May 2019 - Devel - libvirt List Archives

[libvirt] [PATCH] build: fix linking libqemutestdriver with LTO enabled
by Jim Fehlig 12 Jun '19

12 Jun '19

openSUSE Factory is in the process of enabling Link Time Optimization [0] and stumbled upon missing symbols when linking libqemutestdriver libtool: link: gcc -shared -fPIC -DPIC ../src/libvirt_qemu_probes.o \ -Wl,--whole-archive ../src/.libs/libvirt_driver_qemu_impl.a \ ../src/.libs/libvirt_driver_network_impl.a \ ../src/.libs/libvirt_driver_storage_impl.a -Wl,--no-whole-archive \ -lcap-ng -lgnutls -lnl-route-3 -lnl-3 -ldbus-1 -lselinux -lapparmor \ /usr/lib/libxml2.so -ldl -lz -llzma -lm -lblkid -ltirpc \ -fstack-protector-strong -O2 -fstack-protector-strong -flto=16 -g -O2 \ -fstack-protector-strong -flto=16 -g \ -Wl,-soname -Wl,libqemutestdriver.so -o .libs/libqemutestdriver.so ... /usr/lib/gcc/i586-suse-linux/9/../../../../i586-suse-linux/bin/ld: ./.libs/libqemutestdriver.so: undefined reference to `mdir_name' /usr/lib/gcc/i586-suse-linux/9/../../../../i586-suse-linux/bin/ld: ./.libs/libqemutestdriver.so: undefined reference to `rpl_ioctl' /usr/lib/gcc/i586-suse-linux/9/../../../../i586-suse-linux/bin/ld: ./.libs/libqemutestdriver.so: undefined reference to `rpl_pipe2' /usr/lib/gcc/i586-suse-linux/9/../../../../i586-suse-linux/bin/ld: ./.libs/libqemutestdriver.so: undefined reference to `recvfd' /usr/lib/gcc/i586-suse-linux/9/../../../../i586-suse-linux/bin/ld: ./.libs/libqemutestdriver.so: undefined reference to `c_strcasecmp' The missing symbols are provided by GNUlib. Add it when linking libqemutestdriver. [0] https://en.opensuse.org/openSUSE:LTO Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- tests/Makefile.am | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/Makefile.am b/tests/Makefile.am index 46d94d2236..1d80647c9f 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -557,7 +557,7 @@ endif WITH_DTRACE_PROBES libqemutestdriver_la_SOURCES = libqemutestdriver_la_LDFLAGS = $(DRIVERLIB_LDFLAGS) -libqemutestdriver_la_LIBADD = $(qemu_LDADDS) +libqemutestdriver_la_LIBADD = $(qemu_LDADDS) $(GNULIB_LIBS) qemucpumock_la_SOURCES = \ qemucpumock.c testutilshostcpus.h @@ -729,7 +729,7 @@ lxcconf2xmltest_SOURCES = \ lxcconf2xmltest_LDADD = $(lxc_LDADDS) else ! WITH_LXC EXTRA_DIST += lxcxml2xmltest.c testutilslxc.c testutilslxc.h -endif ! WITH_LXC +-endif ! WITH_LXC if WITH_OPENVZ openvzutilstest_SOURCES = \ -- 2.21.0

3 8

[libvirt] [jenkins-ci PATCH] lcitool: Don't warn when using package manager directly
by Andrea Bolognani 11 Jun '19

11 Jun '19

We only do this when performing operations that the corresponding Ansible module doesn't support, so we know what we're doing and don't want warnings to show up. Note that while only the dnf and yum modules complain at the moment, we might as well use warn=no everywhere so that we're already covered in case in the future the pkgng module starts detecting this as well. Signed-off-by: Andrea Bolognani <abologna(a)redhat.com> --- guests/playbooks/update/tasks/base.yml | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/guests/playbooks/update/tasks/base.yml b/guests/playbooks/update/tasks/base.yml index 8fe114e..e0efe5d 100644 --- a/guests/playbooks/update/tasks/base.yml +++ b/guests/playbooks/update/tasks/base.yml @@ -52,11 +52,15 @@ - name: Update installed packages shell: '{{ package_manager }} update && {{ package_manager }} upgrade -y' + args: + warn: no when: - package_format == 'pkg' - name: Clean up packages after update shell: '{{ package_manager }} clean packages -y && {{ package_manager }} autoremove -y' + args: + warn: no when: - package_format == 'rpm' @@ -69,6 +73,8 @@ - name: Clean up packages after update shell: '{{ package_manager }} clean -y && {{ package_manager }} autoremove -y' + args: + warn: no when: - package_format == 'pkg' -- 2.21.0

2 4

[libvirt] [PATCHv2] virt-xml-validate: Allow input to be read from stdin
by Johannes Holmberg 10 Jun '19

10 Jun '19

Signed-off-by: Johannes Holmberg <johannes.holmberg(a)dataductus.se> --- Changes from v1: - Quotes around $TMPFILE everywhere. - Explicit -n checks in if statements - Fixed one instance of incorrect indentation - Signed-off-by line in commit message tools/virt-xml-validate.in | 46 ++++++++++++++++++++++++++++---------- 1 file changed, 34 insertions(+), 12 deletions(-) diff --git a/tools/virt-xml-validate.in b/tools/virt-xml-validate.in index 64aeaaaa33..5cb7dcd276 100644 --- a/tools/virt-xml-validate.in +++ b/tools/virt-xml-validate.in @@ -16,7 +16,17 @@ set -e -case $1 in +TMPFILE= + +cleanup() { + if [ -n "$TMPFILE" ]; then + rm -f "$TMPFILE" + fi +} + +trap cleanup EXIT + +case "$1" in -h | --h | --he | --hel | --help) cat <<EOF Usage: @@ -34,7 +44,7 @@ $0 (libvirt) @VERSION@ EOF exit ;; --) shift ;; - -*) + -?*) echo "$0: unrecognized option '$1'" >&2 exit 1 ;; esac @@ -42,18 +52,27 @@ esac XMLFILE="$1" TYPE="$2" -if [ -z "$XMLFILE" ]; then - echo "syntax: $0 XMLFILE [TYPE]" >&2 - exit 1 -fi +if [ "$XMLFILE" = "-" ]; then + TMPFILE=`mktemp --tmpdir virt-xml.XXXX` + cat > "$TMPFILE" +else + if [ -z "$XMLFILE" ]; then + echo "syntax: $0 XMLFILE [TYPE]" >&2 + exit 1 + fi -if [ ! -f "$XMLFILE" ]; then - echo "$0: document $XMLFILE does not exist" >&2 - exit 2 + if [ ! -f "$XMLFILE" ]; then + echo "$0: document $XMLFILE does not exist" >&2 + exit 2 + fi fi if [ -z "$TYPE" ]; then - ROOT=`xmllint --stream --debug "$XMLFILE" 2>/dev/null | grep "^0 1 " | awk '{ print $3 }'` + if [ -n "$TMPFILE" ]; then + ROOT=`xmllint --stream --debug - < "$TMPFILE" 2>/dev/null | grep "^0 1 " | awk '{ print $3 }'` + else + ROOT=`xmllint --stream --debug "$XMLFILE" 2>/dev/null | grep "^0 1 " | awk '{ print $3 }'` + fi case "$ROOT" in *domainsnapshot*) # Must come first, since *domain* is a substring TYPE="domainsnapshot" @@ -101,6 +120,9 @@ if [ ! -f "$SCHEMA" ]; then exit 4 fi -xmllint --noout --relaxng "$SCHEMA" "$XMLFILE" - +if [ -n "$TMPFILE" ]; then + xmllint --noout --relaxng "$SCHEMA" - < "$TMPFILE" +else + xmllint --noout --relaxng "$SCHEMA" "$XMLFILE" +fi exit -- 2.17.1

2 1

[libvirt] [PATCH 0/4] test_driver: implement virDomainSaveImageGetXMLDesc and virDomainSaveImageDefineXML
by Ilias Stamatis 07 Jun '19

07 Jun '19

While implementing virDomainSaveImageGetXMLDesc and virDomainSaveImageDefineXML for the test driver, I realized that there exists already code for saving and loading test images which can be reused. However, it needed to be extracted from testDomainSaveFlags and testDomainRestoreFlags into separate functions. The new functions are inspired by the corresponding QEMU driver code where e.g. qemuDomainSaveImageOpen serves as a helper used by other functions. This series of patches initially extracts the code mentioned above into separate functions and then provides the test driver with implementations for virDomainSaveImageGetXMLDesc and virDomainSaveImageDefineXML which make use of the newly introduced functions. Ilias Stamatis (4): test_driver: extract image saving code into a separate function test_driver: extract image loading code into a separate function test_driver: implement virDomainSaveImageDefineXML test_driver: implement virDomainSaveImageGetXMLDesc src/test/test_driver.c | 281 ++++++++++++++++++++++++++++------------- 1 file changed, 193 insertions(+), 88 deletions(-) -- 2.21.0

3 15

[libvirt] QMP; unsigned 64-bit ints; JSON standards compliance
by Daniel P. Berrangé 05 Jun '19

05 Jun '19

The QEMU QMP service is based on JSON which is nice because that is a widely supported "standard" data format..... ....except QEMU's implementation (and indeed most impls) are not strictly standards compliant. Specifically the problem is around representing 64-bit integers, whether signed or unsigned. The JSON standard declares that largest integer is 2^53-1 and the likewise the smallest is -(2^53-1): http://www.ecma-international.org/ecma-262/6.0/index.html#sec-number.max_sa… A crazy limit inherited from its javascript origins IIUC. QEMU, and indeed many applications, want to handle 64-bit integers. The C JSON library impls have traditionally mapped integers to the data type 'long long int' which gives a min/max of -(2^63) / 2^63-1. QEMU however /really/ needs 64-bit unsigned integers, ie a max 2^64-1. Libvirt has historically used the YAJL library which uses 'long long int' and thus can't officially go beyond 2^63-1 values. Fortunately it lets libvirt get at the raw json string, so libvirt can re-parse the value to get an 'unsigned long long'. We recently tried to switch to Jansson because YAJL has a dead upstream for many years and countless unanswered bugs & patches. Unfortunately we forgot about this need for 2^64-1 max, and Jansson also uses 'long long int' and raises a fatal parse error for unsigned 64-bit values above 2^63-1. It also provides no backdoor for libvirt todo its own integer parsing. Thus we had to abort our switch to jansson as it broke parsing QEMU's JSON: https://bugzilla.redhat.com/show_bug.cgi?id=1614569 Other JSON libraries we've investigated have similar problems. I imagine the same may well be true of non-C based JOSN impls, though I've not investigated in any detail. Essentially libvirt is stuck with either using the dead YAJL library forever, or writing its own JSON parser (most likely copying QEMU's JSON code into libvirt's git). This feels like a very unappealing situation to be in as not being able to use a JSON library of our choice is loosing one of the key benefits of using a standard data format. Thus I'd like to see a solution to this to allow QMP to be reliably consumed by any JSON library that exists. I can think of some options: 1. Encode unsigned 64-bit integers as signed 64-bit integers. This follows the example that most C libraries map JSON ints to 'long long int'. This is still relying on undefined behaviour as apps don't need to support > 2^53-1. Apps would need to cast back to 'unsigned long long' for those QMP fields they know are supposed to be unsigned. 2. Encode all 64-bit integers as a pair of 32-bit integers. This is fully compliant with the JSON spec as each half is fully within the declared limits. App has to split or assemble the 2 pieces from/to a signed/unsigned 64-bit int as needed. 3. Encode all 64-bit integers as strings The application has todo all parsing/formatting client side. None of these changes are backwards compatible, so I doubt we could make the change transparently in QMP. Instead we would have to have a QMP greeting message capability where the client can request enablement of the enhanced integer handling. Any of the three options above would likely work for libvirt, but I would have a slight preference for either 2 or 3, so that we become 100% standards compliant. Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|

5 22

[libvirt] [PATCH 0/9] Grab modify job for changing domain XML
by Michal Privoznik 04 Jun '19

04 Jun '19

This is an alternative proposal to: https://www.redhat.com/archives/libvir-list/2019-May/msg00830.html The problem I'm trying to fix is described here: https://www.redhat.com/archives/libvir-list/2019-May/msg00810.html Michal Prívozník (9): virDomainObjListAddLocked: Drop useless @cleanup label virDomainObjListAddObjLocked: Don't expect vm->def to be set virDomainObjListAddLocked: Set vm->def only in success path virDomainObjIsActive: Allow vm->def to be NULL virDomainObjListAdd: Leave def assigning as an exercise for caller qemu: Allow vm->def == NULL in job control APIs qemu: Grab modify job for changing domain XML lxc: Grab modify job for changing domain XML libxl: Grab modify job for changing domain XML src/bhyve/bhyve_driver.c | 10 +++++--- src/conf/domain_conf.h | 2 +- src/conf/virdomainobjlist.c | 48 ++++++++++++++---------------------- src/conf/virdomainobjlist.h | 3 +-- src/libxl/libxl_domain.c | 3 ++- src/libxl/libxl_driver.c | 48 ++++++++++++++++++++++++------------ src/libxl/libxl_migration.c | 14 +++++------ src/lxc/lxc_domain.c | 3 ++- src/lxc/lxc_driver.c | 23 +++++++++++------ src/openvz/openvz_conf.c | 12 ++++----- src/openvz/openvz_driver.c | 17 +++++++------ src/qemu/qemu_domain.c | 30 ++++++++++++++--------- src/qemu/qemu_driver.c | 49 ++++++++++++++++++++++++++----------- src/qemu/qemu_migration.c | 13 +++++++--- src/test/test_driver.c | 21 +++++++++------- src/vmware/vmware_conf.c | 4 +-- src/vmware/vmware_driver.c | 9 +++---- src/vz/vz_sdk.c | 4 ++- 18 files changed, 183 insertions(+), 130 deletions(-) -- 2.21.0

2 20

[libvirt] [PATCH 0/4] qemu: Fix qemuProcessInitCpuAffinity()
by Andrea Bolognani 04 Jun '19

04 Jun '19

See detailed explanation in the commit message for patch 1/4 and the corresponding bug report. Andrea Bolognani (4): util: Introduce virBitmapUnion() fixup? util: Optimize virBitmapUnion() util: Introduce virNumaNodesetToCPUset() qemu: Fix qemuProcessInitCpuAffinity() src/libvirt_private.syms | 2 ++ src/qemu/qemu_process.c | 7 ++++- src/util/virbitmap.c | 32 +++++++++++++++++++++++ src/util/virbitmap.h | 4 +++ src/util/virnuma.c | 55 ++++++++++++++++++++++++++++++++++++++++ src/util/virnuma.h | 2 ++ tests/virbitmaptest.c | 37 +++++++++++++++++++++++++++ 7 files changed, 138 insertions(+), 1 deletion(-) -- 2.21.0

3 16

[libvirt] [PATCH 0/5] Check TSC frequency before starting QEMU
by Jiri Denemark 03 Jun '19

03 Jun '19

When migrating a domain with invtsc CPU feature enabled, the TSC frequency of the destination host must match the frequency used when the domain was started on the source host or the destination host has to support TSC scaling. If the frequencies do not match and the destination host does not support TSC scaling, QEMU will fail to set the right TSC frequency when starting vCPUs on the destination and thus migration will fail. However, this is quite late since both host might have spent significant time transferring memory and perhaps even storage data. By adding the check to libvirt we can let migration fail before any data starts to be sent over. If for some reason libvirt is unable to detect the host's TSC frequency or scaling support, we'll just let QEMU try and the migration will either succeed or fail later. Luckily, we mandate TSC frequency to be explicitly set in the domain XML to even allow migration of domains with invtsc. We can just check whether the requested frequency is compatible with the current host before starting QEMU. And to let libvirt client decide whether it should even start the migration to a specific host, this series adds host's TSC frequency and scaling support to the host CPU capabilities XML. https://bugzilla.redhat.com/show_bug.cgi?id=1641702 Jiri Denemark (5): util: Add virHostCPUGetTscInfo conf: Report TSC frequency in host CPU capabilities cpu_x86: Fix placement of *CheckFeature functions cpu_x86: Probe TSC frequency and scaling support qemu: Check TSC frequency before starting QEMU src/conf/cpu_conf.c | 48 ++++++++++++++++++++++++ src/conf/cpu_conf.h | 2 + src/cpu/cpu_x86.c | 81 ++++++++++++++++++++++------------------- src/qemu/qemu_process.c | 53 +++++++++++++++++++++++++++ src/util/virhostcpu.c | 71 ++++++++++++++++++++++++++++++++++++ src/util/virhostcpu.h | 11 ++++++ 6 files changed, 229 insertions(+), 37 deletions(-) -- 2.21.0

2 7

[libvirt] [PATCH] test_driver: implement virDomainGetHostname
by Ilias Stamatis 03 Jun '19

03 Jun '19

Always return "domain_name" + "host". Signed-off-by: Ilias Stamatis <stamatis.iliass(a)gmail.com> --- src/test/test_driver.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index 2f58a1da95..aad7bb6036 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -1910,6 +1910,27 @@ static int testDomainReboot(virDomainPtr domain, return ret; } +static char * +testDomainGetHostname(virDomainPtr domain, + unsigned int flags) +{ + char *ret = NULL; + virDomainObjPtr vm = NULL; + + virCheckFlags(0, NULL); + + if (!(vm = testDomObjFromDomain(domain))) + goto cleanup; + + if (virDomainObjCheckActive(vm) < 0) + goto cleanup; + + ignore_value(virAsprintf(&ret, "%shost", domain->name)); + + cleanup: + return ret; +} + static int testDomainGetInfo(virDomainPtr domain, virDomainInfoPtr info) { @@ -6950,6 +6971,7 @@ static virHypervisorDriver testHypervisorDriver = { .domainGetMaxMemory = testDomainGetMaxMemory, /* 0.1.4 */ .domainSetMaxMemory = testDomainSetMaxMemory, /* 0.1.1 */ .domainSetMemory = testDomainSetMemory, /* 0.1.4 */ + .domainGetHostname = testDomainGetHostname, /* 5.5.0 */ .domainGetInfo = testDomainGetInfo, /* 0.1.1 */ .domainGetState = testDomainGetState, /* 0.9.2 */ .domainGetTime = testDomainGetTime, /* 5.4.0 */ -- 2.21.0

2 2

[libvirt] [PATCH] qemu: fix pr-helper0 remain
by Jie Wang 01 Jun '19

01 Jun '19

if libvirt receive DISCONNECTED event and set prDaemonRunning to false, and qemuDomainRemoveDiskDevice is performing in the meantime. qemuDomainRemoveDiskDevice will return directly by prDaemonRunning check, so the pr-helper0 object will remain. I think it is no need to check prDaemonRunning in qemuHotplugRemoveManagedPR. Signed-off-by: Jie Wang <wangjie88(a)huawei.com> --- src/qemu/qemu_hotplug.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 34249bd030..5e4a929738 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -465,8 +465,7 @@ qemuHotplugRemoveManagedPR(virQEMUDriverPtr driver, virErrorPtr orig_err; int ret = -1; - if (!priv->prDaemonRunning || - virDomainDefHasManagedPR(vm->def)) + if (virDomainDefHasManagedPR(vm->def)) return 0; virErrorPreserveLast(&orig_err); -- 2.16.2.windows.1

3 6