- Devel - libvirt List Archives

[libvirt] [sandbox][PATCH] Add module directory prefix selection feature
by Radu Caragea 20 Jun '12

20 Jun '12

Add module directory prefix selection feature This is useful when running as a non-privileged user if we want to boot a custom compiled kernel: we might not have rights to install in /lib/modules/<kernel release> so when compiling the kernel we can use "make modules_install INSTALL_MOD_PATH=/path" which installs in /path/lib/modules/<kernel release>. By setting with gvir_sandbox_config_set_moddirprefix(cfg, "/path") we can now achieve just that.

2 5

[libvirt] [qpid PATCH] doc: fixed search form action
by Martin Kletzander 20 Jun '12

20 Jun '12

Search was giving 404 because of missing {$href_base} in "action" property. Also one whitespace character got removed from end of the line because my before-save-hook. --- Pushed under the 'trivial' rule. doc/architecture.html | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/doc/architecture.html b/doc/architecture.html index a564fa8..5b0f2bd 100644 --- a/doc/architecture.html +++ b/doc/architecture.html @@ -1,12 +1,12 @@ <?xml version="1.0" encoding="ISO-8859-1"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> -<html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1" /><link rel="stylesheet" type="text/css" href="http://libvirt.org/qpid/main.css" /><link rel="SHORTCUT ICON" href="/32favicon.png" /><title>Architecture</title><meta name="description" content="libvirt, virtualization, virtualization API, qpid" /></head><body><div id="header"><div id="headerLogo"></div><div id="headerSearch"><form action="search.php" enctype="application/x-www-form-urlencoded" method="get"><div><input id="query" name="query" type="text" size="12" value="" /><input id="submit" name="submit" type="submit" value="Search" /></div></form></div></div><div id="body"><div id="menu"><ul class="l0"><li><div><a href="index.html" class="inactive">Home</a></div></li><li><a href="http://libvirt.org/" class="inactive">libvirt</a></li><li><div><a href="releases.html" class="inactive">Releases</a></div></li><li><div><a href="architecture.html" class="inac! tive">Architecture</a></div></li><li><a href="https://www.redhat.com/mailman/listinfo/libvir-list" class="inactive">Mailing list</a></li></ul></div><div id="content"><h1 class="style1">Architecture</h1><p> +<html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1" /><link rel="stylesheet" type="text/css" href="http://libvirt.org/qpid/main.css" /><link rel="SHORTCUT ICON" href="/32favicon.png" /><title>Architecture</title><meta name="description" content="libvirt, virtualization, virtualization API, qpid" /></head><body><div id="header"><div id="headerLogo"></div><div id="headerSearch"><form action="{$href_base}search.php" enctype="application/x-www-form-urlencoded" method="get"><div><input id="query" name="query" type="text" size="12" value="" /><input id="submit" name="submit" type="submit" value="Search" /></div></form></div></div><div id="body"><div id="menu"><ul class="l0"><li><div><a href="index.html" class="inactive">Home</a></div></li><li><a href="http://libvirt.org/" class="inactive">libvirt</a></li><li><div><a href="releases.html" class="inactive">Releases</a></div></li><li><div><a href="architecture.html"! class="inactive">Architecture</a></div></li><li><a href="https://www.redhat.com/mailman/listinfo/libvir-list" class="inactive">Mailing list</a></li></ul></div><div id="content"><h1 class="style1">Architecture</h1><p> libvirt-qpid is an agent that runs on a given host. It will connect to both the local libvirtd and to a qpid broker. It then queries libvirtd every 5 seconds and maintains objects of various classes that reflect the activity and configuration of libvirtd. </p><p> - The classes include the node (the host), domains, storage pools, + The classes include the node (the host), domains, storage pools, and storage volumes. A complete synchronization takes place each iteration so that any state changes that happen outside of the channel provided by libvirt-qpid will be reflected in the objects -- 1.7.8.6

1 0

[libvirt] NEED HELP on libvirt guest resource control
by Zhimou Peng 20 Jun '12

20 Jun '12

Hi all, I'm from libvirt Section 3 I'm tracking "guest resource control" test from kvm, and i found kvm support more than libvirt. they support these: https://riddler.bne.redhat.com/IEEE_Test_Plans/en-US/RHEL/6.3/html-single/K… guest blkio bandwidth and blkio throttle limitation, guest memory limitation, migrate guest from one memory to another, guest CPU bandwidth limitation, guest freezing, pin guest to certain host CPU, guest net limitation. I want to know what the current libvirt support status and if libvirt will support all these feature in the future? BR zhpeng

1 0

[libvirt] [PATCH] conf: Format numatune XML correctly while placement is none
by Osier Yang 20 Jun '12

20 Jun '12

setNumaParameters tunes the numa setting using cgroup, it's another entry except libnuma/numad for numa tuning. And it doesn't set the placement, and further more, the formating codes doesn't take this into consideration. How to reproduce: conn = libvirt.open(None) dom = conn.lookupByName('linux') param = {'numa_nodeset': '0', 'numa_mode': 1} dom.setNumaParameters(param, 2) % virsh start linux error: Failed to start domain rhel6.3rc error: (domain_definition):8: error parsing attribute name <memory mode='preferred' </numatune> -------------------------------^ --- By the way, I see problems of setNumaParameters too. conn = libvirt.open(None) dom = conn.lookupByName('linux') param = {'numa_mode': 1} dom.setNumaParameters(param, 2) The numa 'mode' will be just ignored, and no 'numatune' XML is formated, as neither 'nodeset' nor 'placement' is specified. I'd think it's right to ignore it when formating, it's meaningless to only specify the 'mode'. However, we might have to fix setNumaParameters to prevent setting the numa mode without nodeset, and error out, as it's really a bad user experience to see the API call succeeded, but the expected XML doesn't show up in the end. --- src/conf/domain_conf.c | 17 ++++++++++------- 1 files changed, 10 insertions(+), 7 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 81c6308..c44d89d 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -12795,23 +12795,26 @@ virDomainDefFormatInternal(virDomainDefPtr def, const char *placement; mode = virDomainNumatuneMemModeTypeToString(def->numatune.memory.mode); - virBufferAsprintf(buf, " <memory mode='%s' ", mode); + virBufferAsprintf(buf, " <memory mode='%s'", mode); - if (def->numatune.memory.placement_mode == - VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_STATIC) { + if (def->numatune.memory.nodemask) { nodemask = virDomainCpuSetFormat(def->numatune.memory.nodemask, - VIR_DOMAIN_CPUMASK_LEN); + VIR_DOMAIN_CPUMASK_LEN); if (nodemask == NULL) { virDomainReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("failed to format nodeset for " "NUMA memory tuning")); goto cleanup; } - virBufferAsprintf(buf, "nodeset='%s'/>\n", nodemask); + virBufferAsprintf(buf, " nodeset='%s'/>\n", nodemask); VIR_FREE(nodemask); - } else if (def->numatune.memory.placement_mode) { + } else if (def->numatune.memory.placement_mode == + VIR_DOMAIN_NUMATUNE_MEM_PLACEMENT_MODE_AUTO) { placement = virDomainNumatuneMemPlacementModeTypeToString(def->numatune.memory.placement_mode); - virBufferAsprintf(buf, "placement='%s'/>\n", placement); + virBufferAsprintf(buf, " placement='%s'/>\n", placement); + } else { + /* Should not hit here. */ + virBufferAddLit(buf, "/>\n"); } virBufferAddLit(buf, " </numatune>\n"); } -- 1.7.7.3

2 1

[libvirt] [PATCH] openvz: Fix wordsize on 64 bit architectures
by Guido Günther 20 Jun '12

20 Jun '12

The word size there is 64 bit not 8. --- Came across this while browsing the source. O.k. to apply? -- Guido src/openvz/openvz_conf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/openvz/openvz_conf.c b/src/openvz/openvz_conf.c index ad4ed74..007f9fe 100644 --- a/src/openvz/openvz_conf.c +++ b/src/openvz/openvz_conf.c @@ -195,7 +195,7 @@ virCapsPtr openvzCapsInit(void) if ((guest = virCapabilitiesAddGuest(caps, "exe", utsname.machine, - sizeof(int) == 4 ? 32 : 8, + sizeof(int) == 4 ? 32 : 64, NULL, NULL, 0, -- 1.7.10.4

3 7

[libvirt] [PATCH 1/2] Fix default USB controller for ppc64
by Dipankar Sarma 20 Jun '12

20 Jun '12

From: Dipankar Sarma <dipankar(a)in.ibm.com> Fix the default usb controller for pseries systems if none specified. Signed-off-by: Dipankar Sarma <dipankar(a)in.ibm.com> --- src/qemu/qemu_command.c | 13 +++++++++---- 1 files changed, 9 insertions(+), 4 deletions(-) diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index a34c707..bd4f96a 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -2605,7 +2605,8 @@ qemuControllerModelUSBToCaps(int model) static int -qemuBuildUSBControllerDevStr(virDomainControllerDefPtr def, +qemuBuildUSBControllerDevStr(virDomainDefPtr domainDef, + virDomainControllerDefPtr def, virBitmapPtr qemuCaps, virBuffer *buf) { @@ -2614,8 +2615,12 @@ qemuBuildUSBControllerDevStr(virDomainControllerDefPtr def, model = def->model; - if (model == -1) - model = VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI; + if (model == -1) { + if (STREQ(domainDef->os.arch, "ppc64")) + model = VIR_DOMAIN_CONTROLLER_MODEL_USB_PCI_OHCI; + else + model = VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI; + } smodel = qemuControllerModelUSBTypeToString(model); caps = qemuControllerModelUSBToCaps(model); @@ -2701,7 +2706,7 @@ qemuBuildControllerDevStr(virDomainDefPtr domainDef, break; case VIR_DOMAIN_CONTROLLER_TYPE_USB: - if (qemuBuildUSBControllerDevStr(def, qemuCaps, &buf) == -1) + if (qemuBuildUSBControllerDevStr(domainDef, def, qemuCaps, &buf) == -1) goto error; if (nusbcontroller)

2 4

[libvirt] Fwd: libvirt secret support password or encryption keys?
by Zhimou Peng 20 Jun '12

20 Jun '12

FWD Hi, all PLS help, i'm reviewing man doc of libvirt secrets # man virsh ... SECRET COMMMANDS The following commands manipulate "secrets" (e.g. passwords, passphrases -----> secret-set-value only support base64 "passphrases" and encryption keys). Libvirt can store secrets independently from their use, and other objects (e.g. volumes or domains) can refer to the secrets for encryption or possibly other uses. Secrets are identified using an UUID. See <http://libvirt.org/formatsecret.html> for documentation of the XML format used to represent properties of secrets. ... So, I want to know that if libvirt support other 2 kinds of "secrets". And if so, How to use password and encrytion keys ? If libvirt not support that, i will file a bug of this man doc. BR zhpeng ----- Forwarded Message ----- From: "Zhimou Peng" <zhpeng(a)redhat.com> To: "Daniel Berrange" <berrange(a)redhat.com> Sent: Thursday, June 14, 2012 1:38:09 PM Subject: Fwd: libvirt secret support password or encryption keys? Hi, Daniel I saw your blog:http://berrange.com/posts/2009/12/02/using-qcow2-disk-encryption-with-… and there is only "type=passphrases", so how can i use password or encrytion keys? zhpeng BR ----- Forwarded Message ----- From: "Zhimou Peng" <zhpeng(a)redhat.com> To: "Daniel Berrange" <berrange(a)redhat.com> Cc: "s3-bug-review" <s3-bug-review(a)redhat.com> Sent: Tuesday, June 12, 2012 2:03:05 PM Subject: libvirt secret support password or encryption keys? Hi all, # man virsh ... SECRET COMMMANDS The following commands manipulate "secrets" (e.g. passwords, passphrases -----> secret-set-value only support base64 "passphrases" and encryption keys). Libvirt can store secrets independently from their use, and other objects (e.g. volumes or domains) can refer to the secrets for encryption or possibly other uses. Secrets are identified using an UUID. See <http://libvirt.org/formatsecret.html> for documentation of the XML format used to represent properties of secrets. ... So, I want to know that if libvirt support other 2 kinds of "secrets". And if so, How to use password and encrytion keys ? If libvirt not support that, i will file a bug of this man doc. zhpeng BR

2 2

[libvirt] [PATCHv2 00/13] list all snapshots
by Eric Blake 19 Jun '12

19 Jun '12

This is a v2 of the remaining patches in: https://www.redhat.com/archives/libvir-list/2012-June/msg00273.html https://www.redhat.com/archives/libvir-list/2012-June/msg00284.html It assumes that Peter's list all domain patches have been applied already, at least through patch 6/12: https://www.redhat.com/archives/libvir-list/2012-June/msg00316.html In other words, this series will not apply to current libvirt.git at the time of this mail. And since I developed against v3 of Peter's patches and he has some tweaks to make, I will probably have to make minor tweaks to rebase on top of his eventual v4. I will follow up with a link to a git repo with Peter's v3 and my series, and/or my rebase on top of his v4. Overall changes since v1: actually tested this time address review comments lots of bug fixes due to corner case testing rearrange the patch series so that metaroot comes before filtering pick up some good ideas from Peter's v3 More details in some of the individual patches, if I can remember everything I did. In particular, I'm quite pleased that the overall diffstat for domain_conf.c reduced in size in spite of adding features :) Eric Blake (13): snapshot: new virsh function factored from snapshot-list snapshot: use new virsh function for snapshot-list snapshot: use metaroot node to simplify management snapshot: merge domain and snapshot computation snapshot: merge count and name collectoin snapshot: add additional filters when getting lists snapshot: expose new flags in virsh list: add virDomainListAllSnapshots API list: use the new snapshot API in virsh when possible list: provide python bindings for snapshots list: provide RPC call for snapshots list: new helper function to collect snapshots list: add qemu snapshot list support daemon/remote.c | 126 ++++++ include/libvirt/libvirt.h.in | 35 +- python/Makefile.am | 2 + python/generator.py | 2 + python/libvirt-override-api.xml | 16 +- python/libvirt-override-virDomain.py | 11 + python/libvirt-override-virDomainSnapshot.py | 11 + python/libvirt-override.c | 92 +++++ src/conf/domain_conf.c | 250 ++++------- src/conf/domain_conf.h | 19 +- src/conf/virdomainlist.c | 44 +- src/conf/virdomainlist.h | 18 + src/driver.h | 12 + src/libvirt.c | 275 +++++++++++-- src/libvirt_private.syms | 3 +- src/libvirt_public.syms | 2 + src/qemu/qemu_driver.c | 137 +++++-- src/qemu/qemu_migration.c | 3 +- src/remote/remote_driver.c | 126 ++++++ src/remote/remote_protocol.x | 26 +- src/remote_protocol-structs | 26 ++ tools/virsh.c | 570 +++++++++++++++++--------- tools/virsh.pod | 14 +- 23 files changed, 1371 insertions(+), 449 deletions(-) create mode 100644 python/libvirt-override-virDomain.py create mode 100644 python/libvirt-override-virDomainSnapshot.py -- 1.7.10.2

2 37

[libvirt] [PATCH] domain_conf: fix possible memory leak
by Martin Kletzander 19 Jun '12

19 Jun '12

Until now, it was possible to crash libvirtd when defining domain with channel device with missing source element. When creating new virDomainChrDef, target.port is set to -1, but unfortunately it is an union with addresses that virDomainChrDefFree tries to free in case the deviceType is channel. Having the port set to -1 is intended, however the cleanest way to get around the problems with the crash seems to be renumbering the VIR_DOMAIN_CHR_CHANNEL_ target types to cover new NONE type (with value 0) being the default (no target type yet). --- src/conf/domain_conf.c | 1 + src/conf/domain_conf.h | 3 ++- 2 files changed, 3 insertions(+), 1 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 7bd61a5..41f22d2 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -307,6 +307,7 @@ VIR_ENUM_IMPL(virDomainNetInterfaceLinkState, VIR_DOMAIN_NET_INTERFACE_LINK_STAT VIR_ENUM_IMPL(virDomainChrChannelTarget, VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_LAST, + "none", "guestfwd", "virtio") diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index d627ad8..531176c 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -855,7 +855,8 @@ enum virDomainChrDeviceType { }; enum virDomainChrChannelTargetType { - VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_GUESTFWD = 0, + VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_NONE = 0, + VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_GUESTFWD, VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO, VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_LAST, -- 1.7.8.6

2 2

[libvirt] [PATCH] virsh: fix few typos on desc command
by Martin Kletzander 19 Jun '12

19 Jun '12

virsh help fix: - <--title> can also /get/ the title virsh man page: - missing <domain-id> - <new_desc> should be <new-desc> --- tools/virsh.c | 2 +- tools/virsh.pod | 7 ++++--- 2 files changed, 5 insertions(+), 4 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 4d34d49..036e018 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -1227,7 +1227,7 @@ static const vshCmdOptDef opts_desc[] = { {"live", VSH_OT_BOOL, 0, N_("modify/get running state")}, {"config", VSH_OT_BOOL, 0, N_("modify/get persistent configuration")}, {"current", VSH_OT_BOOL, 0, N_("modify/get current state configuration")}, - {"title", VSH_OT_BOOL, 0, N_("modify the title instead of description")}, + {"title", VSH_OT_BOOL, 0, N_("modify/get the title instead of description")}, {"edit", VSH_OT_BOOL, 0, N_("open an editor to modify the description")}, {"new-desc", VSH_OT_ARGV, 0, N_("message")}, {NULL, 0, 0, NULL} diff --git a/tools/virsh.pod b/tools/virsh.pod index 910a187..252512a 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -497,8 +497,9 @@ Define a domain from an XML <file>. The domain definition is registered but not started. If domain is already running, the changes will take effect on the next boot. -=item B<desc> [[I<--live>] [I<--config>] | [I<--current>]] [I<--title>] - [I<--edit>] [I<--new-desc> New description or title message] +=item B<desc> I<domain-id> [[I<--live>] [I<--config>] | + [I<--current>]] [I<--title>] [I<--edit>] [I<--new-desc> + New description or title message] Show or modify description and title of a domain. These values are user fields that allow to store arbitrary textual data to allow easy @@ -516,7 +517,7 @@ description or title should be opened and the contents saved back afterwards. Flag I<--title> selects operation on the title field instead of description. -If neither of I<--edit> and I<--new_desc> are specified the note or description +If neither of I<--edit> and I<--new-desc> are specified the note or description is displayed instead of being modified. =item B<destroy> I<domain-id> [I<--graceful>] -- 1.7.8.6

2 2