Hi All,
I recently received an internal bug report of VM "crashing" due to hitting
thread limits. Seems there was an assert in pthread_create within the VM when
hitting the limit enforced by pids controller on the host
Apr 28 07:45:46 lpcomp02007 kernel: cgroup: fork rejected by pids
On 5/22/19 6:29 AM, Daniel P. Berrangé wrote:
Creating firewall rules for the virtual networks causes the kernel to
load the conntrack module. This imposes a significant performance
penalty on Linux network traffic. Thus we want to only take that hit if
we actually have virtual networks running.
On 5/22/19 6:29 AM, Daniel P. Berrangé wrote:
Signed-off-by: Daniel P. Berrangé
---
src/network/bridge_driver_linux.c | 18 --
1 file changed, 16 insertions(+), 2 deletions(-)
The additional debug messages proved helpful while testing the patches!
Reviewed-by: Jim Fehlig
On 5/22/19 6:29 AM, Daniel P. Berrangé wrote:
Pull the logic for creating global iptables chains into a separate
method and protect its invokation with virOnce, to make it possible
to reuse it in non-startup paths.
Signed-off-by: Daniel P. Berrangé
---
src/network/bridge_driver_linux.c | 39
Ignore @source in the case of the test driver and return fixed private
IPv4 addresses for all the interfaces defined in the domain.
Signed-off-by: Ilias Stamatis
---
The default config of the test driver has no guest interfaces defined,
so this must be tested with a custom config.
Maybe it
Now that all C examples are neatly sorted into only three
categories, getting rid of our custom installation machinery
and replacing it with the standard autotools mechanism
finally becomes feasible.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 44
All other examples are organized using the either the format/
or the format/category/ hierarchy already, and grouping all
C programs together removes the last remaining outliers.
Signed-off-by: Andrea Bolognani
---
.gitignore | 30 -
examples/Makefile.am
Most C examples live in their own directory, which seems a
bit unnecessary especially considering that all virt-admin
related examples share a single admin/ directory. Reorganize
non-admin C examples in two categories: domain/ for those
that act on a domain, and misc/ for everything else.
$(mkinstalldirs) works like 'mkdir -p' in that it will
create all the necessary parts of the path leading up to
the actual directory, which means creating $(examplesdir)
beforehand is not necessary.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 1 -
1 file changed, 1 deletion(-)
Right now we install the files in RPMs only, and we include
them in the -daemon package which is probably not the best
option either. Start installing them via autotools; the RPMs
will get them automatically in the -docs package.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 9
We already install nwfilters separately.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 19 +--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/examples/Makefile.am b/examples/Makefile.am
index 7a15decc00..c0d0100c9f 100644
--- a/examples/Makefile.am
The virt-lxc-convert shell script is at this point the
only example we don't install on the target system.
Create a sh/ subdirectory, following the example set by
the existing polkit/, systemtap/ and xml/, and move the
script there; then add rules that will install all example
shell scripts as
$(AM_CPPFLAGS) is for passing options to the C preprocessor,
not the C compiler, and the stuff in $(WARN_CFLAGS) belongs
to the latter category.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 3 +++
1 file changed, 3 insertions(+)
diff --git a/examples/Makefile.am
$(LDADD) is for object files that should be added during
linking, not for options that should be passed to the
linker: that's what $(AM_LDFLAGS) is for.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git
$(WARN_CFLAGS) contains options intended for the compiler,
whereas $(LDADD) is supposed to list additional objects
required during linking, so the former clearly doesn't
belong in the latter.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 1 -
1 file changed, 1 deletion(-)
diff
We're doing nothing more than copying files to a target
directory, so we don't need any custom shell commands and
can just use the standard autotools data installation
support instead.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 23 ---
1 file changed, 4
The C programs in this directory are supposed to be only
using public functions, so having $(top_srcdir) in the
header search path is unnecessary at best and actively
harmful at worst.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 1 -
1 file changed, 1 deletion(-)
diff --git
$(AM_CPPFLAGS) already includes $(WARN_CFLAGS), so this is
not doing anything useful.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 3 ---
1 file changed, 3 deletions(-)
diff --git a/examples/Makefile.am b/examples/Makefile.am
index 11c9f44917..9a32fd5d4a 100644
---
This will make further changes easier to review.
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 16 +++-
1 file changed, 11 insertions(+), 5 deletions(-)
diff --git a/examples/Makefile.am b/examples/Makefile.am
index 9a32fd5d4a..8f50149812 100644
---
Signed-off-by: Andrea Bolognani
---
examples/Makefile.am | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/examples/Makefile.am b/examples/Makefile.am
index 4122219e22..7a15decc00 100644
--- a/examples/Makefile.am
+++ b/examples/Makefile.am
@@ -22,10 +22,14 @@
This series changes our install procedure so that we install all
available examples on the target system; in the process, it also
removes all custom shell code from the installation procedure,
replacing it with standard autotools usage, and fixes a few
admittedly very minor issues.
Andrea
We already install the file, along with its driver-specific
counterparts, into ${datadir}/libvirt/api/ where language
bindings will actually look for them.
Signed-off-by: Andrea Bolognani
---
libvirt.spec.in | 2 --
1 file changed, 2 deletions(-)
diff --git a/libvirt.spec.in b/libvirt.spec.in
On Tue, 2019-05-21 at 16:26 +0200, Ján Tomko wrote:
> Rename the DOMAIN_DEVICE_ITERATE_GRAPHICS flag.
> It was introduced by commit dd45c2710f6fd2d4f8a47f97960532d0e0091e7d
> with the intention to run the Validate callback even on the graphics
> device.
>
> However, enumerating every single
On Tue, 2019-05-21 at 16:25 +0200, Ján Tomko wrote:
[...]
> @@ -5793,10 +5806,9 @@ virDomainDefPostParse(virDomainDefPtr def,
> }
>
> /* iterate the devices */
> -ret = virDomainDeviceInfoIterateInternal(def,
> -
>
On Wed, 2019-05-22 at 16:12 +0200, Ján Tomko wrote:
> On Wed, May 22, 2019 at 03:36:59PM +0200, Andrea Bolognani wrote:
> > I don't much like the idea of adding a separate function that does
> > almost the same thing but not quite, because that will ultimately
> > result in the same situation we
On 5/22/19 6:29 AM, Daniel P. Berrangé wrote:
On Mon, May 20, 2019 at 09:44:04PM -0400, Laine Stump wrote:
On 5/14/19 11:48 AM, Daniel P. Berrangé wrote:
Introduce a virNetworkPortDefPtr struct to represent the data associated
with a virtual network port. Add APIs for parsing/formatting XML
On 5/22/19 7:29 AM, Daniel P. Berrangé wrote:
> Pull the logic for creating global iptables chains into a separate
> method and protect its invokation with virOnce, to make it possible
invocation
> to reuse it in non-startup paths.
>
> Signed-off-by: Daniel P. Berrangé
> ---
>
On Wed, May 22, 2019 at 03:36:59PM +0200, Andrea Bolognani wrote:
On Tue, 2019-05-21 at 16:28 +0200, Ján Tomko wrote:
On Mon, May 20, 2019 at 03:24:06PM +0200, Andrea Bolognani wrote:
> On Mon, 2019-05-20 at 14:59 +0200, Ján Tomko wrote:
> > Of course, there is a huge overlap between iterating
On Tue, 2019-05-21 at 16:28 +0200, Ján Tomko wrote:
> On Mon, May 20, 2019 at 03:24:06PM +0200, Andrea Bolognani wrote:
> > On Mon, 2019-05-20 at 14:59 +0200, Ján Tomko wrote:
> > > Of course, there is a huge overlap between iterating over all devices
> > > and just those with an info, but since
On Wed, 2019-05-22 at 13:45 +0200, Peter Krempa wrote:
> On Wed, May 22, 2019 at 12:34:18 +0200, Ján Tomko wrote:
> > On Wed, May 22, 2019 at 10:33:39AM +0200, Andrea Bolognani wrote:
> > > Personally I like the idea of using the same input file for different
> > > DO_TEST*() calls, highlighting
Signed-off-by: Daniel P. Berrangé
---
src/network/bridge_driver_linux.c | 18 --
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/src/network/bridge_driver_linux.c
b/src/network/bridge_driver_linux.c
index 0d849173b2..2b29363f3c 100644
---
Pull the logic for creating global iptables chains into a separate
method and protect its invokation with virOnce, to make it possible
to reuse it in non-startup paths.
Signed-off-by: Daniel P. Berrangé
---
src/network/bridge_driver_linux.c | 39 +++
1 file changed,
Creating firewall rules for the virtual networks causes the kernel to
load the conntrack module. This imposes a significant performance
penalty on Linux network traffic. Thus we want to only take that hit if
we actually have virtual networks running.
We need to create global firewall rules during
This is an series to delay creating libvirt firewall chains if nothing
needs them during startup.
Daniel P. Berrangé (3):
network: pull global chain init into separate method
network: add more debugging of firewall chain creation
network: delay global firewall setup if no networks are
On Wed, May 22, 2019 at 12:34:18 +0200, Ján Tomko wrote:
> On Wed, May 22, 2019 at 10:33:39AM +0200, Andrea Bolognani wrote:
> > On Tue, 2019-05-21 at 16:15 +0200, Ján Tomko wrote:
> > > On Mon, May 20, 2019 at 01:37:47PM +0200, Andrea Bolognani wrote:
> > > > diff --git
On Wed, May 22, 2019 at 10:33:39AM +0200, Andrea Bolognani wrote:
On Tue, 2019-05-21 at 16:15 +0200, Ján Tomko wrote:
On Mon, May 20, 2019 at 01:37:47PM +0200, Andrea Bolognani wrote:
> diff --git a/tests/qemuxml2argvdata/intel-iommu-machine.xml
b/tests/qemuxml2argvdata/intel-iommu-machine.xml
Signed-off-by: Martin Kletzander
---
docs/news.xml | 12
1 file changed, 12 insertions(+)
diff --git a/docs/news.xml b/docs/news.xml
index 2ecb835e45f0..9922cd4de0e2 100644
--- a/docs/news.xml
+++ b/docs/news.xml
@@ -45,6 +45,18 @@
+
+
+
Partially-related-to: https://bugzilla.redhat.com/show_bug.cgi?id=1712781
Martin Kletzander (2):
qemu: Set emulator thread scheduler only after QEMU starts
docs: Add emulatorsched fix info to news.xml
docs/news.xml | 12
src/qemu/qemu_process.c | 10 +-
2
On Tue, May 21, 2019 at 10:46:27AM -0400, Laine Stump wrote:
> On 5/20/19 9:44 PM, Laine Stump wrote:
> > On 5/14/19 11:48 AM, Daniel P. Berrangé wrote:
> > > Introduce a virNetworkPortDefPtr struct to represent the data associated
> > > with a virtual network port. Add APIs for parsing/formatting
On Mon, May 20, 2019 at 09:44:04PM -0400, Laine Stump wrote:
> On 5/14/19 11:48 AM, Daniel P. Berrangé wrote:
> > Introduce a virNetworkPortDefPtr struct to represent the data associated
> > with a virtual network port. Add APIs for parsing/formatting XML docs
> > with the data.
> >
> >
If the scheduler is set before vCPU0 cannot be moved into its cpu,cpuacct
cgroup. While it is not yet known whether this is a bug or not, it makes sense
for us to do that later as otherwise the scheduler would be inherited by vCPU
and I/O Threads even when they do not have any such setting
On Wed, May 22, 2019 at 10:08:44AM +0200, Markus Armbruster wrote:
> Andrea Bolognani writes:
>
> > On Mon, 2019-05-13 at 13:19 +0100, Daniel P. Berrangé wrote:
> >> On Mon, May 13, 2019 at 02:00:28PM +0200, Andrea Bolognani wrote:
> >> > One possible complication is that we would not be able to
On Tue, 2019-05-21 at 16:15 +0200, Ján Tomko wrote:
> On Mon, May 20, 2019 at 01:37:47PM +0200, Andrea Bolognani wrote:
> > diff --git a/tests/qemuxml2argvdata/intel-iommu-machine.xml
> > b/tests/qemuxml2argvdata/intel-iommu-machine.xml
> > index 90aba16156..85027cd938 100644
> > ---
Eric Blake writes:
> On 5/13/19 8:28 AM, Michal Privoznik wrote:
>> On 5/13/19 12:17 PM, Daniel P. Berrangé wrote:
>>> This is a long mail about ENOMEM (OOM) handling in libvirt. The executive
>>> summary is that it is not worth the maint cost because:
>>>
>
>>> The long answer follows...
>>
>>
Andrea Bolognani writes:
> On Mon, 2019-05-13 at 13:19 +0100, Daniel P. Berrangé wrote:
>> On Mon, May 13, 2019 at 02:00:28PM +0200, Andrea Bolognani wrote:
>> > One possible complication is that we would not be able to use any
>> > of the GLib types in our public API... I think the way we
45 matches
Mail list logo
