July 2011 - Devel - libvirt List Archives

[libvirt] virNetClientPtr leak in remote driver
by Matthias Bolte 05 Aug '11

05 Aug '11

doRemoteClose doesn't free the virNetClientPtr and this creates a 260kb leak per remote connection. This happens because virNetClientFree doesn't remove the last ref, because virNetClientNew creates the virNetClientPtr with a refcount of 2. static virNetClientPtr virNetClientNew(virNetSocketPtr sock, const char *hostname) { [...] client->refs = 1; [...] /* Set up a callback to listen on the socket data */ client->refs++; if (virNetSocketAddIOCallback(client->sock, VIR_EVENT_HANDLE_READABLE, virNetClientIncomingEvent, client, virNetClientEventFree) < 0) { client->refs--; VIR_DEBUG("Failed to add event watch, disabling events"); } [...] } virNetClientNew adds a ref before calling virNetSocketAddIOCallback but only removes it when virNetSocketAddIOCallback fails. This seems wrong too me and the ref should be removed in the success case too. The same pattern was added in 0302391ee643ad91fdc6d2ecf7e4cf0fc9724516 (Fix race in ref counting when handling RPC jobs) --- a/src/rpc/virnetserverclient.c +++ b/src/rpc/virnetserverclient.c @@ -763,10 +763,12 @@ readmore: /* Send off to for normal dispatch to workers */ if (msg) { + client->refs++; if (!client->dispatchFunc || client->dispatchFunc(client, msg, client->dispatchOpaque) < 0) { virNetMessageFree(msg); client->wantClose = true; + client->refs--; return; } } Again, this seems wrong and the ref should be removed in the success case here too. Before I spent time to figure out how the refcounting is supposed to work here I just report it and hope that Dan can easily fix this. -- Matthias Bolte http://photron.blogspot.com

3 6

[libvirt] Start of freeze for libvirt-0.9.4 and availability of rc1
by Daniel Veillard 02 Aug '11

02 Aug '11

As plannned we are entering the freeze for libvirt-0.9.4 . I have made a release candidate 1 tarball (and associated rpms) at ftp://libvirt.org/libvirt/libvirt-0.9.4-rc1.tar.gz This seems to pass my minimal tests without problems, but please give it a try too and report problems, I will probably make an rc2 on Thursday or Friday, and then the final release next Monday or Tuesday, thanks ! Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/

13 39

[libvirt] ANNOUNCE: virt-manager 0.9.0 and virtinst 0.600.0 released
by Cole Robinson 02 Aug '11

02 Aug '11

I'm happy to announce two new releases: virt-manager 0.9.0: virt-manager is a desktop application for managing KVM and Xen virtual machines via libvirt. virtinst 0.600.0: virtinst is a collection of command line tools for provisioning libvirt virtual machines, including virt-install and virt-clone. The releases can be downloaded from: http://virt-manager.org/download.html The direct download links are: http://virt-manager.org/download/sources/virt-manager/virt-manager-0.9.0.ta… http://virt-manager.org/download/sources/virtinst/virtinst-0.600.0.tar.gz The virt-manager release includes: - Use a hiding toolbar for fullscreen mode - Use libguestfs to show guest packagelist and more (Richard W.M. Jones) - Basic 'New VM' wizard support for LXC guests - Remote serial console access (with latest libvirt) - Remote URL guest installs (with latest libvirt) - Add Hardware: Support <filesystem> devices - Add Hardware: Support <smartcard> devices (Marc-André Lureau) - Enable direct interface selection for qemu/kvm (Gerhard Stenzel) - Allow viewing and changing disk serial number The virtinst release includes: - virt-install: Various improvements to enable LXC/container guests: - New --filesystem option for <filesystem> devices - New --init option for container <init> path - New --container option (similar to --paravirt or --hvm) - virt-install: Make --location remotely (with latest libvirt) - virt-install: New --smartcard option for <smartcard> devices (Marc-André Lureau) - virt-install: New --numatune option for building guest <numatune> XML - virt-install: option to set --disk error_policy= - virt-install: option to set --disk serial= Thanks to everyone who has contributed to this release through testing, bug reporting, submitting patches, and otherwise sending in feedback! Thanks, Cole

5 7

[libvirt] [PATCH] freebsd: Add gnulib environ module for the commandtest
by Matthias Bolte 02 Aug '11

02 Aug '11

--- bootstrap.conf | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/bootstrap.conf b/bootstrap.conf index f006a47..3b105b1 100644 --- a/bootstrap.conf +++ b/bootstrap.conf @@ -36,6 +36,7 @@ configmake count-one-bits crypto/md5 dirname-lgpl +environ fclose fcntl-h ffs -- 1.7.4.1

3 5

[libvirt] [PATCH] Fix bug #611823 prohibit pools with duplicate storage
by Lei Li 02 Aug '11

02 Aug '11

To make sure the unique storage pool defined and created from different directory to avoid inconsistent version of volume pool created, I add two API be called by storage driver to check for the probable duplicate pools and refused the duplicate pool. virStoragePoolObjFindByPath() provide a method to find pool object by target path in pool list. virStoragePoolTargetDuplicate() implement the function to check if there is duplicate pool. Add judgement for storagePoolCreate&storagePoolDefine by calling virStoragePoolTargetDuplicate() to avoid both transient storage pool and persistent storage pool be created repeatedly in storage driver. Signed-off-by: Lei Li<lilei(a)linux.vnet.ibm.com> --- src/conf/storage_conf.c | 39 +++++++++++++++++++++++++++++++++++++++ src/conf/storage_conf.h | 4 ++++ src/libvirt_private.syms | 2 ++ src/storage/storage_driver.c | 6 ++++++ 4 files changed, 51 insertions(+), 0 deletions(-) diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c index 995f9a6..a499e82 100644 --- a/src/conf/storage_conf.c +++ b/src/conf/storage_conf.c @@ -1317,6 +1317,22 @@ virStoragePoolObjFindByName(virStoragePoolObjListPtr pools, return NULL; } +virStoragePoolObjPtr +virStoragePoolObjFindByPath(virStoragePoolObjListPtr pools, + const char *path) +{ + unsigned int i; + + for (i = 0 ; i< pools->count ; i++) { + virStoragePoolObjLock(pools->objs[i]); + if (STREQ(pools->objs[i]->def->target.path, path)) + return pools->objs[i]; + virStoragePoolObjUnlock(pools->objs[i]); + } + + return NULL; +} + void virStoragePoolObjClearVols(virStoragePoolObjPtr pool) { @@ -1707,6 +1723,29 @@ cleanup: return ret; } +int virStoragePoolTargetDuplicate(virStoragePoolObjListPtr pools, + virStoragePoolDefPtr def) +{ + int ret = 1; + virStoragePoolObjPtr pool = NULL; + + /*check pool list if defined target path already exist*/ + pool = virStoragePoolObjFindByPath(pools, def->target.path); + + if (pool) { + virStorageReportError(VIR_ERR_OPERATION_FAILED, + _("target path '%s' is already in use"), + pool->def->target.path); + ret = -1; + goto cleanup; + } + + +cleanup: + if (pool) + virStoragePoolObjUnlock(pool); + return ret; +} void virStoragePoolObjLock(virStoragePoolObjPtr obj) { diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h index 271441a..454c43d 100644 --- a/src/conf/storage_conf.h +++ b/src/conf/storage_conf.h @@ -335,6 +335,8 @@ virStoragePoolObjPtr virStoragePoolObjFindByUUID(virStoragePoolObjListPtr pools, const unsigned char *uuid); virStoragePoolObjPtr virStoragePoolObjFindByName(virStoragePoolObjListPtr pools, const char *name); +virStoragePoolObjPtr virStoragePoolObjFindByPath(virStoragePoolObjListPtr pools, + const char *path); virStorageVolDefPtr virStorageVolDefFindByKey(virStoragePoolObjPtr pool, const char *key); @@ -387,6 +389,8 @@ char *virStoragePoolSourceListFormat(virStoragePoolSourceListPtr def); int virStoragePoolObjIsDuplicate(virStoragePoolObjListPtr pools, virStoragePoolDefPtr def, unsigned int check_active); +int virStoragePoolTargetDuplicate(virStoragePoolObjListPtr pools, + virStoragePoolDefPtr def); void virStoragePoolObjLock(virStoragePoolObjPtr obj); void virStoragePoolObjUnlock(virStoragePoolObjPtr obj); diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 853ee62..ef323f5 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -930,7 +930,9 @@ virStoragePoolObjClearVols; virStoragePoolObjDeleteDef; virStoragePoolObjFindByName; virStoragePoolObjFindByUUID; +virStoragePoolObjFindByPath; virStoragePoolObjIsDuplicate; +virStoragePoolTargetDuplicate; virStoragePoolObjListFree; virStoragePoolObjLock; virStoragePoolObjRemove; diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c index 9c353e3..8ee63f6 100644 --- a/src/storage/storage_driver.c +++ b/src/storage/storage_driver.c @@ -536,6 +536,9 @@ storagePoolCreate(virConnectPtr conn, if (virStoragePoolObjIsDuplicate(&driver->pools, def, 1)< 0) goto cleanup; + if (virStoragePoolTargetDuplicate(&driver->pools, def)< 0) + goto cleanup; + if ((backend = virStorageBackendForType(def->type)) == NULL) goto cleanup; @@ -589,6 +592,9 @@ storagePoolDefine(virConnectPtr conn, if (virStoragePoolObjIsDuplicate(&driver->pools, def, 0)< 0) goto cleanup; + if (virStoragePoolTargetDuplicate(&driver->pools, def)< 0) + goto cleanup; + if (virStorageBackendForType(def->type) == NULL) goto cleanup; -- 1.7.1

3 2

[libvirt] [PATCHv4 0/2] vnc/spice listen address as a network name
by Laine Stump 02 Aug '11

02 Aug '11

This is the latest reincarnation of this patch series: https://www.redhat.com/archives/libvir-list/2011-July/msg01717.html In the previous version, I had created a new subelement of <graphics>, called <listen>, and placed all listen address and port attributes there. After discussion on the list and in IRC with Dan Berrange and Eric Blake, I became convinced that was overdoing things, and that really only the address or network name should be in the <listen> element, but that port/tlsPort/autoport should stay put. This version removes all changes to port-related attributes, paring it down to just the listen address (and the new listen network) attribute. This greatly simplifies the patch, so new review of it should be easier. Based on Eric's review, the nwe version tries to be more concise about what is accepted, what is ignored, and what is rejected in the XML, both by the parser and by the RNG. (since there is less to deal with, this is also easier.) Having a single port value but multiple addresses does create some potential future restrictions, which we can deal with an eliminate later if they become problematic. In particular, each graphics device will have only a single port, but could have multiple listen addresses; in this case, all the listen addresses would have to listen on the same port. (This isn't a problem right now, because all of our drivers can only listen on a single address anyway). (Note to Eric: I did consider putting the switch to helper functions into its own patch, but after removing all the ports from <listen>, the number of helper functions (and thus the number of changes created by switching to using them, has been reduced considerably, so the gain wouldn't be as much. (also, it's very late, and I'd like to get some sleep tonight :-)

3 10

[libvirt] [PATCH] openvz: detect when a domain was shut down from the inside
by Jean-Baptiste Rouault 02 Aug '11

02 Aug '11

This patch adds an internal function openvzGetVEStatus to get the real state of the domain. This function is used in various places in the driver, in particular to detect when the domain has been shut down by the user with the "halt" command. --- src/openvz/openvz_driver.c | 76 +++++++++++++++++++++++++++++++++++++++----- 1 files changed, 68 insertions(+), 8 deletions(-) diff --git a/src/openvz/openvz_driver.c b/src/openvz/openvz_driver.c index 4e7cb03..3b3b079 100644 --- a/src/openvz/openvz_driver.c +++ b/src/openvz/openvz_driver.c @@ -72,6 +72,7 @@ static int openvzDomainSetVcpusInternal(virDomainObjPtr vm, unsigned int nvcpus); static int openvzDomainSetMemoryInternal(virDomainObjPtr vm, unsigned long memory); +static int openvzGetVEStatus(virDomainObjPtr vm, int *status, int *reason); static void openvzDriverLock(struct openvz_driver *driver) { @@ -340,6 +341,7 @@ static int openvzDomainGetInfo(virDomainPtr dom, virDomainInfoPtr info) { struct openvz_driver *driver = dom->conn->privateData; virDomainObjPtr vm; + int state; int ret = -1; openvzDriverLock(driver); @@ -352,9 +354,11 @@ static int openvzDomainGetInfo(virDomainPtr dom, goto cleanup; } - info->state = virDomainObjGetState(vm, NULL); + if (openvzGetVEStatus(vm, &state, NULL) == -1) + goto cleanup; + info->state = state; - if (!virDomainObjIsActive(vm)) { + if (info->state != VIR_DOMAIN_RUNNING) { info->cpuTime = 0; } else { if (openvzGetProcessInfo(&(info->cpuTime), dom->id) < 0) { @@ -398,8 +402,7 @@ openvzDomainGetState(virDomainPtr dom, goto cleanup; } - *state = virDomainObjGetState(vm, reason); - ret = 0; + ret = openvzGetVEStatus(vm, state, reason); cleanup: if (vm) @@ -584,6 +587,7 @@ openvzDomainShutdownFlags(virDomainPtr dom, virDomainObjPtr vm; const char *prog[] = {VZCTL, "--quiet", "stop", PROGRAM_SENTINAL, NULL}; int ret = -1; + int status; virCheckFlags(0, -1); @@ -596,9 +600,12 @@ openvzDomainShutdownFlags(virDomainPtr dom, _("no domain with matching uuid")); goto cleanup; } + + if (openvzGetVEStatus(vm, &status, NULL) == -1) + goto cleanup; openvzSetProgramSentinal(prog, vm->def->name); - if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_RUNNING) { + if (status != VIR_DOMAIN_RUNNING) { openvzError(VIR_ERR_INTERNAL_ERROR, "%s", _("domain is not in running state")); goto cleanup; @@ -631,6 +638,7 @@ static int openvzDomainReboot(virDomainPtr dom, virDomainObjPtr vm; const char *prog[] = {VZCTL, "--quiet", "restart", PROGRAM_SENTINAL, NULL}; int ret = -1; + int status; virCheckFlags(0, -1); @@ -644,8 +652,11 @@ static int openvzDomainReboot(virDomainPtr dom, goto cleanup; } + if (openvzGetVEStatus(vm, &status, NULL) == -1) + goto cleanup; + openvzSetProgramSentinal(prog, vm->def->name); - if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_RUNNING) { + if (status != VIR_DOMAIN_RUNNING) { openvzError(VIR_ERR_INTERNAL_ERROR, "%s", _("domain is not in running state")); goto cleanup; @@ -1052,6 +1063,7 @@ openvzDomainCreateWithFlags(virDomainPtr dom, unsigned int flags) virDomainObjPtr vm; const char *prog[] = {VZCTL, "--quiet", "start", PROGRAM_SENTINAL, NULL }; int ret = -1; + int status; virCheckFlags(0, -1); @@ -1064,8 +1076,11 @@ openvzDomainCreateWithFlags(virDomainPtr dom, unsigned int flags) _("no domain with matching id")); goto cleanup; } + + if (openvzGetVEStatus(vm, &status, NULL) == -1) + goto cleanup; - if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_SHUTOFF) { + if (status != VIR_DOMAIN_SHUTOFF) { openvzError(VIR_ERR_OPERATION_DENIED, "%s", _("domain is not in shutoff state")); goto cleanup; @@ -1102,6 +1117,7 @@ openvzDomainUndefineFlags(virDomainPtr dom, virDomainObjPtr vm; const char *prog[] = { VZCTL, "--quiet", "destroy", PROGRAM_SENTINAL, NULL }; int ret = -1; + int status; virCheckFlags(0, -1); @@ -1113,7 +1129,10 @@ openvzDomainUndefineFlags(virDomainPtr dom, goto cleanup; } - if (virDomainObjIsActive(vm)) { + if (openvzGetVEStatus(vm, &status, NULL) == -1) + goto cleanup; + + if (status != VIR_DOMAIN_SHUTOFF) { openvzError(VIR_ERR_OPERATION_INVALID, "%s", _("cannot delete active domain")); goto cleanup; @@ -1610,6 +1629,47 @@ cleanup: return -1; } +static int +openvzGetVEStatus(virDomainObjPtr vm, int *status, int *reason) +{ + virCommandPtr cmd; + char *outbuf; + char *line; + int state; + int ret = -1; + + cmd = virCommandNewArgList(VZLIST, vm->def->name, "-ostatus", "-H", NULL); + virCommandSetOutputBuffer(cmd, &outbuf); + if (virCommandRun(cmd, NULL) < 0) + goto cleanup; + + if ((line = strchr(outbuf, '\n')) == NULL) { + openvzError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Failed to parse vzlist output")); + goto cleanup; + } + *line++ = '\0'; + + state = virDomainObjGetState(vm, reason); + + if (STREQ(outbuf, "running")) { + // There is no way to detect whether a domain is paused or not with vzlist + if (state == VIR_DOMAIN_PAUSED) + *status = state; + else + *status = VIR_DOMAIN_RUNNING; + } else { + *status = VIR_DOMAIN_SHUTOFF; + } + + ret = 0; + +cleanup: + virCommandFree(cmd); + VIR_FREE(outbuf); + return ret; +} + static virDriver openvzDriver = { .no = VIR_DRV_OPENVZ, .name = "OPENVZ", -- 1.7.5.4

2 1

[libvirt] [PATCH RFC] virsh: Add option to undefine storage with domains
by Peter Krempa 02 Aug '11

02 Aug '11

Adds an option to virsh undefine command to undefine managed storage volumes along with (inactive) domains. Storage volumes are enumerated and the user may interactivly choose volumes to delete. Unmanaged volumes are listed and the user shall delete them manualy. --- I marked this as a RFC because I am concerned about my "naming scheme" of the added parameters. I couldn't decide which of the following "volumes/storage/disks/..." to use. I'd appreciate your comments on this. This is my second approach to this problem after I got some really good critique from Eric, Daniel and Dave. The user has the choice to activate an interactive mode, that allows to select on a per-device basis which volumes/disks to remove along with the domain. To avoid possible problems, I only allowed to remove storage for inactive domains and unmanaged images (which sidetracked me a lot on my previous attempt) are left to a action of the user. (the user is notified about any unmanaged image for the domain). My next concern is about interactive of the user. I tried to implement a boolean query function, but I'd like to know if I took the right path, as I couldn't find any example in virsh from which I could learn. Thanks for your comments (and time) :) Peter Krempa tools/virsh.c | 265 +++++++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 files changed, 259 insertions(+), 6 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 61f69f0..3795d2b 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -295,6 +295,9 @@ static int vshCommandOptULongLong(const vshCmd *cmd, const char *name, static bool vshCommandOptBool(const vshCmd *cmd, const char *name); static const vshCmdOpt *vshCommandOptArgv(const vshCmd *cmd, const vshCmdOpt *opt); +static int vshInteractiveBoolPrompt(vshControl *ctl, + const char *prompt, + bool *confirm); #define VSH_BYID (1 << 1) #define VSH_BYUUID (1 << 2) @@ -1422,6 +1425,8 @@ static const vshCmdInfo info_undefine[] = { static const vshCmdOptDef opts_undefine[] = { {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, N_("domain name or uuid")}, {"managed-save", VSH_OT_BOOL, 0, N_("remove domain managed state file")}, + {"disks", VSH_OT_BOOL, 0, N_("remove associated disk images managed in storage pools (interactive)")}, + {"disks-all", VSH_OT_BOOL, 0, N_("remove all associated disk images managed in storage pools")}, {NULL, 0, 0, NULL} }; @@ -1434,9 +1439,25 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) int id; int flags = 0; int managed_save = vshCommandOptBool(cmd, "managed-save"); + int remove_disks = vshCommandOptBool(cmd, "disks"); + int remove_all_disks = vshCommandOptBool(cmd, "disks-all"); int has_managed_save = 0; int rc = -1; + char *domxml; + xmlDocPtr xml = NULL; + xmlXPathContextPtr ctxt = NULL; + xmlXPathObjectPtr obj = NULL; + xmlNodePtr cur = NULL; + int i = 0; + char *source = NULL; + char *target = NULL; + char *type = NULL; + xmlBufferPtr xml_buf = NULL; + virStorageVolPtr volume = NULL; + int state; + bool confirm = false; + if (managed_save) flags |= VIR_DOMAIN_UNDEFINE_MANAGED_SAVE; @@ -1475,15 +1496,172 @@ cmdUndefine(vshControl *ctl, const vshCmd *cmd) } } - if (flags == -1) { - if (has_managed_save == 1) { + + if (flags == -1 && has_managed_save == 1) { + vshError(ctl, + _("Refusing to undefine while domain managed save " + "image exists")); + virDomainFree(dom); + return false; + } + + if (remove_disks || remove_all_disks) { + if ((state = vshDomainState(ctl, dom, NULL)) < 0) { + vshError(ctl, _("Failed to get domain state")); + goto disk_error; + } + + /* removal of storage is possible only for inactive domains */ + if (!((state == VIR_DOMAIN_SHUTOFF) || + (state == VIR_DOMAIN_CRASHED))) { vshError(ctl, - _("Refusing to undefine while domain managed save " - "image exists")); - virDomainFree(dom); - return false; + _("Domain needs to be inactive to delete it with associated storage")); + goto disk_error; + } + + if (remove_disks && !ctl->imode) { + vshError(ctl, "%s\n", _("Option --disks is available only in interactive mode")); + goto disk_error; + } + + domxml = virDomainGetXMLDesc(dom, 0); + if (!domxml) { + vshError(ctl, "%s", _("Failed to get disk information")); + goto disk_error; + } + + xml = xmlReadDoc((const xmlChar *) domxml, "domain.xml", NULL, + XML_PARSE_NOENT | + XML_PARSE_NONET | + XML_PARSE_NOWARNING); + VIR_FREE(domxml); + + if (!xml) { + vshError(ctl, "%s", _("Failed to get disk information")); + goto disk_error; + } + + ctxt = xmlXPathNewContext(xml); + if (!ctxt) { + vshError(ctl, "%s", _("Failed to get disk information")); + goto disk_error; + } + + obj = xmlXPathEval(BAD_CAST "/domain/devices/disk", ctxt); + if ((obj == NULL) || (obj->type != XPATH_NODESET) || + (obj->nodesetval == NULL)) { + vshError(ctl, "%s", _("Failed to get disk information")); + goto disk_error; + } + + for (i = 0; i < obj->nodesetval->nodeNr; i++) { + cur = obj->nodesetval->nodeTab[i]->children; + + type = virXMLPropString(cur, "device"); + + while (cur != NULL) { + if (cur->type == XML_ELEMENT_NODE) { + if (xmlStrEqual(cur->name, BAD_CAST "target")) + target = virXMLPropString(cur, "dev"); + else if (xmlStrEqual(cur->name, BAD_CAST "source")) + source = virXMLPropString(cur, "file"); + } + cur = cur->next; + } + + if (!source) { + VIR_FREE(target); + VIR_FREE(type); + } + + volume = virStorageVolLookupByPath(ctl->conn, (const char *) source); + if (!volume) { + vshPrint(ctl, "%s %s %s %s\n", + _("Volume: Source:"), (const char *)source, + _("Target:"), (const char *) target); + vshPrint(ctl, _("This volume isn't managed by any storage pool, " + "please delete it manualy\n\n")); + /* remove error indication */ + virFreeError(last_error); + last_error = NULL; + } else { + vshPrint(ctl, "%s %s %s %s\n", + _("Volume: Source:"), (const char *)source, + _("Target:"), (const char *) target); + + if (remove_all_disks) { + confirm = true; + } else { + if (vshInteractiveBoolPrompt(ctl, + _("Do you want to undefine this volume?"), + &confirm) < 0) { + vshError(ctl, _("\nError while geting response from user")); + virStorageVolFree(volume); + goto disk_error; + } + } + + /* removal of volume */ + if (confirm) { + if (virStorageVolDelete(volume, 0) == 0) { + virStorageVolFree(volume); + + vshPrint(ctl, _("Volume deleted\n\n")); + + /* remove definition of volume from xml */ + xml_buf = xmlBufferCreate(); + if (!xml_buf) { + vshPrint(ctl, _("Failed to create XML buffer. " + "If domain undefinition fails, domain will be left in inconsistent state.\n\n")); + goto disk_next; + } + + if (xmlNodeDump(xml_buf, xml, obj->nodesetval->nodeTab[i], 0, 0) < 0) { + vshPrint(ctl, _("Failed to extract XML volume description. " + "If domain undefinition fails, domain will be left in inconsistent state.\n\n")); + + xmlBufferFree(xml_buf); + xml_buf = NULL; + goto disk_next; + } + + if (virDomainDetachDeviceFlags(dom, + (char *) xmlBufferContent(xml_buf), + VIR_DOMAIN_AFFECT_CONFIG) < 0) { + vshPrint(ctl, + _("Failed to remove volume \"%s\" from configuration. " + "If domain undefinition fails, domain will be left in inconsistent state.\n\n"), + source); + + xmlBufferFree(xml_buf); + xml_buf = NULL; + goto disk_next; + } + + xmlBufferFree(xml_buf); + xml_buf = NULL; + + } else { + virStorageVolFree(volume); + + vshError(ctl, _("Failed to delete volume.")); + goto disk_error; + } + } + } + +disk_next: + VIR_FREE(source); + VIR_FREE(target); + VIR_FREE(type); } + xmlXPathFreeObject(obj); + xmlXPathFreeContext(ctxt); + xmlFreeDoc(xml); + } /* end of disk undefine stuff */ + + if (flags == -1) { rc = virDomainUndefine(dom); } else { rc = virDomainUndefineFlags(dom, flags); @@ -1520,6 +1698,21 @@ end: virDomainFree(dom); return ret; + +disk_error: + VIR_FREE(source); + VIR_FREE(target); + VIR_FREE(type); + + xmlXPathFreeObject(obj); + xmlXPathFreeContext(ctxt); + xmlFreeDoc(xml); + xmlBufferFree(xml_buf); + + virDomainFree(dom); + + vshError(ctl, _("\nFailed to undefine domain %s"), name); + return false; } @@ -14736,6 +14929,66 @@ vshReadline (vshControl *ctl, const char *prompt) #endif /* !USE_READLINE */ + +/* + * Propmpt for boolean question (yes/no) + * + * Returns "true" on positive answer, "false" on negative + * -1 on error. + */ +static int +vshInteractiveBoolPrompt(vshControl *ctl, + const char *prompt, + bool *confirm) { + const char *positive = _("yes"); + const char *negative = _("no"); + char *r; + char buff[100]; + int ret = -1; + int len; + int c; + + if (confirm == NULL) + return ret; + + if (!ctl->imode) + return ret; + + while (ret == -1) { + vshPrint(ctl, "%s (%s/%s)? ", prompt, positive, negative); + + r = fgets(buff, sizeof(buff), stdin); + if (r == NULL) + break; + len = strlen(buff); + if (len > 0 && buff[len-1] == '\n') + buff[len-1] = '\0'; + else + /* eat rest of line */ + while ((c = fgetc(stdin) != EOF)) + if (c == '\n') + break; + + /* compare */ + if (STRCASEEQLEN(buff, positive, strlen(positive))) { + ret = 0; + *confirm = true; + break; + } + + if (STRCASEEQLEN(buff, negative, strlen(negative))) { + ret = 0; + *confirm = false; + break; + } + + /* errorneus response - warn and ask again*/ + vshPrint(ctl, "\"%s\" %s\n", buff, _("Response not understood")); + + } + return ret; +} + /* * Deinitialize virsh */ -- 1.7.6

3 3

[libvirt] [PATCH] support qemuMonitorSend() to be called at the same time
by Wen Congyang 02 Aug '11

02 Aug '11

Current, we support run sync job and async job at the same time. It means that the monitor commands for two jobs can be run in any order. In the function qemuDomainObjEnterMonitorInternal(): if (priv->job.active == QEMU_JOB_NONE && priv->job.asyncJob) { if (qemuDomainObjBeginNestedJob(driver, obj) < 0) We check whether the caller is an async job by priv->job.active and priv->job.asynJob. But when an async job is running, priv->job.active is QEMU_JOB_NONE if no sync job is running, or priv->job.active is not QEMU_JOB_NONE if a sync job is running. So we cannot check whether the caller is an async job in the function qemuDomainObjEnterMonitorInternal(). Unfortunately, if sync job and async job are running at the same time, we may try to send monitor command at the same time in two threads. It's very dangerous, and it will cause libvirtd crashed. We should enhance the function qemuMonitorSend() to support it to be called at the same time. If the last monitor command does not finish, the other monitor commands should wait it to finish. --- src/qemu/qemu_monitor.c | 22 +++++++++++++++++++++- 1 files changed, 21 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_monitor.c b/src/qemu/qemu_monitor.c index db6107c..a10f53f 100644 --- a/src/qemu/qemu_monitor.c +++ b/src/qemu/qemu_monitor.c @@ -46,6 +46,8 @@ struct _qemuMonitor { virMutex lock; /* also used to protect fd */ virCond notify; + virCond send_notify; + int refs; int fd; @@ -675,7 +677,8 @@ qemuMonitorOpen(virDomainObjPtr vm, VIR_FREE(mon); return NULL; } - if (virCondInit(&mon->notify) < 0) { + if ((virCondInit(&mon->notify) < 0) || + (virCondInit(&mon->send_notify) < 0)) { qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("cannot initialize monitor condition")); virMutexDestroy(&mon->lock); @@ -795,6 +798,22 @@ int qemuMonitorSend(qemuMonitorPtr mon, return -1; } + while (mon->msg) { + if (virCondWait(&mon->send_notify, &mon->lock) < 0) { + qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Unable to wait on monitor condition")); + goto cleanup; + } + } + + /* Check whether qemu quited unexpectedly again */ + if (mon->lastError.code != VIR_ERR_OK) { + VIR_DEBUG("Attempt to send command while error is set %s", + NULLSTR(mon->lastError.message)); + virSetError(&mon->lastError); + return -1; + } + mon->msg = msg; qemuMonitorUpdateWatch(mon); @@ -818,6 +837,7 @@ int qemuMonitorSend(qemuMonitorPtr mon, cleanup: mon->msg = NULL; qemuMonitorUpdateWatch(mon); + virCondBroadcast(&mon->send_notify); return ret; } -- 1.7.1

3 11

[libvirt] [PATCH] Correct the default value of lock_manager in qemu.conf
by Guannan Ren 01 Aug '11

01 Aug '11

* src/qemu/qemu.conf lock_manager = "sanlock" --- src/qemu/qemu.conf | 7 +++---- 1 files changed, 3 insertions(+), 4 deletions(-) diff --git a/src/qemu/qemu.conf b/src/qemu/qemu.conf index 8058352..79c6e85 100644 --- a/src/qemu/qemu.conf +++ b/src/qemu/qemu.conf @@ -304,9 +304,8 @@ # # max_processes = 0 -# To enable strict 'fcntl' based locking of the file +# To enable 'Sanlock' project based locking of the file # content (to prevent two VMs writing to the same -# disk), start the 'virtlockd' service, and uncomment -# this +# disk), uncomment this # -# lock_manager = "fcntl" +# lock_manager = "sanlock" -- 1.7.1

2 1