- Devel - libvirt List Archives

[libvirt] Current cpu and memory usage for Host and Domains
by anuj rampal 19 Jan '10

19 Jan '10

Hi, Is there a way to get the Current CPU and Memory usage of the Host machine and the Guest Domains..?? Regards Anuj

4 5

[libvirt] virsh list after libvirt restart "forgets" existing guests (0.7.5)
by Thomas Sjolshagen 19 Jan '10

19 Jan '10

After the upgrade to 0.7.5 from the preview repository, I've noticed that a restart of libvirt does not appear to "re-acquire" any guests already running on the system. # virsh list Compiled against library: libvir 0.7.5 Using library: libvir 0.7.5 Using API: QEMU 0.7.5 Running hypervisor: QEMU 0.12.1 # ps -ef | grep qemu-kvm qemu 1445 1 0 Jan17 ? 00:03:32 /usr/bin/qemu-kvm -S -M pc-0.11 -enable-kvm -m 512 -smp 1 -name imap [SNIP] qemu 1516 1 0 Jan17 ? 00:04:11 /usr/bin/qemu-kvm -S -M pc-0.11 -cpu qemu32 -enable-kvm -m 512 -smp 1 -name mail [SNIP] qemu 21409 1 16 Jan16 ? 07:29:32 /usr/bin/qemu-kvm -S -M pc-0.11 -enable-kvm -m 1640 -smp 1 -name imap2-cluster [SNIP] # virsh list Id Name State ---------------------------------- 1 imap running 2 mail running Additionally, something doesn't appear to be getting either parsed right or set correctly when the guest is started as the /var/log/libvirt/qemu/<guestname>.log file contains: "Option 'ipv4: Use 'on' or 'off'" "Failed to parse "yes" for "dummy.ipv4"

2 1

[libvirt] [PATCH] xen_driver: don't leak a parsed-config buffer
by Jim Meyering 19 Jan '10

19 Jan '10

Another leak: >From 10a428a2ed1ddbe4da2b4fbd7690da5a0fe0420b Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 18 Jan 2010 18:02:17 +0100 Subject: [PATCH] xen_driver: don't leak a parsed-config buffer * src/xen/xen_driver.c (xenUnifiedDomainXMLFromNative): Also free "conf" before returning. --- src/xen/xen_driver.c | 4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/src/xen/xen_driver.c b/src/xen/xen_driver.c index 4911c9e..72f56ae 100644 --- a/src/xen/xen_driver.c +++ b/src/xen/xen_driver.c @@ -1,7 +1,7 @@ /* * xen_driver.c: Unified Xen driver. * - * Copyright (C) 2007, 2008, 2009 Red Hat, Inc. + * Copyright (C) 2007-2010 Red Hat, Inc. * * See COPYING.LIB for the License of this software * @@ -1199,6 +1199,8 @@ xenUnifiedDomainXMLFromNative(virConnectPtr conn, cleanup: virDomainDefFree(def); + if (conf) + virConfFree(conf); return ret; } -- 1.6.6.638.g2bc54

2 1

[libvirt] [PATCH] qemu_driver: don't leak a virDomainDeviceDef buffer
by Jim Meyering 19 Jan '10

19 Jan '10

At first I was going to call virDomainDeviceDefFree only "if (dev)", but saw that it handles a NULL "dev" just fine, so it's better to skip the test altogether, just as we do for many other free-like functions. >From ea8511d709492f5cdc152a1eaccbccd05f036648 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 18 Jan 2010 16:55:36 +0100 Subject: [PATCH] qemu_driver: don't leak a virDomainDeviceDef buffer * src/qemu/qemu_driver.c (qemudDomainAttachDevice): Don't leak "dev". --- src/qemu/qemu_driver.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 365921f..1aa8af6 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -6062,11 +6062,11 @@ cleanup: if (cgroup) virCgroupFree(&cgroup); - if (ret < 0 && dev != NULL) { + if (ret < 0 && dev != NULL) if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 1) < 0) VIR_WARN0("Fail to restore disk device ownership"); - virDomainDeviceDefFree(dev); - } + virDomainDeviceDefFree(dev); + if (vm) virDomainObjUnlock(vm); qemuDriverUnlock(driver); -- 1.6.6.638.g2bc54

3 4

[libvirt] 0.7.5 errors in configure script
by Avi Weit 19 Jan '10

19 Jan '10

Hi, I am using Ubuntu 9.10 and downloaded and extracted libvirt.0.7.5. I have issued: ./configure --prefix=$HOME/usr --without-xen but I get: You must install device-mapper-devel/libdevmapper >= 1.0.0 to compile libvirt Below is a snippet of the final configure output: checking for showmount... no checking for pvcreate... no checking for vgcreate... no checking for lvcreate... no checking for pvremove... no checking for vgremove... no checking for lvremove... no checking for vgchange... no checking for vgscan... no checking for pvs... no checking for vgs... no checking for lvs... no checking for iscsiadm... no checking for DEVMAPPER... no checking libdevmapper.h usability... no checking libdevmapper.h presence... no checking for libdevmapper.h... no checking for dm_task_run in -ldevmapper... no configure: error: You must install device-mapper-devel/libdevmapper >= 1.0.0 to compile libvirt /dev/mapper exists on the system. Any help is appreciated. Thanks, - Avi

4 4

[libvirt] [PATCH 00/34] Add device addressing and disk controller support
by Daniel P. Berrange 18 Jan '10

18 Jan '10

This series is a merge of two previous series I posted http://www.redhat.com/archives/libvir-list/2009-December/msg00232.html http://www.redhat.com/archives/libvir-list/2009-December/msg00392.html It accomplishes quite a lot of things, having major impact on the QEMU driver, hopefully all in a postive way :-) In particular it does * Add standard XML syntax for addressing of PCI devices, and disk drives * Add support for disk controllers as a managed device in XML * Add support for disk controller hotplug/unplug * Convert everything over to use QEMU's -device flag where available * Add PCI addressing when using -device * Introduce a way to give every device a unique 'alias' name in the XML format I can't promise it works perfectly, but i've done quite alot of positive testing with it now & believe all the back comptability stuff is working right.

4 85

[libvirt] [PATCH] storage_backend_fs.c: do not ignore probe failure
by Jim Meyering 18 Jan '10

18 Jan '10

When virStorageBackendProbeTarget fails, it returns -1 or -2. The two uses below obviously intended to handle those cases differently, but due to a wrong comparison, they always treated a "real" (e.g., open) failure (-1) just like an ignorable "wrong file type" failure (-2). FYI, coverity reported that the "goto cleanup" statement was unreachable, which was true in each case, but hardly the real problem. >From c9ba79daf124afbab351741192bbaf7110a199dd Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 18 Jan 2010 10:34:53 +0100 Subject: [PATCH] storage_backend_fs.c: do not ignore probe failure * src/storage/storage_backend_fs.c (virStorageBackendFileSystemRefresh): Correct two error-handling bugs. The documented intent is to ignore non-regular files, yet due to a comparison "< 0" (rather than "!= 0") all errors were handled that way. (vboxStorageVolDelete): Likewise. (vboxDomainDefineXML): Likewise. --- src/storage/storage_backend_fs.c | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/storage/storage_backend_fs.c b/src/storage/storage_backend_fs.c index 4fe40b3..fa72f9c 100644 --- a/src/storage/storage_backend_fs.c +++ b/src/storage/storage_backend_fs.c @@ -1,7 +1,7 @@ /* * storage_backend_fs.c: storage backend for FS and directory handling * - * Copyright (C) 2007-2009 Red Hat, Inc. + * Copyright (C) 2007-2010 Red Hat, Inc. * Copyright (C) 2007-2008 Daniel P. Berrange * * This library is free software; you can redistribute it and/or @@ -562,7 +562,7 @@ virStorageBackendFileSystemRefresh(virConnectPtr conn, &backingStore, &vol->allocation, &vol->capacity, - &vol->target.encryption) < 0)) { + &vol->target.encryption) != 0)) { if (ret == -1) goto cleanup; else { @@ -603,7 +603,7 @@ virStorageBackendFileSystemRefresh(virConnectPtr conn, if ((ret = virStorageBackendProbeTarget(conn, &vol->backingStore, NULL, NULL, NULL, - NULL)) < 0) { + NULL)) != 0) { if (ret == -1) goto cleanup; else { -- 1.6.6.638.g2bc54

2 3

[libvirt] [PATCH] remove useless comparisons in lxc_driver, vbox_tmpl.c
by Jim Meyering 18 Jan '10

18 Jan '10

Here are two change sets. They remove useless comparisons of the style - if (def->nets[i]->mac) { where the mac member is declared as an array. >From d22b04d2684d5a16f797d8df0c2ef036b5b1b2ce Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 18 Jan 2010 09:49:57 +0100 Subject: [PATCH 1/2] lxc_driver: remove useless comparison * src/lxc/lxc_driver.c (lxcSetupInterfaces): Remove always-true array-is-non-NULL test. git grep 'mac\[.*\];'|grep -F .h src/conf/domain_conf.h: unsigned char mac[VIR_MAC_BUFLEN]; --- src/lxc/lxc_driver.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c index 86606c7..94ec874 100644 --- a/src/lxc/lxc_driver.c +++ b/src/lxc/lxc_driver.c @@ -834,7 +834,7 @@ static int lxcSetupInterfaces(virConnectPtr conn, goto error_exit; } - if (def->nets[i]->mac) { + { char macaddr[VIR_MAC_STRING_BUFLEN]; virFormatMacAddr(def->nets[i]->mac, macaddr); if (0 != (rc = setMacAddr(containerVeth, macaddr))) { -- 1.6.6.638.g2bc54 >From c751c0cf3352b08d674bb59efd987f106fc1fffb Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 18 Jan 2010 09:58:57 +0100 Subject: [PATCH 2/2] vbox_tmpl.c: remove useless array-is-non-NULL comparisons * src/vbox/vbox_tmpl.c (vboxStorageVolDelete): Remove always-true array-is-non-NULL test. git grep 'key\[.*\];'|grep -F .h src/datatypes.h: char key[PATH_MAX]; (vboxStorageVolGetInfo): Likewise. (vboxStorageVolGetXMLDesc): Likewise. (vboxStorageVolGetPath): Likewise. (vboxDomainDefineXML): Likewise. (but now with "mac[]") --- src/vbox/vbox_tmpl.c | 16 ++-------------- 1 files changed, 2 insertions(+), 14 deletions(-) diff --git a/src/vbox/vbox_tmpl.c b/src/vbox/vbox_tmpl.c index 07696c0..e40c848 100644 --- a/src/vbox/vbox_tmpl.c +++ b/src/vbox/vbox_tmpl.c @@ -4084,9 +4084,7 @@ static virDomainPtr vboxDomainDefineXML(virConnectPtr conn, const char *xml) { } VBOX_UTF8_TO_UTF16(macaddrvbox, &MACAddress); - if (def->nets[i]->mac) { - adapter->vtbl->SetMACAddress(adapter, MACAddress); - } + adapter->vtbl->SetMACAddress(adapter, MACAddress); VBOX_UTF16_FREE(MACAddress); } } @@ -6645,9 +6643,6 @@ static int vboxStorageVolDelete(virStorageVolPtr vol, int i = 0; int j = 0; - if (!vol->key) - return ret; - vboxUtf8toIID(vol->conn, vol->key, &hddIID); if (!hddIID) return ret; @@ -6774,8 +6769,7 @@ static int vboxStorageVolGetInfo(virStorageVolPtr vol, virStorageVolInfoPtr info vboxIID *hddIID = NULL; nsresult rc; - if ( !vol->key - || !info) + if (!info) return ret; vboxUtf8toIID(vol->conn, vol->key, &hddIID); @@ -6824,9 +6818,6 @@ static char *vboxStorageVolGetXMLDesc(virStorageVolPtr vol, unsigned int flags A int defOk = 0; nsresult rc; - if (!vol->key) - return ret; - memset(&pool, 0, sizeof(pool)); memset(&def, 0, sizeof(def)); @@ -6919,9 +6910,6 @@ static char *vboxStorageVolGetPath(virStorageVolPtr vol) { vboxIID *hddIID = NULL; nsresult rc; - if (!vol->key) - return ret; - vboxUtf8toIID(vol->conn, vol->key, &hddIID); if (!hddIID) return ret; -- 1.6.6.638.g2bc54

2 1

[libvirt] [PATCH] secret_driver.c: remove dead cleanup code
by Jim Meyering 18 Jan '10

18 Jan '10

Here's another example of appropriate use of assert. The while loop removed below is currently guaranteed never to be executed, since "list" is always NULL at that point. However, you can argue that simply removing the loop is a little risky: what if someone changes things to "goto cleanup" before "list" reaches the final NULL? That's why I added the assertion: to catch the potential (albeit unlikely) future coding error. It would be a shame to include "real" run-time code to detect a situation like that that will probably never arise. And if you do, technically you'd have to document that the function would return some sort of error code if a developer changes its guts in a way that violates this invariant. That would be nonsensical. >From 6baa2b0b8f9963fa60ff0946afcefd57586e2720 Mon Sep 17 00:00:00 2001 From: Jim Meyering <meyering(a)redhat.com> Date: Mon, 18 Jan 2010 10:46:57 +0100 Subject: [PATCH] secret_driver.c: remove dead cleanup code * src/secret/secret_driver.c (loadSecrets): Remove dead code after "cleanup:". At that point, "list" is known to be NULL. Instead, add an assertion. Include <assert.h>. --- src/secret/secret_driver.c | 10 +++------- 1 files changed, 3 insertions(+), 7 deletions(-) diff --git a/src/secret/secret_driver.c b/src/secret/secret_driver.c index 1eef468..b4b4f8a 100644 --- a/src/secret/secret_driver.c +++ b/src/secret/secret_driver.c @@ -1,7 +1,7 @@ /* * secret_driver.c: local driver for secret manipulation API * - * Copyright (C) 2009 Red Hat, Inc. + * Copyright (C) 2009-2010 Red Hat, Inc. * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public @@ -28,6 +28,7 @@ #include <string.h> #include <sys/stat.h> #include <unistd.h> +#include <assert.h> #include "internal.h" #include "base64.h" @@ -518,12 +519,7 @@ loadSecrets(virConnectPtr conn, virSecretDriverStatePtr driver, ret = 0; cleanup: - while (list != NULL) { - virSecretEntryPtr s; - - s = listUnlink(&list); - secretFree(s); - } + assert (list == NULL); if (dir != NULL) closedir(dir); return ret; -- 1.6.6.638.g2bc54

2 1

[libvirt] [PATCH] Use closest CPU model when decoding from CPUID
by Jiri Denemark 18 Jan '10

18 Jan '10

Current implementation of x86Decode() used for CPUID -> model+features translation does not always select the closest CPU model. When walking through all models from cpu_map.xml the function considers a new candidate as a better choice than a previously selected candidate only if the new one is a superset of the old one. In case the new candidate is closer to host CPU but lacks some feature comparing to the old candidate, the function does not choose well. This patch changes the algorithm so that the closest model is always selected. That is, the model which requires the lowest number of additional features to describe host CPU. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/cpu/cpu_x86.c | 122 ++++++++++++++++++++++++++++++++++------------------- 1 files changed, 78 insertions(+), 44 deletions(-) diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c index 298b632..dae7c90 100644 --- a/src/cpu/cpu_x86.c +++ b/src/cpu/cpu_x86.c @@ -238,6 +238,55 @@ x86DataFromModel(const struct x86_model *model) } +static virCPUDefPtr +x86DataToCPU(const union cpuData *data, + const struct x86_model *model, + const struct x86_map *map) +{ + virCPUDefPtr cpu; + union cpuData *tmp = NULL; + struct cpuX86cpuid *cpuid; + const struct x86_feature *feature; + int i; + + if (VIR_ALLOC(cpu) < 0 || + (cpu->model = strdup(model->name)) == NULL || + (tmp = x86DataCopy(data)) == NULL) + goto no_memory; + + for (i = 0; i < model->ncpuid; i++) { + x86cpuidClearBits(x86DataCpuid(tmp, model->cpuid[i].function), + model->cpuid + i); + } + + feature = map->features; + while (feature != NULL) { + for (i = 0; i < feature->ncpuid; i++) { + if ((cpuid = x86DataCpuid(tmp, feature->cpuid[i].function)) + && x86cpuidMatchMasked(cpuid, feature->cpuid + i)) { + x86cpuidClearBits(cpuid, feature->cpuid + i); + if (virCPUDefAddFeature(NULL, cpu, feature->name, + VIR_CPU_FEATURE_REQUIRE) < 0) + goto error; + } + } + + feature = feature->next; + } + +cleanup: + x86DataFree(tmp); + return cpu; + +no_memory: + virReportOOMError(NULL); +error: + virCPUDefFree(cpu); + cpu = NULL; + goto cleanup; +} + + static void x86FeatureFree(struct x86_feature *feature) { @@ -896,10 +945,9 @@ x86Decode(virCPUDefPtr cpu, { int ret = -1; struct x86_map *map; - const struct x86_feature *feature; - const struct x86_model *model = NULL; const struct x86_model *candidate; - union cpuData *tmp = NULL; + virCPUDefPtr cpuCandidate; + virCPUDefPtr cpuModel = NULL; struct cpuX86cpuid *cpuid; int i; @@ -908,6 +956,8 @@ x86Decode(virCPUDefPtr cpu, candidate = map->models; while (candidate != NULL) { + bool allowed = (models == NULL); + for (i = 0; i < candidate->ncpuid; i++) { cpuid = x86DataCpuid(data, candidate->cpuid[i].function); if (cpuid == NULL @@ -915,65 +965,49 @@ x86Decode(virCPUDefPtr cpu, goto next; } - if (model == NULL - || x86ModelCompare(model, candidate) == SUBSET) { - bool found = false; - for (i = 0; i < nmodels; i++) { - if (STREQ(models[i], candidate->name)) { - found = true; - break; - } + for (i = 0; i < nmodels; i++) { + if (STREQ(models[i], candidate->name)) { + allowed = true; + break; } + } - if (nmodels > 0 && !found) { - VIR_DEBUG("CPU model %s not allowed by hypervisor; ignoring", - candidate->name); - } - else - model = candidate; + if (!allowed) { + VIR_DEBUG("CPU model %s not allowed by hypervisor; ignoring", + candidate->name); + goto next; } + if (!(cpuCandidate = x86DataToCPU(data, candidate, map))) + goto out; + + if (cpuModel == NULL + || cpuModel->nfeatures > cpuCandidate->nfeatures) { + virCPUDefFree(cpuModel); + cpuModel = cpuCandidate; + } else + virCPUDefFree(cpuCandidate); + next: candidate = candidate->next; } - if (model == NULL) { + if (cpuModel == NULL) { virCPUReportError(NULL, VIR_ERR_INTERNAL_ERROR, "%s", _("Cannot find suitable CPU model for given data")); goto out; } - if ((cpu->model = strdup(model->name)) == NULL - || (tmp = x86DataCopy(data)) == NULL) { - virReportOOMError(NULL); - goto out; - } - - for (i = 0; i < model->ncpuid; i++) { - x86cpuidClearBits(x86DataCpuid(tmp, model->cpuid[i].function), - model->cpuid + i); - } - - feature = map->features; - while (feature != NULL) { - for (i = 0; i < feature->ncpuid; i++) { - if ((cpuid = x86DataCpuid(tmp, feature->cpuid[i].function)) - && x86cpuidMatchMasked(cpuid, feature->cpuid + i)) { - x86cpuidClearBits(cpuid, feature->cpuid + i); - if (virCPUDefAddFeature(NULL, cpu, feature->name, - VIR_CPU_FEATURE_REQUIRE) < 0) - goto out; - } - } - - feature = feature->next; - } + cpu->model = cpuModel->model; + cpu->nfeatures = cpuModel->nfeatures; + cpu->features = cpuModel->features; + VIR_FREE(cpuModel); ret = 0; out: - x86DataFree(tmp); x86MapFree(map); + virCPUDefFree(cpuModel); return ret; } -- 1.6.6

3 2