- Devel - libvirt List Archives

[libvirt] gdb works abnormally when debug the libvirt.
by ShaoHe Feng 29 Nov '11

29 Nov '11

When I to use gdb to check whether qemuMonitorJSONIOProcessEvent processes the Event that I expect. but the gdb works abnormally. I list the the position of the source file by gdb. It is the correct functionqemuMonitorJSONIOProcessEvent. However I set break point on this position. then "info b", it is the wrong function qemuMonitorJSONIOProcess. I debug my libvirt on Ubuntu 11.10 this time.But I have set a break point at this function on redhat 6.0 before, it can work well. My gcc version is: gcc (Ubuntu/Linaro 4.6.1-9ubuntu3) 4.6.1 My gdb version is: GNU gdb (Ubuntu/Linaro 7.3-0ubuntu2) 7.3-2011.08 here is the info: (gdb) list qemu/qemu_monitor_json.c:88 83 qemuMonitorJSONIOProcessEvent(qemuMonitorPtr mon, 84 virJSONValuePtr obj) 85 { 86 const char *type; 87 int i; 88 int findEventFlag = -1; 89 VIR_DEBUG("mon=%p obj=%p", mon, obj); 90 91 type = virJSONValueObjectGetString(obj, "event"); 92 if (!type) { (gdb) b qemu/qemu_monitor_json.c:88 Breakpoint 5 at 0x4a18d4: file qemu/qemu_monitor_json.c, line 88. (gdb) info b Num Type Disp Enb Address What 5 breakpoint keep y 0x00000000004a18d4 in qemuMonitorJSONIOProcess at qemu/qemu_monitor_json.c:88 (gdb)

3 3

[libvirt] [PATCH] build: fix typo in last patch
by Eric Blake 29 Nov '11

29 Nov '11

* src/remote_protocol-structs: Use correct RPC name. --- Pushing under the build-breaker rule. Osier attempted to fix this already, but added a typo. src/remote_protocol-structs | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/src/remote_protocol-structs b/src/remote_protocol-structs index 62c0248..927b6c6 100644 --- a/src/remote_protocol-structs +++ b/src/remote_protocol-structs @@ -2019,5 +2019,5 @@ enum remote_procedure { REMOTE_PROC_DOMAIN_EVENT_DISK_CHANGE = 248, REMOTE_PROC_DOMAIN_OPEN_GRAPHICS = 249, REMOTE_PROC_NODE_SUSPEND_FOR_DURATION = 250, - REMOTE_PROC_NODE_SUSPEND_FOR_DURATION = 251, + REMOTE_PROC_DOMAIN_BLOCK_RESIZE = 251, }; -- 1.7.7.3

1 0

[libvirt] [PATCH] Fix leak build config file path
by Daniel P. Berrange 29 Nov '11

29 Nov '11

From: "Daniel P. Berrange" <berrange(a)redhat.com> * src/libvirt.c: Free user directory path --- src/libvirt.c | 5 ++++- 1 files changed, 4 insertions(+), 1 deletions(-) diff --git a/src/libvirt.c b/src/libvirt.c index cdd111c..55ef743 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -974,8 +974,11 @@ virConnectConfigFile(void) goto error; if (virAsprintf(&path, "%s/.libvirt/libvirt.conf", - userdir) < 0) + userdir) < 0) { + VIR_FREE(userdir); goto no_memory; + } + VIR_FREE(userdir); } return path; -- 1.7.6.4

2 1

[libvirt] [PATCH 0/5] Interface pools and passthrough mode
by Shradha Shah 29 Nov '11

29 Nov '11

Interface Pools and Passthrough mode: Current Method: The passthrough mode uses a macvtap âdirectâ connection to connect each guest to the network. The physical interface to be used is picked from among those listed in <interface> sub elements of the <forward> element. The current specification for <forward> extends to allow 0 or more <interface> sub-elements: Example: <forward mode='passthrough' dev='eth10'/> <interface dev='eth10'/> <interface dev='eth12'/> <interface dev='eth18'/> <interface dev='eth20'/> </forward> However with an ethernet card with 64 VF's or more, the above method gets tedious on the system. On the other hand, just parameterizing a string (eth%d) is inadequate, eg, when there are multiple non-contiguous ranges. Proposed Method: The 5 patches provided along with this introductory e-mail i) Introduce a structure to store the state of all the virtual functions attached to each physical function ii) Find a free virtual function given the physical function. The forward specification will hence only mention the physical function as the interface sub element: Example: <forward mode='passthrough' dev='eth2'/> <interface dev='eth2'/> </forward> Shradha Shah (5): Moving the declaration of _pciDevice and _pciDeviceList to pci.h Added function pciSysfsFile to enable access to the PCI SYSFS files. Adding functions virNetDevGetVirtualFunctions and virNetDevGetNetName virNetDevGetVirtualFunctions: Gets the BDF of all the Virtual Functions given a physical function virNetDevGetNetName: Gets the interface name of the PCI Device. Addition of a new device structure to store the state of a Virtual Function Modifications to the Physical Device Structure to store state of its Virtual Functions If a system has 64 or more VF's, it is quite tedious to mention each VF in the interface pool. The following modification find a Free VF given a Physical Function when mode=passthrough. We also save the state of each VF. Hence modifications to networkAllocateActualDevice, networkNotifyActualDevice and networkReleaseActualDevice were required. The following patch does just that. src/conf/network_conf.c | 19 ++++- src/conf/network_conf.h | 10 ++ src/network/bridge_driver.c | 208 +++++++++++++++++++++++++++++++++++++------ src/util/pci.c | 52 ++++------- src/util/pci.h | 36 ++++++++ src/util/virnetdev.c | 89 ++++++++++++++++++ src/util/virnetdev.h | 10 ++ 7 files changed, 359 insertions(+), 65 deletions(-) -- 1.7.4.4

1 0

[libvirt] [libvirt-glib] gir: Use --pkg=libxml-2.0 instead of LIBXML2_CFLAGS
by Christophe Fergeau 29 Nov '11

29 Nov '11

g-ir-scanner has a standard way of getting the flags for libraries providing .pc files, this commit uses this --pkg flag for libxml instead of using LIBXML2_CFLAGS we computed ourselves. --- libvirt-gconfig/Makefile.am | 2 +- libvirt-gobject/Makefile.am | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/libvirt-gconfig/Makefile.am b/libvirt-gconfig/Makefile.am index b14bd92..fe5b1a5 100644 --- a/libvirt-gconfig/Makefile.am +++ b/libvirt-gconfig/Makefile.am @@ -119,9 +119,9 @@ LibvirtGConfig-1.0.gir: libvirt-gconfig-1.0.la $(G_IR_SCANNER) Makefile.am --output $@ \ -I$(top_builddir) \ -I$(top_srcdir) \ - $(LIBXML2_CFLAGS) \ --verbose \ --c-include="libvirt-gconfig/libvirt-gconfig.h" \ + --pkg=libxml-2.0 \ --pkg-export=libvirt-gconfig-1.0 \ $(srcdir)/libvirt-gconfig.h \ $(GCONFIG_SOURCE_FILES:%=$(srcdir)/%) \ diff --git a/libvirt-gobject/Makefile.am b/libvirt-gobject/Makefile.am index 5aa1080..ad746bb 100644 --- a/libvirt-gobject/Makefile.am +++ b/libvirt-gobject/Makefile.am @@ -133,10 +133,10 @@ LibvirtGObject-1.0.gir: libvirt-gobject-1.0.la $(G_IR_SCANNER) Makefile.am --output $@ \ -I$(top_builddir) \ -I$(top_srcdir) \ - $(LIBXML2_CFLAGS) \ --verbose \ - --pkg=gthread-2.0 \ --c-include="libvirt-gobject/libvirt-gobject.h" \ + --pkg=gthread-2.0 \ + --pkg=libxml-2.0 \ --pkg-export=libvirt-gobject-1.0 \ $(srcdir)/libvirt-gobject.h \ $(GOBJECT_SOURCE_FILES:%=$(srcdir)/%) \ -- 1.7.7.3

2 1

[libvirt] [PATCH 0/5 v2] Support block device online resizing
by Osier Yang 29 Nov '11

29 Nov '11

This patch series introduce new API "virDomainBlockResize" to expose qemu monitor command "block_size", which is for resizing the a block device while the domain is running. The prototype for the new API is: int virDomainBlockResize (virDomainPtr dom, const char *disk, unsigned long long size, unsigned int flags) * "@disk" is the absolute source path or device target shorthand of the block device. * The units for "@size" is kilobytes. * "@flags" is unused currently. [PATCH 1/5] block_resize: Define the new API [PATCH 2/5] block_resize: Wire up the remote protocol [PATCH 3/5] block_resize: Implement qemu monitor functions [PATCH 4/5] block_resize: Implement qemu driver method [PATCH 5/5] block_resize: Expose the new API to virsh v1: http://www.redhat.com/archives/libvir-list/2011-July/msg01788.html v1 ~ v2: * Nearly just rebasing, with small improvements, nothing new except updating the API definitions to be consistent with commit 3ac26e26. Regards, Osier

2 11

[libvirt] All virsh commands testing failed with git head
by Osier Yang 29 Nov '11

29 Nov '11

Part of the error log is pasted, looks like it's caused by the calling of virNodeSuspendInit in virInitialize. ================================ cd tests; LIBVIRT_DEBUG=1 VIR_TEST_DEBUG=1 ./virshtest; cd - TEST: virshtest 1) virsh list (default) ... Offset 0 Expect [ ] Actual [2011-11-29 18:44:38.153: 24448: info : libvirt version: 0.9.7 2011-11-29 18:44:38.153: 24448: error : virNodeSuspendInit:86 : Failed to get host power management features ] ... FAILED 2) virsh list (custom) ... Offset 0 Expect [ ] Actual [2011-11-29 18:44:38.180: 24463: info : libvirt version: 0.9.7 2011-11-29 18:44:38.180: 24463: error : virNodeSuspendInit:86 : Failed to get host power management features ] ... FAILED 3) virsh nodeinfo (default) ... Offset 0 Expect [C] Actual [2011-11-29 18:44:38.212: 24478: info : libvirt version: 0.9.7 2011-11-29 18:44:38.212: 24478: error : virNodeSuspendInit:86 : Failed to get host power management features C] ... FAILED Regards, Osier

3 4

[libvirt] [libvirt-glib 1/5] Add various missing packages to libvirt-gobject-1.0.pc
by Christophe Fergeau 29 Nov '11

29 Nov '11

--- libvirt-gobject-1.0.pc.in | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/libvirt-gobject-1.0.pc.in b/libvirt-gobject-1.0.pc.in index 5442ba4..3141e87 100644 --- a/libvirt-gobject-1.0.pc.in +++ b/libvirt-gobject-1.0.pc.in @@ -6,7 +6,7 @@ includedir=@includedir@ Name: libvirt-gobject Version: @VERSION@ Description: libvirt GObject library -Requires: gobject-2.0 libvirt-gconfig-1.0 +Requires: gio-2.0 gobject-2.0 libvirt-gconfig-1.0 Libs: -L${libdir} -lvirt-gobject-1.0 -Libs.private: @LIBVIRT_LIBS@ +Requires.private: gthread-2.0 libvirt Cflags: -I${includedir}/libvirt-gobject-1.0 -- 1.7.7.3

2 12

[libvirt] [PATCH 0/3] Improve async signal safety of timestamp APIs
by Daniel P. Berrange 29 Nov '11

29 Nov '11

There are rare reports of libvirt deadlocking itself which show a fork'd, but not exec'd, child process stuck in localtime_r(). This is because localtime_r() is not an async signal, so if we happened to fork() at the exact time another thread was calling localtime_r() we could get stuck with a stale lock in the child. Although we go to pains to protect the libvirt logging code by acquiring the mutex before forking, this is not sufficient, because some non-logging related code could invoke virTimestamp() which calls localtime_r() too. The only sane approach to fix this is to make our timestamp generation APIs thread safe, by doing a clean impl of gmtime() and strftime() inside libvirt using static buffers and only async safe POSIX apis.

2 5

[libvirt] [libvirt-glib] Ensure 'domain-added' signal on transient domain creation
by Zeeshan Ali (Khattak) 29 Nov '11

29 Nov '11

From: "Zeeshan Ali (Khattak)" <zeeshanak(a)gnome.org> --- libvirt-gobject/libvirt-gobject-connection.c | 10 ++++++++-- 1 files changed, 8 insertions(+), 2 deletions(-) diff --git a/libvirt-gobject/libvirt-gobject-connection.c b/libvirt-gobject/libvirt-gobject-connection.c index a258847..02b54a1 100644 --- a/libvirt-gobject/libvirt-gobject-connection.c +++ b/libvirt-gobject/libvirt-gobject-connection.c @@ -264,6 +264,7 @@ static int domain_event_cb(virConnectPtr conn G_GNUC_UNUSED, GVirConnection *gconn = opaque; GVirDomain *gdom; GVirConnectionPrivate *priv = gconn->priv; + gboolean was_unknown = FALSE; if (virDomainGetUUIDString(dom, uuid) < 0) { g_warning("Failed to get domain UUID on %p", dom); @@ -282,6 +283,8 @@ static int domain_event_cb(virConnectPtr conn G_GNUC_UNUSED, g_mutex_lock(priv->lock); g_hash_table_insert(priv->domains, (gpointer)gvir_domain_get_uuid(gdom), gdom); g_mutex_unlock(priv->lock); + + was_unknown = TRUE; } switch (event) { @@ -307,9 +310,12 @@ static int domain_event_cb(virConnectPtr conn G_GNUC_UNUSED, break; case VIR_DOMAIN_EVENT_STARTED: - if (detail == VIR_DOMAIN_EVENT_STARTED_BOOTED) + if (detail == VIR_DOMAIN_EVENT_STARTED_BOOTED) { + if (was_unknown) + /* Most probably a transient domain */ + g_signal_emit(gconn, signals[VIR_DOMAIN_ADDED], 0, gdom); g_signal_emit_by_name(gdom, "started::booted"); - else if (detail == VIR_DOMAIN_EVENT_STARTED_MIGRATED) + } else if (detail == VIR_DOMAIN_EVENT_STARTED_MIGRATED) g_signal_emit_by_name(gdom, "started::migrated"); else if (detail == VIR_DOMAIN_EVENT_STARTED_RESTORED) g_signal_emit_by_name(gdom, "started::restored"); -- 1.7.7.3

2 1