- Devel - libvirt List Archives

Re: [libvirt] virsh question
by Eric Blake 23 Aug '10

23 Aug '10

[again, adding the list] On 08/23/2010 07:35 AM, Matthew Whitehead wrote: > I want to enable gdb debugging for the qemu or qemu-kvm executables. The -s flag tells it to run qemu's gdbserver and -p tells the gdbserver to listen on a non-default (1234) port. Yes, this sounds like a perfect fit for libvirt 0.8.3's qemu monitor support. However, I did not implement that patch, so I'm not the best person to ask on how to modify your XML to inject those options. However, commit a71be01f04e7cac in libvirt.git includes some example XML that uses the new elements, such as: +<domain type='qemu' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'> + <name>QEMUGuest1</name> + <uuid>c7a5fdbd-edaf-9455-926a-d65c16db1809</uuid> + <memory>219200</memory> + <currentMemory>219200</currentMemory> + <vcpu>1</vcpu> + <os> + <type arch='i686' machine='pc'>hvm</type> + <boot dev='hd'/> + </os> + <clock offset='utc'/> + <on_poweroff>destroy</on_poweroff> + <on_reboot>restart</on_reboot> + <source dev='/dev/HostVG/QEMUGuest1'/> + <target dev='hda' bus='ide'/> + <address type='drive' controller='0' bus='0' unit='0'/> + </disk> + <controller type='ide' index='0'/> + <memballoon model='virtio'/> + </devices> + <qemu:commandline> + <qemu:arg value='-unknown'/> + <qemu:arg value='parameter'/> + </qemu:commandline> +</domain> So, I think it is as simple as passing <qemu:arg value='-p'/> from within your <domain> XML. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

Re: [libvirt] virsh question
by Eric Blake 23 Aug '10

23 Aug '10

[Adding the list] On 08/23/2010 06:26 AM, Matthew Whitehead wrote: > Eric, > I'm trying to get "virsh domxml-from-native " working for the -s and -p flags. Do you know if support for those flags is possible? If not, is there a way I could hand code them into the xml to be passed through to the command line? I'm not sure in which context you are referring to the -s and -p flags. Which command line are you trying to modify so that those flags are added? Also, as of libvirt 0.8.3, you can modify qemu command lines issued by libvirt using appropriate XML, although such modifications go behind libvirt's back and are therefore at your own risk. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

1 0

[libvirt] [PATCH] xenapi: Fix compile error in previous commit
by Matthias Bolte 23 Aug '10

23 Aug '10

--- I'm pushing this under the trivial compile error fix rule. src/xenapi/xenapi_utils.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/xenapi/xenapi_utils.c b/src/xenapi/xenapi_utils.c index c161b22..f61e09d 100644 --- a/src/xenapi/xenapi_utils.c +++ b/src/xenapi/xenapi_utils.c @@ -225,7 +225,7 @@ xenapiNormalExitEnum2virDomainLifecycle(enum xen_on_normal_exit action) enum virDomainLifecycleCrashAction xenapiCrashExitEnum2virDomainLifecycle(enum xen_on_crash_behaviour action) { - enum virDomainLifecycleCrashAction num = VIR_DOMAIN_LIFECYCLE_CRASH__RESTART; + enum virDomainLifecycleCrashAction num = VIR_DOMAIN_LIFECYCLE_CRASH_RESTART; if (action == XEN_ON_CRASH_BEHAVIOUR_DESTROY) num = VIR_DOMAIN_LIFECYCLE_CRASH_DESTROY; else if (action == XEN_ON_CRASH_BEHAVIOUR_RESTART) -- 1.7.0.4

4 6

Re: [libvirt] using sync_manager with libvirt
by Andrew Cathrow 22 Aug '10

22 Aug '10

----- "Perry Myers" <pmyers(a)redhat.com> wrote: > From: "Perry Myers" <pmyers(a)redhat.com> > To: "David Teigland" <teigland(a)redhat.com> > Cc: libvir-list(a)redhat.com > Sent: Sunday, August 22, 2010 12:13:16 PM GMT -05:00 US/Canada Eastern > Subject: Re: [libvirt] using sync_manager with libvirt > > On 08/19/2010 01:23 PM, David Teigland wrote: > > On Thu, Aug 19, 2010 at 11:12:25AM -0400, David Teigland wrote: > >> I'm only aware of one goal, and the current plan is to implement it > >> correctly and completely. That goal is to lock vm images so if the vm > >> happens to run on two hosts, only one instance can access the image. > > Ok. So for the first implementation of sync_manager it will still be > possible for someone to corrupt data by configuring two separate vms to > accidentally use the same storage volumes. That's fine for the first > pass, just something to keep in mind for later. Presumably it'll use support the <shareable/> flag for shared disks. > > > (That's slightly misleading; technically, the lock prevents a second qemu > > process from even being started.) > > ack > > Perry > > > -- > libvir-list mailing list > libvir-list(a)redhat.com > https://www.redhat.com/mailman/listinfo/libvir-list >

1 0

[libvirt] Add spice support
by Songwen Liao 21 Aug '10

21 Aug '10

Hi,everyone, I used the patch file to modify the libvirt,but it seems not to work right. It looks like that libvirt accpet the graphices type of spice,but it's can use "virsh start xp" to start the machine,with the error is: internal error unable to start guest:Supported machines are: pc Standard PC(default) isapc ISA-only PC So,anyone have ideas about that? thanks, PATCH file: This adds an element <graphics type='spice' port='5903' tlsPort='5904' listen='127.0.0.1'/> This is the bare minimum that should be exposed in the guest config for SPICE. Other parameters are better handled as per host level configuration tunables * docs/schemas/domain.rng: Define the SPICE <graphics> schema * src/domain_conf.h, src/domain_conf.c: Add parsing and formatting for SPICE graphics config * src/qemu_conf.c: Complain about unsupported graphics types --- docs/schemas/domain.rng | 30 +++++++++++++++++++++ src/conf/domain_conf.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++- src/conf/domain_conf.h | 8 ++++++ src/qemu/qemu_conf.c | 9 ++++++ 4 files changed, 111 insertions(+), 1 deletions(-) diff --git a/docs/schemas/domain.rng b/docs/schemas/domain.rng index f9d6d7e..8aa2079 100644 --- a/docs/schemas/domain.rng +++ b/docs/schemas/domain.rng @@ -746,6 +746,36 @@ </group> <group> <attribute name="type"> + <value>spice</value> + </attribute> + <optional> + <attribute name="port"> + <ref name="PortNumber"/> + </attribute> + </optional> + <optional> + <attribute name="tlsPort"> + <ref name="PortNumber"/> + </attribute> + </optional> + <optional> + <attribute name="listen"> + <ref name="addrIP"/> + </attribute> + </optional> + <optional> + <attribute name="passwd"> + <text/> + </attribute> + </optional> + <optional> + <attribute name="keymap"> + <text/> + </attribute> + </optional> + </group> + <group> + <attribute name="type"> <value>rdp</value> </attribute> <optional> diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 4930c38..ca5dfc3 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -165,7 +165,8 @@ VIR_ENUM_IMPL(virDomainGraphics, VIR_DOMAIN_GRAPHICS_TYPE_LAST, "sdl", "vnc", "rdp", - "desktop") + "desktop", + "spice") VIR_ENUM_IMPL(virDomainHostdevMode, VIR_DOMAIN_HOSTDEV_MODE_LAST, "subsystem", @@ -267,6 +268,12 @@ void virDomainGraphicsDefFree(virDomainGraphicsDefPtr def) case VIR_DOMAIN_GRAPHICS_TYPE_DESKTOP: VIR_FREE(def->data.desktop.display); break; + + case VIR_DOMAIN_GRAPHICS_TYPE_SPICE: + VIR_FREE(def->data.spice.listenAddr); + VIR_FREE(def->data.spice.keymap); + VIR_FREE(def->data.spice.passwd); + break; } VIR_FREE(def); @@ -1692,6 +1699,38 @@ virDomainGraphicsDefParseXML(virConnectPtr conn, def->data.desktop.fullscreen = 0; def->data.desktop.display = virXMLPropString(node, "display"); + } else if (def->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) { + char *port = virXMLPropString(node, "port"); + char *tlsPort; + + if (port) { + if (virStrToLong_i(port, NULL, 10, &def->data.spice.port) < 0) { + virDomainReportError(conn, VIR_ERR_INTERNAL_ERROR, + _("cannot parse spice port %s"), port); + VIR_FREE(port); + goto error; + } + VIR_FREE(port); + } else { + def->data.spice.port = 5900; + } + + tlsPort = virXMLPropString(node, "tlsPort"); + if (tlsPort) { + if (virStrToLong_i(tlsPort, NULL, 10, &def->data.spice.tlsPort) < 0) { + virDomainReportError(conn, VIR_ERR_INTERNAL_ERROR, + _("cannot parse spice tlsPort %s"), tlsPort); + VIR_FREE(tlsPort); + goto error; + } + VIR_FREE(tlsPort); + } else { + def->data.spice.tlsPort = 0; + } + + def->data.spice.listenAddr = virXMLPropString(node, "listen"); + def->data.spice.passwd = virXMLPropString(node, "passwd"); + def->data.spice.keymap = virXMLPropString(node, "keymap"); } cleanup: @@ -4100,6 +4139,30 @@ virDomainGraphicsDefFormat(virConnectPtr conn, break; + case VIR_DOMAIN_GRAPHICS_TYPE_SPICE: + if (def->data.spice.port) + virBufferVSprintf(buf, " port='%d'", + def->data.spice.port); + + if (def->data.spice.tlsPort) + virBufferVSprintf(buf, " tlsPort='%d'", + def->data.spice.tlsPort); + + if (def->data.spice.listenAddr) + virBufferVSprintf(buf, " listen='%s'", + def->data.spice.listenAddr); + + if (def->data.spice.keymap) + virBufferEscapeString(buf, " keymap='%s'", + def->data.spice.keymap); + + if (def->data.spice.passwd && + (flags & VIR_DOMAIN_XML_SECURE)) + virBufferEscapeString(buf, " passwd='%s'", + def->data.spice.passwd); + + break; + } virBufferAddLit(buf, "/>\n"); diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index 7ea1152..4cb10d3 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -334,6 +334,7 @@ enum virDomainGraphicsType { VIR_DOMAIN_GRAPHICS_TYPE_VNC, VIR_DOMAIN_GRAPHICS_TYPE_RDP, VIR_DOMAIN_GRAPHICS_TYPE_DESKTOP, + VIR_DOMAIN_GRAPHICS_TYPE_SPICE, VIR_DOMAIN_GRAPHICS_TYPE_LAST, }; @@ -366,6 +367,13 @@ struct _virDomainGraphicsDef { char *display; int fullscreen : 1; } desktop; + struct { + int port; + int tlsPort; + char *listenAddr; + char *keymap; + char *passwd; + } spice; } data; }; diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 7ba0ac2..ae171bc 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2067,6 +2067,12 @@ int qemudBuildCommandLine(virConnectPtr conn, } } + if (def->ngraphics > 1) { + qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, + "%s", _("only one graphics output is currently supported")); + goto error; + } + if ((def->ngraphics == 1) && def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC) { virBuffer opt = VIR_BUFFER_INITIALIZER; @@ -2155,6 +2161,9 @@ int qemudBuildCommandLine(virConnectPtr conn, */ ADD_ENV_COPY("QEMU_AUDIO_DRV"); ADD_ENV_COPY("SDL_AUDIODRIVER"); + } else if (def->ngraphics) { + qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, + "%s", _("unsupported graphics output requested")); } if (def->nvideos) { -- -- Name: Ben ------------------------------------------- Msn: liaosongwen(a)live.cn

2 2

[libvirt] [PATCH] esx: Use MD5 sum of mount path as storage pool UUID
by Matthias Bolte 21 Aug '10

21 Aug '10

With the previous storage pool UUID source not all storage pools had a proper UUID, especially GSX storage pools. The mount path is unique per host and cannot change during the lifetime of the datastore. Therefore, its MD5 sum can be used as UUID. Use gnulib's crypto/md5 module to generate the MD5 sum. --- bootstrap.conf | 1 + src/esx/esx_storage_driver.c | 107 +++++++++++------------------------------- 2 files changed, 28 insertions(+), 80 deletions(-) diff --git a/bootstrap.conf b/bootstrap.conf index 5af77a7..ca31a6e 100644 --- a/bootstrap.conf +++ b/bootstrap.conf @@ -26,6 +26,7 @@ canonicalize-lgpl close connect count-one-bits +crypto/md5 dirname-lgpl fcntl-h getaddrinfo diff --git a/src/esx/esx_storage_driver.c b/src/esx/esx_storage_driver.c index 644e66c..adbe8c6 100644 --- a/src/esx/esx_storage_driver.c +++ b/src/esx/esx_storage_driver.c @@ -24,6 +24,7 @@ #include <config.h> +#include "md5.h" #include "internal.h" #include "util.h" #include "memory.h" @@ -197,10 +198,7 @@ esxStoragePoolLookupByName(virConnectPtr conn, const char *name) esxPrivate *priv = conn->storagePrivateData; esxVI_ObjectContent *datastore = NULL; esxVI_DatastoreHostMount *hostMount = NULL; - char *suffix = NULL; - int suffixLength; - char uuid_string[VIR_UUID_STRING_BUFLEN] = "00000000-00000000-0000-000000000000"; - unsigned char uuid[VIR_UUID_BUFLEN]; + unsigned char md5[MD5_DIGEST_SIZE]; /* MD5_DIGEST_SIZE = VIR_UUID_BUFLEN = 16 */ virStoragePoolPtr pool = NULL; if (esxVI_EnsureSession(priv->primary) < 0) { @@ -213,53 +211,22 @@ esxStoragePoolLookupByName(virConnectPtr conn, const char *name) } /* - * Datastores don't have a UUID. We can use the 'host.mountInfo.path' - * property as source for a "UUID" on ESX, because the property value has - * this format: + * Datastores don't have a UUID, but we can use the 'host.mountInfo.path' + * property as source for a UUID. The mount path is unique per host and + * cannot change during the lifetime of the datastore. * - * host.mountInfo.path = /vmfs/volumes/4b0beca7-7fd401f3-1d7f-000ae484a6a3 - * host.mountInfo.path = /vmfs/volumes/b24b7a78-9d82b4f5 (short format) - * - * The 'host.mountInfo.path' property comes in two forms, with a complete - * "UUID" and a short "UUID". - * - * But this trailing "UUID" is not guaranteed to be there. On the other - * hand we already rely on another implementation detail of the ESX server: - * The object name of virtual machine contains an integer, we use that as - * domain ID. + * The MD5 sum of the mount path can be used as UUID, assuming MD5 is + * considered to be collision-free enough for this use case. */ if (esxVI_LookupDatastoreHostMount(priv->primary, datastore->obj, &hostMount) < 0) { goto cleanup; } - if ((suffix = STRSKIP(hostMount->mountInfo->path, "/vmfs/volumes/")) != NULL) { - suffixLength = strlen(suffix); + md5_buffer(hostMount->mountInfo->path, + strlen(hostMount->mountInfo->path), md5); - if ((suffixLength == 35 && /* = strlen("4b0beca7-7fd401f3-1d7f-000ae484a6a3") */ - suffix[8] == '-' && suffix[17] == '-' && suffix[22] == '-') || - (suffixLength == 17 && /* = strlen("b24b7a78-9d82b4f5") */ - suffix[8] == '-')) { - /* - * Intentionally use memcpy here, because we want to be able to - * replace a prefix of the initial Zero-UUID. virStrncpy would - * null-terminate the string in an unwanted place. - */ - memcpy(uuid_string, suffix, suffixLength); - } else { - VIR_WARN("Datastore host mount path suffix '%s' has unexpected " - "format, cannot deduce a UUID from it", suffix); - } - } - - if (virUUIDParse(uuid_string, uuid) < 0) { - ESX_ERROR(VIR_ERR_INTERNAL_ERROR, - _("Could not parse UUID from string '%s'"), - uuid_string); - goto cleanup; - } - - pool = virGetStoragePool(conn, name, uuid); + pool = virGetStoragePool(conn, name, md5); cleanup: esxVI_ObjectContent_Free(&datastore); @@ -275,9 +242,11 @@ esxStoragePoolLookupByUUID(virConnectPtr conn, const unsigned char *uuid) { esxPrivate *priv = conn->storagePrivateData; esxVI_String *propertyNameList = NULL; + esxVI_ObjectContent *datastoreList = NULL; esxVI_ObjectContent *datastore = NULL; + esxVI_DatastoreHostMount *hostMount = NULL; + unsigned char md5[MD5_DIGEST_SIZE]; /* MD5_DIGEST_SIZE = VIR_UUID_BUFLEN = 16 */ char uuid_string[VIR_UUID_STRING_BUFLEN] = ""; - char *absolutePath = NULL; char *name = NULL; virStoragePoolPtr pool = NULL; @@ -285,48 +254,26 @@ esxStoragePoolLookupByUUID(virConnectPtr conn, const unsigned char *uuid) return NULL; } - /* - * Convert UUID to 'host.mountInfo.path' form by stripping the second '-': - * - * <---- 14 ----><-------- 22 --------> <---- 13 ---><-------- 22 --------> - * 4b0beca7-7fd4-01f3-1d7f-000ae484a6a3 -> 4b0beca7-7fd401f3-1d7f-000ae484a6a3 - */ - virUUIDFormat(uuid, uuid_string); - memmove(uuid_string + 13, uuid_string + 14, 22 + 1); - - if (virAsprintf(&absolutePath, "/vmfs/volumes/%s", uuid_string) < 0) { - virReportOOMError(); - goto cleanup; - } - if (esxVI_String_AppendValueToList(&propertyNameList, "summary.name") < 0 || - esxVI_LookupDatastoreByAbsolutePath(priv->primary, absolutePath, - propertyNameList, &datastore, - esxVI_Occurrence_OptionalItem) < 0) { + esxVI_LookupDatastoreList(priv->primary, propertyNameList, + &datastoreList) < 0) { goto cleanup; } - /* - * If the first try didn't succeed and the trailing 16 digits are zero then - * the "UUID" could be a short one. Strip the 16 zeros and try again: - * - * <------ 17 -----> <------ 17 -----> - * b24b7a78-9d82b4f5-0000-000000000000 -> b24b7a78-9d82b4f5 - */ - if (datastore == NULL && STREQ(uuid_string + 17, "-0000-000000000000")) { - uuid_string[17] = '\0'; + for (datastore = datastoreList; datastore != NULL; + datastore = datastore->_next) { + esxVI_DatastoreHostMount_Free(&hostMount); - VIR_FREE(absolutePath); - - if (virAsprintf(&absolutePath, "/vmfs/volumes/%s", uuid_string) < 0) { - virReportOOMError(); + if (esxVI_LookupDatastoreHostMount(priv->primary, datastore->obj, + &hostMount) < 0) { goto cleanup; } - if (esxVI_LookupDatastoreByAbsolutePath(priv->primary, absolutePath, - propertyNameList, &datastore, - esxVI_Occurrence_RequiredItem) < 0) { - goto cleanup; + md5_buffer(hostMount->mountInfo->path, + strlen(hostMount->mountInfo->path), md5); + + if (memcmp(uuid, md5, VIR_UUID_BUFLEN) == 0) { + break; } } @@ -348,9 +295,9 @@ esxStoragePoolLookupByUUID(virConnectPtr conn, const unsigned char *uuid) pool = virGetStoragePool(conn, name, uuid); cleanup: - VIR_FREE(absolutePath); esxVI_String_Free(&propertyNameList); - esxVI_ObjectContent_Free(&datastore); + esxVI_ObjectContent_Free(&datastoreList); + esxVI_DatastoreHostMount_Free(&hostMount); return pool; } -- 1.7.0.4

5 7

[libvirt] [PATCH] esx: Make sure dumpxml outputs proper ID for active domains
by Matthias Bolte 21 Aug '10

21 Aug '10

--- src/esx/esx_driver.c | 13 +++++++++++-- 1 files changed, 11 insertions(+), 2 deletions(-) diff --git a/src/esx/esx_driver.c b/src/esx/esx_driver.c index 4fb357b..c6bd3b8 100644 --- a/src/esx/esx_driver.c +++ b/src/esx/esx_driver.c @@ -2624,6 +2624,8 @@ esxDomainDumpXML(virDomainPtr domain, int flags) esxPrivate *priv = domain->conn->privateData; esxVI_String *propertyNameList = NULL; esxVI_ObjectContent *virtualMachine = NULL; + esxVI_VirtualMachinePowerState powerState; + int id; char *vmPathName = NULL; char *datastoreName = NULL; char *directoryName = NULL; @@ -2640,11 +2642,14 @@ esxDomainDumpXML(virDomainPtr domain, int flags) return NULL; } - if (esxVI_String_AppendValueToList(&propertyNameList, - "config.files.vmPathName") < 0 || + if (esxVI_String_AppendValueListToList(&propertyNameList, + "config.files.vmPathName\0" + "runtime.powerState\0") < 0 || esxVI_LookupVirtualMachineByUuid(priv->primary, domain->uuid, propertyNameList, &virtualMachine, esxVI_Occurrence_RequiredItem) < 0 || + esxVI_GetVirtualMachinePowerState(virtualMachine, &powerState) < 0 || + esxVI_GetVirtualMachineIdentity(virtualMachine, &id, NULL, NULL) < 0 || esxVI_GetStringValue(virtualMachine, "config.files.vmPathName", &vmPathName, esxVI_Occurrence_RequiredItem) < 0) { goto cleanup; @@ -2693,6 +2698,10 @@ esxDomainDumpXML(virDomainPtr domain, int flags) priv->primary->productVersion); if (def != NULL) { + if (powerState != esxVI_VirtualMachinePowerState_PoweredOff) { + def->id = id; + } + xml = virDomainDefFormat(def, flags); } -- 1.7.0.4

2 2

[libvirt] [PATCH] Add actions to virDomainLifecycle enum
by Jim Fehlig 20 Aug '10

20 Aug '10

I'm not too fond of the schema change but thus far haven't found a way to condense it. Suggestions welcomed :-). Regards, Jim

3 7

[libvirt] inability to open local read-only connection
by Tavares, John 20 Aug '10

20 Aug '10

I have been experimenting with using libvirt (0.3.3) on a variety of systems (RHEL, CentOS and Oracle VM). I have run into an issue when I try to open a local read-only connection to the hypervisor that is failing only on Oracle VM server release 2.2.0. I have created a root owned setuid executable that is effectively running as root, but even so, still cannot open the local read-only connection of the hypervisor. It only works if I run it directly as root. This is not an option. I do not understand why it works as is on my RHEL and CentOS machines, but not my Oracle machine. It would seem as thought it is not checking if the effective uid is root, just the uid. Has anyone run into a similar issue or have any suggestions of what I might try to fix this issue or can tell me that this is a defect that needs (is) fixed?? Thanks.

4 23

[libvirt] [PATCH] qemu: Remove code duplication
by Jiri Denemark 20 Aug '10

20 Aug '10

We already filled the PCI address structure when we checked whether it's free or not, so let's just use the structure here instead of filling it again. --- src/qemu/qemu_conf.c | 4 +--- 1 files changed, 1 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index c7ed0a3..1c98447 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2258,9 +2258,7 @@ int qemuDomainPCIAddressSetNextAddr(qemuDomainPCIAddressSetPtr addrs, } dev->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI; - dev->addr.pci.domain = 0; - dev->addr.pci.bus = 0; - dev->addr.pci.slot = i; + dev->addr.pci = maybe.addr.pci; addrs->nextslot = i + 1; if (QEMU_PCI_ADDRESS_LAST_SLOT < addrs->nextslot) -- 1.7.2

2 2

[libvirt] [PATCH] qemu: Check for errors when converting PCI address to string
by Jiri Denemark 20 Aug '10

20 Aug '10

--- src/qemu/qemu_conf.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index fb85220..c7ed0a3 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -2096,6 +2096,8 @@ static int qemuCollectPCIAddress(virDomainDefPtr def ATTRIBUTE_UNUSED, if (dev->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) { char *addr = qemuPCIAddressAsString(dev); + if (!addr) + return -1; VIR_DEBUG("Remembering PCI addr %s", addr); @@ -2239,7 +2241,8 @@ int qemuDomainPCIAddressSetNextAddr(qemuDomainPCIAddressSetPtr addrs, maybe.addr.pci.bus = 0; maybe.addr.pci.slot = i; - addr = qemuPCIAddressAsString(&maybe); + if (!(addr = qemuPCIAddressAsString(&maybe))) + return -1; if (virHashLookup(addrs->used, addr)) { VIR_DEBUG("PCI addr %s already in use", addr); -- 1.7.2

2 2

[libvirt] [PATCH] qemu: Fix JSON migrate_set_downtime command
by Jiri Denemark 20 Aug '10

20 Aug '10

--- src/qemu/qemu_monitor_json.c | 9 ++------- 1 files changed, 2 insertions(+), 7 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index e8609aa..8a586bc 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -1481,17 +1481,12 @@ int qemuMonitorJSONSetMigrationDowntime(qemuMonitorPtr mon, unsigned long long downtime) { int ret; - char *downtimestr; virJSONValuePtr cmd; virJSONValuePtr reply = NULL; - if (virAsprintf(&downtimestr, "%llums", downtime) < 0) { - virReportOOMError(); - return -1; - } + cmd = qemuMonitorJSONMakeCommand("migrate_set_downtime", - "s:value", downtimestr, + "d:value", downtime / 1000.0, NULL); - VIR_FREE(downtimestr); if (!cmd) return -1; -- 1.7.2

3 5

[libvirt] how to Create virtual machine
by Saravanan S 20 Aug '10

20 Aug '10

Hi, I am working on a virtualization project that involves setting up virtual appliances to implement the same, which tool is required to 1) create virtual appliances 2) deploy virtual appliances 3) manage virtual appliances/virtual machines Could you please guide me with some pointers or resources to get started with the creation of virtual machines? I am little confused with libvirt with Xen and qemu. how does xen or qemu related with libvirt ? Thank you. Saravanan Sundaramoorthy Red Hat Certified Engineer http://www.google.com/profiles/dearsaravanan#about +91 99404 32545

1 0

[libvirt] [PATCH 0/2] qemu: Update next usable PCI slot on domain reconnect
by Jiri Denemark 20 Aug '10

20 Aug '10

When libvirtd restarted and reconnected to existing domains, we didn't update nextslot so that hotplug starts with a slot next to the last used one. Jiri Denemark (2): qemu: Use structured PCI address as hash payload qemu: Update next usable PCI slot on domain reconnect src/qemu/qemu_conf.c | 118 +++++++++++++++++++++++++++++++++++++----------- src/qemu/qemu_conf.h | 1 + src/qemu/qemu_driver.c | 2 + 3 files changed, 94 insertions(+), 27 deletions(-) -- 1.7.2

3 14

[libvirt] [PATCH 00/10] start tackling *printf issues
by Eric Blake 20 Aug '10

20 Aug '10

Based on the long thread about whether %zu and %llu are safe to use, here's a series of patches to start tackling the problems. I've got more work to do, but had enough ready that it was worth posting this much for review, and can push in pieces as individual patches are ACK'd. Eric Blake (10): uml: fix logic bug in checking reply length nwfilter: use consistent OOM reporting build: delete dead comment maint: whitespace cleanups storage: avoid s[n]printf squash to dead comment xenapi: avoid sprintf vbox: add location used in rpmfusion release vbox: factor a large function vbox: avoid sprintf configure.ac | 1 + src/nwfilter/nwfilter_driver.c | 6 +- src/qemu/qemu_driver.c | 1 - src/storage/storage_backend.c | 16 ++- src/storage/storage_backend_disk.c | 124 +++++++++----- src/uml/uml_driver.c | 8 +- src/vbox/vbox_tmpl.c | 327 +++++++++++++++++++----------------- src/xen/sexpr.c | 8 - src/xenapi/xenapi_utils.c | 15 +- src/xenapi/xenapi_utils.h | 4 - 10 files changed, 277 insertions(+), 233 deletions(-) -- 1.7.2.1

3 30

[libvirt] [PATCH] Add support for Vendor and Model in Storage Pool XML
by Patrick Dignan 20 Aug '10

20 Aug '10

Hi all, I wrote a patch to add support for listing the Vendor and Model of a storage pool in the storage pool XML. This would allow vendor extensions of specific devices. The patch includes a test for the new attributes as well. I'd appreciate some feedback on it, and would like get it merged when it's ready. Patch added at the bottom of the message. Best, Patrick Dignan diff -Naurb libvirt-0.8.2.mod/docs/schemas/storagepool.rng libvirt-0.8.2.ml/docs/schemas/storagepool.rng --- libvirt-0.8.2.mod/docs/schemas/storagepool.rng 2010-06-16 17:27:22.000000000 -0500 +++ libvirt-0.8.2.ml/docs/schemas/storagepool.rng 2010-07-30 17:19:32.835531455 -0500 @@ -8,6 +8,10 @@ <define name='pool'> <element name='pool'> + <optional> + <ref name='poolvendor'/> + <ref name='poolmodel'/> + </optional> <choice> <ref name='pooldir'/> <ref name='poolfs'/> @@ -103,6 +107,18 @@ <ref name='target'/> </define> + <define name='poolvendor'> + <attribute name='vendor'> + <text/> + </attribute> + </define> + + <define name='poolmodel'> + <attribute name='model'> + <text/> + </attribute> + </define> + <define name='commonmetadata'> <element name='name'> <ref name='name'/> diff -Naurb libvirt-0.8.2.mod/include/libvirt/libvirt.h libvirt-0.8.2.ml/include/libvirt/libvirt.h --- libvirt-0.8.2.mod/include/libvirt/libvirt.h 2010-07-22 10:26:05.000000000 -0500 +++ libvirt-0.8.2.ml/include/libvirt/libvirt.h 2010-07-30 11:56:48.726415607 -0500 @@ -1141,6 +1141,8 @@ typedef struct _virStoragePoolInfo virStoragePoolInfo; struct _virStoragePoolInfo { + char *vendor; + char *model; int state; /* virStoragePoolState flags */ unsigned long long capacity; /* Logical size bytes */ unsigned long long allocation; /* Current allocation bytes */ diff -Naurb libvirt-0.8.2.mod/src/conf/storage_conf.c libvirt-0.8.2.ml/src/conf/storage_conf.c --- libvirt-0.8.2.mod/src/conf/storage_conf.c 2010-06-16 17:27:22.000000000 -0500 +++ libvirt-0.8.2.ml/src/conf/storage_conf.c 2010-08-02 11:52:48.618533010 -0500 @@ -298,6 +298,12 @@ VIR_FREE(def->name); + if (def->vendor != "generic") + VIR_FREE(def->vendor); + + if (def->model != "generic") + VIR_FREE(def->model); + virStoragePoolSourceFree(&def->source); VIR_FREE(def->target.path); @@ -601,6 +607,8 @@ virStoragePoolDefPtr ret; xmlNodePtr source_node; char *type = NULL; + char *vendor = NULL; + char *model = NULL; char *uuid = NULL; char *tmppath; @@ -616,6 +624,22 @@ goto cleanup; } + /* Get the vendor and model from the XML and set them in the struct */ + vendor = virXPathString("string(./@vendor)", ctxt); + model = virXPathString("string(./@model)", ctxt); + + if (vendor != NULL) { + ret->vendor = vendor; + } else { + ret->vendor = "generic"; + } + + if (model != NULL) { + ret->model = model; + } else { + ret->model = "generic"; + } + xmlFree(type); type = NULL; @@ -861,7 +885,13 @@ "%s", _("unexpected pool type")); goto cleanup; } + + if (def->vendor == NULL || def->model == NULL || STREQ(def->vendor,"generic") || STREQ(def->model,"generic")) { virBufferVSprintf(&buf, "<pool type='%s'>\n", type); + } else { + virBufferVSprintf(&buf, "<pool type='%s' vendor='%s' model='%s'>\n", type, def->vendor, def->model); + } + virBufferVSprintf(&buf," <name>%s</name>\n", def->name); virUUIDFormat(def->uuid, uuid); diff -Naurb libvirt-0.8.2.mod/src/conf/storage_conf.h libvirt-0.8.2.ml/src/conf/storage_conf.h --- libvirt-0.8.2.mod/src/conf/storage_conf.h 2010-06-16 17:27:22.000000000 -0500 +++ libvirt-0.8.2.ml/src/conf/storage_conf.h 2010-07-29 16:37:28.333458251 -0500 @@ -256,6 +256,8 @@ char *name; unsigned char uuid[VIR_UUID_BUFLEN]; int type; /* virStoragePoolType */ + char *vendor; + char *model; unsigned long long allocation; unsigned long long capacity; diff -Naurb libvirt-0.8.2.mod/tests/storagepoolxml2xmlin/pool-iscsi-vendor-model.xml libvirt-0.8.2.ml/tests/storagepoolxml2xmlin/pool-iscsi-vendor-model.xml --- libvirt-0.8.2.mod/tests/storagepoolxml2xmlin/pool-iscsi-vendor-model.xml 1969-12-31 18:00:00.000000000 -0600 +++ libvirt-0.8.2.ml/tests/storagepoolxml2xmlin/pool-iscsi-vendor-model.xml 2010-08-02 12:00:37.015538583 -0500 @@ -0,0 +1,17 @@ +<pool type='iscsi' vendor='test-vendor' model='test-model'> + <name>virtimages</name> + <uuid>e9392370-2917-565e-692b-d057f46512d6</uuid> + <source> + <host name="iscsi.example.com"/> + <device path="demo-target"/> + <auth type='chap' login='foobar' passwd='frobbar'/> + </source> + <target> + <path>/dev/disk/by-path</path> + <permissions> + <mode>0700</mode> + <owner>0</owner> + <group>0</group> + </permissions> + </target> +</pool> diff -Naurb libvirt-0.8.2.mod/tests/storagepoolxml2xmlout/pool-iscsi-vendor-model.xml libvirt-0.8.2.ml/tests/storagepoolxml2xmlout/pool-iscsi-vendor-model.xml --- libvirt-0.8.2.mod/tests/storagepoolxml2xmlout/pool-iscsi-vendor-model.xml 1969-12-31 18:00:00.000000000 -0600 +++ libvirt-0.8.2.ml/tests/storagepoolxml2xmlout/pool-iscsi-vendor-model.xml 2010-08-02 12:16:19.571537734 -0500 @@ -0,0 +1,20 @@ +<pool type='iscsi' vendor='test-vendor' model='test-model'> + <name>virtimages</name> + <uuid>e9392370-2917-565e-692b-d057f46512d6</uuid> + <capacity>0</capacity> + <allocation>0</allocation> + <available>0</available> + <source> + <host name='iscsi.example.com'/> + <device path='demo-target'/> + <auth type='chap' login='foobar' passwd='frobbar'/> + </source> + <target> + <path>/dev/disk/by-path</path> + <permissions> + <mode>0700</mode> + <owner>0</owner> + <group>0</group> + </permissions> + </target> +</pool> diff -Naurb libvirt-0.8.2.mod/tests/storagepoolxml2xmltest.c libvirt-0.8.2.ml/tests/storagepoolxml2xmltest.c --- libvirt-0.8.2.mod/tests/storagepoolxml2xmltest.c 2010-06-16 17:27:22.000000000 -0500 +++ libvirt-0.8.2.ml/tests/storagepoolxml2xmltest.c 2010-08-02 12:32:47.838532225 -0500 @@ -96,6 +96,7 @@ DO_TEST("pool-scsi"); DO_TEST("pool-mpath"); DO_TEST("pool-iscsi-multiiqn"); + DO_TEST("pool-iscsi-vendor-model"); return (ret==0 ? EXIT_SUCCESS : EXIT_FAILURE); }

5 11

[libvirt] [PATCH] Support virDomainAttachDevice and virDomainDetachDevice for disks in UML
by soren＠linux2go.dk 19 Aug '10

19 Aug '10

From: Soren Hansen <soren(a)linux2go.dk> UML supports hot plugging and unplugging of various devices. This patch exposes this functionality for disks. Signed-off-by: Soren Hansen <soren(a)linux2go.dk> --- src/uml/uml_driver.c | 223 +++++++++++++++++++++++++++++++++++++++++++++++++- 1 files changed, 221 insertions(+), 2 deletions(-) diff --git a/src/uml/uml_driver.c b/src/uml/uml_driver.c index 04493ba..a09bc60 100644 --- a/src/uml/uml_driver.c +++ b/src/uml/uml_driver.c @@ -1686,6 +1686,225 @@ cleanup: } +static inline void umlShrinkDisks(virDomainDefPtr def, size_t i) +{ + if (def->ndisks > 1) { + memmove(def->disks + i, + def->disks + i + 1, + sizeof(*def->disks) * + (def->ndisks - (i + 1))); + def->ndisks--; + if (VIR_REALLOC_N(def->disks, def->ndisks) < 0) { + /* ignore, harmless */ + } + } else { + VIR_FREE(def->disks); + def->ndisks = 0; + } +} + + +static int umlDomainAttachUmlDisk(struct uml_driver *driver, + virDomainObjPtr vm, + virDomainDiskDefPtr disk) +{ + int i, ret; + char *cmd = NULL; + char *reply; + + for (i = 0 ; i < vm->def->ndisks ; i++) { + if (STREQ(vm->def->disks[i]->dst, disk->dst)) { + umlReportError(VIR_ERR_OPERATION_FAILED, + _("target %s already exists"), disk->dst); + return -1; + } + } + + if (!disk->src) { + umlReportError(VIR_ERR_INTERNAL_ERROR, + "%s", _("disk source path is missing")); + goto error; + } + + if (virAsprintf(&cmd, "config %s=%s", disk->dst, disk->src) < 0) { + virReportOOMError(); + return -1; + } + + if (umlMonitorCommand(driver, vm, cmd, &reply) < 0) + return -1; + + if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks+1) < 0) { + virReportOOMError(); + goto error; + } + + if (ret < 0) + goto error; + + virDomainDiskInsertPreAlloced(vm->def, disk); + + VIR_FREE(cmd); + + return 0; + +error: + + VIR_FREE(cmd); + + return -1; +} + + +static int umlDomainAttachDevice(virDomainPtr dom, const char *xml) +{ + struct uml_driver *driver = dom->conn->privateData; + virDomainObjPtr vm; + virDomainDeviceDefPtr dev = NULL; + int ret = -1; + + umlDriverLock(driver); + + vm = virDomainFindByUUID(&driver->domains, dom->uuid); + if (!vm) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(dom->uuid, uuidstr); + umlReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching uuid '%s'"), uuidstr); + goto cleanup; + } + + if (!virDomainObjIsActive(vm)) { + umlReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("cannot attach device on inactive domain")); + goto cleanup; + } + + dev = virDomainDeviceDefParse(driver->caps, vm->def, xml, + VIR_DOMAIN_XML_INACTIVE); + + if (dev == NULL) + goto cleanup; + + if (dev->type == VIR_DOMAIN_DEVICE_DISK) { + if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_UML) { + ret = umlDomainAttachUmlDisk(driver, vm, dev->data.disk); + if (ret == 0) + dev->data.disk = NULL; + } else { + umlReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("disk bus '%s' cannot be hotplugged."), + virDomainDiskBusTypeToString(dev->data.disk->bus)); + } + } else { + umlReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("device type '%s' cannot be attached"), + virDomainDeviceTypeToString(dev->type)); + goto cleanup; + } + +cleanup: + + virDomainDeviceDefFree(dev); + if (vm) + virDomainObjUnlock(vm); + umlDriverUnlock(driver); + return ret; +} + + +static int umlDomainDetachUmlDisk(struct uml_driver *driver, + virDomainObjPtr vm, + virDomainDeviceDefPtr dev) +{ + int i, ret = -1; + virDomainDiskDefPtr detach = NULL; + char *cmd; + char *reply; + + for (i = 0 ; i < vm->def->ndisks ; i++) { + if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) { + break; + } + } + + if (i == vm->def->ndisks) { + umlReportError(VIR_ERR_OPERATION_FAILED, + _("disk %s not found"), dev->data.disk->dst); + return -1; + } + + detach = vm->def->disks[i]; + + if (virAsprintf(&cmd, "remove %s", detach->dst) < 0) { + virReportOOMError(); + return -1; + } + + if (umlMonitorCommand(driver, vm, cmd, &reply) < 0) + goto cleanup; + + umlShrinkDisks(vm->def, i); + + virDomainDiskDefFree(detach); + + ret = 0; + +cleanup: + VIR_FREE(cmd); + + return ret; +} + + +static int umlDomainDetachDevice(virDomainPtr dom, const char *xml) { + struct uml_driver *driver = dom->conn->privateData; + virDomainObjPtr vm; + virDomainDeviceDefPtr dev = NULL; + int ret = -1; + + umlDriverLock(driver); + vm = virDomainFindByUUID(&driver->domains, dom->uuid); + if (!vm) { + char uuidstr[VIR_UUID_STRING_BUFLEN]; + virUUIDFormat(dom->uuid, uuidstr); + umlReportError(VIR_ERR_NO_DOMAIN, + _("no domain with matching uuid '%s'"), uuidstr); + goto cleanup; + } + + if (!virDomainObjIsActive(vm)) { + umlReportError(VIR_ERR_OPERATION_INVALID, + "%s", _("cannot detach device on inactive domain")); + goto cleanup; + } + + dev = virDomainDeviceDefParse(driver->caps, vm->def, xml, + VIR_DOMAIN_XML_INACTIVE); + if (dev == NULL) + goto cleanup; + + if (dev->type == VIR_DOMAIN_DEVICE_DISK && + dev->data.disk->device == VIR_DOMAIN_DISK_DEVICE_DISK) { + if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_UML) + ret = umlDomainDetachUmlDisk(driver, vm, dev); + else { + umlReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("This type of disk cannot be hot unplugged")); + } + } else { + umlReportError(VIR_ERR_CONFIG_UNSUPPORTED, + "%s", _("This type of device cannot be hot unplugged")); + } + +cleanup: + virDomainDeviceDefFree(dev); + if (vm) + virDomainObjUnlock(vm); + umlDriverUnlock(driver); + return ret; +} + static int umlDomainGetAutostart(virDomainPtr dom, int *autostart) { @@ -1900,9 +2119,9 @@ static virDriver umlDriver = { umlDomainStartWithFlags, /* domainCreateWithFlags */ umlDomainDefine, /* domainDefineXML */ umlDomainUndefine, /* domainUndefine */ - NULL, /* domainAttachDevice */ + umlDomainAttachDevice, /* domainAttachDevice */ NULL, /* domainAttachDeviceFlags */ - NULL, /* domainDetachDevice */ + umlDomainDetachDevice, /* domainDetachDevice */ NULL, /* domainDetachDeviceFlags */ NULL, /* domainUpdateDeviceFlags */ umlDomainGetAutostart, /* domainGetAutostart */ -- 1.7.0.4

3 2

[libvirt] [PATCH] build: fix compiler warning
by Eric Blake 18 Aug '10

18 Aug '10

node_device/node_device_driver.c: In function 'nodeDeviceVportCreateDelete': node_device/node_device_driver.c:423: error: implicit declaration of function 'stat' [-Wimplicit-function-declaration] * src/node_device/node_device_driver.c (includes): Add <sys/stat.h>. --- Pushing as obvious under the build-breaker rule. src/node_device/node_device_driver.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/node_device/node_device_driver.c b/src/node_device/node_device_driver.c index a6ac80b..448cfd3 100644 --- a/src/node_device/node_device_driver.c +++ b/src/node_device/node_device_driver.c @@ -1,6 +1,7 @@ /* * node_device.c: node device enumeration * + * Copyright (C) 2010 Red Hat, Inc. * Copyright (C) 2008 Virtual Iron Software, Inc. * Copyright (C) 2008 David F. Lively * @@ -27,6 +28,7 @@ #include <errno.h> #include <fcntl.h> #include <time.h> +#include <sys/stat.h> #include "virterror_internal.h" #include "datatypes.h" -- 1.7.2.1

1 0

[libvirt] virsh attach-disk using files and KVM
by Ryan Harper 18 Aug '10

18 Aug '10

Currently virsh attach-disk interface only accepts 'file' or 'tap' for driver type when attaching files as disks. One can succesfully attach a file as disk with: virsh attach-disk <vm> <file> <drive> --driver file --type disk which generates the following xml which is passed to libvirt: <disk type='file' device='disk'> <driver name='file' type='raw'/> <source file='/images/test02.img'/> <target dev='vdc' bus='virtio'/> <alias name='virtio-disk2'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x07' function='0x0'/> </disk> Now, if you shutdown the guest and restart, libvirt complains that the driver type 'file' isn't supported. This is from src/qemu/qemu_conf.c:4146 where if the driver name isn't 'qemu' it rejects the configuration. How best to resolve this? Update qemu_conf.c to accept 'file' type? update virsh to allow specifying 'qemu' as a driver type for files? -- Ryan Harper Software Engineer; Linux Technology Center IBM Corp., Austin, Tx ryanh(a)us.ibm.com

2 2

[libvirt] [PATCH 0/5] Salvage old fixes
by Jiri Denemark 18 Aug '10

18 Aug '10

The following fixes did never make it to upstream libvirt Daniel Berrange (2): remote: Fix incorrect use of private data field xen: Fix device count on detach Daniel Veillard (1): xen: Fix scheduler setting problems Dave Allan (1): nodedev: Fix sysfs paths for vport operations Jiri Denemark (1): nodedev: Free the right pointers when getting WWNs fails src/conf/node_device_conf.c | 4 +- src/node_device/node_device_driver.c | 21 +++++++++++++++++++ src/node_device/node_device_driver.h | 2 +- src/node_device/node_device_linux_sysfs.c | 31 +++++++++++++++++++++++----- src/remote/remote_driver.c | 10 +++++++- src/xen/xen_driver.c | 3 +- src/xen/xen_hypervisor.c | 4 +- src/xen/xm_internal.c | 2 + 8 files changed, 63 insertions(+), 14 deletions(-) -- 1.7.2

3 11

[libvirt] [PATCH 0/3] Managed save updates
by Chris Lalancette 18 Aug '10

18 Aug '10

A couple of updates to the managed save code in qemu. Details are in the individual patches.

3 13

[libvirt] cygwin: #include netinet/inet.h -- needed only on cygwin
by Stefan Berger 18 Aug '10

18 Aug '10

When doing './autogen.sh --with-sasl --with-libvirtd=no' on cygwin, I needed to #include netinet/in.h for a successful build. Now authentication using SASL works. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com> diff --git a/src/remote/qemu_protocol.h b/src/remote/qemu_protocol.h index b822187..8d806fc 100644 --- a/src/remote/qemu_protocol.h +++ b/src/remote/qemu_protocol.h @@ -6,6 +6,9 @@ #ifndef _RP_QEMU_H_RPCGEN #define _RP_QEMU_H_RPCGEN +#ifdef __CYGWIN__ +# include <netinet/in.h> +#endif #include <rpc/rpc.h> diff --git a/src/remote/remote_protocol.h b/src/remote/remote_protocol.h index afe9287..113818d 100644 --- a/src/remote/remote_protocol.h +++ b/src/remote/remote_protocol.h @@ -6,6 +6,9 @@ #ifndef _RP_H_RPCGEN #define _RP_H_RPCGEN +#ifdef __CYGWIN__ +# include <netinet/in.h> +#endif #include <rpc/rpc.h>

4 6

Re: [libvirt] how to print size_t in LGPLv2+ program
by Eric Blake 18 Aug '10

18 Aug '10

On 08/17/2010 03:47 PM, Paul Eggert wrote: > On 08/17/10 23:17, Eric Blake wrote: > >> libvirt is currently LGPLv2+ > > Can this be changed to LPGLv3+? That'd solve the problem. True, but it would have knock-on effects on all sorts of other clients of libvirt. I'm re-adding the libvirt list for thoughts from anyone else. > >> Is it worth relaxing the license on the *printf-posix family of modules >> to LGPLv2+ from their current LGPLv3+, or is this too big of a request? > > I dunno, at some point we should be encouraging people to switch > to LGPLv3+. It makes sense for the FSF to request that GNU projects use better licenses. Unfortunately, libvirt is not a GNU project; so it is more a question as to how Red Hat wants to license it, rather than what the FSF prefers. At any rate, I think the ultimate decision on how libvirt is licensed is rather political in nature, and will involve quite a bit of discussion among more than just the software developers contributing to libvirt. But at least libvirt is LGPLv2+; unlike some projects (like git) that are explicitly v2-only and can't do the upgrade. Meanwhile, I tend to agree that relaxing printf-posix to LGPLv2+ is a rather hefty sledge-hammer, and probably not appropriate for gnulib. > >> since using umaxtostr penalizes 32-bit machines for converting to the >> 64-bit intermediary, maybe it's worth adding a size_t variant? > > That would be fine. The *tostr routines were mainly written > for two reasons. First, for platforms that can't/couldn't portably > and reliably output wide integers. Second, speed. > >> What a shame that POSIX omitted an <inttypes.h> PRIu* for size_t. > > We could add one in gnulib, no? In a quick google search, I found this use of PRIdS (which is geared more for ssize_t, but the analog PRIuS would apply to size_t): http://bytes.com/topic/c/answers/506972-64-bit-portability-size_t-printf-fo… However, S is awfully short; I'd prefer something a bit longer like PRIuSIZE. Is there any other project already using gnulib that has invented a PRIuXXX name for size_t? It would also be nice if we had some prior art from glibc or even one of the BSD systems to copy a commonly-used name. Meanwhile, the idea is slightly more appealing than relicensing lots of gnulib or using the inttostr code; but it still requires auditing code and forbidding "%zu" in favor of "%"PRIuSIZE (or whatever other name we settle on). Also, it seems like we might want to create a new intttypes-plus (or inttypes-gnu) module, and leave the existing inttypes module as providing just the namespace guaranteed by glibc, while requiring the use of the new module to pollute the namespace with these new (but useful) macros. -- Eric Blake eblake(a)redhat.com +1-801-349-2682 Libvirt virtualization library http://libvirt.org

4 3

[libvirt] [PATCH] esx: Fix memory leak when looking up an non-existing domain by name
by Matthias Bolte 17 Aug '10

17 Aug '10

In case an optional object cannot be found the lookup function is left early and the cleanup code is not executed. Add a success label and goto instead of an early return. This pattern occurs in some other functions too. --- src/esx/esx_vi.c | 14 ++++++++------ 1 files changed, 8 insertions(+), 6 deletions(-) diff --git a/src/esx/esx_vi.c b/src/esx/esx_vi.c index c6b2b63..b064b11 100644 --- a/src/esx/esx_vi.c +++ b/src/esx/esx_vi.c @@ -2254,7 +2254,7 @@ esxVI_LookupVirtualMachineByUuid(esxVI_Context *ctx, const unsigned char *uuid, if (managedObjectReference == NULL) { if (occurrence == esxVI_Occurrence_OptionalItem) { - return 0; + goto success; } else { ESX_VI_ERROR(VIR_ERR_NO_DOMAIN, _("Could not find domain with UUID '%s'"), @@ -2270,6 +2270,7 @@ esxVI_LookupVirtualMachineByUuid(esxVI_Context *ctx, const unsigned char *uuid, goto cleanup; } + success: result = 0; cleanup: @@ -2327,7 +2328,7 @@ esxVI_LookupVirtualMachineByName(esxVI_Context *ctx, const char *name, if (*virtualMachine == NULL) { if (occurrence == esxVI_Occurrence_OptionalItem) { - return 0; + goto success; } else { ESX_VI_ERROR(VIR_ERR_NO_DOMAIN, _("Could not find domain with name '%s'"), name); @@ -2335,6 +2336,7 @@ esxVI_LookupVirtualMachineByName(esxVI_Context *ctx, const char *name, } } + success: result = 0; cleanup: @@ -2890,7 +2892,7 @@ esxVI_LookupCurrentSnapshotTree if (currentSnapshot == NULL) { if (occurrence == esxVI_Occurrence_OptionalItem) { - return 0; + goto success; } else { ESX_VI_ERROR(VIR_ERR_NO_DOMAIN_SNAPSHOT, "%s", _("Domain has no current snapshot")); @@ -2911,6 +2913,7 @@ esxVI_LookupCurrentSnapshotTree goto cleanup; } + success: result = 0; cleanup: @@ -3050,9 +3053,7 @@ esxVI_LookupFileInfoByDatastorePath(esxVI_Context *ctx, /* Interpret search result */ if (searchResults->file == NULL) { if (occurrence == esxVI_Occurrence_OptionalItem) { - result = 0; - - goto cleanup; + goto success; } else { ESX_VI_ERROR(VIR_ERR_NO_STORAGE_VOL, _("No storage volume with key or path '%s'"), @@ -3064,6 +3065,7 @@ esxVI_LookupFileInfoByDatastorePath(esxVI_Context *ctx, *fileInfo = searchResults->file; searchResults->file = NULL; + success: result = 0; cleanup: -- 1.7.0.4

2 4

[libvirt] [PATCH 0/4]
by Daniel P. Berrange 17 Aug '10

17 Aug '10

For https://bugzilla.redhat.com/show_bug.cgi?id=620847 We have had sporadic reports of # virsh capabilities error: failed to get capabilities error: server closed connection: This normally means that libvirtd has crashed, closing the connection but in this case libvirtd has always remained running. It turns out that the capabilities XML was too large for the remote RPC message size. This caused XDR serialization to fail. This caused libvirtd to close the client connection immediately. The cause of the large XML was node handling an edge case in libnuma where it returns a CPU mask of all-1s to indicate a non-existant node. Machines that exhibit this problem will show this as a symptom in the logs # grep NUMA /var/log/messages Aug 16 10:30:34 sgi-xe270-01 libvirtd: 10:30:34.933: warning : nodeCapsInitNUMA:388 : NUMA topology for cell 1 of 2 not available, ignoring And have sparse NUMA topology (ie empty nodes) This series does many things: - Adds explicit warnings in places where XDR serialization fails, so we see an indication of problem in /var/log/messages - Try to send a real remote_error back to client, instead of closing its connection - Add logging of capabilities XML in libvirt.c so we can identify the too large doc in libvirtd - Add fix to cope with all-1s node mask This may also fix some other unexplained bug reports we've had with 'server closed connection' messages, or at least make it possible to diagnose them Daniel

3 6

[libvirt] [PATCH] Add support for -enable-kqemu flag
by Daniel P. Berrange 17 Aug '10

17 Aug '10

Previously QEMU enabled KQEMU by default and had -no-kqemu. 0.11.x switched to requiring -enable-kqemu. 0.12.x dropped kqemu entirely. This patch adds support for -enable-kqemu so 0.11.x works. It replaces a huge set of if() with a switch() to make the code a bit more readable. This patch was inspired by the one John Lumby send to the list a while ago, but the way I've implemented it is more in line with the logic we have for KVM. http://www.redhat.com/archives/libvir-list/2010-July/msg00368.html * src/qemu/qemu_conf.c, src/qemu/qemu_conf.h: Support -enable-kqemu --- src/qemu/qemu_conf.c | 84 +++++++++++++++++++++++++++++++++---------------- src/qemu/qemu_conf.h | 1 + 2 files changed, 57 insertions(+), 28 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index fb85220..0ddf128 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -1156,6 +1156,8 @@ static unsigned long long qemudComputeCmdFlags(const char *help, if (strstr(help, "-no-kqemu")) flags |= QEMUD_CMD_FLAG_KQEMU; + if (strstr(help, "-enable-kqemu")) + flags |= QEMUD_CMD_FLAG_ENABLE_KQEMU; if (strstr(help, "-no-kvm")) flags |= QEMUD_CMD_FLAG_KVM; if (strstr(help, "-enable-kvm")) @@ -3662,6 +3664,7 @@ int qemudBuildCommandLine(virConnectPtr conn, char boot[VIR_DOMAIN_BOOT_LAST]; struct utsname ut; int disableKQEMU = 0; + int enableKQEMU = 0; int disableKVM = 0; int enableKVM = 0; int qargc = 0, qarga = 0; @@ -3711,38 +3714,59 @@ int qemudBuildCommandLine(virConnectPtr conn, emulator = def->emulator; - /* Need to explicitly disable KQEMU if - * 1. Guest domain is 'qemu' - * 2. The qemu binary has the -no-kqemu flag - */ - if ((qemuCmdFlags & QEMUD_CMD_FLAG_KQEMU) && - def->virtType == VIR_DOMAIN_VIRT_QEMU) - disableKQEMU = 1; - - /* Need to explicitly disable KVM if - * 1. Guest domain is 'qemu' - * 2. The qemu binary has the -no-kvm flag + /* + * do not use boot=on for drives when not using KVM since this + * is not supported at all in upstream QEmu. */ if ((qemuCmdFlags & QEMUD_CMD_FLAG_KVM) && - def->virtType == VIR_DOMAIN_VIRT_QEMU) { - disableKVM = 1; + (def->virtType == VIR_DOMAIN_VIRT_QEMU) && + (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE_BOOT)) + qemuCmdFlags -= QEMUD_CMD_FLAG_DRIVE_BOOT; + + switch (def->virtType) { + case VIR_DOMAIN_VIRT_QEMU: + if (qemuCmdFlags & QEMUD_CMD_FLAG_KQEMU) + disableKQEMU = 1; + if (qemuCmdFlags & QEMUD_CMD_FLAG_KVM) + disableKVM = 1; + break; - /* - * do not use boot=on for drives when not using KVM since this - * is not supported at all in upstream QEmu. - */ - if (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE_BOOT) - qemuCmdFlags -= QEMUD_CMD_FLAG_DRIVE_BOOT; - } + case VIR_DOMAIN_VIRT_KQEMU: + if (qemuCmdFlags & QEMUD_CMD_FLAG_KVM) + disableKVM = 1; - /* Should explicitly enable KVM if - * 1. Guest domain is 'kvm' - * 2. The qemu binary has the -enable-kvm flag - * NOTE: user must be responsible for loading the kvm modules - */ - if ((qemuCmdFlags & QEMUD_CMD_FLAG_ENABLE_KVM) && - def->virtType == VIR_DOMAIN_VIRT_KVM) - enableKVM = 1; + if (qemuCmdFlags & QEMUD_CMD_FLAG_ENABLE_KQEMU) { + enableKQEMU = 1; + } else if (!(qemuCmdFlags & QEMUD_CMD_FLAG_KQEMU)) { + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("the QEMU binary %s does not support kqemu"), + emulator); + } + break; + + case VIR_DOMAIN_VIRT_KVM: + if (qemuCmdFlags & QEMUD_CMD_FLAG_KQEMU) + disableKQEMU = 1; + + if (qemuCmdFlags & QEMUD_CMD_FLAG_ENABLE_KVM) { + enableKVM = 1; + } else if (!(qemuCmdFlags & QEMUD_CMD_FLAG_KVM)) { + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("the QEMU binary %s does not support kvm"), + emulator); + } + break; + + case VIR_DOMAIN_VIRT_XEN: + /* XXX better check for xenner */ + break; + + default: + qemuReportError(VIR_ERR_CONFIG_UNSUPPORTED, + _("the QEMU binary %s does not support %s"), + emulator, virDomainVirtTypeToString(def->virtType)); + break; + } #define ADD_ARG_SPACE \ do { \ @@ -3857,6 +3881,10 @@ int qemudBuildCommandLine(virConnectPtr conn, if (disableKQEMU) ADD_ARG_LIT("-no-kqemu"); + else if (enableKQEMU) { + ADD_ARG_LIT("-enable-kqemu"); + ADD_ARG_LIT("-kernel-kqemu"); + } if (disableKVM) ADD_ARG_LIT("-no-kvm"); if (enableKVM) diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h index 1aa9d2e..4f9c714 100644 --- a/src/qemu/qemu_conf.h +++ b/src/qemu/qemu_conf.h @@ -92,6 +92,7 @@ enum qemud_cmd_flags { QEMUD_CMD_FLAG_PCI_CONFIGFD = (1LL << 36), /* pci-assign.configfd */ QEMUD_CMD_FLAG_NODEFCONFIG = (1LL << 37), /* -nodefconfig */ QEMUD_CMD_FLAG_BOOT_MENU = (1LL << 38), /* -boot menu=on support */ + QEMUD_CMD_FLAG_ENABLE_KQEMU = (1LL << 39), /* Is the -enable-kqemu flag available to "enable KQEMU full virtualization support" */ }; /* Main driver state */ -- 1.7.2.1

2 1

[libvirt] [PATCH] docs: mention domain <clock> improvements
by Eric Blake 17 Aug '10

17 Aug '10

Add documentation for features added a while ago. * docs/formatdomain.html.in (Time keeping): Update documentation of <clock> element to match 0.8.0 addition. --- I've had this sitting on my tree for a while; any suggestions for improvements, or should I at least check it in as-is so that we have a start on this documentation? docs/formatdomain.html.in | 76 ++++++++++++++++++++++++++++++++++++++++++--- 1 files changed, 71 insertions(+), 5 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 5269cc5..f9a153b 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -411,13 +411,18 @@ <pre> ... - <clock offset="localtime"/> + <clock offset="localtime"> + <timer name="rtc" tickpolicy="catchup" track="guest"> + <catchup threshold=123 slew=120 limit=10000/> + </timer> + <timer name="pit" tickpolicy="none"/> + </clock> ...</pre> <dl> <dt><code>clock</code></dt> <dd> - The <code>offset</code> attribute takes three possible + The <code>offset</code> attribute takes four possible values, allowing fine grained control over how the guest clock is synchronized to the host. NB, not all hypervisors support all modes. @@ -435,6 +440,7 @@ <dd> The guest clock will be synchronized to the requested timezone using the <code>timezone</code> attribute. + Since 0.7.7 </dd> <dt><code>variable</code></dt> <dd> @@ -444,14 +450,74 @@ The guest is free to adjust the RTC over time an expect that it will be honoured at next reboot. This is in contrast to 'utc' mode, where the RTC adjustments are - lost at each reboot. + lost at each reboot. Since 0.7.7 </dd> </dl> - NB, at time of writing, only QEMU supports the variable - clock mode, or custom timezones. + A <code>clock</code> may have zero or more + <code>timer</code>sub-elements. Since + 0.8.0 </dd> + <dt><code>timer</code></dt> + <dd> + + Each timer element requires a <code>name</code> attribute, + and has other optional attributes that depend on + the <code>name</code> specified. Various hypervisors + support different combinations of attributes. + + <dl> + <dt><code>name</code></dt> + <dd> + The <code>name</code> attribute selects which timer is + being modified, and can be one of "platform", "pit", + "rtc", "hpet", or "tsc". + </dd> + <dt><code>track</code></dt> + <dd> + The <code>track</code> attribute specifies what the timer + tracks, and can be "boot", "guest", or "wall". + Only valid for <code>name="rtc"</code> + or <code>name="platform"</code>. + </dd> + <dt><code>tickpolicy</code></dt> + <dd> + The <code>tickpolicy</code> attribute determines how + missed ticks in the guest are handled, and can be "delay", + "catchup", "merge", or "discard". If the policy is + "catchup", there can be further details in + the <code>catchup</code> sub-element. + <dl> + <dt><code>catchup</code></dt> + <dd> + The <code>catchup</code> element has three optional + attributes, each a positive integer. The attributes + are <code>threshold</code>, <code>slew</code>, + and <code>limit</code>. + </dd> + </dl> + </dd> + <dt><code>frequency</code></dt> + <dd> + The <code>frequency</code> attribute is an unsigned + integer specifying the frequency at + which <code>name="tsc"</code> runs. + </dd> + <dt><code>mode</code></dt> + <dd> + The <code>mode</code> attribute controls how + the <code>name="tsc"</code> timer is managed, and can be + "auto", "native", "emulate", "paravirt", or "smpsafe". + Other timers are always emulated. + </dd> + <dt><code>present</code></dt> + <dd> + The <code>present</code> attribute can be "yes" or "no" to + specify whether a particular timer is available to the guest. + </dd> + </dl> + </dd> </dl> <h3><a name="elementsDevices">Devices</a></h3> -- 1.7.2.1

2 2

[libvirt] [PATCH 0/4] Fix problem with capabilities XML generation
by Daniel P. Berrange 17 Aug '10

17 Aug '10

We have had sporadic reports of # virsh capabilities error: failed to get capabilities error: server closed connection: This normally means that libvirtd has crashed, closing the connection but in this case libvirtd has always remained running. It turns out that the capabilities XML was too large for the remote RPC message size. This caused XDR serialization to fail. This caused libvirtd to close the client connection immediately. The cause of the large XML was node handling an edge case in libnuma where it returns a CPU mask of all-1s to indicate a non-existant node. This series does many things: - Adds explicit warnings in places where XDR serialization fails, so we see an indication of problem in /var/log/messages - Try to send a real remote_error back to client, instead of closing its connection - Add logging of capabilities XML in libvirt.c so we can identify the too large doc in libvirtd - Add fix to cope with all-1s node mask

3 13

[libvirt] [PATCH] cygwin: build fix
by Stefan Berger 17 Aug '10

17 Aug '10

Fixing a problem in the build on cygwin due to missing #define's. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com> --- src/remote/remote_protocol.c | 6 ++++++ src/remote/remote_protocol.h | 6 ++++++ src/remote/remote_protocol.x | 6 ++++++ 3 files changed, 18 insertions(+) Index: libvirt-acl/src/remote/remote_protocol.h =================================================================== --- libvirt-acl.orig/src/remote/remote_protocol.h +++ libvirt-acl/src/remote/remote_protocol.h @@ -24,6 +24,12 @@ extern "C" { #ifndef IXDR_GET_INT32 # define IXDR_GET_INT32 IXDR_GET_LONG #endif +#ifndef IXDR_PUT_U_INT32 +# define IXDR_PUT_U_INT32 IXDR_PUT_U_LONG +#endif +#ifndef IXDR_GET_U_INT32 +# define IXDR_GET_U_INT32 IXDR_GET_U_LONG +#endif #define REMOTE_MESSAGE_MAX 262144 #define REMOTE_MESSAGE_HEADER_MAX 24 #define REMOTE_MESSAGE_PAYLOAD_MAX 262120 Index: libvirt-acl/src/remote/remote_protocol.x =================================================================== --- libvirt-acl.orig/src/remote/remote_protocol.x +++ libvirt-acl/src/remote/remote_protocol.x @@ -51,6 +51,12 @@ %#ifndef IXDR_GET_INT32 %# define IXDR_GET_INT32 IXDR_GET_LONG %#endif +%#ifndef IXDR_PUT_U_INT32 +%# define IXDR_PUT_U_INT32 IXDR_PUT_U_LONG +%#endif +%#ifndef IXDR_GET_U_INT32 +%# define IXDR_GET_U_INT32 IXDR_GET_U_LONG +%#endif /*----- Data types. -----*/ Index: libvirt-acl/src/remote/remote_protocol.c =================================================================== --- libvirt-acl.orig/src/remote/remote_protocol.c +++ libvirt-acl/src/remote/remote_protocol.c @@ -16,6 +16,12 @@ #ifndef IXDR_GET_INT32 # define IXDR_GET_INT32 IXDR_GET_LONG #endif +#ifndef IXDR_PUT_U_INT32 +# define IXDR_PUT_U_INT32 IXDR_PUT_U_LONG +#endif +#ifndef IXDR_GET_U_INT32 +# define IXDR_GET_U_INT32 IXDR_GET_U_LONG +#endif bool_t xdr_remote_nonnull_string (XDR *xdrs, remote_nonnull_string *objp)

2 2

[libvirt] Potential gotcha in virsh migration with --persistent?
by Justin Clift 17 Aug '10

17 Aug '10

Hi all, Just noticed that when using the virsh --persistent option for migration, that the "autostart" attribute isn't mirrored across to the destination host. This seems like a "gotcha", in that a person migrating an autostarted guest to another host with --persistent will be expecting the migration to persist _all_ of the (important at least) guest settings. Without mirroring this attribute, the person doing the migration needs to manually follow up the migration by subsequently connecting to the destination host, then turning on autostart attribute for the domain there. Guessing it's not a design decision, and is just something we should fix? Regards and best wishes, Justin Clift

2 2

[libvirt] [PATCH 3/3] Remove wrong check for uml monitor response size
by Soren Hansen 16 Aug '10

16 Aug '10

The current check for the size of the response from the uml monitor is problematic for a couple of reasons: First of all, for me, the call to recvfrom returns 0, even though the buffer actually is populated with the response from the monitor. Second, it seems to me that (assuming recvfrom actually returned the number of bytes read) it should be compared against the size of the response header plus the actual data length rather than the max size of the datagram. At any rate, the only way I can get anything useful to happen here is to completely remove the check, so this patch does just that. Signed-off-by: Soren Hansen <soren(a)linux2go.dk> --- src/uml/uml_driver.c | 4 ---- 1 files changed, 0 insertions(+), 4 deletions(-) diff --git a/src/uml/uml_driver.c b/src/uml/uml_driver.c index 04493ba..9cf669f 100644 --- a/src/uml/uml_driver.c +++ b/src/uml/uml_driver.c @@ -737,10 +737,6 @@ static int umlMonitorCommand(const struct uml_driver *driver, virReportSystemError(errno, _("cannot read reply %s"), cmd); goto error; } - if (nbytes < sizeof res) { - virReportSystemError(0, _("incomplete reply %s"), cmd); - goto error; - } if (sizeof res.data < res.length) { virReportSystemError(0, _("invalid length in reply %s"), cmd); goto error; -- 1.7.0.4

4 12

[libvirt] [PATCH 0/3] qemu: Several PCI device assignment fixes
by Jiri Denemark 16 Aug '10

16 Aug '10

Jiri Denemark (3): qemu: Re-reserve all PCI addresses on libvirtd restart qemu: Release PCI slot when detaching disk and net devices qemu: Fix copy&paste error in warning message src/qemu/qemu_driver.c | 18 ++++++++++++++---- 1 files changed, 14 insertions(+), 4 deletions(-) -- 1.7.2

2 7

[libvirt] PATCH 4/4: AppArmor updates
by Jamie Strandboge 16 Aug '10

16 Aug '10

Attached is 0004-apparmor-fix-warn.patch -- Jamie Strandboge | http://www.canonical.com

3 2

[libvirt] [PATCH v3] nwfilter: extend nwfilter reload support
by Stefan Berger 16 Aug '10

16 Aug '10

This time with a proper title. v3: - Fixed an indentation problem - added bool parameter to function terminating the IP address learner threads to determine whether future threads may still run (needed in case of driver reload) or all must terminate (need in case of libvirtd termination) v2: - Fixes to the nwfilter driver reload function that also needs a valid virConnectPtr. In this patch I am extending and fixing the nwfilter module's reload support to stop all ongoing threads (for learning IP addresses of interfaces) and rebuild the filtering rules of all interfaces of all VMs when libvirt is started. Now libvirtd rebuilds the filters upon the SIGHUP signal and libvirtd restart. About the patch: The nwfilter functions require a virConnectPtr. Therefore I am opening a connection in qemudStartup, which later on needs to be closed outside where the driver lock is held since otherwise it ends up in a deadlock due to virConnectClose() trying to lock the driver as well. I have tested this now for a while with several machines running and needing the IP address learner thread(s). The rebuilding of the firewall rules seems to work fine following libvirtd restart or a SIGHUP. Also the termination of libvirtd worked fine. Signed-off-by: Stefan Berger<stefanb(a)us.ibm.com> --- src/nwfilter/nwfilter_driver.c | 21 +++++++++++--- src/nwfilter/nwfilter_learnipaddr.c | 16 ++++++++--- src/nwfilter/nwfilter_learnipaddr.h | 1 src/qemu/qemu_driver.c | 52 +++++++++++++++++++++++++++++++++--- 4 files changed, 77 insertions(+), 13 deletions(-) Index: libvirt-acl/src/nwfilter/nwfilter_driver.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_driver.c +++ libvirt-acl/src/nwfilter/nwfilter_driver.c @@ -143,15 +143,26 @@ conf_init_err: */ static int nwfilterDriverReload(void) { + virConnectPtr conn; + if (!driverState) { return -1; } - nwfilterDriverLock(driverState); - virNWFilterPoolLoadAllConfigs(NULL, - &driverState->pools, - driverState->configDir); - nwfilterDriverUnlock(driverState); + conn = virConnectOpen("qemu:///system"); + + if (conn) { + /* shut down all threads -- qemud for example will restart them */ + virNWFilterLearnThreadsTerminate(true); + + nwfilterDriverLock(driverState); + virNWFilterPoolLoadAllConfigs(conn, + &driverState->pools, + driverState->configDir); + nwfilterDriverUnlock(driverState); + + virConnectClose(conn); + } return 0; } Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.c +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c @@ -857,6 +857,17 @@ virNWFilterLearnInit(void) { } +void +virNWFilterLearnThreadsTerminate(bool allowNewThreads) { + threadsTerminate = true; + + while (virHashSize(pendingLearnReq) != 0) + usleep((PKT_TIMEOUT_MS * 1000) / 3); + + if (allowNewThreads) + threadsTerminate = false; +} + /** * virNWFilterLearnShutdown * Shutdown of this layer @@ -864,10 +875,7 @@ virNWFilterLearnInit(void) { void virNWFilterLearnShutdown(void) { - threadsTerminate = true; - - while (virHashSize(pendingLearnReq) != 0) - usleep((PKT_TIMEOUT_MS * 1000) / 3); + virNWFilterLearnThreadsTerminate(false); virHashFree(pendingLearnReq, freeLearnReqEntry); pendingLearnReq = NULL; Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.h =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.h +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.h @@ -71,5 +71,6 @@ void virNWFilterUnlockIface(const char * int virNWFilterLearnInit(void); void virNWFilterLearnShutdown(void); +void virNWFilterLearnThreadsTerminate(bool allowNewThreads); #endif /* __NWFILTER_LEARNIPADDR_H */ Index: libvirt-acl/src/qemu/qemu_driver.c =================================================================== --- libvirt-acl.orig/src/qemu/qemu_driver.c +++ libvirt-acl/src/qemu/qemu_driver.c @@ -170,6 +170,9 @@ static int qemuDetectVcpuPIDs(struct qem static int qemuUpdateActivePciHostdevs(struct qemud_driver *driver, virDomainDefPtr def); +static int qemudVMFiltersInstantiate(virConnectPtr conn, + virDomainDefPtr def); + static struct qemud_driver *qemu_driver = NULL; @@ -1423,6 +1426,10 @@ error: return ret; } +struct virReconnectDomainData { + virConnectPtr conn; + struct qemud_driver *driver; +}; /* * Open an existing VM's monitor, re-detect VCPU threads * and re-reserve the security labels in use @@ -1431,9 +1438,11 @@ static void qemuReconnectDomain(void *payload, const char *name ATTRIBUTE_UNUSED, void *opaque) { virDomainObjPtr obj = payload; - struct qemud_driver *driver = opaque; + struct virReconnectDomainData *data = opaque; + struct qemud_driver *driver = data->driver; qemuDomainObjPrivatePtr priv; unsigned long long qemuCmdFlags; + virConnectPtr conn = data->conn; virDomainObjLock(obj); @@ -1467,6 +1476,9 @@ qemuReconnectDomain(void *payload, const obj) < 0) goto error; + if (qemudVMFiltersInstantiate(conn, obj->def)) + goto error; + if (obj->def->id >= driver->nextvmid) driver->nextvmid = obj->def->id + 1; @@ -1491,9 +1503,10 @@ error: * about. */ static void -qemuReconnectDomains(struct qemud_driver *driver) +qemuReconnectDomains(virConnectPtr conn, struct qemud_driver *driver) { - virHashForEach(driver->domains.objs, qemuReconnectDomain, driver); + struct virReconnectDomainData data = {conn, driver}; + virHashForEach(driver->domains.objs, qemuReconnectDomain, &data); } @@ -1691,6 +1704,7 @@ qemudStartup(int privileged) { char *base = NULL; char driverConf[PATH_MAX]; int rc; + virConnectPtr conn = NULL; if (VIR_ALLOC(qemu_driver) < 0) return -1; @@ -1912,7 +1926,11 @@ qemudStartup(int privileged) { 1, NULL, NULL) < 0) goto error; - qemuReconnectDomains(qemu_driver); + conn = virConnectOpen(qemu_driver->privileged ? + "qemu:///system" : + "qemu:///session"); + + qemuReconnectDomains(conn, qemu_driver); /* Then inactive persistent configs */ if (virDomainLoadAllConfigs(qemu_driver->caps, @@ -1930,6 +1948,8 @@ qemudStartup(int privileged) { qemudAutostartConfigs(qemu_driver); + if (conn) + virConnectClose(conn); return 0; @@ -1938,6 +1958,8 @@ out_of_memory: error: if (qemu_driver) qemuDriverUnlock(qemu_driver); + if (conn) + virConnectClose(conn); VIR_FREE(base); qemudShutdown(); return -1; @@ -12738,6 +12760,28 @@ qemudVMFilterRebuild(virConnectPtr conn return 0; } +static int +qemudVMFiltersInstantiate(virConnectPtr conn, + virDomainDefPtr def) +{ + int err = 0; + int i; + + if (!conn) + return 1; + + for (i = 0 ; i < def->nnets ; i++) { + virDomainNetDefPtr net = def->nets[i]; + if ((net->filter) && (net->ifname)) { + if (virDomainConfNWFilterInstantiate(conn, net)) { + err = 1; + break; + } + } + } + + return err; +} static virNWFilterCallbackDriver qemuCallbackDriver = { .name = "QEMU",

4 3

[libvirt] [RFC: PATCH 0/4] Update array allocation macros
by Eric Blake 16 Aug '10

16 Aug '10

I'm posing this series as an RFC for the moment. The basic premise is that I got bit trying to use VIR_REALLOC_N when working on the virCommand API, assuming that it zeroed out new memory because the documentation said so, only to hit a core dump. On looking closer, it became obvious to me that the VIR_REALLOC_N API is deficient - there's no way to know how much memory, if any, to zeroize without knowing the original size, so the new memory has to be uninitialized (explaining my crash), and the documentation is wrong. So, after some IRC discussions, two ideas emerged in addition to the obvious documentation update. One is the simpler VIR_EXPAND_N, where the array size always matches the array allocation, along with a VIR_SHRINK_N counterpart which can trim an array back down (and failure to trim need not be fatal). But this can scale quadratically if you end up with a large array, as the existing contents must be copied on every realloc. See patches 1 and 2 for an implementation and use. The other idea is VIR_RESIZE_N, which tracks allocation separately from usage, and has better scaling by doing geometric growth only when needed rather than on every array usage increase. But it takes more parameters (making it slightly harder to remember how to use correctly), along with possible modifications of the struct tracking an array (so it might not be possible to modify all uses of VIR_REALLOC_N to use this, if it would end up altering a public struct layout). See patches 3 and 4 for an implementation and use. Note that event.c is an example of code that already tracked allocation separately from usage, so the new VIR_RESIZE_N was very easy to use. On the other hand, libvirtd.c is an example of code where I had to add allocation tracking; notice the difference between patches 2/4 and 4/4 of the logic differences needed when using VIR_EXPAND_N vs. VIR_RESIZE_N. And notice that both patch 2 and patch 4 provide a net reduction in lines outside of memory.[ch]. Either way, I'm willing to audit all remaining uses of VIR_REALLOC_N (in fact, libvirtd.c has another use of VIR_REALLOC_N to allocate a buffer for getgrnam_r, where the uninitialized memory aspect was just fine and not worth converting). The only reason I'm posting this series as an RFC is to get a feel for which of the two styles I should prefer when doing my audit on the rest of the code base (rather than doing style A now only to find that style B would be preferred). Also, as written, the implementation forces all array count variables to be size_t if using VIR_EXPAND_N or VIR_RESIZE_N; I may come across a public API where I can't change the array count type, in which case I will have to rewrite the macros to be type-agnostic. Well, more like making the macro choose between a 4-byte or 8-byte implementation, based on sizeof(count), and adding a compiler verify() that count must be one of those two sizes, whether on 32-bit or 64-bit machines. Eric Blake (4): memory: make it safer to expand arrays daemon: use safer memory growth macros memory: make it easier to avoid quadratic scaling of arrays daemon: use more efficient memory growth macros daemon/event.c | 44 ++++++++++------------- daemon/libvirtd.c | 8 ++--- daemon/libvirtd.h | 11 +++--- docs/hacking.html.in | 61 ++++++++++++++++++++++++++------ src/util/memory.c | 94 +++++++++++++++++++++++++++++++++++++++++++++++++- src/util/memory.h | 77 +++++++++++++++++++++++++++++++++++++--- 6 files changed, 242 insertions(+), 53 deletions(-) -- 1.7.2.1

2 10

[libvirt] Updated application development guide on libvirt.org
by David Jorm 16 Aug '10

16 Aug '10

I have published an updated draft of the application development guide at: http://libvirt.org/guide/ It includes expanded content, editing and formatting for: * Chapter 3 - Connections * Chapter 4 - Guest Domains * Chapter 7 - Network Interfaces Additional content for Chapter 3 has been provided to me, and will be included soon. Thanks David

1 0

[libvirt] xen live migration with libvirt
by 姚远 16 Aug '10

16 Aug '10

The format of live migration command is "virsh migrate --live domain xen:/// xenmigr://remotehost" or another? My source and destination is rhel5.4 and xen is 3.1.0, libvirt is 0.8.2. I configured a shared storage between source and destination and modified /etc/xen/xend-config.sxp. I test the command "xm migrate --live domain remotehost" and it's going well. But when I use "virsh migrate --live domain xen:/// xenmigr://remotehost", a error appeared. error:post operation failed:xend_post:error from xen daemon:(xend.err '/usr/lib64/xen/bin/xc_save 23 2 0 0 5 failed') Regards, Best wishes.

1 0

[libvirt] [PATCH v3] nwfilter:
by Stefan Berger 14 Aug '10

14 Aug '10

v3: - Fixed an indentation problem - added bool parameter to function terminating the IP address learner threads to determine whether future threads may still run (needed in case of driver reload) or all must terminate (need in case of libvirtd termination) v2: - Fixes to the nwfilter driver reload function that also needs a valid virConnectPtr. In this patch I am extending and fixing the nwfilter module's reload support to stop all ongoing threads (for learning IP addresses of interfaces) and rebuild the filtering rules of all interfaces of all VMs when libvirt is started. Now libvirtd rebuilds the filters upon the SIGHUP signal and libvirtd restart. About the patch: The nwfilter functions require a virConnectPtr. Therefore I am opening a connection in qemudStartup, which later on needs to be closed outside where the driver lock is held since otherwise it ends up in a deadlock due to virConnectClose() trying to lock the driver as well. I have tested this now for a while with several machines running and needing the IP address learner thread(s). The rebuilding of the firewall rules seems to work fine following libvirtd restart or a SIGHUP. Also the termination of libvirtd worked fine. Signed-off-by: Stefan Berger <stefanb us ibm com> --- src/nwfilter/nwfilter_driver.c | 21 +++++++++++--- src/nwfilter/nwfilter_learnipaddr.c | 16 ++++++++--- src/nwfilter/nwfilter_learnipaddr.h | 1 src/qemu/qemu_driver.c | 52 +++++++++++++++++++++++++++++++++--- 4 files changed, 77 insertions(+), 13 deletions(-) Index: libvirt-acl/src/nwfilter/nwfilter_driver.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_driver.c +++ libvirt-acl/src/nwfilter/nwfilter_driver.c @@ -143,15 +143,26 @@ conf_init_err: */ static int nwfilterDriverReload(void) { + virConnectPtr conn; + if (!driverState) { return -1; } - nwfilterDriverLock(driverState); - virNWFilterPoolLoadAllConfigs(NULL, - &driverState->pools, - driverState->configDir); - nwfilterDriverUnlock(driverState); + conn = virConnectOpen("qemu:///system"); + + if (conn) { + /* shut down all threads -- qemud for example will restart them */ + virNWFilterLearnThreadsTerminate(true); + + nwfilterDriverLock(driverState); + virNWFilterPoolLoadAllConfigs(conn, + &driverState->pools, + driverState->configDir); + nwfilterDriverUnlock(driverState); + + virConnectClose(conn); + } return 0; } Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.c +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c @@ -857,6 +857,17 @@ virNWFilterLearnInit(void) { } +void +virNWFilterLearnThreadsTerminate(bool allowNewThreads) { + threadsTerminate = true; + + while (virHashSize(pendingLearnReq) != 0) + usleep((PKT_TIMEOUT_MS * 1000) / 3); + + if (allowNewThreads) + threadsTerminate = false; +} + /** * virNWFilterLearnShutdown * Shutdown of this layer @@ -864,10 +875,7 @@ virNWFilterLearnInit(void) { void virNWFilterLearnShutdown(void) { - threadsTerminate = true; - - while (virHashSize(pendingLearnReq) != 0) - usleep((PKT_TIMEOUT_MS * 1000) / 3); + virNWFilterLearnThreadsTerminate(false); virHashFree(pendingLearnReq, freeLearnReqEntry); pendingLearnReq = NULL; Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.h =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.h +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.h @@ -71,5 +71,6 @@ void virNWFilterUnlockIface(const char * int virNWFilterLearnInit(void); void virNWFilterLearnShutdown(void); +void virNWFilterLearnThreadsTerminate(bool allowNewThreads); #endif /* __NWFILTER_LEARNIPADDR_H */ Index: libvirt-acl/src/qemu/qemu_driver.c =================================================================== --- libvirt-acl.orig/src/qemu/qemu_driver.c +++ libvirt-acl/src/qemu/qemu_driver.c @@ -170,6 +170,9 @@ static int qemuDetectVcpuPIDs(struct qem static int qemuUpdateActivePciHostdevs(struct qemud_driver *driver, virDomainDefPtr def); +static int qemudVMFiltersInstantiate(virConnectPtr conn, + virDomainDefPtr def); + static struct qemud_driver *qemu_driver = NULL; @@ -1423,6 +1426,10 @@ error: return ret; } +struct virReconnectDomainData { + virConnectPtr conn; + struct qemud_driver *driver; +}; /* * Open an existing VM's monitor, re-detect VCPU threads * and re-reserve the security labels in use @@ -1431,9 +1438,11 @@ static void qemuReconnectDomain(void *payload, const char *name ATTRIBUTE_UNUSED, void *opaque) { virDomainObjPtr obj = payload; - struct qemud_driver *driver = opaque; + struct virReconnectDomainData *data = opaque; + struct qemud_driver *driver = data->driver; qemuDomainObjPrivatePtr priv; unsigned long long qemuCmdFlags; + virConnectPtr conn = data->conn; virDomainObjLock(obj); @@ -1467,6 +1476,9 @@ qemuReconnectDomain(void *payload, const obj) < 0) goto error; + if (qemudVMFiltersInstantiate(conn, obj->def)) + goto error; + if (obj->def->id >= driver->nextvmid) driver->nextvmid = obj->def->id + 1; @@ -1491,9 +1503,10 @@ error: * about. */ static void -qemuReconnectDomains(struct qemud_driver *driver) +qemuReconnectDomains(virConnectPtr conn, struct qemud_driver *driver) { - virHashForEach(driver->domains.objs, qemuReconnectDomain, driver); + struct virReconnectDomainData data = {conn, driver}; + virHashForEach(driver->domains.objs, qemuReconnectDomain, &data); } @@ -1691,6 +1704,7 @@ qemudStartup(int privileged) { char *base = NULL; char driverConf[PATH_MAX]; int rc; + virConnectPtr conn = NULL; if (VIR_ALLOC(qemu_driver) < 0) return -1; @@ -1912,7 +1926,11 @@ qemudStartup(int privileged) { 1, NULL, NULL) < 0) goto error; - qemuReconnectDomains(qemu_driver); + conn = virConnectOpen(qemu_driver->privileged ? + "qemu:///system" : + "qemu:///session"); + + qemuReconnectDomains(conn, qemu_driver); /* Then inactive persistent configs */ if (virDomainLoadAllConfigs(qemu_driver->caps, @@ -1930,6 +1948,8 @@ qemudStartup(int privileged) { qemudAutostartConfigs(qemu_driver); + if (conn) + virConnectClose(conn); return 0; @@ -1938,6 +1958,8 @@ out_of_memory: error: if (qemu_driver) qemuDriverUnlock(qemu_driver); + if (conn) + virConnectClose(conn); VIR_FREE(base); qemudShutdown(); return -1; @@ -12738,6 +12760,28 @@ qemudVMFilterRebuild(virConnectPtr conn return 0; } +static int +qemudVMFiltersInstantiate(virConnectPtr conn, + virDomainDefPtr def) +{ + int err = 0; + int i; + + if (!conn) + return 1; + + for (i = 0 ; i < def->nnets ; i++) { + virDomainNetDefPtr net = def->nets[i]; + if ((net->filter) && (net->ifname)) { + if (virDomainConfNWFilterInstantiate(conn, net)) { + err = 1; + break; + } + } + } + + return err; +} static virNWFilterCallbackDriver qemuCallbackDriver = { .name = "QEMU",

4 5

[libvirt] [PATCH] build: allow mingw VPATH build
by Eric Blake 14 Aug '10

14 Aug '10

* .gnulib: Update to latest. Reported by Matthias Bolte. --- Mainly for the top one, but there are a couple of relevant upstream patches in this list. * .gnulib 1629006...7ba06c8 (152): > pthread: fix pthread.h creation for srcdir != builddir > avoid some overlong lines from posix urls, etc. > autoupdate > missed commit > autoupdate > strtod: fix const diagnostic > copy-acl: ignore ENOTSUP on HP-UX > open, chown: relax license > autocheck new ar-lib script from automake > readlinkat: adjust client modules > mknod: be more vocal about danger of running tests as root > readlinkat: split into its own module > memxfrm: Speed up. > erroneous commas inside @var > remove spurious leading i > missing @item inside @itemize > autoupdate > Integrate the regex documentation. > Whitespace cleanup. > Add regex documentation. > link: Update documentation. > ansi-c++-opt: Mention last change in NEWS. > Update modules list. > Improve doc in MODULES.html. > ansi-c++-opt: Provide option --enable-c++/--disable-c++ when possible. > readlink, areadlink: Relax test a bit. > unistr/u8-strstr, unistr/u16-strstr: Optimize the one-character case. > unistr/u*-strstr: Fix dependencies. > unistr/u8-chr, unistr/u8-strchr: Optimize and add comments. > unistr/u8-strchr: Fix several bugs. > More tests for unistr/u8-strchr. > Oops, fix last commit so that it also works with modf(). > posix-modules: Ignore backup files of documentation files. > symlinkat: Fix documentation. > fchownat: Replace also when chown has the trailing slash bug. > Complete last ChangeLog entry. > linkat: Work around AIX 7.1 bug. > Correctly determine whether pow is available in libc on AIX 7 with xlc. > iconv: Work around AIX 6.1..7.1 bug. > readlink: Relax test a bit. > copysign: Does not require -lm on glibc systems. > duplocale: Work around AIX 7.1 bug. > dirfd: Avoid link error on AIX 7.1. > strtod: next round of AIX fixes > futimens: fix configure check > getline: Update regarding AIX. > wcwidth: Drop replacement on AIX 7. > strtok_r: Avoid triggering bug in AIX 7.1 xlc compiler. > unlink: Update regarding AIX. > symlink: Update regarding AIX. > strndup: Update regarding AIX. > stat: Update regarding AIX. > truncl: Fix autoconf test. > round: Update regarding AIX. > rename: Update regarding AIX. > printf.m4: Update regarding AIX. > iconv: Update regarding AIX. > getopt: Update regarding AIX. > ldexpl; Update regarding AIX. > frexpl: Update regarding AIX. > open, fopen: Update regarding AIX. > chown: Update doc regarding AIX. > autoupdate > strtod: fix bug in replacement function on AIX > mbrlen: Fix cross-compilation guess for AIX. > mbrtowc: Fix cross-compilation guess for AIX. > * build-aux/gendocs.sh: restore lost x bit > init.sh: work around trap limitation of some shells > strtod: aid debugging > unistr/u*-chr, unistr/u*-strchr: Fix link errors and warnings. > Add ChangeLog entries for recent commits. > Use spaces for indentation, not tabs. > mbspcasecmp: Fix function specification. > autoupdate > timespec: use cast and not conditional, as truncation isn't possible > unistr/u8-chr, unistr/u8-strchr: use Boyer-Moore like algorithm. > unistr/u*-strchr: add tests > unistr/u*-chr: test multibyte sequences more > unistr/u*-chr: test multibyte sequences > unistr/u*-chr: prepare for multibyte tests > unistr/u8-strchr: Optimize non-ASCII argument case. > fdl.texi from gnustandards > getcwd: on Solaris, work better if ancestors are inaccessible > pthread: Add enough so that coreutils/src/sort.c compiles. > striconveh: Simplify last commit. > striconveh: Don't malloc memory if the result buffer is sufficient. > strtod: Add safety check. > Unify tests that set gl_cv_func_ldexpl_no_libm. > Unify tests that set gl_cv_func_ldexp_no_libm. > Unify tests that set gl_cv_func_frexpl_no_libm. > Unify tests that set gl_cv_func_frexp_no_libm. > memcoll: clarify sizes versus lengths, document better, and tweak perf > Tests for module '_Exit'. > New module '_Exit'. > strtod: make it more-accurate typically, and don't require libm > autoupdate > unistr/u8-strchr: Optimize ASCII argument case. > (x)memcoll: minor tweaks > (x)memcoll: speedup when input is known to be NUL delimited > bootstrap: discard non translation project po files > fsusage: Clarify which code applies to which platforms. > havelib: Fix bug when AC_LIB_FROMPACKAGE is used more than twice. > hash: once again explicitly disallow insertion of NULL > stdbool: Update doc. > hash: extend module to deal with non-pointer keys > gettext: Use AC_GNU_SOURCE as a fallback for AC_USE_SYSTEM_EXTENSIONS. > idpriv-drop: Fix tests. > string: Fix syntax error with g++ 2.96. > unitypes: Fix bug introduced on 2010-05-18. > autoupdate > autoupdates > memmem: slight optimization > Fix HAVE_CALLOC_POSIX misnomer. > Use modern idiom for calloc() replacement. > Fix HAVE_REALLOC_POSIX misnomer. > Use modern idiom for realloc() replacement. > Fix HAVE_MALLOC_POSIX misnomer. > Use modern idiom for malloc() replacement. > stdio.in.h: fix compilation failure when using HP-UX 11's C compiler > update from texinfo > update from texinfo > update from texinfo > ipv6: fix detection under mingw > strtod: Assume strtod() works when cross-compiling to new-enough glibc. > Factor out common code in gl_FUNC_STRTOD. > strtod: Stop using AC_FUNC_STRTOD. > select: Correct timeout. > git-version-gen: init shell var to avoid env var influence > priv-set: Don't assume that priv.h exists merely because getppriv does. > relocatable: Make it easier to test whether to install wrappers. > gnulib-tool: Display specified modules and dependencies differently. > gnulib-tool: Align code of func_import and func_create_testdir. > test-inttostr: avoid spurious failure on Solaris 9 > update from texinfo > test-sys_socket: mark variables as used more readably > Avoid some more warnings from "gcc -Wwrite-strings". > init.sh: change framework_failure_ to fail with status 99, not 1 > test-inttostr: avoid warnings about 4-6KB literal strings > init.sh: don't use $ME_ or skip_ before they are defined > test-sys_socket: avoid set-but-not-used warnings from gcc > test-xvasprintf: avoid 'const' discard warnings > tests: avoid compilation warnings in argmatch and exclude tests... > tests: avoid 'const' discard warnings in mbsstr tests > test-verify: avoid warning from gcc's -Wmissing-declarations > test-inttostr.c: include <string.h> for use of strcmp > test-linkat: avoid failed assertion on "other" architectures > printf.m4: avoid autoconf's "Expanded Before Required" warning > Replacement header templates are now named with ".in", not "_". > inttostr-tests: depend on snprintf, not snprintf-posix > autoupdate > inttostr: add a new function, inttostr, and tests > Add "Extending Gnulib" chapter to manual. .gnulib | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/.gnulib b/.gnulib index 1629006..7ba06c8 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit 1629006348e1f66f07ce3ddcf3ebd2d14556cfce +Subproject commit 7ba06c8ff37d8725cb824c64b94be41b9294ffb7 -- 1.7.2.1

2 3

[libvirt] [PATCH v2] esx: Explicitly disable unused floppy devices
by Matthias Bolte 14 Aug '10

14 Aug '10

floppy0.present defaults to true. Therefore, it needs to be explicitly set to false when the XML config doesn't specify the corresponding floppy device. --- v2: - This issue affects ESX too, the assumed default was wrong in general - Explicitly disable unused floppy devices independent of the product version src/esx/esx_vmx.c | 19 +++++++++++++++---- src/esx/esx_vmx.h | 2 +- 2 files changed, 16 insertions(+), 5 deletions(-) diff --git a/src/esx/esx_vmx.c b/src/esx/esx_vmx.c index d5d9ff0..12cd005 100644 --- a/src/esx/esx_vmx.c +++ b/src/esx/esx_vmx.c @@ -215,7 +215,7 @@ def->disks[0]... ## disks: floppy from .flp image ############################################### - floppy0.present = "true" # defaults to "false" + floppy0.present = "true" # defaults to "true" floppy0.startConnected = "true" # defaults to "true" floppy0.clientDevice = "false" # defaults to "false" @@ -235,7 +235,7 @@ def->disks[0]... ## disks: floppy from host device ############################################## - floppy0.present = "true" # defaults to "false" + floppy0.present = "true" # defaults to "true" floppy0.startConnected = "true" # defaults to "true" floppy0.clientDevice = "false" # defaults to "false" @@ -2320,6 +2320,7 @@ esxVMX_FormatConfig(esxVMX_Context *ctx, virCapsPtr caps, virDomainDefPtr def, virBuffer buffer = VIR_BUFFER_INITIALIZER; bool scsi_present[4] = { false, false, false, false }; int scsi_virtualDev[4] = { -1, -1, -1, -1 }; + bool floppy_present[2] = { false, false }; if (ctx->formatFileName == NULL) { ESX_ERROR(VIR_ERR_INTERNAL_ERROR, "%s", @@ -2525,7 +2526,8 @@ esxVMX_FormatConfig(esxVMX_Context *ctx, virCapsPtr caps, virDomainDefPtr def, break; case VIR_DOMAIN_DISK_DEVICE_FLOPPY: - if (esxVMX_FormatFloppy(ctx, def->disks[i], &buffer) < 0) { + if (esxVMX_FormatFloppy(ctx, def->disks[i], &buffer, + floppy_present) < 0) { goto failure; } @@ -2539,6 +2541,13 @@ esxVMX_FormatConfig(esxVMX_Context *ctx, virCapsPtr caps, virDomainDefPtr def, } } + for (i = 0; i < 2; ++i) { + /* floppy[0..1].present defaults to true, disable it explicitly */ + if (! floppy_present[i]) { + virBufferVSprintf(&buffer, "floppy%d.present = \"false\"\n", i); + } + } + /* def:fss */ /* FIXME */ @@ -2810,7 +2819,7 @@ esxVMX_FormatCDROM(esxVMX_Context *ctx, virDomainDiskDefPtr def, int esxVMX_FormatFloppy(esxVMX_Context *ctx, virDomainDiskDefPtr def, - virBufferPtr buffer) + virBufferPtr buffer, bool floppy_present[2]) { int unit; char *fileName = NULL; @@ -2824,6 +2833,8 @@ esxVMX_FormatFloppy(esxVMX_Context *ctx, virDomainDiskDefPtr def, return -1; } + floppy_present[unit] = true; + virBufferVSprintf(buffer, "floppy%d.present = \"true\"\n", unit); if (def->type == VIR_DOMAIN_DISK_TYPE_FILE) { diff --git a/src/esx/esx_vmx.h b/src/esx/esx_vmx.h index a77264a..12fc5af 100644 --- a/src/esx/esx_vmx.h +++ b/src/esx/esx_vmx.h @@ -137,7 +137,7 @@ esxVMX_FormatCDROM(esxVMX_Context *ctx, virDomainDiskDefPtr def, int esxVMX_FormatFloppy(esxVMX_Context *ctx, virDomainDiskDefPtr def, - virBufferPtr buffer); + virBufferPtr buffer, bool floppy_present[2]); int esxVMX_FormatEthernet(virDomainNetDefPtr def, int controller, -- 1.7.0.4

2 2

[libvirt] [PATCH] Generate libvirt_qemu.def from libvirt_qemu.syms for MinGW builds
by Matthias Bolte 14 Aug '10

14 Aug '10

--- configure.ac | 1 + src/.gitignore | 1 + src/Makefile.am | 9 ++++++++- 3 files changed, 10 insertions(+), 1 deletions(-) diff --git a/configure.ac b/configure.ac index cd48c1b..a762bae 100644 --- a/configure.ac +++ b/configure.ac @@ -1866,6 +1866,7 @@ case "$host" in # Also set the symbol file to .def, so src/Makefile generates libvirt.def # from libvirt.syms and passes libvirt.def instead of libvirt.syms to the linker LIBVIRT_SYMBOL_FILE=libvirt.def + LIBVIRT_QEMU_SYMBOL_FILE='$(srcdir)/libvirt_qemu.def' # mingw's ld has the --version-script parameter, but it requires a .def file # instead to work properly, therefore clear --version-script here and use # -Wl, to pass the .def file to the linker diff --git a/src/.gitignore b/src/.gitignore index 5d114c9..7ea8d89 100644 --- a/src/.gitignore +++ b/src/.gitignore @@ -11,6 +11,7 @@ libvirt_parthelper libvirt_lxc libvirt.def libvirt.syms +libvirt_qemu.def *.i *.s virt-aa-helper diff --git a/src/Makefile.am b/src/Makefile.am index a66eb2a..d7858dd 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -1010,7 +1010,7 @@ EXTRA_DIST += \ libvirt_daemon.syms \ libvirt_nwfilter.syms -BUILT_SOURCES += libvirt.syms libvirt.def +BUILT_SOURCES += libvirt.syms libvirt.def libvirt_qemu.def libvirt.syms: libvirt_public.syms $(USED_SYM_FILES) $(AM_V_GEN)rm -f $@-tmp $@ ; \ @@ -1033,6 +1033,13 @@ libvirt.def: libvirt.syms chmod a-w $@-tmp && \ mv $@-tmp libvirt.def +libvirt_qemu.def: libvirt_qemu.syms + $(AM_V_GEN)rm -f -- $@-tmp $@ ; \ + printf 'EXPORTS\n' > $@-tmp && \ + sed -e '/^$$/d; /#/d; /:/d; /\}/d; /\*/d; /LIBVIRT_/d; s/[ \t]*$.*$\;/ \1/g' $^ >> $@-tmp && \ + chmod a-w $@-tmp && \ + mv $@-tmp libvirt_qemu.def + # Empty source list - it merely links a bunch of convenience libs together libvirt_la_SOURCES = libvirt_la_LDFLAGS = $(VERSION_SCRIPT_FLAGS)$(LIBVIRT_SYMBOL_FILE) \ -- 1.7.0.4

2 4

[libvirt] [PATCH] PHYP: Add rudimentary network driver
by Eduardo Otubo 14 Aug '10

14 Aug '10

I changed virStorage[Open|Close] to virVIOSDriver[Open|Close] so the network driver can use it - since the network driver deals with Open/Close in the same way. --- src/phyp/phyp_driver.c | 33 +++++++++++++++++++++++++++++---- 1 files changed, 29 insertions(+), 4 deletions(-) diff --git a/src/phyp/phyp_driver.c b/src/phyp/phyp_driver.c index aa4722e..975d941 100644 --- a/src/phyp/phyp_driver.c +++ b/src/phyp/phyp_driver.c @@ -3864,7 +3864,7 @@ phypDomainSetCPU(virDomainPtr dom, unsigned int nvcpus) } static virDrvOpenStatus -phypStorageOpen(virConnectPtr conn, +phypVIOSDriverOpen(virConnectPtr conn, virConnectAuthPtr auth ATTRIBUTE_UNUSED, int flags) { @@ -3877,7 +3877,7 @@ phypStorageOpen(virConnectPtr conn, } static int -phypStorageClose(virConnectPtr conn ATTRIBUTE_UNUSED) +phypVIOSDriverClose(virConnectPtr conn ATTRIBUTE_UNUSED) { return 0; } @@ -3984,8 +3984,8 @@ static virDriver phypDriver = { static virStorageDriver phypStorageDriver = { .name = "PHYP", - .open = phypStorageOpen, - .close = phypStorageClose, + .open = phypVIOSDriverOpen, + .close = phypVIOSDriverClose, .numOfPools = phypNumOfStoragePools, .listPools = phypListStoragePools, @@ -4023,6 +4023,29 @@ static virStorageDriver phypStorageDriver = { .poolIsPersistent = NULL }; +static virNetworkDriver phypNetworkDriver = { + .name = "PHYP", + .open = phypVIOSDriverOpen, + .close = phypVIOSDriverClose, + .numOfNetworks = NULL, + .listNetworks = NULL, + .numOfDefinedNetworks = NULL, + .listDefinedNetworks = NULL, + .networkLookupByUUID = NULL, + .networkLookupByName = NULL, + .networkCreateXML = NULL, + .networkDefineXML = NULL, + .networkUndefine = NULL, + .networkCreate = NULL, + .networkDestroy = NULL, + .networkDumpXML = NULL, + .networkGetBridgeName = NULL, + .networkGetAutostart = NULL, + .networkSetAutostart = NULL, + .networkIsActive = NULL, + .networkIsPersistent = NULL +}; + int phypRegister(void) { @@ -4030,6 +4053,8 @@ phypRegister(void) return -1; if (virRegisterStorageDriver(&phypStorageDriver) < 0) return -1; + if (virRegisterNetworkDriver(&phypNetworkDriver) < 0) + return -1; return 0; } -- 1.7.0.4

2 1

[libvirt] [PATCH 2/3] Make umlConnectTapDevice ask brAddTap for a persistent tap device.
by Soren Hansen 14 Aug '10

14 Aug '10

This patch does two things: * It makes umlConnectTapDevice ask brAddTap for a persistent tap by passing it a NULL tapfd argument. * Stops umlConnectTapDevice from immediately dismantling the bridge it just set up. Signed-off-by: Soren Hansen <soren(a)linux2go.dk> --- src/uml/uml_conf.c | 6 +----- 1 files changed, 1 insertions(+), 5 deletions(-) diff --git a/src/uml/uml_conf.c b/src/uml/uml_conf.c index bc8cbce..06543cb 100644 --- a/src/uml/uml_conf.c +++ b/src/uml/uml_conf.c @@ -112,7 +112,6 @@ umlConnectTapDevice(virDomainNetDefPtr net, const char *bridge) { brControl *brctl = NULL; - int tapfd = -1; int template_ifname = 0; int err; unsigned char tapmac[VIR_MAC_BUFLEN]; @@ -140,7 +139,7 @@ umlConnectTapDevice(virDomainNetDefPtr net, &net->ifname, tapmac, 0, - &tapfd))) { + NULL))) { if (err == ENOTSUP) { /* In this particular case, give a better diagnostic. */ umlReportError(VIR_ERR_INTERNAL_ERROR, @@ -164,9 +163,6 @@ umlConnectTapDevice(virDomainNetDefPtr net, VIR_FREE(net->ifname); goto error; } - close(tapfd); - - brShutdown(brctl); return 0; -- 1.7.0.4

4 4

[libvirt] [PATCH 1/3] Close fd's of persistent tap devices
by Soren Hansen 14 Aug '10

14 Aug '10

When passing a NULL tapfd argument to brAddTap, we need to close the fd of the tap device. If we don't, libvirt will keep the fd open indefinitely and renders the the guest unable to configure its side of the tap device. Signed-off-by: Soren Hansen <soren(a)linux2go.dk> --- src/util/bridge.c | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/src/util/bridge.c b/src/util/bridge.c index 7d0caae..da62c5e 100644 --- a/src/util/bridge.c +++ b/src/util/bridge.c @@ -538,6 +538,8 @@ brAddTap(brControl *ctl, goto error; if (tapfd) *tapfd = fd; + else + close(fd); return 0; error: -- 1.7.0.4

2 1

[libvirt] [PATCH] Make sure all command line arguments get passed to UML
by soren＠linux2go.dk 14 Aug '10

14 Aug '10

From: Soren Hansen <soren(a)linux2go.dk> If umlBuildCommandLineChr fails (e.g. due to an unsupported chardev type), it returns NULL. umlBuildCommandLine does not check for this and sets this as an argument on the comand line, effectively ending the argument list. This patch checks for this case and sets the chardev to "none". Signed-off-by: Soren Hansen <soren(a)linux2go.dk> --- src/uml/uml_conf.c | 8 ++++---- 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/src/uml/uml_conf.c b/src/uml/uml_conf.c index 1623a78..42193e4 100644 --- a/src/uml/uml_conf.c +++ b/src/uml/uml_conf.c @@ -511,10 +511,10 @@ int umlBuildCommandLine(virConnectPtr conn, } for (i = 0 ; i < UML_MAX_CHAR_DEVICE ; i++) { - char *ret; + char *ret = NULL; if (i == 0 && vm->def->console) ret = umlBuildCommandLineChr(vm->def->console, "con"); - else + if (!ret) if (virAsprintf(&ret, "con%d=none", i) < 0) goto no_memory; ADD_ARG(ret); @@ -522,13 +522,13 @@ int umlBuildCommandLine(virConnectPtr conn, for (i = 0 ; i < UML_MAX_CHAR_DEVICE ; i++) { virDomainChrDefPtr chr = NULL; - char *ret; + char *ret = NULL; for (j = 0 ; j < vm->def->nserials ; j++) if (vm->def->serials[j]->target.port == i) chr = vm->def->serials[j]; if (chr) ret = umlBuildCommandLineChr(chr, "ssl"); - else + if (!ret) if (virAsprintf(&ret, "ssl%d=none", i) < 0) goto no_memory; ADD_ARG(ret); -- 1.7.0.4

2 1

[libvirt] PATCH 0/4: AppArmor updates
by Jamie Strandboge 13 Aug '10

13 Aug '10

This patchset consists of various small updates to the AppArmor security driver: 0001-apparmor-dont-ignore-open.patch: Exit with error if virDomainDiskDefForeachPath() fails, unless the disk doesn't exist, at which point we skip it without error. Also add several tests to virt-aa-helper-test for '-p' option. 0002-apparmor-chardev.patch: fix serial ports, parallel ports and channels 0003-apparmor-examples.patch: update to example AppArmor profile 0004-apparmor-fix-warn.patch: fix a compiler warning The updates in 0001-apparmor-dont-ignore-open.patch are needed to ensure that it works consistently with and without '-p 1' and that the security fix works as advertised. The other patches should speak for themselves. -- Jamie Strandboge | http://www.canonical.com

1 0

[libvirt] [PATCH v2] nwfilter: Discard class D and E IP addresses when sniffing packets
by Stefan Berger 13 Aug '10

13 Aug '10

V2: Corrected comment and simplified mask to check for class D and E IP addresses When sniffing the network traffic, discard class D and E IP addresses when sniffing traffic. This was a reason why filters were not correctly rebuilt on VMs on the local 192.* network when libvirt was restarted and those VMs did not use a DHCP request to get its IP address. Signed-off-by: Stefan Berger<stefanb(a)us.ibm.com> --- src/nwfilter/nwfilter_learnipaddr.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.c +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c @@ -546,9 +546,11 @@ learnIPAddressThread(void *arg) struct iphdr *iphdr = (struct iphdr*)(packet + ethHdrSize); vmaddr = iphdr->saddr; - // skip eth. bcast and mcast addresses, - // and zero address in DHCP Requests - if ((ntohl(vmaddr)& 0xc0000000) || vmaddr == 0) { + // skip mcast addresses (224.0.0.0 - 239.255.255.255), + // class E (240.0.0.0 - 255.255.255.255, includes eth. + // bcast) and zero address in DHCP Requests + if ( (ntohl(vmaddr)& 0xe0000000) == 0xe0000000 || + vmaddr == 0) { vmaddr = 0; continue; }

3 2

[libvirt] [PATCH[ nwfilter: Discard class D and E IP addresses when sniffing
by Stefan Berger 13 Aug '10

13 Aug '10

When sniffing the network traffic, discard class D and E IP addresses when sniffing traffic. This was a reason why filters were not correctly rebuilt on VMs on the local 192.* network when libvirt was restarted and those VMs did not use a DHCP request to get its IP address. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com> --- src/nwfilter/nwfilter_learnipaddr.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) Index: libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_learnipaddr.c +++ libvirt-acl/src/nwfilter/nwfilter_learnipaddr.c @@ -546,9 +546,12 @@ learnIPAddressThread(void *arg) struct iphdr *iphdr = (struct iphdr*)(packet + ethHdrSize); vmaddr = iphdr->saddr; - // skip eth. bcast and mcast addresses, + // skip eth. bcast and mcast addresses (224.0.0.0 - + // 239.255.255.255), class E (255.*) // and zero address in DHCP Requests - if ((ntohl(vmaddr) & 0xc0000000) || vmaddr == 0) { + if ( (ntohl(vmaddr) & 0xe0000000) == 0xe0000000 || + (ntohl(vmaddr) & 0xf0000000) == 0xf0000000 || + vmaddr == 0) { vmaddr = 0; continue; }

3 4

[libvirt] [PATCH] nwfilter: fix a memory leak
by Stefan Berger 13 Aug '10

13 Aug '10

Fixing a memory leak. Signed-off-by: Stefan Berger <stefanb(a)us.ibm.com> --- src/nwfilter/nwfilter_ebiptables_driver.c | 3 +++ 1 file changed, 3 insertions(+) Index: libvirt-acl/src/nwfilter/nwfilter_ebiptables_driver.c =================================================================== --- libvirt-acl.orig/src/nwfilter/nwfilter_ebiptables_driver.c +++ libvirt-acl/src/nwfilter/nwfilter_ebiptables_driver.c @@ -2973,6 +2973,9 @@ static int ebtablesCleanAll(const char * ebtablesRemoveTmpRootChain(&buf, 0, ifname); ebiptablesExecCLI(&buf, &cli_status); + + virBufferFreeAndReset(&buf); + return 0; }

3 2

[libvirt] [PATCH] Only require XDR when building libvirtd or the remote driver
by Matthias Bolte 13 Aug '10

13 Aug '10

--- configure.ac | 31 +++++++++++++++++++------------ 1 files changed, 19 insertions(+), 12 deletions(-) diff --git a/configure.ac b/configure.ac index a762bae..764405c 100644 --- a/configure.ac +++ b/configure.ac @@ -111,18 +111,6 @@ dnl Availability of various common headers (non-fatal if missing). AC_CHECK_HEADERS([pwd.h paths.h regex.h sys/syslimits.h \ termios.h sys/poll.h syslog.h mntent.h net/ethernet.h linux/magic.h]) -dnl Where are the XDR functions? -dnl If portablexdr is installed, prefer that. -dnl Otherwise try -lrpc (Cygwin) -lxdr (some MinGW), -lnsl (Solaris) -dnl or none (most Unix) -AC_CHECK_LIB([portablexdr],[xdrmem_create],[],[ - AC_SEARCH_LIBS([xdrmem_create],[rpc xdr nsl],[], - [AC_MSG_ERROR([Cannot find a XDR library])]) - ]) - -dnl check for cygwin's variation in xdr function names -AC_CHECK_FUNCS([xdr_u_int64_t],[],[],[#include <rpc/xdr.h>]) - AC_CHECK_LIB([intl],[gettext],[]) dnl Do we have rpcgen? @@ -319,6 +307,25 @@ AM_CONDITIONAL([WITH_OPENVZ], [test "$with_openvz" = "yes"]) dnl +dnl check for XDR +dnl + +if test x"$with_remote" = x"yes" || test x"$with_libvirtd" = x"yes"; then + dnl Where are the XDR functions? + dnl If portablexdr is installed, prefer that. + dnl Otherwise try -lrpc (Cygwin) -lxdr (some MinGW), -lnsl (Solaris) + dnl or none (most Unix) + AC_CHECK_LIB([portablexdr],[xdrmem_create],[],[ + AC_SEARCH_LIBS([xdrmem_create],[rpc xdr nsl],[], + [AC_MSG_ERROR([Cannot find a XDR library])]) + ]) + + dnl check for cygwin's variation in xdr function names + AC_CHECK_FUNCS([xdr_u_int64_t],[],[],[#include <rpc/xdr.h>]) +fi + + +dnl dnl check for VirtualBox XPCOMC location dnl -- 1.7.0.4

2 2