set(&b_info->u.hvm.sdl.enable, 0);
+
+ for (i = 0; i < def->ngraphics; i++) {
+ switch (def->graphics[i]->type) {
+ case VIR_DOMAIN_GRAPHICS_TYPE_VNC:
+ if (libxl_defbool_val(b_info->u.hvm.vnc.enable))
+ continue;
+ if (libxlMakeVNCInfo(driver, def->graphics[i],
+ &b_info->u.hvm.vnc) < 0)
+ goto error;
+ if (def->graphics[i]->data.vnc.keymap &&
+ (b_info->u.hvm.keymap =
+ strdup(def->graphics[i]->data.vnc.keymap)) == NULL) {
+ virReportOOMError();
+ goto error;
+ }
+ break;
+ case VIR_DOMAIN_GRAPHICS_TYPE_SDL:
+ if (libxl_defbool_val(b_info->u.hvm.sdl.enable))
+ continue;
+ if (libxlMakeSDLInfo(def->graphics[i], &b_info->u.hvm.sdl)
< 0)
+ goto error;
+ break;
+ }
+ }
+
/*
* The following comment and calculation were taken directly from
* libxenlight's internal function libxl_get_required_shadow_memory():
@@ -706,50 +791,14 @@ libxlMakeVfb(libxlDriverPrivatePtr driver,
virDomainGraphicsDefPtr l_vfb,
libxl_device_vfb *x_vfb)
{
- unsigned short port;
- const char *listenAddr;
-
switch (l_vfb->type) {
case VIR_DOMAIN_GRAPHICS_TYPE_SDL:
- libxl_defbool_set(&x_vfb->sdl.enable, 1);
- if (l_vfb->data.sdl.display &&
- (x_vfb->sdl.display = strdup(l_vfb->data.sdl.display)) == NULL) {
- virReportOOMError();
- return -1;
- }
- if (l_vfb->data.sdl.xauth &&
- (x_vfb->sdl.xauthority =
- strdup(l_vfb->data.sdl.xauth)) == NULL) {
- virReportOOMError();
+ if (libxlMakeSDLInfo(l_vfb, &x_vfb->sdl) < 0)
return -1;
- }
break;
case VIR_DOMAIN_GRAPHICS_TYPE_VNC:
- libxl_defbool_set(&x_vfb->vnc.enable, 1);
- /* driver handles selection of free port */
- libxl_defbool_set(&x_vfb->vnc.findunused, 0);
- if (l_vfb->data.vnc.autoport) {
-
- if (virPortAllocatorAcquire(driver->reservedVNCPorts, &port) <
0)
- return -1;
- if (port == 0) {
- virReportError(VIR_ERR_INTERNAL_ERROR,
- "%s", _("Unable to find an unused VNC
port"));
- return -1;
- }
- l_vfb->data.vnc.port = port;
- }
- x_vfb->vnc.display = l_vfb->data.vnc.port - LIBXL_VNC_PORT_MIN;
-
- listenAddr = virDomainGraphicsListenGetAddress(l_vfb, 0);
- if (listenAddr) {
- /* libxl_device_vfb_init() does strdup("127.0.0.1") */
- VIR_FREE(x_vfb->vnc.listen);
- if ((x_vfb->vnc.listen = strdup(listenAddr)) == NULL) {
- virReportOOMError();
- return -1;
- }
- }
+ if (libxlMakeVNCInfo(driver, l_vfb, &x_vfb->vnc) < 0)
+ return -1;
if (l_vfb->data.vnc.keymap &&
(x_vfb->keymap =
strdup(l_vfb->data.vnc.keymap)) == NULL) {
@@ -920,7 +969,7 @@ libxlBuildDomainConfig(libxlDriverPrivatePtr driver,
if (libxlMakeDomCreateInfo(driver, def, &d_config->c_info) < 0)
return -1;
- if (libxlMakeDomBuildInfo(def, d_config) < 0) {
+ if (libxlMakeDomBuildInfo(driver, def, d_config) < 0) {
goto error;
}
--
1.8.1.4
12:15 p.m.
New subject: [libvirt] [PATCH RESENT 09/12] libxl: fill HVM SDL and VNC settings based on <graphics/> entries
Marek Marczykowski wrote:
Vfb entries in domain config are used only by PV drivers. Qemu
parameters are build based on b_info struct. So fill it with the same
s/build/built/
data as vfb entries (actually the first one).
This will additionally allow graphic-less domain, when no <graphics/>
entries are present in domain XML (previously VNC was always enabled).
---
src/libxl/libxl_conf.c | 131 +++++++++++++++++++++++++++++++++----------------
1 file changed, 90 insertions(+), 41 deletions(-)
diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
index 477e46d..068a97a 100644
--- a/src/libxl/libxl_conf.c
+++ b/src/libxl/libxl_conf.c
@@ -347,7 +347,64 @@ error:
}
static int
-libxlMakeDomBuildInfo(virDomainDefPtr def, libxl_domain_config *d_config)
+libxlMakeVNCInfo(libxlDriverPrivatePtr driver,
+ virDomainGraphicsDefPtr l_vfb,
+ libxl_vnc_info *x_vnc)
Indentation.
+{
+ unsigned short port;
+ const char *listenAddr;
+
+ libxl_defbool_set(&x_vnc->enable, 1);
+ /* driver handles selection of free port */
+ libxl_defbool_set(&x_vnc->findunused, 0);
+ if (l_vfb->data.vnc.autoport) {
+
+ if (virPortAllocatorAcquire(driver->reservedVNCPorts, &port) < 0)
+ return -1;
+ if (port == 0) {
+ virReportError(VIR_ERR_INTERNAL_ERROR,
+ "%s", _("Unable to find an unused VNC port"));
+ return -1;
+ }
+ l_vfb->data.vnc.port = port;
+ }
+ x_vnc->display = l_vfb->data.vnc.port - LIBXL_VNC_PORT_MIN;
+
+ listenAddr = virDomainGraphicsListenGetAddress(l_vfb, 0);
+ if (listenAddr) {
+ /* libxl_device_vfb_init() does strdup("127.0.0.1") */
+ VIR_FREE(x_vnc->listen);
+ if ((x_vnc->listen = strdup(listenAddr)) == NULL) {
+ virReportOOMError();
+ return -1;
+ }
Thanks to my delay in reviewing this, you'll need to rebase and use
VIR_STRDUP :(.
+ }
+ return 0;
+}
+
+static int
+libxlMakeSDLInfo(virDomainGraphicsDefPtr l_vfb,
+ libxl_sdl_info *x_sdl)
Indentation.
+{
+ libxl_defbool_set(&x_sdl->enable, 1);
+ if (l_vfb->data.sdl.display &&
+ (x_sdl->display = strdup(l_vfb->data.sdl.display)) == NULL) {
+ virReportOOMError();
+ return -1;
+ }
+ if (l_vfb->data.sdl.xauth &&
+ (x_sdl->xauthority =
+ strdup(l_vfb->data.sdl.xauth)) == NULL) {
+ virReportOOMError();
+ return -1;
+ }
+ return 0;
+}
+
+static int
+libxlMakeDomBuildInfo(libxlDriverPrivatePtr driver,
+ virDomainDefPtr def,
+ libxl_domain_config *d_config)
Indentation.
Looks good but will need a rebase.
Regards,
Jim
{
libxl_domain_build_info *b_info = &d_config->b_info;
int hvm = STREQ(def->os.type, "hvm");
@@ -421,6 +478,34 @@ libxlMakeDomBuildInfo(virDomainDefPtr def, libxl_domain_config
*d_config)
goto error;
}
+ /* Disable VNC and SDL until explicitly enabled */
+ libxl_defbool_set(&b_info->u.hvm.vnc.enable, 0);
+ libxl_defbool_set(&b_info->u.hvm.sdl.enable, 0);
+
+ for (i = 0; i < def->ngraphics; i++) {
+ switch (def->graphics[i]->type) {
+ case VIR_DOMAIN_GRAPHICS_TYPE_VNC:
+ if (libxl_defbool_val(b_info->u.hvm.vnc.enable))
+ continue;
+ if (libxlMakeVNCInfo(driver, def->graphics[i],
+ &b_info->u.hvm.vnc) < 0)
+ goto error;
+ if (def->graphics[i]->data.vnc.keymap &&
+ (b_info->u.hvm.keymap =
+ strdup(def->graphics[i]->data.vnc.keymap)) == NULL)
{
+ virReportOOMError();
+ goto error;
+ }
+ break;
+ case VIR_DOMAIN_GRAPHICS_TYPE_SDL:
+ if (libxl_defbool_val(b_info->u.hvm.sdl.enable))
+ continue;
+ if (libxlMakeSDLInfo(def->graphics[i], &b_info->u.hvm.sdl)
< 0)
+ goto error;
+ break;
+ }
+ }
+
/*
* The following comment and calculation were taken directly from
* libxenlight's internal function libxl_get_required_shadow_memory():
@@ -706,50 +791,14 @@ libxlMakeVfb(libxlDriverPrivatePtr driver,
virDomainGraphicsDefPtr l_vfb,
libxl_device_vfb *x_vfb)
{
- unsigned short port;
- const char *listenAddr;
-
switch (l_vfb->type) {
case VIR_DOMAIN_GRAPHICS_TYPE_SDL:
- libxl_defbool_set(&x_vfb->sdl.enable, 1);
- if (l_vfb->data.sdl.display &&
- (x_vfb->sdl.display = strdup(l_vfb->data.sdl.display)) == NULL) {
- virReportOOMError();
- return -1;
- }
- if (l_vfb->data.sdl.xauth &&
- (x_vfb->sdl.xauthority =
- strdup(l_vfb->data.sdl.xauth)) == NULL) {
- virReportOOMError();
+ if (libxlMakeSDLInfo(l_vfb, &x_vfb->sdl) < 0)
return -1;
- }
break;
case VIR_DOMAIN_GRAPHICS_TYPE_VNC:
- libxl_defbool_set(&x_vfb->vnc.enable, 1);
- /* driver handles selection of free port */
- libxl_defbool_set(&x_vfb->vnc.findunused, 0);
- if (l_vfb->data.vnc.autoport) {
-
- if (virPortAllocatorAcquire(driver->reservedVNCPorts, &port) <
0)
- return -1;
- if (port == 0) {
- virReportError(VIR_ERR_INTERNAL_ERROR,
- "%s", _("Unable to find an unused VNC
port"));
- return -1;
- }
- l_vfb->data.vnc.port = port;
- }
- x_vfb->vnc.display = l_vfb->data.vnc.port - LIBXL_VNC_PORT_MIN;
-
- listenAddr = virDomainGraphicsListenGetAddress(l_vfb, 0);
- if (listenAddr) {
- /* libxl_device_vfb_init() does strdup("127.0.0.1") */
- VIR_FREE(x_vfb->vnc.listen);
- if ((x_vfb->vnc.listen = strdup(listenAddr)) == NULL) {
- virReportOOMError();
- return -1;
- }
- }
+ if (libxlMakeVNCInfo(driver, l_vfb, &x_vfb->vnc) < 0)
+ return -1;
if (l_vfb->data.vnc.keymap &&
(x_vfb->keymap =
strdup(l_vfb->data.vnc.keymap)) == NULL) {
@@ -920,7 +969,7 @@ libxlBuildDomainConfig(libxlDriverPrivatePtr driver,
if (libxlMakeDomCreateInfo(driver, def, &d_config->c_info) < 0)
return -1;
- if (libxlMakeDomBuildInfo(def, d_config) < 0) {
+ if (libxlMakeDomBuildInfo(driver, def, d_config) < 0) {
goto error;
}
9:44 p.m.
New subject: [libvirt] [PATCH RESENT 11/12] conf: virDomainObjListRemoveLocked function
While iterating with virDomainObjListForEach it is safe to remove
current element. But while iterating, 'doms' lock is already taken, so
can't use standard virDomainObjListRemove. So introduce
virDomainObjListRemoveLocked for this purpose.
---
src/conf/domain_conf.c | 17 +++++++++++++++++
src/conf/domain_conf.h | 2 ++
src/libvirt_private.syms | 1 +
3 files changed, 20 insertions(+)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index bf1fec6..e4f7288 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -2199,6 +2199,23 @@ void virDomainObjListRemove(virDomainObjListPtr doms,
virObjectUnlock(doms);
}
+/* The caller must hold lock on 'doms' in addition to
'virDomainObjListRemove'
+ * requirements
+ *
+ * Can be used to remove current element while iterating with
+ * virDomainObjListForEach
+ */
+void virDomainObjListRemoveLocked(virDomainObjListPtr doms,
+ virDomainObjPtr dom)
+{
+ char uuidstr[VIR_UUID_STRING_BUFLEN];
+
+ virUUIDFormat(dom->def->uuid, uuidstr);
+ virObjectUnlock(dom);
+
+ virHashRemoveEntry(doms->objs, uuidstr);
+}
+
static int
virDomainDeviceCCWAddressIsValid(virDomainDeviceCCWAddressPtr addr)
{
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index db3002b..e34143b 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -2073,6 +2073,8 @@ virDomainObjCopyPersistentDef(virCapsPtr caps, virDomainXMLConfPtr
xmlconf,
void virDomainObjListRemove(virDomainObjListPtr doms,
virDomainObjPtr dom);
+void virDomainObjListRemoveLocked(virDomainObjListPtr doms,
+ virDomainObjPtr dom);
virDomainDeviceDefPtr virDomainDeviceDefParse(virCapsPtr caps,
virDomainDefPtr def,
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
index 96eea0a..e6b6b1b 100644
--- a/src/libvirt_private.syms
+++ b/src/libvirt_private.syms
@@ -270,6 +270,7 @@ virDomainObjListLoadAllConfigs;
virDomainObjListNew;
virDomainObjListNumOfDomains;
virDomainObjListRemove;
+virDomainObjListRemoveLocked;
virDomainObjNew;
virDomainObjSetDefTransient;
virDomainObjSetState;
--
1.8.1.4
12:45 p.m.
New subject: [libvirt] [PATCH RESENT 11/12] conf: virDomainObjListRemoveLocked function
Marek Marczykowski wrote:
While iterating with virDomainObjListForEach it is safe to remove
current element. But while iterating, 'doms' lock is already taken, so
can't use standard virDomainObjListRemove. So introduce
virDomainObjListRemoveLocked for this purpose.
This seems sane to me, and the code looks good with the exception of my
usual indentation nit, but I'd like to get an opinion from someone more
familiar with this code.
Regards,
Jim
---
src/conf/domain_conf.c | 17 +++++++++++++++++
src/conf/domain_conf.h | 2 ++
src/libvirt_private.syms | 1 +
3 files changed, 20 insertions(+)
diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
index bf1fec6..e4f7288 100644
--- a/src/conf/domain_conf.c
+++ b/src/conf/domain_conf.c
@@ -2199,6 +2199,23 @@ void virDomainObjListRemove(virDomainObjListPtr doms,
virObjectUnlock(doms);
}
+/* The caller must hold lock on 'doms' in addition to
'virDomainObjListRemove'
+ * requirements
+ *
+ * Can be used to remove current element while iterating with
+ * virDomainObjListForEach
+ */
+void virDomainObjListRemoveLocked(virDomainObjListPtr doms,
+ virDomainObjPtr dom)
+{
+ char uuidstr[VIR_UUID_STRING_BUFLEN];
+
+ virUUIDFormat(dom->def->uuid, uuidstr);
+ virObjectUnlock(dom);
+
+ virHashRemoveEntry(doms->objs, uuidstr);
+}
+
static int
virDomainDeviceCCWAddressIsValid(virDomainDeviceCCWAddressPtr addr)
{
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index db3002b..e34143b 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -2073,6 +2073,8 @@ virDomainObjCopyPersistentDef(virCapsPtr caps, virDomainXMLConfPtr
xmlconf,
void virDomainObjListRemove(virDomainObjListPtr doms,
virDomainObjPtr dom);
+void virDomainObjListRemoveLocked(virDomainObjListPtr doms,
+ virDomainObjPtr dom);
virDomainDeviceDefPtr virDomainDeviceDefParse(virCapsPtr caps,
virDomainDefPtr def,
diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
index 96eea0a..e6b6b1b 100644
--- a/src/libvirt_private.syms
+++ b/src/libvirt_private.syms
@@ -270,6 +270,7 @@ virDomainObjListLoadAllConfigs;
virDomainObjListNew;
virDomainObjListNumOfDomains;
virDomainObjListRemove;
+virDomainObjListRemoveLocked;
virDomainObjNew;
virDomainObjSetDefTransient;
virDomainObjSetState;
4226
days inactive
4276
days old
44 comments
7 participants
participants (7)
-
Chunyan Liu -
Daniel P. Berrange -
Ian Campbell -
Jim Fehlig -
Laine Stump -
Marek Marczykowski -
Stefano Stabellini