July 2015 - Devel - libvirt List Archives

[libvirt] [PATCH 0/2] nodeinfo: Various cleanups
by Andrea Bolognani 14 Jul '15

14 Jul '15

The first patch builds on the sysfs_prefix work by John, while the the second one contains unrelated formatting changes that increase internal consistency. Neither introduces behavioral changes. Andrea Bolognani (2): nodeinfo: Make sysfs_prefix usage more consistent nodeinfo: Formatting changes src/nodeinfo.c | 73 ++++++++++++++++++++++++++-------------------------- src/nodeinfopriv.h | 4 +-- tests/nodeinfotest.c | 14 +++++----- 3 files changed, 46 insertions(+), 45 deletions(-) -- 2.4.3

2 3

[libvirt] [PATCH] virsh: Teach cmdFreepages to work with lxc driver
by Michal Privoznik 14 Jul '15

14 Jul '15

Some drivers don't expose available huge page sizes in the capabilities XML. For instance, LXC driver is one of those. This has a downside that when virsh is trying to get aggregated info on free pages per all NUMA nodes, it fails. The problem is that the virNodeGetFreePages() API expects caller to pass an array of page sizes he is interested in. In virsh, this array is filled from the capabilities from '/capabilities/host/cpu/pages' XPath. As said, in LXC there's no such XPath and therefore virsh fails currently. But hey, we can fallback: the page sizes are exposed under '/capabilities/host/topology/cells/cell/pages'. The page size can be collected from there, and voilà the command works again. But now we must make sure that there are no duplicates in the array passed to the public API. Otherwise we won't get as beautiful output as we are getting now. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- tools/virsh-host.c | 37 ++++++++++++++++++++++++++++++++++--- 1 file changed, 34 insertions(+), 3 deletions(-) diff --git a/tools/virsh-host.c b/tools/virsh-host.c index 66f7fd9..28f6da1 100644 --- a/tools/virsh-host.c +++ b/tools/virsh-host.c @@ -285,6 +285,15 @@ static const vshCmdOptDef opts_freepages[] = { {.name = NULL} }; +static int +vshPageSizeSorter(const void *a, const void *b) +{ + unsigned int pa = *(unsigned int *)a; + unsigned int pb = *(unsigned int *)b; + + return pa - pb; +} + static bool cmdFreepages(vshControl *ctl, const vshCmd *cmd) { @@ -326,9 +335,15 @@ cmdFreepages(vshControl *ctl, const vshCmd *cmd) nodes_cnt = virXPathNodeSet("/capabilities/host/cpu/pages", ctxt, &nodes); if (nodes_cnt <= 0) { - vshError(ctl, "%s", _("could not get information about " - "supported page sizes")); - goto cleanup; + /* Some drivers don't export page sizes under the + * XPath above. Do another trick to get them */ + nodes_cnt = virXPathNodeSet("/capabilities/host/topology/cells/cell/pages", + ctxt, &nodes); + if (nodes_cnt <= 0) { + vshError(ctl, "%s", _("could not get information about " + "supported page sizes")); + goto cleanup; + } } pagesize = vshCalloc(ctl, nodes_cnt, sizeof(*pagesize)); @@ -345,6 +360,22 @@ cmdFreepages(vshControl *ctl, const vshCmd *cmd) VIR_FREE(val); } + /* Here, if we've done the trick few lines above, + * @pagesize array will contain duplicates. We should + * remove them otherwise not very nice output will be + * produced. */ + qsort(pagesize, nodes_cnt, sizeof(*pagesize), vshPageSizeSorter); + + for (i = 0; i < nodes_cnt - 1; ) { + if (pagesize[i] == pagesize[i + 1]) { + memmove(pagesize + i, pagesize + i + 1, + (nodes_cnt - i + 1) * sizeof(*pagesize)); + nodes_cnt--; + } else { + i++; + } + } + npages = nodes_cnt; VIR_FREE(nodes); } else { -- 2.3.6

2 1

[libvirt] Serial connection between guests
by Marcelo Ricardo Leitner 14 Jul '15

14 Jul '15

Hi, (I'm not subscribed to the list, please keep me on Cc) I'm attempting to get a serial link between two guests, same hypervisor. The only practical way I could find is to add a serial port using a pty to a guest and then manually connecting to the serial (console in my case) of the other guest using socat in the hypervisor. Then it made me think.. we could have this implemented at libvirt level. We could have a serial port on which we choose pty, udp, tcp, etc, and also a serial port from another guest, so that libvirt would handle socat start/stop automatically as both guests come up/down. Maybe libvirt could even do something smarter than that, maybe it can avoid socat somehow. What do you think? My usage is for virtualizing TAHI: http://networktest.sourceforge.net/usage.html I need 2 ethernet links plus a serial one, which TAHI can't break while running the tests. I didn't think this regarding multi-platform & all.. just sharing the idea/need. Thanks, Marcelo

2 4

[libvirt] [PATCH] libxl: rework reference counting
by Jim Fehlig 14 Jul '15

14 Jul '15

Similar to commit 540c339a for the QEMU driver, rework reference counting in the libxl driver to make it more deterministic and the code a bit cleaner. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- I've been testing this patch on and off for a few weeks now using libvirt-tck domain tests, local test scripts, and some manual tests for good measure. I sent the patch to Anthony Perard (cc'd) nearly two weeks ago for testing in his OpenStack+Xen+libvirt CI loop, although I haven't received any feedback thus far. Also included Martin in the cc since he encouraged this patch https://www.redhat.com/archives/libvir-list/2015-April/msg00014.html src/libxl/libxl_domain.c | 32 ++---- src/libxl/libxl_domain.h | 5 +- src/libxl/libxl_driver.c | 274 ++++++++++++++++++-------------------------- src/libxl/libxl_migration.c | 6 +- 4 files changed, 128 insertions(+), 189 deletions(-) diff --git a/src/libxl/libxl_domain.c b/src/libxl/libxl_domain.c index 0652270..ce188ee 100644 --- a/src/libxl/libxl_domain.c +++ b/src/libxl/libxl_domain.c @@ -96,13 +96,12 @@ libxlDomainObjFreeJob(libxlDomainObjPrivatePtr priv) #define LIBXL_JOB_WAIT_TIME (1000ull * 30) /* - * obj must be locked before calling, libxlDriverPrivatePtr must NOT be locked + * obj must be locked before calling * * This must be called by anything that will change the VM state - * in any way + * in any way. * - * Upon successful return, the object will have its ref count increased, - * successful calls must be followed by EndJob eventually + * Successful calls must eventually result in a call to EndJob. */ int libxlDomainObjBeginJob(libxlDriverPrivatePtr driver ATTRIBUTE_UNUSED, @@ -117,8 +116,6 @@ libxlDomainObjBeginJob(libxlDriverPrivatePtr driver ATTRIBUTE_UNUSED, return -1; then = now + LIBXL_JOB_WAIT_TIME; - virObjectRef(obj); - while (priv->job.active) { VIR_DEBUG("Wait normal job condition for starting job: %s", libxlDomainJobTypeToString(job)); @@ -149,21 +146,16 @@ libxlDomainObjBeginJob(libxlDriverPrivatePtr driver ATTRIBUTE_UNUSED, virReportSystemError(errno, "%s", _("cannot acquire job mutex")); - virObjectUnref(obj); return -1; } /* - * obj must be locked before calling + * obj must be locked and have a reference before calling * * To be called after completing the work associated with the * earlier libxlDomainBeginJob() call - * - * Returns true if the remaining reference count on obj is - * non-zero, false if the reference count has dropped to zero - * and obj is disposed. */ -bool +void libxlDomainObjEndJob(libxlDriverPrivatePtr driver ATTRIBUTE_UNUSED, virDomainObjPtr obj) { @@ -175,8 +167,6 @@ libxlDomainObjEndJob(libxlDriverPrivatePtr driver ATTRIBUTE_UNUSED, libxlDomainObjResetJob(priv); virCondSignal(&priv->job.cond); - - return virObjectUnref(obj); } static void * @@ -485,12 +475,11 @@ libxlDomainShutdownThread(void *opaque) } endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + virObjectUnlock(vm); + virObjectUnref(vm); if (dom_event) libxlDomainEventQueue(driver, dom_event); libxl_event_free(cfg->ctx, ev); @@ -528,6 +517,7 @@ libxlDomainEventHandler(void *data, VIR_LIBXL_EVENT_CONST libxl_event *event) VIR_INFO("Received event for unknown domain ID %d", event->domid); goto error; } + virObjectRef(vm); /* * Start a thread to handle shutdown. We don't want to be tying up @@ -558,8 +548,10 @@ libxlDomainEventHandler(void *data, VIR_LIBXL_EVENT_CONST libxl_event *event) /* Cast away any const */ libxl_event_free(cfg->ctx, (libxl_event *)event); virObjectUnref(cfg); - if (vm) + if (vm) { virObjectUnlock(vm); + virObjectUnref(vm); + } VIR_FREE(shutdown_info); } diff --git a/src/libxl/libxl_domain.h b/src/libxl/libxl_domain.h index 8c73cc4..714ed91 100644 --- a/src/libxl/libxl_domain.h +++ b/src/libxl/libxl_domain.h @@ -83,10 +83,9 @@ libxlDomainObjBeginJob(libxlDriverPrivatePtr driver, enum libxlDomainJob job) ATTRIBUTE_RETURN_CHECK; -bool +void libxlDomainObjEndJob(libxlDriverPrivatePtr driver, - virDomainObjPtr obj) - ATTRIBUTE_RETURN_CHECK; + virDomainObjPtr obj); void libxlDomainEventQueue(libxlDriverPrivatePtr driver, diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index a7be745..c0061b3 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -282,7 +282,7 @@ libxlDomObjFromDomain(virDomainPtr dom) libxlDriverPrivatePtr driver = dom->conn->privateData; char uuidstr[VIR_UUID_STRING_BUFLEN]; - vm = virDomainObjListFindByUUID(driver->domains, dom->uuid); + vm = virDomainObjListFindByUUIDRef(driver->domains, dom->uuid); if (!vm) { virUUIDFormat(dom->uuid, uuidstr); virReportError(VIR_ERR_NO_DOMAIN, @@ -294,6 +294,25 @@ libxlDomObjFromDomain(virDomainPtr dom) return vm; } +/* + * Finish working with a domain object in an API. This function + * clears whatever was left of a domain that was gathered using + * libxlDomObjFromDomain(). Currently that means only unlocking and + * decrementing the reference counter of that domain. And in order to + * make sure the caller does not access the domain, the pointer is + * cleared. + */ +static void +libxlDomObjEndAPI(virDomainObjPtr *vm) +{ + if (!*vm) + return; + + virObjectUnlock(*vm); + virObjectUnref(*vm); + *vm = NULL; +} + static int libxlAutostartDomain(virDomainObjPtr vm, void *opaque) @@ -303,6 +322,7 @@ libxlAutostartDomain(virDomainObjPtr vm, int ret = -1; virObjectLock(vm); + virObjectRef(vm); virResetLastError(); if (libxlDomainObjBeginJob(driver, vm, LIBXL_JOB_MODIFY) < 0) { @@ -322,8 +342,10 @@ libxlAutostartDomain(virDomainObjPtr vm, ret = 0; endjob: - if (libxlDomainObjEndJob(driver, vm)) - virObjectUnlock(vm); + libxlDomainObjEndJob(driver, vm); + + virObjectUnlock(vm); + virObjectUnref(vm); return ret; } @@ -908,19 +930,19 @@ libxlDomainCreateXML(virConnectPtr conn, const char *xml, VIR_DOMAIN_OBJ_LIST_ADD_CHECK_LIVE, NULL))) goto cleanup; + virObjectRef(vm); def = NULL; if (libxlDomainObjBeginJob(driver, vm, LIBXL_JOB_MODIFY) < 0) { - if (!vm->persistent) { + if (!vm->persistent) virDomainObjListRemove(driver->domains, vm); - vm = NULL; - } goto cleanup; } if (libxlDomainStart(driver, vm, (flags & VIR_DOMAIN_START_PAUSED) != 0, -1) < 0) { - virDomainObjListRemove(driver->domains, vm); + if (!vm->persistent) + virDomainObjListRemove(driver->domains, vm); goto endjob; } @@ -929,13 +951,11 @@ libxlDomainCreateXML(virConnectPtr conn, const char *xml, dom->id = vm->def->id; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: virDomainDefFree(def); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return dom; } @@ -1060,12 +1080,10 @@ libxlDomainSuspend(virDomainPtr dom) ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); if (event) libxlDomainEventQueue(driver, event); virObjectUnref(cfg); @@ -1117,12 +1135,10 @@ libxlDomainResume(virDomainPtr dom) ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); if (event) libxlDomainEventQueue(driver, event); virObjectUnref(cfg); @@ -1183,8 +1199,7 @@ libxlDomainShutdownFlags(virDomainPtr dom, unsigned int flags) } cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -1232,8 +1247,7 @@ libxlDomainReboot(virDomainPtr dom, unsigned int flags) } cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -1281,12 +1295,10 @@ libxlDomainDestroyFlags(virDomainPtr dom, ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); if (event) libxlDomainEventQueue(driver, event); virObjectUnref(cfg); @@ -1315,8 +1327,7 @@ libxlDomainGetOSType(virDomainPtr dom) goto cleanup; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return type; } @@ -1335,8 +1346,7 @@ libxlDomainGetMaxMemory(virDomainPtr dom) ret = virDomainDefGetMemoryActual(vm->def); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -1455,12 +1465,10 @@ libxlDomainSetMemoryFlags(virDomainPtr dom, unsigned long newmem, ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -1513,8 +1521,7 @@ libxlDomainGetInfo(virDomainPtr dom, virDomainInfoPtr info) ret = 0; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -1540,8 +1547,7 @@ libxlDomainGetState(virDomainPtr dom, ret = 0; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -1638,7 +1644,6 @@ libxlDomainSaveFlags(virDomainPtr dom, const char *to, const char *dxml, libxlDriverPrivatePtr driver = dom->conn->privateData; virDomainObjPtr vm; int ret = -1; - bool remove_dom = false; #ifdef LIBXL_HAVE_NO_SUSPEND_RESUME virReportUnsupportedError(); @@ -1670,21 +1675,15 @@ libxlDomainSaveFlags(virDomainPtr dom, const char *to, const char *dxml, goto endjob; if (!vm->persistent) - remove_dom = true; + virDomainObjListRemove(driver->domains, vm); ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (remove_dom && vm) { - virDomainObjListRemove(driver->domains, vm); - vm = NULL; - } - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -1735,10 +1734,8 @@ libxlDomainRestoreFlags(virConnectPtr conn, const char *from, def = NULL; if (libxlDomainObjBeginJob(driver, vm, LIBXL_JOB_MODIFY) < 0) { - if (!vm->persistent) { + if (!vm->persistent) virDomainObjListRemove(driver->domains, vm); - vm = NULL; - } goto cleanup; } @@ -1746,15 +1743,13 @@ libxlDomainRestoreFlags(virConnectPtr conn, const char *from, if (ret < 0 && !vm->persistent) virDomainObjListRemove(driver->domains, vm); - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: if (VIR_CLOSE(fd) < 0) virReportSystemError(errno, "%s", _("cannot close file")); virDomainDefFree(def); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -1772,7 +1767,6 @@ libxlDomainCoreDump(virDomainPtr dom, const char *to, unsigned int flags) libxlDriverConfigPtr cfg = libxlDriverConfigGet(driver); virDomainObjPtr vm; virObjectEventPtr event = NULL; - bool remove_dom = false; bool paused = false; int ret = -1; @@ -1828,7 +1822,7 @@ libxlDomainCoreDump(virDomainPtr dom, const char *to, unsigned int flags) event = virDomainEventLifecycleNewFromObj(vm, VIR_DOMAIN_EVENT_STOPPED, VIR_DOMAIN_EVENT_STOPPED_CRASHED); if (!vm->persistent) - remove_dom = true; + virDomainObjListRemove(driver->domains, vm); } ret = 0; @@ -1846,16 +1840,10 @@ libxlDomainCoreDump(virDomainPtr dom, const char *to, unsigned int flags) } endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (remove_dom && vm) { - virDomainObjListRemove(driver->domains, vm); - vm = NULL; - } - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); if (event) libxlDomainEventQueue(driver, event); virObjectUnref(cfg); @@ -1869,7 +1857,6 @@ libxlDomainManagedSave(virDomainPtr dom, unsigned int flags) virDomainObjPtr vm = NULL; char *name = NULL; int ret = -1; - bool remove_dom = false; virCheckFlags(0, -1); @@ -1902,21 +1889,15 @@ libxlDomainManagedSave(virDomainPtr dom, unsigned int flags) goto endjob; if (!vm->persistent) - remove_dom = true; + virDomainObjListRemove(driver->domains, vm); ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (remove_dom && vm) { - virDomainObjListRemove(driver->domains, vm); - vm = NULL; - } - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); VIR_FREE(name); return ret; } @@ -1960,8 +1941,7 @@ libxlDomainHasManagedSaveImage(virDomainPtr dom, unsigned int flags) ret = vm->hasManagedSave; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -1990,8 +1970,7 @@ libxlDomainManagedSaveRemove(virDomainPtr dom, unsigned int flags) cleanup: VIR_FREE(name); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -2119,13 +2098,11 @@ libxlDomainSetVcpusFlags(virDomainPtr dom, unsigned int nvcpus, ret = virDomainSaveConfig(cfg->configDir, def); endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: VIR_FREE(bitmask); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -2187,8 +2164,7 @@ libxlDomainGetVcpusFlags(virDomainPtr dom, unsigned int flags) ret = (flags & VIR_DOMAIN_VCPU_MAXIMUM) ? def->maxvcpus : def->vcpus; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -2270,12 +2246,10 @@ libxlDomainPinVcpuFlags(virDomainPtr dom, unsigned int vcpu, } endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virBitmapFree(pcpumap); virObjectUnref(cfg); return ret; @@ -2354,8 +2328,7 @@ libxlDomainGetVcpuPinInfo(virDomainPtr dom, int ncpumaps, cleanup: virBitmapFree(allcpumap); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -2418,8 +2391,7 @@ libxlDomainGetVcpus(virDomainPtr dom, virVcpuInfoPtr info, int maxinfo, ret = maxinfo; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -2442,8 +2414,7 @@ libxlDomainGetXMLDesc(virDomainPtr dom, unsigned int flags) virDomainDefFormatConvertXMLFlags(flags)); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -2618,12 +2589,10 @@ libxlDomainCreateWithFlags(virDomainPtr dom, dom->id = vm->def->id; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -2760,8 +2729,7 @@ libxlDomainUndefineFlags(virDomainPtr dom, cleanup: VIR_FREE(name); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); if (event) libxlDomainEventQueue(driver, event); virObjectUnref(cfg); @@ -3618,14 +3586,12 @@ libxlDomainAttachDeviceFlags(virDomainPtr dom, const char *xml, } endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: virDomainDefFree(vmdef); virDomainDeviceDefFree(dev); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -3726,14 +3692,12 @@ libxlDomainDetachDeviceFlags(virDomainPtr dom, const char *xml, } endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: virDomainDefFree(vmdef); virDomainDeviceDefFree(dev); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -3833,8 +3797,7 @@ libxlDomainUpdateDeviceFlags(virDomainPtr dom, const char *xml, cleanup: virDomainDefFree(vmdef); virDomainDeviceDefFree(dev); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -3964,8 +3927,7 @@ libxlDomainGetAutostart(virDomainPtr dom, int *autostart) ret = 0; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -4029,14 +3991,12 @@ libxlDomainSetAutostart(virDomainPtr dom, int autostart) ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: VIR_FREE(configFile); VIR_FREE(autostartLink); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -4091,8 +4051,7 @@ libxlDomainGetSchedulerType(virDomainPtr dom, int *nparams) ignore_value(VIR_STRDUP(ret, name)); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -4157,8 +4116,7 @@ libxlDomainGetSchedulerParametersFlags(virDomainPtr dom, ret = 0; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -4241,12 +4199,10 @@ libxlDomainSetSchedulerParametersFlags(virDomainPtr dom, ret = 0; endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -4316,8 +4272,7 @@ libxlDomainOpenConsole(virDomainPtr dom, } cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -4452,8 +4407,7 @@ libxlDomainGetNumaParameters(virDomainPtr dom, VIR_FREE(nodeset); virBitmapFree(nodes); libxl_bitmap_dispose(&nodemap); - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); virObjectUnref(cfg); return ret; } @@ -4474,8 +4428,7 @@ libxlDomainIsActive(virDomainPtr dom) ret = virDomainObjIsActive(obj); cleanup: - if (obj) - virObjectUnlock(obj); + libxlDomObjEndAPI(&obj); return ret; } @@ -4494,8 +4447,7 @@ libxlDomainIsPersistent(virDomainPtr dom) ret = obj->persistent; cleanup: - if (obj) - virObjectUnlock(obj); + libxlDomObjEndAPI(&obj); return ret; } @@ -4514,8 +4466,7 @@ libxlDomainIsUpdated(virDomainPtr dom) ret = vm->updated; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -4780,6 +4731,7 @@ libxlDomainMigrateBegin3Params(virDomainPtr domain, { const char *xmlin = NULL; virDomainObjPtr vm = NULL; + char *ret = NULL; #ifdef LIBXL_HAVE_NO_SUSPEND_RESUME virReportUnsupportedError(); @@ -4798,19 +4750,20 @@ libxlDomainMigrateBegin3Params(virDomainPtr domain, if (!(vm = libxlDomObjFromDomain(domain))) return NULL; - if (virDomainMigrateBegin3ParamsEnsureACL(domain->conn, vm->def) < 0) { - virObjectUnlock(vm); - return NULL; - } + if (virDomainMigrateBegin3ParamsEnsureACL(domain->conn, vm->def) < 0) + goto cleanup; if (!virDomainObjIsActive(vm)) { virReportError(VIR_ERR_OPERATION_INVALID, "%s", _("domain is not running")); - virObjectUnlock(vm); - return NULL; + goto cleanup; } - return libxlDomainMigrationBegin(domain->conn, vm, xmlin); + ret = libxlDomainMigrationBegin(domain->conn, vm, xmlin); + + cleanup: + libxlDomObjEndAPI(&vm); + return ret; } static int @@ -4919,8 +4872,7 @@ libxlDomainMigratePerform3Params(virDomainPtr dom, ret = 0; cleanup: - if (vm) - virObjectUnlock(vm); + libxlDomObjEndAPI(&vm); return ret; } @@ -4963,23 +4915,20 @@ libxlDomainMigrateFinish3Params(virConnectPtr dconn, NULLSTR(dname)); return NULL; } + virObjectRef(vm); - if (virDomainMigrateFinish3ParamsEnsureACL(dconn, vm->def) < 0) { - virDomainObjEndAPI(&vm); - return NULL; - } + if (virDomainMigrateFinish3ParamsEnsureACL(dconn, vm->def) < 0) + goto cleanup; - if (libxlDomainObjBeginJob(driver, vm, LIBXL_JOB_MODIFY) < 0) { - virDomainObjEndAPI(&vm); - return NULL; - } + if (libxlDomainObjBeginJob(driver, vm, LIBXL_JOB_MODIFY) < 0) + goto cleanup; ret = libxlDomainMigrationFinish(dconn, vm, flags, cancelled); - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); - virDomainObjEndAPI(&vm); + cleanup: + libxlDomObjEndAPI(&vm); return ret; } @@ -4995,6 +4944,7 @@ libxlDomainMigrateConfirm3Params(virDomainPtr domain, { libxlDriverPrivatePtr driver = domain->conn->privateData; virDomainObjPtr vm = NULL; + int ret = -1; #ifdef LIBXL_HAVE_NO_SUSPEND_RESUME virReportUnsupportedError(); @@ -5008,12 +4958,14 @@ libxlDomainMigrateConfirm3Params(virDomainPtr domain, if (!(vm = libxlDomObjFromDomain(domain))) return -1; - if (virDomainMigrateConfirm3ParamsEnsureACL(domain->conn, vm->def) < 0) { - virObjectUnlock(vm); - return -1; - } + if (virDomainMigrateConfirm3ParamsEnsureACL(domain->conn, vm->def) < 0) + goto cleanup; - return libxlDomainMigrationConfirm(driver, vm, flags, cancelled); + ret = libxlDomainMigrationConfirm(driver, vm, flags, cancelled); + + cleanup: + libxlDomObjEndAPI(&vm); + return ret; } static int libxlNodeGetSecurityModel(virConnectPtr conn, diff --git a/src/libxl/libxl_migration.c b/src/libxl/libxl_migration.c index 39e4a65..665b21a 100644 --- a/src/libxl/libxl_migration.c +++ b/src/libxl/libxl_migration.c @@ -257,13 +257,9 @@ libxlDomainMigrationBegin(virConnectPtr conn, xml = virDomainDefFormat(def, VIR_DOMAIN_DEF_FORMAT_SECURE); endjob: - if (!libxlDomainObjEndJob(driver, vm)) - vm = NULL; + libxlDomainObjEndJob(driver, vm); cleanup: - if (vm) - virObjectUnlock(vm); - virDomainDefFree(tmpdef); virObjectUnref(cfg); return xml; -- 2.3.7

3 4

[libvirt] [PATCH v2] Fix qemu-nbd cleanup crashes
by Cédric Bosdonnat 14 Jul '15

14 Jul '15

The virLXCControllerAppendNBDPids function didn't properly initialize pids and npids. In case of failure it was crashing when freeing those. The nbd device pid file doesn't appear immediately after starting qemu-nbd: adding a small loop to wait for it. Diff to v1: * Fixed a typo in a variable name.... working with several repos leads to troubles ;) --- src/lxc/lxc_controller.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 828b8a8..78d3eee 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -533,16 +533,31 @@ static int virLXCControllerAppendNBDPids(virLXCControllerPtr ctrl, const char *dev) { char *pidpath = NULL; - pid_t *pids; - size_t npids; + pid_t *pids = NULL; + size_t npids = 0; size_t i; int ret = -1; + size_t loops = 0; pid_t pid; if (!STRPREFIX(dev, "/dev/") || virAsprintf(&pidpath, "/sys/devices/virtual/block/%s/pid", dev + 5) < 0) goto cleanup; + /* Wait for the pid file to appear */ + while (!virFileExists(pidpath)) { + /* wait for 100ms before checking again, but don't do it for ever */ + if (errno == ENOENT && loop < 10) { + usleep(100 * 1000); + loop++; + } else { + virReportSystemError(errno, + _("Cannot check NBD device %s pid"), + dev + 5); + goto cleanup; + } + } + if (virPidFileReadPath(pidpath, &pid) < 0) goto cleanup; -- 2.1.4

3 2

[libvirt] [PATCH v2] qemu: process: Improve update of maximum baloon state at startup
by Peter Krempa 14 Jul '15

14 Jul '15

In commit 641a145d73fdc3dd9350fd57b3d3247abf101c05 I've added code that resets the ballon memory value to full size prior to resuming the vCPUs since the size certainly was not reduced at that point. Since qemuProcess start is used also in code paths with already booted up guests (migration, save/restore) the assumption is not entirely true since the guest might already been running before. This patch adds a function that queries the monitor rather than using the full size since a balloon event would not be reissued in case we are recovering a saved migration state. Additionally the new function is used also when reconnecting to a VM after libvirtd restart since we might have missed a few ballon events while libvirtd was not running. --- src/qemu/qemu_process.c | 38 +++++++++++++++++++++++++++++++++++++- 1 file changed, 37 insertions(+), 1 deletion(-) diff --git a/src/qemu/qemu_process.c b/src/qemu/qemu_process.c index ba84182..13b1696 100644 --- a/src/qemu/qemu_process.c +++ b/src/qemu/qemu_process.c @@ -2065,6 +2065,38 @@ qemuProcessReconnectRefreshChannelVirtioState(virQEMUDriverPtr driver, static int +qemuProcessRefreshBallonState(virQEMUDriverPtr driver, + virDomainObjPtr vm, + int asyncJob) +{ + unsigned long long balloon; + int rc; + + /* if no balloning is available, the current size equals to the current + * full memory size */ + if (!vm->def->memballoon || + vm->def->memballoon->model == VIR_DOMAIN_MEMBALLOON_MODEL_NONE) { + vm->def->mem.cur_balloon = virDomainDefGetMemoryActual(vm->def); + return 0; + } + + if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0) + return -1; + + rc = qemuMonitorGetBalloonInfo(qemuDomainGetMonitor(vm), &balloon); + if (qemuDomainObjExitMonitor(driver, vm) < 0) + rc = -1; + + if (rc < 0) + return -1; + + vm->def->mem.cur_balloon = balloon; + + return 0; +} + + +static int qemuProcessWaitForMonitor(virQEMUDriverPtr driver, virDomainObjPtr vm, int asyncJob, @@ -3779,6 +3811,9 @@ qemuProcessReconnect(void *opaque) if (qemuProcessReconnectRefreshChannelVirtioState(driver, obj) < 0) goto error; + if (qemuProcessRefreshBallonState(driver, obj, QEMU_ASYNC_JOB_NONE) < 0) + goto error; + if (qemuProcessRecoverJob(driver, obj, conn, &oldjob) < 0) goto error; @@ -4922,7 +4957,8 @@ int qemuProcessStart(virConnectPtr conn, /* Since CPUs were not started yet, the ballon could not return the memory * to the host and thus cur_balloon needs to be updated so that GetXMLdesc * and friends return the correct size in case they can't grab the job */ - vm->def->mem.cur_balloon = virDomainDefGetMemoryActual(vm->def); + if (qemuProcessRefreshBallonState(driver, vm, asyncJob) < 0) + goto cleanup; VIR_DEBUG("Detecting actual memory size for video device"); if (qemuProcessUpdateVideoRamSize(driver, vm, asyncJob) < 0) -- 2.4.1

3 4

[libvirt] How to get the fully qualified path for network disks
by Deepak Shetty 14 Jul '15

14 Jul '15

Some background: Looking @ the comments in https://review.openstack.org/#/c/192736/8/nova/virt/libvirt/driver.py we need a way to construct the fully qualified path for a 'network' type disk which can then be passed to qemu-img for the rebase cmd. Checking with some folks offline and on IRC, i was suggested to bring this up here. Is there a way in libvirt to get the full path of network disk that can work with qemu-img given the virDomain pointer ? thanx, deepak

1 0

[libvirt] [PATCH] qemuDomainSetNumaParamsLive:
by Michal Privoznik 14 Jul '15

14 Jul '15

https://bugzilla.redhat.com/show_bug.cgi?id=1232663 In one of my previous ptaches (bcd9a564) I've tried to fix the problem that we blindly assumed strict NUMA mode for guests. This led to several problems like us pinning a domain onto a nodeset via libnuma among with CGroups. Once the nodeset was changed by user, well, it did not result in desired effect. See the original commit for more info. But, the commit I wrote had a bug: when NUMA parameters are changed on a running domain we require domain to be strictly pinned onto a nodeset. Due to a typo a condition was mis-evaluated. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/qemu/qemu_driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index c8cbd57..8c705c4 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -9954,7 +9954,7 @@ qemuDomainSetNumaParamsLive(virDomainObjPtr vm, size_t i = 0; int ret = -1; - if (virDomainNumatuneGetMode(vm->def->numa, -1, &mode) < 0 || + if (virDomainNumatuneGetMode(vm->def->numa, -1, &mode) == 0 && mode != VIR_DOMAIN_NUMATUNE_MEM_STRICT) { virReportError(VIR_ERR_OPERATION_INVALID, "%s", _("change of nodeset for running domain " -- 2.3.6

2 1

[libvirt] [PATCH 0/2] Check static route collisions
by Martin Kletzander 14 Jul '15

14 Jul '15

Martin Kletzander (2): conf: Add getter for network routes network: Add another collision check into networkCheckRouteCollision src/conf/network_conf.c | 26 ++++++++++++++++++++++++++ src/conf/network_conf.h | 3 +++ src/libvirt_private.syms | 1 + src/network/bridge_driver_linux.c | 29 +++++++++++++++++++++++++++++ 4 files changed, 59 insertions(+) -- 2.4.5

2 4

Re: [libvirt] Supporting remote connections to QEMU session mode
by Spencer Baugh 13 Jul '15

13 Jul '15

Eric Blake <eblake(a)redhat.com> writes: > On 07/03/2015 12:01 AM, sbaugh(a)catern.com wrote: >> Has there been any thought given to supporting remote connections to >> QEMU session instances in virt-manager? >> (i.e. qemu+ssh://hostname/session?socket=path) Manually hacking in the >> URI with dconf seems to mostly work, but (for example) I quickly >> encountered a bug when creating a VM where virt-manager assumed the only >> usable storage pool was my *local* homedir instead of the remote >> homedir. > > Libvirt does not (yet) allow remote connections to qemu://host/session. > You'd first have to figure out how to patch libvirt to allow remote > session connections; part of the problem is that with qemu:///system, > there is a daemon running that is always listening for remote > connections, but with qemu:///session, since libvirtd is only > auto-started as needed, there is no easy way to guarantee that the > remote machine is running a libvirtd process as the correct user. > >> >> virsh and other tools support these connections just fine. > > No, virsh does NOT support remote session connections, at least not > intentionally. If it works for you, it is a pure accident. I'm not sure what you mean by "not support". Connecting to a qemu+ssh://host/session?socket=path url works, and clearly some thought has been put into it because if I leave off the socket argument, I get this message: @sbaugh@earth ~ $ virsh -c qemu+ssh://host/session error: failed to connect to the hypervisor error: Operation not supported: Connecting to session instance without socket path is not supported by the ssh connection driver which is created by this snippet of code: /* Right now we don't support default session connections */ if (STREQ_NULLABLE(conn->uri->path, "/session")) { virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", _("Connecting to session instance without " "socket path is not supported by the libssh2 " "connection driver")); goto failed; } which certainly suggests to me that *non*-default session connections are supported. No patching is necessary to allow remote session connections, they are implicitly allowed. That if statement could easily disable them, but chose instead to allow them when they are sufficiently specified to actually work. All that aside, it's obvious that not much attention has been paid to making remote session connections work well even in core libvirt. Is improving the support for them something that is interesting to the libvirt developers? Would patches to improve the support be accepted? Well, I guess virt-tools-list isn't the list to discuss that, so I've also CC'd libvirt-list. An explanation of why I want to use remote connections to QEMU session mode is in the description of this bug: https://bugzilla.redhat.com/show_bug.cgi?id=1241311 Thanks, Spencer Baugh

1 0