[ruby PATCH] Fix cpumap allocation for virDomainGetVcpus and use return value
by Charlie Smurthwaite
This patch fixes a bug in which only enough memory for one cpumap is
allocated
for virDomainGetVcpus instead of one per virtual CPU. This Fixes an
overflow.
Additionally, it uses the return value of virDomainGetVcpus to determine how
many cpuinfo structs were actually populated rather than assuming they
all are.
Finally, it uses the logical CPU number from the cpuinfo struct in the
retrurn
data instead of assuming CPU numbers are sequential. This should handle
cases
where arbitrary CPUs are offline.
Signed-off-by: Charlie Smurthwaite <charlie(a)atechmedia.com>
---
ext/libvirt/domain.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/ext/libvirt/domain.c b/ext/libvirt/domain.c
index d665907..c6de1bf 100644
--- a/ext/libvirt/domain.c
+++ b/ext/libvirt/domain.c
@@ -803,7 +803,7 @@ static VALUE libvirt_domain_vcpus(VALUE d)
cpumaplen = VIR_CPU_MAPLEN(maxcpus);
- cpumap = alloca(sizeof(unsigned char) * cpumaplen);
+ cpumap = alloca(sizeof(unsigned char) * cpumaplen * dominfo.nrVirtCpu);
r = virDomainGetVcpus(ruby_libvirt_domain_get(d), cpuinfo,
dominfo.nrVirtCpu, cpumap, cpumaplen);
@@ -832,15 +832,16 @@ static VALUE libvirt_domain_vcpus(VALUE d)
result = rb_ary_new();
- for (i = 0; i < dominfo.nrVirtCpu; i++) {
+ for (i = 0; i < r; i++) {
vcpuinfo = rb_class_new_instance(0, NULL, c_domain_vcpuinfo);
- rb_iv_set(vcpuinfo, "@number", UINT2NUM(i));
if (cpuinfo != NULL) {
+ rb_iv_set(vcpuinfo, "@number", INT2NUM(cpuinfo[i].number));
rb_iv_set(vcpuinfo, "@state", INT2NUM(cpuinfo[i].state));
rb_iv_set(vcpuinfo, "@cpu_time", ULL2NUM(cpuinfo[i].cpuTime));
rb_iv_set(vcpuinfo, "@cpu", INT2NUM(cpuinfo[i].cpu));
}
else {
+ rb_iv_set(vcpuinfo, "@number", Qnil);
rb_iv_set(vcpuinfo, "@state", Qnil);
rb_iv_set(vcpuinfo, "@cpu_time", Qnil);
rb_iv_set(vcpuinfo, "@cpu", Qnil);
--
2.25.0
4 years, 8 months
[PATCH 0/3] Couple of memleak fixes
by Michal Privoznik
*** BLURB HERE ***
Michal Prívozník (3):
virDomainFSDefFree: Unref private data
virDomainNetDefClear: Free @persistent name
qemuTestParseCapabilitiesArch: Free @binary
src/conf/domain_conf.c | 2 ++
tests/testutilsqemu.c | 4 ++--
2 files changed, 4 insertions(+), 2 deletions(-)
--
2.24.1
4 years, 8 months
[libvirt PATCH] docs: Expand documentation for the tickpolicy timer attribute
by Andrea Bolognani
The current documentation is fairly terse and not easy to decode
for someone who's not intimately familiar with the inner workings
of timer devices. Expand on it by providing a somewhat verbose
description of what behavior each policy will result in, as seen
from both the guest OS and host point of view.
This is lifted directly from QEMU commit
commit 2a7d957596786404c4ed16b089273de95a9580ad
Author: Andrea Bolognani <abologna(a)redhat.com>
Date: Tue Feb 11 19:37:44 2020 +0100
qapi: Expand documentation for LostTickPolicy
v4.2.0-1442-g2a7d957596
The original text also matched word for word the documentation
found in QEMU.
Signed-off-by: Andrea Bolognani <abologna(a)redhat.com>
---
docs/formatdomain.html.in | 32 +++++++++++++++++++++-----------
1 file changed, 21 insertions(+), 11 deletions(-)
diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in
index f4af65f13f..4fef2a0a97 100644
--- a/docs/formatdomain.html.in
+++ b/docs/formatdomain.html.in
@@ -2487,26 +2487,36 @@
<p>
The <code>tickpolicy</code> attribute determines what
happens when QEMU misses a deadline for injecting a
- tick to the guest:
+ tick to the guest. This can happen, for example, because the
+ guest was paused.
</p>
<dl>
<dt><code>delay</code></dt>
- <dd>Continue to deliver ticks at the normal rate.
- The guest time will be delayed due to the late
- tick</dd>
+ <dd>Continue to deliver ticks at the normal rate. The guest OS
+ will not notice anything is amiss, as from its point of view
+ time will have continued to flow normally. The time in the
+ guest should now be behind the time in the host by exactly
+ the amount of time during which ticks have been missed.</dd>
<dt><code>catchup</code></dt>
- <dd>Deliver ticks at a higher rate to catch up
- with the missed tick. The guest time should
- not be delayed once catchup is complete.</dd>
+ <dd>Deliver ticks at a higher rate to catch up with the missed
+ ticks. The guest OS will not notice anything is amiss, as
+ from its point of view time will have continued to flow
+ normally. Once the timer has managed to catch up with all
+ the missing ticks, the time in the guest and in the host
+ should match.</dd>
<dt><code>merge</code></dt>
<dd>Merge the missed tick(s) into one tick and
inject. The guest time may be delayed, depending
on how the OS reacts to the merging of ticks</dd>
<dt><code>discard</code></dt>
- <dd>Throw away the missed tick(s) and continue
- with future injection normally. The guest time
- may be delayed, unless the OS has explicit
- handling of lost ticks</dd>
+ <dd>Throw away the missed ticks and continue with future
+ injection normally. The guest OS will see the timer jump
+ ahead by a potentially quite significant amount all at once,
+ as if the intervening chunk of time had simply not existed;
+ needless to say, such a sudden jump can easily confuse a
+ guest OS which is not specifically prepared to deal with it.
+ Assuming the guest OS can deal correctly with the time jump,
+ the time in the guest and in the host should now match.</dd>
</dl>
<p>If the policy is "catchup", there can be further details in
the <code>catchup</code> sub-element.</p>
--
2.24.1
4 years, 8 months
[PATCH v1 0/3] qemumonitorjson tests for cpu compare and baseline
by Collin Walling
These patches implement tests for the libvirt qemu_monitor_json API for
the hypervisor-cpu-compare and -baseline commands. The input and output
data is mocked with arbitrary values.
A prerequisite patch is included to load the capabilities schema for
a specific architecture. Originally, only the x86 capabilities were
loaded for the qemu_monitor_json tests. By accepting a string denoting
which architecture's QEMU capabilities we'd like to load, we can now
test the comparison and baseline code that is currently only supported
on s390.
Collin Walling (3):
qemumonitorjsontest: load schema based on specified arch
qemumonitorjsontest: add tests for cpu comparison
qemumonitorjsontest: add test for cpu baseline
tests/qemublocktest.c | 2 +-
tests/qemuhotplugtest.c | 2 +-
tests/qemumonitorjsontest.c | 131 +++++++++++++++++++++++++++++++++++++++++++-
tests/testutilsqemuschema.c | 8 +--
tests/testutilsqemuschema.h | 4 +-
5 files changed, 137 insertions(+), 10 deletions(-)
--
2.7.4
4 years, 8 months
[libvirt PATCH] m4: libxl: properly fail when libxl is required
by Ján Tomko
We specify "true" as the fail-action for LIBVIRT_CHECK_PKG.
This was used when we had a fallback to non-pkg-config detection,
then removed in commit 5bdcef13d13560512c7d6d8c9e8822e456889e0c
later re-introduced in commit dc3d2c9f8c7678a950abedd227b1587ca62335c4
and then left in when removing the old detection again in
commit 18981877d2e20390a79d068861a24e716f8ee422
Remove it to properly error out when libxl was requested but not
detected.
Signed-off-by: Ján Tomko <jtomko(a)redhat.com>
Fixes: 18981877d2e20390a79d068861a24e716f8ee422
---
m4/virt-driver-libxl.m4 | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/m4/virt-driver-libxl.m4 b/m4/virt-driver-libxl.m4
index 2f3565f1d5..a958cb26fa 100644
--- a/m4/virt-driver-libxl.m4
+++ b/m4/virt-driver-libxl.m4
@@ -30,7 +30,7 @@ AC_DEFUN([LIBVIRT_DRIVER_CHECK_LIBXL], [
dnl search for libxl, aka libxenlight
old_with_libxl="$with_libxl"
- LIBVIRT_CHECK_PKG([LIBXL], [xenlight], [4.6.0], [true])
+ LIBVIRT_CHECK_PKG([LIBXL], [xenlight], [4.6.0])
if test "x$with_libxl" = "xyes" ; then
LIBXL_FIRMWARE_DIR=$($PKG_CONFIG --variable xenfirmwaredir xenlight)
LIBXL_EXECBIN_DIR=$($PKG_CONFIG --variable libexec_bin xenlight)
--
2.24.1
4 years, 8 months
[PATCH 0/3] Couple of almost trivial patches
by Michal Privoznik
These stem out from my review of Marc-André's patches:
https://www.redhat.com/archives/libvir-list/2020-January/msg00648.html
Michal Prívozník (3):
virpidfile: Set correct retval in virPidFileReadPath()
qemu: Don't explicitly remove pidfile after
virPidFileForceCleanupPath()
qemu_migration: Rearrange some checks in qemuMigrationSrcIsAllowed()
src/qemu/qemu_migration.c | 66 +++++++++++++++++-----------------
src/qemu/qemu_process.c | 9 +----
src/qemu/qemu_vhost_user_gpu.c | 10 +-----
src/util/virpidfile.c | 2 +-
4 files changed, 36 insertions(+), 51 deletions(-)
--
2.24.1
4 years, 8 months
[libvirt PATCH 00/10] support BR_ISOLATED flag for guest interfaces attached to a Linux host bridge
by Laine Stump
https://bugzilla.redhat.com/1727263
Since Linux kernel 4.18, the Linux host bridge has had a flag
BR_ISOLATED that can be applied to individual ports. When this flag is
set for a port, traffic is blocked between that port and any other
port that also has the BR_ISOLATED flag set. libvirt domain interface
config now supports setting this flag via the <portOptions
isolated='yes'/> setting. It can also be set for all connections to
a particular libvirt network by setting the same option in the network
config - since the port for the host itself does not have BR_ISOLATED
set, the guests can communicate with the host and the outside world,
but guests on that network can't communicate with each other. This
feature works for QEMU and LXC guests with interfaces attached to a
Linux host bridge.
(I had contemplated (and experimented with) putting this new flag in
the <virtualport> element to avoid creating a new element, but that
ended up creating lots of extra code since none of the existing
virtualport types would support this new flag, Linux host bridges
already work with *no* <virtualport> (much less a virtualport type),
and there are some attributes in the <virtualport> parameters
subelement that are always autogenerated if there is no virtualport
type specified, so I would needed to add a new virtualport type for
Linux host bridge, which seems redundant as that information is
already implicit in the interface's connection type. etc. etc. It all
just turned into a big mess, and starting over fresh with something
generic (and hopefully expandable in a sensible way) seemed
cleaner). (I am of course open to suggestions though!)
Laine Stump (10):
schema: trivial indentation fix
schema: add missing vlan element to networkport RNG
qemu: save/restore original error when recovering from failed bridge
attach
util: query/set BR_ISOLATED flag on netdevs attached to bridge
conf: parse/format <portOptions isolated='yes|no'/>
network: propagate <portOptions isolated='yes'/> between network and
domain
qemu/lxc: plumb isolatedPort from config down through bridge
attachment
qemu: support updating <portOptions isolated='yes|no'/> during device
update
conf: extra validation for <portOptions isolated='yes'/>
docs: add info about <portOptions isolated='yes'/> to news file
docs/news.xml | 21 +++++
docs/schemas/domaincommon.rng | 3 +
docs/schemas/network.rng | 9 ++-
docs/schemas/networkcommon.rng | 11 +++
docs/schemas/networkport.rng | 6 ++
src/bhyve/bhyve_command.c | 1 +
src/conf/domain_conf.c | 79 +++++++++++++++++++
src/conf/domain_conf.h | 4 +
src/conf/network_conf.c | 32 ++++++++
src/conf/network_conf.h | 9 +++
src/conf/virnetworkportdef.c | 3 +
src/conf/virnetworkportdef.h | 1 +
src/libvirt_private.syms | 3 +
src/lxc/lxc_process.c | 10 +++
src/network/bridge_driver.c | 4 +
src/qemu/qemu_hotplug.c | 47 +++++++++--
src/qemu/qemu_interface.c | 1 +
src/util/virnetdevbridge.c | 46 +++++++++++
src/util/virnetdevbridge.h | 9 +++
src/util/virnetdevtap.c | 17 +++-
src/util/virnetdevtap.h | 3 +
tests/bhyvexml2argvmock.c | 1 +
tests/networkxml2xmlin/isolated-ports.xml | 7 ++
tests/networkxml2xmlout/isolated-ports.xml | 7 ++
tests/networkxml2xmltest.c | 1 +
tests/qemuxml2argvdata/net-isolated-port.xml | 34 ++++++++
.../net-isolated-port.x86_64-latest.xml | 63 +++++++++++++++
tests/qemuxml2xmltest.c | 1 +
28 files changed, 423 insertions(+), 10 deletions(-)
create mode 100644 tests/networkxml2xmlin/isolated-ports.xml
create mode 100644 tests/networkxml2xmlout/isolated-ports.xml
create mode 100644 tests/qemuxml2argvdata/net-isolated-port.xml
create mode 100644 tests/qemuxml2xmloutdata/net-isolated-port.x86_64-latest.xml
--
2.24.1
4 years, 8 months
[PATCH 0/4] lxc: Add VCPU features for LXC
by Julio Faracco
This series cover a lots of functionalities to LXC VCPUs. It enables
sharing some timer devices between host and LXC guest using `timer`
settings. It still has other improvements related to VCPU and LXC such
as virtual cpuinfo content based on VCPU settings and some better
resource limits. Each patch has the description of the problem and what
it is trying to fix.
Julio Faracco (4):
lxc: Add Real Time Clock device into allowed devices
lxc: Add HPET device into allowed devices
lxc: Implement virtual /proc/cpuinfo via LXC fuse
lxc: Count max VCPUs based on cpuset.cpus in native config.
docs/formatdomain.html.in | 4 +-
src/lxc/lxc_cgroup.c | 76 ++++++++++++++
src/lxc/lxc_container.c | 38 +++++++
src/lxc/lxc_container.h | 2 +
src/lxc/lxc_controller.c | 98 +++++++++++++++++++
src/lxc/lxc_fuse.c | 78 ++++++++++++++-
src/lxc/lxc_native.c | 24 ++++-
.../lxcconf2xml-cpusettune.xml | 2 +-
8 files changed, 313 insertions(+), 9 deletions(-)
--
2.20.1
4 years, 8 months
Re: [PATCH v2 0/2] finish qemu-nbd --partition deprecation
by Eric Blake
ping
On 1/23/20 10:46 AM, Eric Blake wrote:
> Based-on: <20200116141511.16849-1-peter.maydell(a)linaro.org>
> (0/3 convert qemu-nbd, qemu-block-drivers to rST)
>
> In v2:
> - rebased on top of rST doc changes
> - patch 1 added
>
> Eric Blake (2):
> docs: Fix typo in qemu-nbd -P replacement
> qemu-nbd: Removed deprecated --partition option
>
> docs/interop/qemu-nbd.rst | 15 ++---
> qemu-deprecated.texi | 49 ++++++--------
> qemu-nbd.c | 133 +-------------------------------------
> 3 files changed, 24 insertions(+), 173 deletions(-)
>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3226
Virtualization: qemu.org | libvirt.org
4 years, 8 months