April 2019 - Devel - libvirt List Archives

[libvirt] [PATCH v2 0/2] qemu: Honor memory mode='strict'
by Michal Privoznik 16 Apr '19

16 Apr '19

v2 of: https://www.redhat.com/archives/libvir-list/2019-April/msg00658.html diff to v1: - Fixed the reported problem. Basically, even though emulator CGroup was created qemu was not running in it. Now qemu is moved into the CGroup even before exec() Michal Prívozník (2): qemuSetupCpusetMems: Use VIR_AUTOFREE() qemu: Set up EMULATOR thread and cpuset.mems before exec()-ing qemu src/qemu/qemu_cgroup.c | 5 ++--- src/qemu/qemu_process.c | 12 ++++++++---- 2 files changed, 10 insertions(+), 7 deletions(-) -- 2.21.0

4 13

[libvirt] [PATCH v2 0/6] Add support for setting the emulator scheduler parameters
by Martin Kletzander 16 Apr '19

16 Apr '19

I/O threads and vCPU threads already support setting schedulers, but until now it was impossible to do so for the main QEMU thread (emulator thread in the libvirt naming). This is, however, requested for some very specific scenarios, for example when vCPU threads are running at such priority that could starve the main thread. v2: - news.xml addition in a separate patch Martin Kletzander (6): conf: Parse common scheduler attributes in separate function conf: Format thread IDs optionally docs: Mention iothreadsched element in the docs and reword conf: Add support for emulatorsched qemu: Add support for emulatorsched docs: Add emulatorsched support to news.xml docs/formatdomain.html.in | 25 +++-- docs/news.xml | 12 +++ docs/schemas/domaincommon.rng | 5 + src/conf/domain_conf.c | 136 +++++++++++++++++------- src/conf/domain_conf.h | 1 + src/qemu/qemu_process.c | 2 +- tests/genericxml2xmlindata/cputune.xml | 37 +++++++ tests/genericxml2xmloutdata/cputune.xml | 40 +++++++ tests/genericxml2xmltest.c | 2 + 9 files changed, 212 insertions(+), 48 deletions(-) create mode 100644 tests/genericxml2xmlindata/cputune.xml create mode 100644 tests/genericxml2xmloutdata/cputune.xml -- 2.21.0

2 16

[libvirt] [PATCH] qemu: do not set wait:false for client sockets
by marcandre.lureau＠redhat.com 16 Apr '19

16 Apr '19

From: Marc-André Lureau <marcandre.lureau(a)redhat.com> Qemu commit 767abe7 ("chardev: forbid 'wait' option with client sockets") effectively deprecates usage of "wait" with client sockets starting with qemu 4.0, and earlier versions ignored the value. Cc: Daniel P. Berrangé <berrange(a)redhat.com> Signed-off-by: Marc-André Lureau <marcandre.lureau(a)redhat.com> --- src/qemu/qemu_monitor_json.c | 14 ++++++++++---- tests/qemumonitorjsontest.c | 2 -- 2 files changed, 10 insertions(+), 6 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index 71c452b25b..e7d063a5a8 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -6922,8 +6922,11 @@ qemuMonitorJSONAttachCharDevCommand(const char *chrID, telnet = chr->data.tcp.protocol == VIR_DOMAIN_CHR_TCP_PROTOCOL_TELNET; - if (virJSONValueObjectAppendBoolean(data, "wait", false) < 0 || - virJSONValueObjectAppendBoolean(data, "telnet", telnet) < 0 || + if (chr->data.tcp.listen && + virJSONValueObjectAppendBoolean(data, "wait", false) < 0) + goto cleanup; + + if (virJSONValueObjectAppendBoolean(data, "telnet", telnet) < 0 || virJSONValueObjectAppendBoolean(data, "server", chr->data.tcp.listen) < 0) goto cleanup; if (chr->data.tcp.tlscreds) { @@ -6973,8 +6976,11 @@ qemuMonitorJSONAttachCharDevCommand(const char *chrID, goto cleanup; addr = NULL; - if (virJSONValueObjectAppendBoolean(data, "wait", false) < 0 || - virJSONValueObjectAppendBoolean(data, "server", chr->data.nix.listen) < 0) + if (chr->data.nix.listen && + virJSONValueObjectAppendBoolean(data, "wait", false) < 0) + goto cleanup; + + if (virJSONValueObjectAppendBoolean(data, "server", chr->data.nix.listen) < 0) goto cleanup; if (qemuMonitorJSONBuildChrChardevReconnect(data, &chr->data.nix.reconnect) < 0) diff --git a/tests/qemumonitorjsontest.c b/tests/qemumonitorjsontest.c index 055e201611..f5ad3f6a73 100644 --- a/tests/qemumonitorjsontest.c +++ b/tests/qemumonitorjsontest.c @@ -874,7 +874,6 @@ qemuMonitorJSONTestAttachChardev(virDomainXMLOptionPtr xmlopt) "'data':{'addr':{'type':'inet'," "'data':{'host':'example.com'," "'port':'1234'}}," - "'wait':false," "'telnet':false," "'server':false}}}"); @@ -920,7 +919,6 @@ qemuMonitorJSONTestAttachChardev(virDomainXMLOptionPtr xmlopt) "'backend':{'type':'socket'," "'data':{'addr':{'type':'unix'," "'data':{'path':'/path/to/socket'}}," - "'wait':false," "'server':false}}}"); chr = (virDomainChrSourceDef) { .type = VIR_DOMAIN_CHR_TYPE_SPICEVMC }; -- 2.21.0.313.ge35b8cb8e2

2 2

[libvirt] [PATCH v4] socket: allow wait=false for client socket
by Marc-André Lureau 16 Apr '19

16 Apr '19

Commit 767abe7 ("chardev: forbid 'wait' option with client sockets") is a bit too strict. Current libvirt always set wait=false, and will thus fail to add client chardev. Make the code more permissive, allowing wait=false with client socket chardevs. Deprecate usage of 'wait' with client sockets. Fixes: 767abe7f49e8be14d29da5db3527817b5d696a52 Cc: Daniel P. Berrangé <berrange(a)redhat.com> Signed-off-by: Marc-André Lureau <marcandre.lureau(a)redhat.com> --- chardev/char-socket.c | 12 ++++++++---- qemu-deprecated.texi | 5 +++++ 2 files changed, 13 insertions(+), 4 deletions(-) diff --git a/chardev/char-socket.c b/chardev/char-socket.c index 3916505d67..b2cf593107 100644 --- a/chardev/char-socket.c +++ b/chardev/char-socket.c @@ -1263,10 +1263,14 @@ static bool qmp_chardev_validate_socket(ChardevSocket *sock, return false; } if (sock->has_wait) { - error_setg(errp, "%s", - "'wait' option is incompatible with " - "socket in client connect mode"); - return false; + warn_report("'wait' option is deprecated with " + "socket in client connect mode"); + if (sock->wait) { + error_setg(errp, "%s", + "'wait' option is incompatible with " + "socket in client connect mode"); + return false; + } } } diff --git a/qemu-deprecated.texi b/qemu-deprecated.texi index 2219386769..842e71b11d 100644 --- a/qemu-deprecated.texi +++ b/qemu-deprecated.texi @@ -105,6 +105,11 @@ details. The ``query-events'' command has been superseded by the more powerful and accurate ``query-qmp-schema'' command. +@subsection chardev client socket with 'wait' option (since 4.0) + +Character devices creating sockets in client mode should not specify +the 'wait' field, which is only applicable to sockets in server mode + @section Human Monitor Protocol (HMP) commands @subsection The hub_id parameter of 'hostfwd_add' / 'hostfwd_remove' (since 3.1) -- 2.21.0.313.ge35b8cb8e2

3 2

[libvirt] [PATCH] test_driver: provide virDomainGetTime implementation
by Ilias Stamatis 16 Apr '19

16 Apr '19

Implement testDomainGetTime by returning the current time. Signed-off-by: Ilias Stamatis <stamatis.iliass(a)gmail.com> --- I initially implemented this using clock_gettime, but Pavel suggested that this might not be a good idea since it isn't a cross-platform function. So I used virTimeMillisNow instead and set the nanoseconds part to 0 which can be ok for the test driver. src/test/test_driver.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/src/test/test_driver.c b/src/test/test_driver.c index d5eecf4b7f..b6f12e784f 100644 --- a/src/test/test_driver.c +++ b/src/test/test_driver.c @@ -64,6 +64,7 @@ #include "virinterfaceobj.h" #include "virhostcpu.h" #include "virdomainsnapshotobjlist.h" +#include "virtime.h" #define VIR_FROM_THIS VIR_FROM_TEST @@ -1943,6 +1944,23 @@ testDomainGetState(virDomainPtr domain, return 0; } +static int +testDomainGetTime(virDomainPtr dom ATTRIBUTE_UNUSED, + long long *seconds, + unsigned int *nseconds, + unsigned int flags ATTRIBUTE_UNUSED) +{ + unsigned long long now; + + if (virTimeMillisNow(&now) < 0) + return -1; + + *seconds = now / 1000; + *nseconds = 0; + + return 0; +} + #define TEST_SAVE_MAGIC "TestGuestMagic" static int @@ -6786,6 +6804,7 @@ static virHypervisorDriver testHypervisorDriver = { .domainSetMemory = testDomainSetMemory, /* 0.1.4 */ .domainGetInfo = testDomainGetInfo, /* 0.1.1 */ .domainGetState = testDomainGetState, /* 0.9.2 */ + .domainGetTime = testDomainGetTime, /* 5.3.0 */ .domainSave = testDomainSave, /* 0.3.2 */ .domainSaveFlags = testDomainSaveFlags, /* 0.9.4 */ .domainRestore = testDomainRestore, /* 0.3.2 */ -- 2.21.0

4 5

[libvirt] [PATCH] rpc: Segfaults and memory leak in virNetTLSContextNew function
by Adrian Brzezinski 16 Apr '19

16 Apr '19

Failed new gnutls context allocations in virNetTLSContextNew function results in double free and segfault. Occasional memory leaks may also occur. You can read detailed description at: https://bugzilla.redhat.com/show_bug.cgi?id=1699062 Signed-off-by: Adrian Brzezinski <redhat(a)adrb.pl> --- docs/news.xml | 10 ++++++++++ src/rpc/virnettlscontext.c | 6 ++++-- 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/docs/news.xml b/docs/news.xml index 21807f2..f6157ec 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -350,6 +350,16 @@ <section title="Bug fixes"> <change> <summary> + rpc: Segfaults and memory leak in virNetTLSContextNew function + </summary> + <description> + Failed new gnutls context allocations in virNetTLSContextNew function + results in double free and segfault. Occasional memory leaks may also + occur. + </description> + </change> + <change> + <summary> qemu: Use CAP_DAC_OVERRIDE during QEMU capabilities probing </summary> <description> diff --git a/src/rpc/virnettlscontext.c b/src/rpc/virnettlscontext.c index 72e9ed9..8f6ec8f 100644 --- a/src/rpc/virnettlscontext.c +++ b/src/rpc/virnettlscontext.c @@ -703,14 +703,14 @@ static virNetTLSContextPtr virNetTLSContextNew(const char *cacert, return NULL; if (VIR_STRDUP(ctxt->priority, priority) < 0) - goto error; + goto ctxt_init_error; err = gnutls_certificate_allocate_credentials(&ctxt->x509cred); if (err) { virReportError(VIR_ERR_SYSTEM_ERROR, _("Unable to allocate x509 credentials: %s"), gnutls_strerror(err)); - goto error; + goto ctxt_init_error; } if (sanityCheckCert && @@ -759,6 +759,8 @@ static virNetTLSContextPtr virNetTLSContextNew(const char *cacert, if (isServer) gnutls_dh_params_deinit(ctxt->dhParams); gnutls_certificate_free_credentials(ctxt->x509cred); + ctxt_init_error: + if (ctxt->priority) VIR_FREE(ctxt->priority); VIR_FREE(ctxt); return NULL; } -- 1.8.3.1

3 7

[libvirt] [PATCH] news: document dropping support of VirtualBox versions < 5.0
by Daniel P. Berrangé 16 Apr '19

16 Apr '19

Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- docs/news.xml | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/docs/news.xml b/docs/news.xml index 86c7734694..79b5658cc3 100644 --- a/docs/news.xml +++ b/docs/news.xml @@ -49,6 +49,16 @@ release. </description> </change> + <change> + <summary> + Removed support for VirtualBox versions < 5.0 + </summary> + <description> + Support for managing VirtualBox releases prior to 5.0 has + been dropped. The upstream project has declared that these + versions are end of life several years ago. + </description> + </change> </section> <section title="Bug fixes"> </section> -- 2.20.1

2 4

[libvirt] [PATCH 0/2] refactor virDomainSchedulerFormat
by Ján Tomko 16 Apr '19

16 Apr '19

*** BLURB HERE *** Ján Tomko (2): conf: pass the element/attribute names to virSchedulerFormat conf: switch virDomainSchedulerFormat to use virXMLFormatElement src/conf/domain_conf.c | 36 ++++++++++++++++++++---------------- 1 file changed, 20 insertions(+), 16 deletions(-) -- 2.19.2

2 4

[libvirt] [PATCH 0/1] power: add support for vm,uuid file
by Dimitri John Ledkov 16 Apr '19

16 Apr '19

On POWER systems, qemu exports uuid as a firmware devicetree node vm,uuid. Use that in a similar manner to the product_uuid file as on DMI based (x86) machines. Dimitri John Ledkov (1): viruuid: On POWER, qemu exports uuid as 'vm,uuid' devicetree node. src/util/viruuid.c | 2 ++ 1 file changed, 2 insertions(+) -- 2.20.1

2 2

[libvirt] [dbus PATCH] tests: fix test_connect
by Pavel Hrdina 16 Apr '19

16 Apr '19

Calling fixtures directly was removed in pytest 4.0, we can change the fixture to be a wrapper around the original function and use the original fixture name. Signed-off-by: Pavel Hrdina <phrdina(a)redhat.com> --- tests/libvirttest.py | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) diff --git a/tests/libvirttest.py b/tests/libvirttest.py index 14baf5b..a442196 100644 --- a/tests/libvirttest.py +++ b/tests/libvirttest.py @@ -84,13 +84,7 @@ class BaseTestClass(): interface_obj.Create(0) return path, interface_obj - @pytest.fixture def node_device_create(self): - """ Fixture to create dummy node device on the test driver - - This fixture should be used in the setup of every test manipulating - with node devices. - """ # We need a usable parent nodedev: possible candidates are # scsi_host2 (available since libvirt 3.1.0) and # test-scsi-host-vport (available until libvirt 3.0.0). @@ -109,6 +103,15 @@ class BaseTestClass(): path = self.connect.NodeDeviceCreateXML(xml, 0) return path + @pytest.fixture(name="node_device_create") + def fixture_node_device_create(self): + """ Fixture to create dummy node device on the test driver + + This fixture should be used in the setup of every test manipulating + with node devices. + """ + return self.node_device_create() + @pytest.fixture def storage_volume_create(self): """ Fixture to create dummy storage volume on the test driver -- 2.20.1

3 3