Hi Eric,
On Thu, 2013-12-19 at 13:42 -0700, Eric Blake wrote:
-DO_CTX_TEST(true, cacert4req.filename, servercert4req.filename,
GNUTLS_VERSION_MAJOR = 3);
+DO_CTX_TEST(true, cacert4req.filename, servercert4req.filename,
+GNUTLS_VERSION_MAJOR = 3
On 12/19/2013 07:41 PM, Ján Tomko wrote:
On 12/19/2013 06:18 PM, Jiri Denemark wrote:
On Thu, Dec 19, 2013 at 17:43:33 +0100, Jano Tomko wrote:
Don't set the PCI address of primary video to the default
if it's already occupied by another device.
Without this, a primary video card that
Sorry, I must have taken some bad pills. Just realized that I replied
to a message already more than a year old. I swear my MUA displayed it
as new...
At Thu, 19 Dec 2013 16:55:58 +0100,
Claudio Bley wrote:
At Thu, 18 Oct 2012 15:48:22 +,
Benjamin Wang (gendwang) wrote:
--
AV-Test GmbH,
thanks to all and that is working with libvirt-1.2.0 and xen-4.3.1.
thanks to all people for their valuable suggestions.
regards,
cooldharma06.
On Thu, Dec 19, 2013 at 5:01 PM, Dario Faggioli
dario.faggi...@citrix.comwrote:
[Moving this to libvir, libvir-users in Bcc. Also, added xen-devel]
hi , I am confused by pinvcpu before migration.
I have two physical machine. One has 12 CPUs (E5645 2*6 ) as src and the other
has 8 CPUs (2*4) as des.
Step 1 : VM is running on src host (E5645 2*6 CPUs) with its VCPUs pinned on
some PCPUs
Step 2 :Because PCPU count is different between src
[snip]
for (i = 0; i *nparams; i++) {
virNodeCPUStatsPtr param = params[i];
What about this?
diff --git a/src/nodeinfo.c b/src/nodeinfo.c
index 1838547..aa1ad81 100644
--- a/src/nodeinfo.c
+++ b/src/nodeinfo.c
@@ -637,8 +637,9 @@ linuxNodeGetCPUStats(FILE
On Thu, 2013-12-19 at 11:39 -0700, Jim Fehlig wrote:
Stefan Bader wrote:
Oh, just while talking about setdefault. Jim, this is one of the odd things
when
moving from xm to xl stack from libvirt: libvirt defaults to the netfront
NIC
when no model is specified and sets the type. The
On 19.12.2013 18:57, Ian Campbell wrote:
On Thu, 2013-12-19 at 18:06 +0100, Stefan Bader wrote:
How about we:
* move the init to setdefault to catch the single NIC added via
hotplug case
Init of devid?
Yes, sorry for not being clear.
Hm, would that work as I am not sure
On Thu, Dec 19, 2013 at 01:42:37PM -0700, Eric Blake wrote:
On 12/19/2013 09:23 AM, Cédric Bosdonnat wrote:
Changed the constraints on gnutls to 3.1+
---
tests/virnettlscontexttest.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/tests/virnettlscontexttest.c
On Fri, Dec 20, 2013 at 10:17:22AM +0200, Laine Stump wrote:
On 12/19/2013 07:41 PM, Ján Tomko wrote:
On 12/19/2013 06:18 PM, Jiri Denemark wrote:
On Thu, Dec 19, 2013 at 17:43:33 +0100, Jano Tomko wrote:
Don't set the PCI address of primary video to the default
if it's already occupied
On Fri, 2013-12-20 at 11:16 +0100, Stefan Bader wrote:
One issue is that xl stored the guest config and then retrieves it for
use in xl list -l, but libvirt != xl and therefore has no config file to
save.
Right, that kind of was what I tried to say in many words. :) Oh, hm probably
with
From: Martin Kletzander mklet...@redhat.com
The function doesn't check whether the request is made for active or
inactive domain. Thus when the domain is not running it still tries
accessing non-existing cgroups (priv-cgroup, which is NULL).
I re-made the function in order for it to work the
The first of these 2 issues is a security issue CVE-2013-6436
Both patches have been pushed to GIT master.
Martin Kletzander (2):
CVE-2013-6436: fix crash in lxcDomainGetMemoryParameters
Fix crash in lxcDomainSetMemoryParameters
src/lxc/lxc_driver.c | 153
From: Martin Kletzander mklet...@redhat.com
The function doesn't check whether the request is made for active or
inactive domain. Thus when the domain is not running it still tries
accessing non-existing cgroups (priv-cgroup, which is NULL).
I re-made the function in order for it to work the
On 12/20/2013 03:16 AM, Daniel P. Berrange wrote:
What distro were you on when you hit this failure? I'm a little bit
reluctant to bump the minimum requirement without knowing a bit more
about how common 3.0 is in practice. Adding more details in your commit
log about why you needed it (not
On Thu, Dec 19, 2013 at 16:53:17 +0200, Laine Stump wrote:
These were all found as a result of researching Bug 1035490. Patch 1
fixes the originally reported problem, Patch 2 fixes the cause of
incorrect error reporting in a later comment of that bug, and Patch 3
fixes a behavior that I
On ven, 2013-12-20 at 09:09 +, Wangyufei (James) wrote:
hi , I am confused by pinvcpu before migration.
I have two physical machine. One has 12 CPUs (E5645 2*6 ) as src and the
other has 8 CPUs (2*4) as des.
Step 1 : VM is running on src host (E5645 2*6 CPUs) with its VCPUs pinned on
On gio, 2013-12-19 at 21:32 -0700, Jim Fehlig wrote:
Signed-off-by: Dario Faggioli dario.faggi...@citrix.com
Cc: Jim Fehlig jfeh...@suse.com
Cc: Ian Jackson ian.jack...@eu.citrix.com
---
src/libxl/libxl_driver.c | 83
++
1 file changed,
On 12/09/2013 04:11 AM, Hu Tao wrote:
...snip...
+static bool
+virDomainPanicCheckABIStability(virDomainPanicDefPtr src,
+virDomainPanicDefPtr dst)
+{
+return virDomainDeviceInfoCheckABIStability(src-info, dst-info);
+}
+
These changes have
On 12/20/2013 06:59 AM, John Ferlan wrote:
On 12/09/2013 04:11 AM, Hu Tao wrote:
...snip...
+static bool
+virDomainPanicCheckABIStability(virDomainPanicDefPtr src,
+virDomainPanicDefPtr dst)
+{
+return
Libvirt Security Notice
===
Summary: libvirtd daemon crash when reading memory tunables
for LXC guest in shutoff status
Reported on: 20131209
Published on: 20131220
Fixed on: 20131220
Reported by: Martin Kletzander mklet
On Fri, Dec 20, 2013 at 07:11:00AM -0700, Eric Blake wrote:
On 12/20/2013 06:59 AM, John Ferlan wrote:
On 12/09/2013 04:11 AM, Hu Tao wrote:
...snip...
+static bool
+virDomainPanicCheckABIStability(virDomainPanicDefPtr src,
+
Hi,
Take two, with the comments from Jim on v1 fixed. Basically, that means I've
added a few comments and removed some non necessary safety checking.
Series available here:
git://xenbits.xen.org/people/dariof/libvirt.git libxl/VcpuPinX-v2
Regards,
Dario
---
Dario Faggioli (3):
libxl:
So that it is possible to query vcpu related information of
a persistent but not running domain, like it is for the QEMU
driver.
In fact, before this patch, we have:
# virsh list --all
IdName State
5
By actually removing the vcpupin element (from within the
cputune section) from the XML, rather than jus update it with
a fully set vcpu affinity mask.
Signed-off-by: Dario Faggioli dario.faggi...@citrix.com
Cc: Jim Fehlig jfeh...@suse.com
Cc: Ian Jackson ian.jack...@eu.citrix.com
---
And use it to implement libxlDomainPinVcpu(), similarly to what
happens in the QEMU driver. This way, it is possible to both
query and change the vcpu affinity of a persistent but not
running domain.
In face, before this patch, we have:
# virsh list --all
IdName
by, in libxlDomainGetNumaParameters(), calling libxl_bitmap_init() as soon as
possible, which avoids getting to 'cleanup:', where libxl_bitmap_dispose()
happens, without having initialized the nodemap, and hence crashing after some
invalid free()-s:
# ./daemon/libvirtd -v
*** Error in
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
---
src/libvirt_private.syms |1 +
src/util/virprocess.c| 63 ++
src/util/virprocess.h|6 +
3 files
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot.
---
src/lxc/lxc_driver.c | 44
1 file changed, 28 insertions(+), 16 deletions(-)
diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c
index e5298d1..7f4acbe
On Fri, Dec 20, 2013 at 07:39:21PM +0400, Reco wrote:
diff --git a/src/util/virprocess.c b/src/util/virprocess.c
index 9fc3207..2e8535e 100644
--- a/src/util/virprocess.c
+++ b/src/util/virprocess.c
@@ -31,6 +31,7 @@
# include sys/resource.h
#endif
#include sched.h
+#include stdlib.h
On Fri, Dec 20, 2013 at 07:42:28PM +0400, Reco wrote:
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot.
---
src/lxc/lxc_driver.c | 44
1 file changed, 28 insertions(+), 16 deletions(-)
diff --git
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot.
---
src/lxc/lxc_driver.c | 44
1 file changed, 28 insertions(+), 16 deletions(-)
diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c
index e5298d1..2385f5b
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
Hope it'll nail it this time.
---
src/libvirt_private.syms |1 +
src/util/virprocess.c| 63 ++
src/util/virprocess.h
On Fri, Dec 20, 2013 at 08:24:41PM +0400, Reco wrote:
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
Hope it'll nail it this time.
---
src/libvirt_private.syms |1 +
src/util/virprocess.c| 63
On Fri, Dec 20, 2013 at 08:24:52PM +0400, Reco wrote:
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot.
---
src/lxc/lxc_driver.c | 44
1 file changed, 28 insertions(+), 16 deletions(-)
diff --git
On 19.12.2013 21:36, Eric Blake wrote:
On 12/19/2013 09:01 AM, Daniel P. Berrange wrote:
+typedef void (*virConnectDomainQemuMonitorEventCallback)(virConnectPtr
conn,
+ virDomainPtr dom,
+
On 12/20/2013 09:41 AM, Michal Privoznik wrote:
Ooh, just noticed that the timestamp is not part of the event data;
probably worth adding another parameter to the callback function to list
the event timestamp (as knowing when qemu fired an event may indeed be
important to a developer using
On 12/20/2013 09:24 AM, Reco wrote:
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
Hope it'll nail it this time.
This comment doesn't fit in the commit message; it's better to put
review comments...
---
...here,
Hi.
On Fri, 20 Dec 2013 10:48:39 -0700
Eric Blake ebl...@redhat.com wrote:
On 12/20/2013 09:24 AM, Reco wrote:
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
Hope it'll nail it this time.
This comment
Commit id '4313fead' added a call to virDomainPanicCheckABIStability()
which did not check whether the panic device existed before making a call
to virDomainDeviceInfoCheckABIStability() which ended up segfaulting:
Thread 1 (Thread 0x7f5332837700 (LWP 10964)):
(src=optimized out,
On 12/20/2013 09:34 AM, Daniel P. Berrange wrote:
On Fri, Dec 20, 2013 at 08:24:41PM +0400, Reco wrote:
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
Hope it'll nail it this time.
Fails 'make syntax-check':
On 12/20/2013 10:53 AM, Reco wrote:
At any rate, since Dan has ack'ed it, and it fixes a CVE (where we're
still waiting for the number to be assigned, but the flaw is real), I'll
go ahead and push this soon.
Thank you for the advices, Eric, sorry I didn't followed'em to the
letter. It
Dario Faggioli wrote:
by, in libxlDomainGetNumaParameters(), calling libxl_bitmap_init() as soon as
possible, which avoids getting to 'cleanup:', where libxl_bitmap_dispose()
happens, without having initialized the nodemap, and hence crashing after some
invalid free()-s:
Yikes! ACK to
You still haven't answered my query from an earlier version: we prefer
to list a legal name in the authorship of git commits (after all,
copyleft licenses work _because_ of copyright law, but the law prefers
working with full names rather than nicknames). Is there something that
I should
On 12/20/2013 11:10 AM, John Ferlan wrote:
Commit id '4313fead' added a call to virDomainPanicCheckABIStability()
which did not check whether the panic device existed before making a call
to virDomainDeviceInfoCheckABIStability() which ended up segfaulting:
Signed-off-by: John Ferlan
On 12/20/2013 11:40 AM, Reco wrote:
You still haven't answered my query from an earlier version: we prefer
to list a legal name in the authorship of git commits (after all,
copyleft licenses work _because_ of copyright law, but the law prefers
working with full names rather than nicknames).
Bamvor Jian Zhang wrote:
there is a segfault in libxl logging in libxl_ctx_free when domain
create fail. because the log output handler vmessage is freed by
xtl_logger_destroy before libxl_ctx_free in virDomainObjListRemove.
move xtl_logger_destroy after libxl_ctx_free could fix this bug.
On 12/20/2013 09:24 AM, Reco wrote:
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot.
This fails to compile, due to [1] below:
CCLD libvirt_lxc
lxc/lxc_driver.c:2783:1: error: return type defaults to 'int' [-Werror]
virDomainRebootCallback(pid_t
On 12/20/2013 01:41 PM, Eric Blake wrote:
On 12/20/2013 09:24 AM, Reco wrote:
Use helper virProcessRunInMountNamespace in lxcDomainShutdownFlags and
lxcDomainReboot.
static int
+virDomainShutdownCallback(pid_t pid ATTRIBUTE_UNUSED,
+ void *opaque ATTRIBUTE_UNUSED)
https://bugzilla.redhat.com/show_bug.cgi?id=1043069
When virDomainDetachDeviceFlags is called concurrently to
virDomainBlockStats: libvirtd may crash because qemuDomainBlockStats
finds a disk in vm-def before getting a job on a domain and uses the
disk pointer after getting the job. However, the
Every API that is going to begin a job should do that before fetching
data from vm-def.
Signed-off-by: Jiri Denemark jdene...@redhat.com
---
src/qemu/qemu_driver.c | 11 +--
1 file changed, 5 insertions(+), 6 deletions(-)
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
Generally, every API that is going to begin a job should do that before
fetching data from vm-def. However, qemuDomainGetBlockInfo does not
know whether it will have to start a job or not before checking vm-def.
To avoid using disk alias that might have been freed while we were
waiting for a job,
When fixing https://bugzilla.redhat.com/show_bug.cgi?id=1043069 I
realized qemuDomainBlockStats is not the only API that does not acquire
a job early enough. Generally, every API that is going to begin a job
should do that before fetching data from vm-def. The following 5 APIs
failed to do so and
Every API that is going to begin a job should do that before fetching
data from vm-def.
Signed-off-by: Jiri Denemark jdene...@redhat.com
---
src/qemu/qemu_driver.c | 22 +++---
1 file changed, 11 insertions(+), 11 deletions(-)
diff --git a/src/qemu/qemu_driver.c
Every API that is going to begin a job should do that before fetching
data from vm-def.
Signed-off-by: Jiri Denemark jdene...@redhat.com
---
src/qemu/qemu_driver.c | 25 ++---
1 file changed, 10 insertions(+), 15 deletions(-)
diff --git a/src/qemu/qemu_driver.c
This patch fixes the secret type checking done in the virDomainDiskDefParseXML
function. Previously, it would not allow any volumes that utilized a secret.
This patch is a simple bypass of the checking code for volumes.
Signed-off-by: Adam Walters a...@pandorasboxen.com
---
This patch modifies the qemuTranslateDiskSourcePool function to add RBD storage
pool support. The code is heavily based off of the existing iSCSI code, but
modified for RBD support. The modification calls the qemuAddRBDPoolSourceHost
from my previous patch, along with setting up the ceph user
Here is a re-based re-submission of my patches to implement RBD storage pool
support for QEMU domains. Nothing in it has changed other than it has been
rebased against the latest. The race condition I located still exists, but I
have some patches forthcoming to address that issue. The code here
This function is a helper function that grabs RBD hosts from a storage pool
definition, and applies them to the domain's disk defi
nitions at runtime. This is a pre-requisite for RBD storage pool support in the
domain XML.
Signed-off-by: Adam Walters a...@pandorasboxen.com
---
This patch adds VIR_FROM_CONFIG to the virErrorDomain enum. Both of these files
must be patched in unison to prevent compilation failures.
Signed-off-by: Adam Walters a...@pandorasboxen.com
---
include/libvirt/virterror.h | 2 ++
src/util/virerror.c | 2 ++
2 files changed, 4
This is the header file for the config driver.
Signed-off-by: Adam Walters a...@pandorasboxen.com
---
src/config/config_driver.h | 44
1 file changed, 44 insertions(+)
create mode 100644 src/config/config_driver.h
diff --git
This patch adds the config driver hooks and moves the secret driver hook
definitions higher on the list. The secret driver move isn't strictly needed,
but the comments state that these should be in preferred load order. Since
other drivers might utilize the secret driver, it makes sense to have
This patchset adds a driver named 'config' that allows access to configuration
data, such as secret and storage definitions. This is a pre-requisite for my
next patchset which resolves the race condition on libvirtd startup and the
circular dependencies between QEMU and the storage driver.
The
This is the source code to the config driver. This driver is a hypervisor
driver that does not support any domain operations. The sole purpose of this
driver is to allow access to various bits of configuration information, such as
secret or network definitions, from the initialization and
This conditionally enables compilation of the config driver based on if we are
building libvirtd or not. Since this is only needed for hypervisor modules
during libvirtd startup, we don't need to bother compiling the config driver
when only building the client.
Signed-off-by: Adam Walters
This implements a new field in the virStateDriver struct. In order to prevent
possible compilation issues, this patch also implements te new field in all of
the existing drivers. Other than in driver.h, the changes are all a single line
addition to the files.
Signed-off-by: Adam Walters
Adding config_driver.c to POTFILES.in to fix a syntax-check error.
Signed-off-by: Adam Walters a...@pandorasboxen.com
---
po/POTFILES.in | 1 +
1 file changed, 1 insertion(+)
diff --git a/po/POTFILES.in b/po/POTFILES.in
index 49dfc9c..0e23610 100644
--- a/po/POTFILES.in
+++ b/po/POTFILES.in
@@
This completes the addition of the config driver to libvirt. The final piece
here is to add the config driver into the Makefile (via automake) so that the
driver is actually compiled and linked.
Signed-off-by: Adam Walters a...@pandorasboxen.com
---
src/Makefile.am | 25
This utilizes the config driver I submitted to resolve the hardcoded qemu
connection string. With this, the storage driver no longer has a circular
dependency with QEMU. Without this patch, when libvirtd is restarted, QEMU
requires storage (when domains are using storage pool backings) and
This patchset implements a two-tier driver loading system. I split the
hypervisor drivers out into their own tier, which is loaded after the other
drivers. This has the net effect of ensuring that things like secrets,
networks, etc., are initialized and auto-started before any hypervisors, such
This implements a two-tier driver loading system into libvirt. The two classes
of drivers are Libvirt drivers and Hypervisor drivers. Hypervisor drivers
are fairly self-explanatory, they provide domain services. Libvirt drivers are
sort of the backend drivers for those, like the secret and
On 12/20/2013 02:36 PM, Jiri Denemark wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=1043069
When virDomainDetachDeviceFlags is called concurrently to
virDomainBlockStats: libvirtd may crash because qemuDomainBlockStats
finds a disk in vm-def before getting a job on a domain and uses the
On 12/20/2013 05:47 PM, Pradipta Kumar Banerjee wrote:
[snip]
for (i = 0; i *nparams; i++) {
virNodeCPUStatsPtr param = params[i];
What about this?
diff --git a/src/nodeinfo.c b/src/nodeinfo.c
index 1838547..aa1ad81 100644
--- a/src/nodeinfo.c
+++
On 12/20/2013 11:17 AM, Eric Blake wrote:
On 12/20/2013 09:34 AM, Daniel P. Berrange wrote:
On Fri, Dec 20, 2013 at 08:24:41PM +0400, Reco wrote:
Implement virProcessRunInMountNamespace, which runs callback of type
virProcessNamespaceCallback in a container namespace.
Hope it'll nail it this
74 matches
Mail list logo