[libvirt] security: the qemu agent command "guest-exec" may cause Insider Access
by Zhangbo (Oscar)
Hi all:
The Host Administrator is capable of running any exec in guests via the qemu-ga command "guest-exec", eg:
virsh qemu-agent-command test_guest '{"execute": "guest-exec", "arguments": {"path": "ifconfig", "arg": [ "eth1", "192.168.0.99" ],"capture-output": true } }'
{"return":{"pid":12425}}
virsh qemu-agent-command test_guest '{"execute": "guest-exec-status", "arguments": { "pid": 12425 } }'
{"return":{"exitcode":0,"exited":true}}
The example above just change the guests' ip address, the Administrator may also change guests' user password, get sensitive information, etc. which causes Insider Access.
The Administrator also can use other commands such as " guest-file-open" that also cause Insider Access.
So, how to avoid this security problem, what's your suggestion?
Thanks!
Best Regrads
Oscar
7 years, 2 months
[libvirt] [discuss] add reconnect support for vhost-user
by lu.zhipeng＠zte.com.cn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7 years, 2 months
[libvirt] Why not support additional machine args in XML such as max-ram-below-4g?
by Wuzongyong (Euler Dept)
Since qemu support the arg -machine like that:
-machine [type=]name[,prop[=value][,...]]
So I'm inquisitive about why libvirt doesn't support specify the prop in xml ?
So I can specify the qemu machine args like -max-raw-below-4g in xml, well, I mean not by specifying the -qemu-command-line in xml.
Thanks,
Zongyong Wu
7 years, 2 months
[libvirt] [PATCH] qemu: undefine is not allowed during domain starting up
by Yi Wang
Start a domain whilst undefine it, if starting failed duing ProcessLaunch,
on which period qemu exited unexpectedly, the operation will lead to failure
of undefine the domain until libvirtd restarted. The reason is that libvirtd
will unlock vm during qemuProcessStart, qemuDomainUndefineFlags can get the
lock and set vm->persistent 0 but not remove the "active" domain.
Signed-off-by: Yi Wang <wang.yi59(a)zte.com.cn>
---
src/conf/domain_conf.h | 1 +
src/qemu/qemu_driver.c | 6 ++++++
src/qemu/qemu_process.c | 3 +++
3 files changed, 10 insertions(+)
diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h
index af15ee8..f339f84 100644
--- a/src/conf/domain_conf.h
+++ b/src/conf/domain_conf.h
@@ -2468,6 +2468,7 @@ struct _virDomainObj {
virDomainSnapshotObjPtr current_snapshot;
bool hasManagedSave;
+ bool starting;
void *privateData;
void (*privateDataFreeFunc)(void *);
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 6568def..5d9acfc 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -7317,6 +7317,12 @@ qemuDomainUndefineFlags(virDomainPtr dom,
if (!(vm = qemuDomObjFromDomain(dom)))
return -1;
+ if (vm->starting) {
+ virReportError(VIR_ERR_OPERATION_INVALID,
+ "%s", _("cannot undefine during domain starting up"));
+ goto cleanup;
+ }
+
cfg = virQEMUDriverGetConfig(driver);
if (virDomainUndefineFlagsEnsureACL(dom->conn, vm->def) < 0)
diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c
index 525521a..7b708be 100644
--- a/src/qemu/qemu_process.c
+++ b/src/qemu/qemu_process.c
@@ -5847,6 +5847,8 @@ qemuProcessStart(virConnectPtr conn,
if (!migrateFrom && !snapshot)
flags |= VIR_QEMU_PROCESS_START_NEW;
+ vm->starting = true;
+
if (qemuProcessInit(driver, vm, updatedCPU,
asyncJob, !!migrateFrom, flags) < 0)
goto cleanup;
@@ -5892,6 +5894,7 @@ qemuProcessStart(virConnectPtr conn,
ret = 0;
cleanup:
+ vm->starting = false;
qemuProcessIncomingDefFree(incoming);
return ret;
--
1.8.3.1
7 years, 2 months
[libvirt] [PATCH go-xml] Add support for domain hostdev and test code
by zhenwei.pi
Signed-off-by: zhenwei.pi <zhenwei.pi(a)youruncloud.com>
---
domain.go | 36 ++++++++++++++++++++++++++++++++++++
domain_test.go | 36 ++++++++++++++++++++++++++++++++++++
2 files changed, 72 insertions(+)
diff --git a/domain.go b/domain.go
index a5d3203..2c63307 100644
--- a/domain.go
+++ b/domain.go
@@ -406,6 +406,29 @@ type DomainRNG struct {
Backend *DomainRNGBackend `xml:"backend"`
}
+type DomainHostdevAdapter struct {
+ Name string `xml:"name,attr,omitempty"`
+}
+
+type DomainHostdevSource struct {
+ Protocol string `xml:"protocol,attr,omitempty"`
+ Name string `xml:"name,attr,omitempty"`
+ Wwpn string `xml:"wwpn,attr,omitempty"`
+ Adapter *DomainHostdevAdapter `xml:"adapter"`
+ Address *DomainAddress `xml:"address"`
+}
+
+type DomainHostdev struct {
+ XMLName xml.Name `xml:"hostdev"`
+ Mode string `xml:"mode,attr"`
+ Type string `xml:"type,attr"`
+ Sgio string `xml:"sgio,attr,omitempty"`
+ Rawio string `xml:"rawio,attr,omitempty"`
+ Managed string `xml:"managed,attr,omitempty"`
+ Source *DomainHostdevSource `xml:"source"`
+ Address *DomainAddress `xml:"address"`
+}
+
type DomainDeviceList struct {
Emulator string `xml:"emulator,omitempty"`
Controllers []DomainController `xml:"controller"`
@@ -421,6 +444,7 @@ type DomainDeviceList struct {
MemBalloon *DomainMemBalloon `xml:"memballoon"`
Sounds []DomainSound `xml:"sound"`
RNGs []DomainRNG `xml:"rng"`
+ Hostdevs []DomainHostdev `xml:"hostdev"`
}
type DomainMemory struct {
@@ -792,3 +816,15 @@ func (d *DomainRNG) Marshal() (string, error) {
}
return string(doc), nil
}
+
+func (d *DomainHostdev) Unmarshal(doc string) error {
+ return xml.Unmarshal([]byte(doc), d)
+}
+
+func (d *DomainHostdev) Marshal() (string, error) {
+ doc, err := xml.MarshalIndent(d, "", " ")
+ if err != nil {
+ return "", err
+ }
+ return string(doc), nil
+}
diff --git a/domain_test.go b/domain_test.go
index 2af451f..e3abbbf 100644
--- a/domain_test.go
+++ b/domain_test.go
@@ -1457,6 +1457,42 @@ var domainTestData = []struct {
`</rng>`,
},
},
+ {
+ Object: &DomainHostdev{
+ Mode: "subsystem",
+ Type: "scsi",
+ Sgio: "unfiltered",
+ Rawio: "yes",
+ Source: &DomainHostdevSource{
+ Adapter: &DomainHostdevAdapter{
+ Name: "scsi_host0",
+ },
+ Address: &DomainAddress{
+ Type: "scsi",
+ Bus: &diskAddr.Bus,
+ Target: &diskAddr.Slot,
+ Unit: &diskAddr.Function,
+ },
+ },
+ Address: &DomainAddress{
+ Type: "drive",
+ Controller: &diskAddr.Domain,
+ Bus: &diskAddr.Bus,
+ Target: &diskAddr.Slot,
+ Unit: &diskAddr.Function,
+ },
+ },
+
+ Expected: []string{
+ `<hostdev mode="subsystem" type="scsi" sgio="unfiltered" rawio="yes">`,
+ ` <source>`,
+ ` <adapter name="scsi_host0"></adapter>`,
+ ` <address type="scsi" bus="0" target="3" unit="0"></address>`,
+ ` </source>`,
+ ` <address type="drive" controller="0" bus="0" target="3" unit="0"></address>`,
+ `</hostdev>`,
+ },
+ },
}
func TestDomain(t *testing.T) {
--
2.7.4
7 years, 2 months
[libvirt] [PATCH 0/8] qemu: Default to video type=virtio for machvirt
by Cole Robinson
This series aim is to change the qemu machvirt video type default to
virtio, but rather than continue to hack things into place in
domain_conf.c, this rearranges things to allow drivers to set a
video type default.
Patches 1-4 are small cleanups/improvements in this area
Patches 5-7 are the plumbing to allow drivers to set their own default
Patch 8 is the actual default change
https://bugzilla.redhat.com/show_bug.cgi?id=1404112
Cole Robinson (8):
qemu: parse: drop redundant video config
qemu: domain: Move some validation out of DeviceDefPostParse
qemu: annotate some VIDEO_TYPE enum switch
conf: add virDomainVideoDefNew
conf: domain: add VIDEO_TYPE_DEFAULT
conf: domain: move video type validation to DeviceDefValidate
qemu: Set default video type in qemu PostParse
qemu: Default to video type=virtio for machvirt
src/conf/domain_conf.c | 54 +++++++++++-------
src/conf/domain_conf.h | 2 +
src/libvirt_private.syms | 1 +
src/qemu/qemu_command.c | 7 +--
src/qemu/qemu_domain.c | 66 ++++++++++++++--------
src/qemu/qemu_domain_address.c | 1 +
src/qemu/qemu_monitor_json.c | 16 ++----
src/qemu/qemu_parse_command.c | 14 +----
src/qemu/qemu_process.c | 7 +--
src/vz/vz_sdk.c | 3 +-
tests/domaincapsschemadata/full.xml | 1 +
.../qemuxml2argv-aarch64-video-default.args | 24 ++++++++
.../qemuxml2argv-aarch64-video-default.xml | 17 ++++++
tests/qemuxml2argvtest.c | 6 ++
.../qemuxml2xmlout-aarch64-video-default.xml | 46 +++++++++++++++
tests/qemuxml2xmltest.c | 6 ++
16 files changed, 191 insertions(+), 80 deletions(-)
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-video-default.args
create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-aarch64-video-default.xml
create mode 100644 tests/qemuxml2xmloutdata/qemuxml2xmlout-aarch64-video-default.xml
--
2.13.0
7 years, 2 months
[libvirt] [PATCH v4 0/4] Implement migrate-getmaxdowntime command
by Scott Garfinkle
Currently, the maximum tolerable downtime for a domain being migrated is
write-only. This patch implements a way to query that value nondestructively.
I've tried to incorporate the feedback from v3; however, I've left alone a
couple of places where the feedback would have diverged from parallelism with
the existing migrate-setdowntime help/info. Also, since the main use of this
new command is likely to be (as in virt-manager) in combination with
migrate-setmaxdowntime, I don't think it's going to really help to add an
options to report this in some other unit other than milliseconds.
To-do: Ought to revisit setmaxdowntime to use the new preferred qemu interface.
Changes from [v3]:
* Incorporated suggestions, including adding an error message
* Added a patch to the news.xml file
[v1} https://www.redhat.com/archives/libvir-list/2017-July/msg00908.html
[v2} https://www.redhat.com/archives/libvir-list/2017-July/msg01010.html
[v3} https://www.redhat.com/archives/libvir-list/2017-July/msg01301.html
Scott Garfinkle (4)
Add virDomainMigrateGetMaxDowntime public API
qemu: Implement virDomainMigrateGetMaxDowntime
virsh: Add support for virDomainMigrateGetMaxDowntime
docs: document migrate-getmaxdowntime support
---
docs/news.xml | 10 +++++++
include/libvirt/libvirt-domain.h | 4 +++
src/driver-hypervisor.h | 6 +++++
src/libvirt-domain.c | 41 ++++++++++++++++++++++++++++
src/libvirt_public.syms | 4 +++
src/qemu/qemu_driver.c | 58 ++++++++++++++++++++++++++++++++++++++++
src/qemu/qemu_monitor.h | 3 +++
src/qemu/qemu_monitor_json.c | 4 +++
src/remote/remote_driver.c | 1 +
src/remote/remote_protocol.x | 16 ++++++++++-
src/remote_protocol-structs | 8 ++++++
tools/virsh-domain.c | 42 +++++++++++++++++++++++++++++
tools/virsh.pod | 6 +++++
13 files changed, 202 insertions(+), 1 deletion(-)
7 years, 2 months
[libvirt] [PATCH 0/2] Some domain name handing improvements
by Martin Kletzander
It's not like anyone would hit any of those limitations, almost...
Martin Kletzander (3):
conf: Properly truncate wide character names in
virDomainObjGetShortName
qemu: Use short domain name in qemuDomainGetPreservedMountPath
docs: Update news with domain name bug fixes
docs/news.xml | 20 +++++++++++++++
src/conf/domain_conf.c | 30 +++++++++++++++++++---
src/qemu/qemu_domain.c | 10 ++++++--
.../qemuxml2argv-aarch64-virt-default-nic.args | 2 +-
.../qemuxml2argv-hugepages-pages2.args | 4 +--
.../qemuxml2argv-hugepages-pages3.args | 5 ++--
.../qemuxml2argv-hugepages-pages5.args | 4 +--
.../qemuxml2argv-hugepages-pages6.args | 2 +-
.../qemuxml2argv-pcie-expander-bus.args | 2 +-
9 files changed, 65 insertions(+), 14 deletions(-)
--
2.14.1
7 years, 3 months
[libvirt] [PATCH 0/4] Make it easier to clean up after using virBuffer
by Martin Kletzander
There are many places in the code where virBufferCheckError() is used
and then, right after that, virBufferContentAndReset() is called. The
former has ATTRIBUTE_RETURN_CHECK, so every occurrence just checks
that. However, if the return value of the latter is also the return
value of the current function, there is no need to duplicate the work
and act upon the error twice.
This series proposes the idea of virCheckError being used for only
reporting the error [1/4] and shows an example commit on how to clean
up existing functions [2/4] so that it can be posted to our wiki under
https://wiki.libvirt.org/page/BiteSizedTasks and linked from there.
Further enhancements could go a step further and create one function
(actually a macro the same way CheckError is done) and wrap those two
lines in one so that it is even shorter. This, however, is not meant
to be part of this series.
Patches [3/4] and [4/4] utilize this for miscellaneous clean-ups in
src/conf/.
Martin Kletzander (4):
util: Umark virBufferCheckErrorInternal as ATTRIBUTE_RETURN_CHECK
util: Use virBufferCheckError to its full potential.
conf: Clean up and report error in virDomainCapsFormat
conf: Clean up and report error in virDomainGenerateMachineName
src/conf/domain_capabilities.c | 68 +++++++++++++++++-------------------------
src/conf/domain_conf.c | 16 +++++-----
src/util/virbitmap.c | 6 +---
src/util/virbuffer.h | 2 +-
4 files changed, 37 insertions(+), 55 deletions(-)
--
2.14.1
7 years, 3 months
[libvirt] [PATCH] qemu: Prevent isolation group-related guest disappearance
by Andrea Bolognani
We can't retrieve the isolation group of a device that's
not present in the system. However, it's very common for
VFs to be created late in the boot, so they might not be
present yet when libvirtd starts, which would cause the
guests using them to disappear.
If a PCI address has already been set for the host device
in question, we can assume that it either existed at some
point in the past or the user is assigning addresses
manually: in both cases, we should not error out and just
ignore the (temporary) failure.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1484254
Signed-off-by: Andrea Bolognani <abologna(a)redhat.com>
---
src/qemu/qemu_domain_address.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/src/qemu/qemu_domain_address.c b/src/qemu/qemu_domain_address.c
index 16bf0cdf9..7f12f186b 100644
--- a/src/qemu/qemu_domain_address.c
+++ b/src/qemu/qemu_domain_address.c
@@ -1012,6 +1012,18 @@ qemuDomainFillDeviceIsolationGroup(virDomainDefPtr def,
tmp = virPCIDeviceAddressGetIOMMUGroupNum(hostAddr);
if (tmp < 0) {
+ /* If there's already a PCI address assigned to the device
+ * we move on instead of erroring out.
+ *
+ * This means we still don't allow non-existing host
+ * devices to be assigned to guests; however, if the host
+ * device existed at some point in the past but no longer
+ * does, which can happen very easily when dealing with VFs,
+ * we prevent the guest from disappearing and give the user
+ * an opportunity to edit its configuration */
+ if (virDeviceInfoPCIAddressPresent(info))
+ goto skip;
+
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Can't look up isolation group for host device "
"%04x:%02x:%02x.%x"),
--
2.13.5
7 years, 3 months
