January 2010 - Devel - libvirt List Archives

[libvirt] [PATCH] qemu: Check for ia64 kvm
by Cole Robinson 14 Jan '10

14 Jan '10

From: Dustin Xiong <x_k_123(a)hotmail.com> Ported to current code. Untested, but builds fine. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/qemu/qemu_conf.c | 7 +++++++ 1 files changed, 7 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index d3da776..0d970d6 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -373,12 +373,19 @@ static const struct qemu_feature_flags const arch_info_x86_64_flags [] = { { "apic", 1, 0 }, }; +static const struct qemu_feature_flags const arch_info_ia64_flags [] = { + { "acpi", 1, 1 }, + { "apic", 1, 0 }, +}; + /* The archicture tables for supported QEMU archs */ static const struct qemu_arch_info const arch_info_hvm[] = { { "i686", 32, NULL, "/usr/bin/qemu", "/usr/bin/qemu-system-x86_64", arch_info_i686_flags, 4 }, { "x86_64", 64, NULL, "/usr/bin/qemu-system-x86_64", NULL, arch_info_x86_64_flags, 2 }, + { "itanium", 64, NULL, "/usr/bin/qemu-system-ia64", + NULL, arch_info_ia64_flags, 2}, { "arm", 32, NULL, "/usr/bin/qemu-system-arm", NULL, NULL, 0 }, { "mips", 32, NULL, "/usr/bin/qemu-system-mips", NULL, NULL, 0 }, { "mipsel", 32, NULL, "/usr/bin/qemu-system-mipsel", NULL, NULL, 0 }, -- 1.6.5.2

3 3

[libvirt] [PATCH] util: Make sure virExec hook failures are raised
by Cole Robinson 14 Jan '10

14 Jan '10

With the introduction virDispatchError, hook function errors are never sent through the error callback, so users will never see these messages. Fix this by calling virDispatchError after hook failure. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/util/util.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/src/util/util.c b/src/util/util.c index ba6b0db..45ca657 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -557,8 +557,11 @@ __virExec(virConnectPtr conn, } if (hook) - if ((hook)(data) != 0) + if ((hook)(data) != 0) { + VIR_DEBUG0("Hook function failed."); + virDispatchError(NULL); _exit(1); + } /* The steps above may need todo something privileged, so * we delay clearing capabilities until the last minute */ -- 1.6.5.2

3 3

[libvirt] [PATCH] Fix linkage of virt-aa-helper to libgnu.a
by Matthias Bolte 14 Jan '10

14 Jan '10

--- src/Makefile.am | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index dbf708b..324030b 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -876,9 +876,9 @@ virt_aa_helper_LDFLAGS = $(WARN_CFLAGS) virt_aa_helper_LDADD = \ $(WARN_CFLAGS) \ $(LIBXML_LIBS) \ - @top_srcdir@/gnulib/lib/libgnu.la \ @top_srcdir@/src/libvirt_conf.la \ - @top_srcdir@/src/libvirt_util.la + @top_srcdir@/src/libvirt_util.la \ + @top_srcdir@/gnulib/lib/libgnu.la virt_aa_helper_CFLAGS = \ -I@top_srcdir@/src/conf \ -I@top_srcdir@/src/security -- 1.6.3.3

2 1

[libvirt] [PATCH] Fix event-test segfault for mixed remote/local driver usage
by Matthias Bolte 14 Jan '10

14 Jan '10

If examples/domain-events/events-c/event-test.c uses a local driver for the hyperviror and a remote driver for storage then the segfault occurs in remoteDomainEventDispatchFunc because conn->privateData is used. The problem is as follows. The doRemoteOpen function registers remoteDomainEventFired as event handle and remoteDomainEventQueueFlush as event timeout. The doRemoteOpen function is used in remoteOpen and remoteOpenSecondaryDriver. So both handlers are registered for all types of remote drivers, but remoteDomainEventQueueFlush always uses conn->privateData. That's wrong when remoteDomainEventQueueFlush is called for any other driver than the hypervisor driver, it should use conn->storagePrivateData for the storage driver for example. In the common case this doesn't result in a segfault because in the common case all drivers for a connection are either all remote or all local and conn->privateData and conn->storagePrivateData are the same. But in the local Xen case ('event-test xen:///' executed in Domain-0, with a libvirtd running in Domain-0) the hypervisor driver is local and the storage driver is remote. This results in a mismatch between conn->privateData and conn->storagePrivateData. Now the call to remoteDomainEventQueueFlush in the context of the storage driver results in a segfault, because it assumes that conn->privateData points to remote driver private data, but it points to Xen driver private data. To fix this, the pointer to the private driver data gets passed as opaque parameter instead of the connection pointer. Because some functions called by the event handler functions need the connection pointer for other purposes than error reporting the connection pointer is passed as member of the private remote driver data struct. Now remoteDomainEventFired and remoteDomainEventQueueFlush have the correct private data pointer for the respective calling context. --- src/remote/remote_driver.c | 16 ++++++++-------- 1 files changed, 8 insertions(+), 8 deletions(-) diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c index d6f5fce..582b85c 100644 --- a/src/remote/remote_driver.c +++ b/src/remote/remote_driver.c @@ -152,6 +152,7 @@ struct private_stream_data { struct private_data { virMutex lock; + virConnectPtr conn; int sock; /* Socket. */ int watch; /* File handle watch */ @@ -893,7 +894,7 @@ doRemoteOpen (virConnectPtr conn, if ((priv->watch = virEventAddHandle(priv->sock, VIR_EVENT_HANDLE_READABLE, remoteDomainEventFired, - conn, NULL)) < 0) { + priv, NULL)) < 0) { DEBUG0("virEventAddHandle failed: No addHandleImpl defined." " continuing without events."); } else { @@ -901,7 +902,7 @@ doRemoteOpen (virConnectPtr conn, DEBUG0("Adding Timeout for remote event queue flushing"); if ( (priv->eventFlushTimer = virEventAddTimeout(-1, remoteDomainEventQueueFlush, - conn, NULL)) < 0) { + priv, NULL)) < 0) { DEBUG0("virEventAddTimeout failed: No addTimeoutImpl defined. " "continuing without events."); virEventRemoveHandle(priv->watch); @@ -980,6 +981,7 @@ remoteAllocPrivateData(virConnectPtr conn) return NULL; } remoteDriverLock(priv); + priv->conn = conn; priv->localUses = 1; priv->watch = -1; priv->sock = -1; @@ -8659,8 +8661,7 @@ remoteDomainEventFired(int watch, int event, void *opaque) { - virConnectPtr conn = opaque; - struct private_data *priv = conn->privateData; + struct private_data *priv = opaque; remoteDriverLock(priv); @@ -8684,7 +8685,7 @@ remoteDomainEventFired(int watch, goto done; } - if (remoteIOHandleInput(conn, priv, 0) < 0) + if (remoteIOHandleInput(priv->conn, priv, 0) < 0) DEBUG0("Something went wrong during async message processing"); done: @@ -8708,8 +8709,7 @@ static void remoteDomainEventDispatchFunc(virConnectPtr conn, void remoteDomainEventQueueFlush(int timer ATTRIBUTE_UNUSED, void *opaque) { - virConnectPtr conn = opaque; - struct private_data *priv = conn->privateData; + struct private_data *priv = opaque; virDomainEventQueue tempQueue; remoteDriverLock(priv); @@ -8731,7 +8731,7 @@ remoteDomainEventQueueFlush(int timer ATTRIBUTE_UNUSED, void *opaque) if ( priv->callbackList->count == 0 ) { /* Tell the server when we are the last callback deregistering */ - if (call (conn, priv, 0, REMOTE_PROC_DOMAIN_EVENTS_DEREGISTER, + if (call (priv->conn, priv, 0, REMOTE_PROC_DOMAIN_EVENTS_DEREGISTER, (xdrproc_t) xdr_void, (char *) NULL, (xdrproc_t) xdr_void, (char *) NULL) == -1) VIR_WARN0("Failed to de-register events"); -- 1.6.3.3

2 1

[libvirt] Set MAC address for host virtual interface?
by Neil Aggarwal 14 Jan '10

14 Jan '10

Hello: Is there a way to set the MAC address for the virtual interface on the host? I have this in my guests's XML file: <interface type='bridge'> <mac address='54:52:00:7d:f8:fc'/> <source bridge='br0'/> <target dev='tkl-joomla'/> </interface> the mac address line controls the guest interface but the host interface seems to get a random MAC address. This is causing a problem for me since I use Cacti to monitor the bandwidth usage of the guest machines. Every time I stop and start a guest, Cacti creates a new interface and I have to reset the graphs to use the new interfaces. Thanks, Neil -- Neil Aggarwal, (281)846-8957, http://UnmeteredVPS.net Host Joomla!, Wordpress, phpBB, or vBulletin for $25/mo Unmetered bandwidth = no overage charges, 7 day free trial

2 1

[libvirt] Man page patch for BZ 548485
by David Jorm 14 Jan '10

14 Jan '10

See https://bugzilla.redhat.com/show_bug.cgi?id=548485 git diff for the patch: ---SNIP--- --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -342,7 +342,9 @@ severed upon restore, as TCP timeouts may have expired. =item B<schedinfo> optional I<--weight> B<number> optional I<--cap> B<number> I<domain-id> -Allows to show (and set) the domain scheduler parameters. +Allows you to show (and set) the domain scheduler parameters. + +B<Note>: The only parameter currently available for all schedulers is cpu_shares. It has a valid value range of 0-262144. B<Note>: The weight and cap parameters are defined only for the XEN_CREDIT scheduler and are now I<DEPRECATED>. ---SNIP--- I hope that's right, I found that BZ a bit confusing to comprehend. Learning curve. David

2 2

[libvirt] [BUG?] "no domain with matching uuid" error, when vm restarts
by su disheng 13 Jan '10

13 Jan '10

Hi, In the following libvirt API calling sequence, I always get an error "no domain with matching uuid" Connect _conn = new Connect("qemu:///system", false); _conn.domainDefineXML(kvm_guest_xml); Domain dm = _conn.domainLookupByName(kvm_guest_name); dm.create(); /* stop, undefine, and re-start the vm*/ dm.shutdown(); dm.undefine(); dm.domainDefineXML(kvm_guest_xml); /****A****/ Domain dm = _conn.domainLookupByName(kvm_guest_name); dm.create() /********Error!!!!**/ if I close the connection, and re-connect qemu at the end of /****A****/, then everything is OK. From the log, seems that uuid is not updated immediately, for this connection, the uuid is in the stale state? I am using libvirt 0.6.3, if it's a bug, does it fixed in the latest code? or Have I need to close the connection for each vm shutdown/re-define? I am a newbie in this API, any help are appreciated.

2 2

[libvirt] [PATCH] virterror: Don't invoke error callback for ERR_OK
by Cole Robinson 13 Jan '10

13 Jan '10

Since virDispatchError is now responsible for invoking the error callback, give it the same semantics as ReportError, which will skip VIR_ERR_OK (which is encountered when no error was raised). This fixes invoking the error callback after every non-erroring API call. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/util/virterror.c | 6 +++++- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/src/util/virterror.c b/src/util/virterror.c index e2128b9..78974ee 100644 --- a/src/util/virterror.c +++ b/src/util/virterror.c @@ -603,8 +603,12 @@ virDispatchError(virConnectPtr conn) if (!err) return; - /* Set a generic error message if none is already set */ + /* We never used to raise ERR_OK, so maintain existing behavior */ if (err->code == VIR_ERR_OK) + return; + + /* Set a generic error message if none is already set */ + if (!err->message) virErrorGenericFailure(err); /* Copy the global error to per-connection error if needed */ -- 1.6.5.2

2 5

[libvirt] [PATCH 0/2] Some new utility functions
by Laine Stump 13 Jan '10

13 Jan '10

I'm planning to use these to fix the problems with creating storage volumes on root-squashing NFS servers, but since they may be of general utility I thought I'd throw them out for comments beforehand - it something different would make them more generally useful, better to change it now than after I've already finished code to use them. My plan is to use virRunWithHook() to call setuid (or possibly some sort of capng calisthenics) prior to execing qemu-img and qcow-create. virFileCreate and virDirCreate will be used when creating raw volumes. I have written test code to verify that virFileCreate works in several different scenarios. Haven't done so with virRun yet, but it's a fairly simple change, just exposing __virExec arguments that were already there. (I'd previously written test code (yes, it worked ;-)) that used a virRun with uid/gid args added on, but have decided against that approach, since this seems more flexible)

4 19

[libvirt] [PATCH] node_device: udev: Fix memory leak
by Cole Robinson 13 Jan '10

13 Jan '10

We are setting the same property two different ways without free'ing in between. Just drop the second assignment. Signed-off-by: Cole Robinson <crobinso(a)redhat.com> --- src/node_device/node_device_udev.c | 6 +----- 1 files changed, 1 insertions(+), 5 deletions(-) diff --git a/src/node_device/node_device_udev.c b/src/node_device/node_device_udev.c index a9e196d..ff6bd46 100644 --- a/src/node_device/node_device_udev.c +++ b/src/node_device/node_device_udev.c @@ -997,11 +997,7 @@ static int udevProcessStorage(struct udev_device *device, goto out; } data->storage.block = strdup(devnode); - if (udevGetStringProperty(device, - "DEVNAME", - &data->storage.block) == PROPERTY_ERROR) { - goto out; - } + if (udevGetStringProperty(device, "ID_BUS", &data->storage.bus) == PROPERTY_ERROR) { -- 1.6.5.2

3 3