- Devel - libvirt List Archives

[libvirt] [PATCH v2 0/2] processor frequency information on S390
by Bjoern Walk 12 Jan '18

12 Jan '18

Since kernel version 4.7, processor frequency information is available on S390. This patch series extends the parser for system information. Let's also add a testcase to the test suite for a S390 CPU configuration running kernel version 4.14 on LPAR. This is basically patch 3 of this series: https://www.redhat.com/archives/libvir-list/2017-December/msg00633.html v1 -> v2: - add debug logging - don't discard the whole hostsysinfo when the parsing fails - don't try to parse a value for external_clock, that information is not available on S390 - add test case for to sysinfotest.c Bjoern Walk (2): util: virsysinfo: parse frequency information on S390 tests: sysinfotest: add new test case for S390 src/util/virsysinfo.c | 41 ++++++- tests/sysinfodata/s390-freqcpuinfo.data | 52 +++++++++ tests/sysinfodata/s390-freqsysinfo.data | 173 ++++++++++++++++++++++++++++++ tests/sysinfodata/s390-freqsysinfo.expect | 63 +++++++++++ tests/sysinfotest.c | 1 + 5 files changed, 325 insertions(+), 5 deletions(-) create mode 100644 tests/sysinfodata/s390-freqcpuinfo.data create mode 100644 tests/sysinfodata/s390-freqsysinfo.data create mode 100644 tests/sysinfodata/s390-freqsysinfo.expect -- 2.13.4

3 9

[libvirt] [jenkins-ci PATCH] guests: Make rpcgen mapping more future-proof
by Andrea Bolognani 12 Jan '18

12 Jan '18

Since whatever is in Fedora Rawhide will end up in future Fedora releases, it makes sense to use the newer package as default and override it for existing releases (as well as CentOS), so that no tweaking will be needed when Fedora 28 and later are added to the CI setup. Suggested-by: Peter Krempa <pkrempa(a)redhat.com> Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- guests/vars/mappings.yml | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/guests/vars/mappings.yml b/guests/vars/mappings.yml index b79757e..0b32f90 100644 --- a/guests/vars/mappings.yml +++ b/guests/vars/mappings.yml @@ -606,8 +606,10 @@ mappings: rpcgen: deb: libc-dev-bin - rpm: glibc-common - FedoraRawhide: glibc-rpcgen + rpm: glibc-rpcgen + CentOS: glibc-common + Fedora26: glibc-common + Fedora27: glibc-common FreeBSD: rpmbuild: -- 2.14.3

2 2

[libvirt] [PATCH 0/2] travis: Fix build failure
by Andrea Bolognani 12 Jan '18

12 Jan '18

It would be nice to have all greens before the release :) Andrea Bolognani (2): travis: Sync packages with libvirt-jenkins-ci travis: Skip nfs-common installation .travis.yml | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) -- 2.14.3

2 3

[libvirt] [PATCH V4] blockjob: Fix error checking of blockjob status
by Jie Wang 12 Jan '18

12 Jan '18

offset and len can also be equal to 0 on failed if blockjob return status:"BLOCK_JOB_COMPLETED" with error:"File descriptor in bad state", so we need to check 'error' in this case. --- src/qemu/qemu_monitor_json.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index e45868b..39c0909 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -886,7 +886,7 @@ qemuMonitorJSONHandleBlockJobImpl(qemuMonitorPtr mon, case VIR_DOMAIN_BLOCK_JOB_COMPLETED: error = virJSONValueObjectGetString(data, "error"); /* Make sure the whole device has been processed */ - if (offset != len) + if (offset != len || error) event = VIR_DOMAIN_BLOCK_JOB_FAILED; break; case VIR_DOMAIN_BLOCK_JOB_CANCELED: -- 1.8.3.1

2 2

[libvirt] [jenkins-ci PATCH] guests: Install rpcgen when building libvirt
by Andrea Bolognani 12 Jan '18

12 Jan '18

The dependency on rpcgen had not been made explicit up until now because the command was part of the base glibc development packages; however, Fedora Rawhide now ships it in a separate package, so we need to list it along with all other dependencies in order for it to get installed. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- guests/vars/mappings.yml | 6 ++++++ guests/vars/projects/libvirt.yml | 1 + 2 files changed, 7 insertions(+) diff --git a/guests/vars/mappings.yml b/guests/vars/mappings.yml index eca8dbe..b79757e 100644 --- a/guests/vars/mappings.yml +++ b/guests/vars/mappings.yml @@ -604,6 +604,12 @@ mappings: pkg: readline rpm: readline-devel + rpcgen: + deb: libc-dev-bin + rpm: glibc-common + FedoraRawhide: glibc-rpcgen + FreeBSD: + rpmbuild: rpm: rpm-build diff --git a/guests/vars/projects/libvirt.yml b/guests/vars/projects/libvirt.yml index 407881d..ba1e4c3 100644 --- a/guests/vars/projects/libvirt.yml +++ b/guests/vars/projects/libvirt.yml @@ -67,6 +67,7 @@ packages: - qemu-img - radvd - readline + - rpcgen - sanlock - scrub - sheepdog -- 2.14.3

2 1

[libvirt] [PATCH] libvirtd: clarify the TLS conf default vaule setting
by Chen Hanxiao 12 Jan '18

12 Jan '18

From: Chen Hanxiao <chenhanxiao(a)gmail.com> As the description of daemon/libvirtd.conf, setting key_file, cert_file or key_file will override the default value. But if we set any one of them, we need to set all the rest of them. This patch clarify that description. Signed-off-by: Chen Hanxiao <chenhanxiao(a)gmail.com> --- daemon/libvirtd.conf | 3 +++ 1 file changed, 3 insertions(+) diff --git a/daemon/libvirtd.conf b/daemon/libvirtd.conf index 8e0c0d96d..7040ff26b 100644 --- a/daemon/libvirtd.conf +++ b/daemon/libvirtd.conf @@ -183,6 +183,9 @@ # +# NB, if the default value of 'key_file', 'cert_file" or +# 'ca_file' would be changed, +# all of them should be changed together. # Override the default server key file path # #key_file = "/etc/pki/libvirt/private/serverkey.pem" -- 2.14.3

1 0

[libvirt] [PATCH] m4: Don't enable bash-completion by default
by Michal Privoznik 12 Jan '18

12 Jan '18

Due to the way that check logic was written we basically enabled bash completion whenever readline was enabled. This is not right because it made bash-completion pkg-config module required. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Even though this would qualify as a build breaker, I'd rather have another pair of eyes to look at it (sorry Cyclops). m4/virt-bash-completion.m4 | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/m4/virt-bash-completion.m4 b/m4/virt-bash-completion.m4 index e1ef58740..c8342176f 100644 --- a/m4/virt-bash-completion.m4 +++ b/m4/virt-bash-completion.m4 @@ -30,14 +30,10 @@ AC_DEFUN([LIBVIRT_CHECK_BASH_COMPLETION], [ AC_REQUIRE([LIBVIRT_CHECK_READLINE]) if test "x$with_readline" != "xyes" ; then - if test "x$with_bash_completion" != "xyes" ; then - with_bash_completion=no - else + if test "x$with_bash_completion" = "xyes" ; then AC_MSG_ERROR([readline is required for bash completion support]) - fi - else - if test "x$with_bash_completion" = "xcheck" ; then - with_bash_completion=yes + else + with_bash_completion=no fi fi -- 2.13.6

2 1

[libvirt] [PATCH] qemu: Fix type of a completed job
by Jiri Denemark 12 Jan '18

12 Jan '18

Libvirt 3.7.0 and earlier libvirt reported a migration job as completed immediately after QEMU finished sending migration data at which point migration was not really complete yet. Commit v3.7.0-29-g3f2d6d829e fixed this, but caused a regression in reporting statistics for completed jobs which started reporting the job as still running. This happened because the completed job statistics including the job status are copied from the running job before we finally mark it as completed. Let's make sure QEMU_DOMAIN_JOB_STATUS_COMPLETED is always set in the completed job info even when the job has not finished yet. https://bugzilla.redhat.com/show_bug.cgi?id=1523036 Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/qemu/qemu_migration.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index c1ceb164d1..a56bc596ff 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -1646,8 +1646,10 @@ qemuMigrationWaitForCompletion(virQEMUDriverPtr driver, qemuDomainJobInfoUpdateTime(jobInfo); qemuDomainJobInfoUpdateDowntime(jobInfo); VIR_FREE(priv->job.completed); - if (VIR_ALLOC(priv->job.completed) == 0) + if (VIR_ALLOC(priv->job.completed) == 0) { *priv->job.completed = *jobInfo; + priv->job.completed->status = QEMU_DOMAIN_JOB_STATUS_COMPLETED; + } if (asyncJob != QEMU_ASYNC_JOB_MIGRATION_OUT && jobInfo->status == QEMU_DOMAIN_JOB_STATUS_QEMU_COMPLETED) @@ -5479,8 +5481,9 @@ qemuMigrationFinish(virQEMUDriverPtr driver, } if (dom) { - priv->job.completed = jobInfo; - jobInfo = NULL; + VIR_STEAL_PTR(priv->job.completed, jobInfo); + priv->job.completed->status = QEMU_DOMAIN_JOB_STATUS_COMPLETED; + if (qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen, QEMU_MIGRATION_COOKIE_STATS) < 0) VIR_WARN("Unable to encode migration cookie"); -- 2.15.1

2 1

[libvirt] [PATCH] qemu: Ignore fallback CPU attribute on reconnect
by Jiri Denemark 12 Jan '18

12 Jan '18

When reconnecting to a running domain with host-model CPU started by old libvirt which did not store the actual CPU in the status XML, we need to ignore the fallback attribute to make sure we can translate the detected host CPU model to a model which is supported by the running QEMU. https://bugzilla.redhat.com/show_bug.cgi?id=1532980 Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/qemu/qemu_process.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index 1a0923af36..25ec464d3e 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -3887,6 +3887,11 @@ qemuProcessUpdateCPU(virQEMUDriverPtr driver, virDomainCapsCPUModelsPtr models = NULL; int ret = -1; + /* The host CPU model comes from host caps rather than QEMU caps so + * fallback must be allowed no matter what the user specified in the XML. + */ + vm->def->cpu->fallback = VIR_CPU_FALLBACK_ALLOW; + if (qemuProcessFetchGuestCPU(driver, vm, asyncJob, &cpu, &disabled) < 0) goto cleanup; -- 2.15.1

2 1

[libvirt] [RFC PATCH] lxc: Up back the veth interfaces by default
by Benjamin Cama 12 Jan '18

12 Jan '18

Upping an interface without configuring it is not a “cardinal sin” but a sensible way to achieve auto-configuration, e.g. with IPv6 SLAAC (RFC 4862). If NetworkManager has troube with interfaces having only a link-local address, this is a bug in NetworkManager, not in libvirt; it should listen for router advertisements to decide if some interface has global connectivity or not. With network interfaces up by default, stateless containers can be easily auto-configured through the network with SLAAC, without any specific configuration from the host system. This reverts commit c3cf3c43a0bb2e0e4909c32821e20f607635ec85. Signed-off-by: Benjamin Cama <b.cama(a)kerlink.fr> --- Hi, The patch that I propose to revert basically broke my workflow for light stateless containers, where they could be auto-configured on IPv6-only network through SLAAC. Of course, fully-fledged containers can bring up the interface themselves, but this behavior had previously been the default for quite some time, and is even indicated as default in src/conf/domain_conf.h ("Default link state (up)"). I cannot find any real justification for the patch I am reverting, and the bugzilla looks private so I can not comment on the NetworkManager behavior, which looks very buggy to me. Please tell me if you think this is wrong. Also, please Cc me, I am not subscribed. Regards. .mailmap | 1 + src/lxc/lxc_container.c | 7 +------ src/lxc/lxc_native.c | 10 ++++++---- 3 files changed, 8 insertions(+), 10 deletions(-) diff --git a/.mailmap b/.mailmap index 2f0fc901e..9dc3bff85 100644 --- a/.mailmap +++ b/.mailmap @@ -36,6 +36,7 @@ <gerd(a)egidy.de> <lists(a)egidy.de> <gerd(a)egidy.de> <gerd.von.egidy(a)intra2net.com> <benoar(a)dolka.fr> <benjamin.cama(a)telecom-bretagne.eu> +<benoar(a)dolka.fr> <b.cama(a)kerlink.fr> <zhlcindy(a)linux.vnet.ibm.com> <zhlcindy(a)gmail.com> <serge.hallyn(a)canonical.com> <serue(a)us.ibm.com> <pritesh.kothari(a)sun.com> <Pritesh.Kothari(a)Sun.COM> diff --git a/src/lxc/lxc_container.c b/src/lxc/lxc_container.c index 96fceaf1b..e546f0aaf 100644 --- a/src/lxc/lxc_container.c +++ b/src/lxc/lxc_container.c @@ -517,12 +517,7 @@ lxcContainerRenameAndEnableInterfaces(virDomainDefPtr vmDef, if (virNetDevSetName(veths[i], newname) < 0) goto cleanup; - /* Only enable this device if there is a reason to do so (either - * at least one IP was specified, or link state was set to up in - * the config) - */ - if (netDef->guestIP.nips || - netDef->linkstate == VIR_DOMAIN_NET_INTERFACE_LINK_STATE_UP) { + if (netDef->linkstate != VIR_DOMAIN_NET_INTERFACE_LINK_STATE_DOWN) { VIR_DEBUG("Enabling %s", newname); if (virNetDevSetOnline(newname, true) < 0) goto cleanup; diff --git a/src/lxc/lxc_native.c b/src/lxc/lxc_native.c index fdc03a57e..f77a2a910 100644 --- a/src/lxc/lxc_native.c +++ b/src/lxc/lxc_native.c @@ -349,10 +349,12 @@ lxcCreateNetDef(const char *type, if (VIR_ALLOC(net) < 0) goto error; - if (STREQ_NULLABLE(flag, "up")) - net->linkstate = VIR_DOMAIN_NET_INTERFACE_LINK_STATE_UP; - else - net->linkstate = VIR_DOMAIN_NET_INTERFACE_LINK_STATE_DOWN; + if (flag) { + if (STREQ(flag, "up")) + net->linkstate = VIR_DOMAIN_NET_INTERFACE_LINK_STATE_UP; + else + net->linkstate = VIR_DOMAIN_NET_INTERFACE_LINK_STATE_DOWN; + } if (VIR_STRDUP(net->ifname_guest, name) < 0) goto error; -- 2.11.0

1 1