- Devel - libvirt List Archives

[libvirt] [PATCHv2] build: avoid -lgcrypt with newer gnutls
by Eric Blake 29 Jul '13

29 Jul '13

https://bugzilla.redhat.com/show_bug.cgi?id=951637 Newer gnutls uses nettle, rather than gcrypt, which is a lot nicer regarding initialization. Yet we were unconditionally initializing gcrypt even when gnutls wouldn't be using it, and having two crypto libraries linked into libvirt.so is pointless. Assume that the switch to gnutls 3.0 is a reliable witness, when pkg-config is present; otherwise be pessimistic and use gcrypt. * configure.ac (WITH_GNUTLS): Probe whether to add -lgcrypt, and define a witness WITH_GNUTLS_GCRYPT. * src/libvirt.c (virTLSMutexInit, virTLSMutexDestroy) (virTLSMutexLock, virTLSMutexUnlock, virTLSThreadImpl) (virGlobalInit): Honor the witness. * libvirt.spec.in (BuildRequires): Make gcrypt usage conditional, no longer needed in Fedora 19. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- v2: use second pkg-config invocation rather than ldd to determine whether gnutls uses gcrypt configure.ac | 27 +++++++++++++++++++-------- libvirt.spec.in | 2 ++ src/libvirt.c | 10 ++++++---- 3 files changed, 27 insertions(+), 12 deletions(-) diff --git a/configure.ac b/configure.ac index cc9942a..eb56b63 100644 --- a/configure.ac +++ b/configure.ac @@ -1076,12 +1076,19 @@ if test "x$with_gnutls" != "xno"; then LIBS="$LIBS $GNUTLS_LIBS" GNUTLS_FOUND=no + GNUTLS_GCRYPT=no if test -x "$PKG_CONFIG" ; then + dnl double probe, since we know that gnutls 3.0 switched to nettle instead of + dnl gcrypt PKG_CHECK_MODULES(GNUTLS, gnutls >= $GNUTLS_REQUIRED, - [GNUTLS_FOUND=yes], [GNUTLS_FOUND=no]) + [GNUTLS_FOUND=yes + PKG_CHECK_MODULES([GNUTLS], [gnutls >= 3.0], [], [GNUTLS_GCRYPT=yes])], + [GNUTLS_FOUND=no]) fi if test "$GNUTLS_FOUND" = "no"; then + dnl pkg-config couldn't help us, assume gcrypt is necessary fail=0 + GNUTLS_GCRYPT=yes AC_CHECK_HEADER([gnutls/gnutls.h], [], [fail=1]) AC_CHECK_LIB([gnutls], [gnutls_handshake],[], [fail=1], [-lgcrypt]) @@ -1098,13 +1105,17 @@ if test "x$with_gnutls" != "xno"; then AC_MSG_ERROR([You must install the GnuTLS library in order to compile and run libvirt]) fi else - dnl Not all versions of gnutls include -lgcrypt, and so we add - dnl it explicitly for the calls to gcry_control/check_version - GNUTLS_LIBS="$GNUTLS_LIBS -lgcrypt" - - dnl We're not using gcrypt deprecated features so define - dnl GCRYPT_NO_DEPRECATED to avoid deprecated warnings - GNUTLS_CFLAGS="$GNUTLS_CFLAGS -DGCRYPT_NO_DEPRECATED" + dnl If gnutls linked against -lgcrypt, then we must initialize gcrypt + dnl prior to using gnutls. Newer versions of gnutls use -lnettle, in + dnl which case we don't want to drag in gcrypt ourselves. + if test "$GNUTLS_GCRYPT" = yes; then + GNUTLS_LIBS="$GNUTLS_LIBS -lgcrypt" + dnl We're not using gcrypt deprecated features so define + dnl GCRYPT_NO_DEPRECATED to avoid deprecated warnings + GNUTLS_CFLAGS="$GNUTLS_CFLAGS -DGCRYPT_NO_DEPRECATED" + AC_DEFINE_UNQUOTED([WITH_GNUTLS_GCRYPT], 1, + [set to 1 if it is known or assumed that GNUTLS uses gcrypt]) + fi dnl gnutls 3.x moved some declarations to a new header AC_CHECK_HEADERS([gnutls/crypto.h], [], [], [[ diff --git a/libvirt.spec.in b/libvirt.spec.in index e0e0004..4320281 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -422,7 +422,9 @@ BuildRequires: readline-devel BuildRequires: ncurses-devel BuildRequires: gettext BuildRequires: libtasn1-devel +%if (0%{?rhel} && 0%{?rhel} < 7) || (0%{?fedora} && 0%{?fedora} < 19) BuildRequires: libgcrypt-devel +%endif BuildRequires: gnutls-devel BuildRequires: libattr-devel %if %{with_libvirtd} diff --git a/src/libvirt.c b/src/libvirt.c index 8157488..66e8248 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -55,7 +55,9 @@ #include "intprops.h" #include "virconf.h" #if WITH_GNUTLS -# include <gcrypt.h> +# if WITH_GNUTLS_GCRYPT +# include <gcrypt.h> +# endif # include "rpc/virnettlscontext.h" #endif #include "vircommand.h" @@ -270,7 +272,7 @@ winsock_init(void) #endif -#ifdef WITH_GNUTLS +#ifdef WITH_GNUTLS_GCRYPT static int virTLSMutexInit(void **priv) { virMutexPtr lock = NULL; @@ -323,7 +325,7 @@ static struct gcry_thread_cbs virTLSThreadImpl = { virTLSMutexUnlock, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL }; -#endif +#endif /* WITH_GNUTLS_GCRYPT */ /* Helper macros to implement VIR_DOMAIN_DEBUG using just C99. This * assumes you pass fewer than 15 arguments to VIR_DOMAIN_DEBUG, but @@ -407,7 +409,7 @@ virGlobalInit(void) virErrorInitialize() < 0) goto error; -#ifdef WITH_GNUTLS +#ifdef WITH_GNUTLS_GCRYPT /* * This sequence of API calls it copied exactly from * gnutls 2.12.23 source lib/gcrypt/init.c, with -- 1.8.3.1

3 6

[libvirt] [PATCH 1/1] cpu: Fix one compile error for PPC.
by Li Zhang 29 Jul '13

29 Jul '13

From: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> CPU data structure is refined, which causes one compile error for PPC. Signed-off-by: Li Zhang <zhlcindy(a)linux.vnet.ibm.com> --- src/cpu/cpu_powerpc.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/cpu/cpu_powerpc.c b/src/cpu/cpu_powerpc.c index 62437d3..55a4153 100644 --- a/src/cpu/cpu_powerpc.c +++ b/src/cpu/cpu_powerpc.c @@ -354,15 +354,15 @@ ppcDataFree(virCPUDataPtr data) static virCPUDataPtr ppcNodeData(void) { - virCPUDataPtr data; + virCPUDataPtr cpuData; - if (VIR_ALLOC(data) < 0) + if (VIR_ALLOC(cpuData) < 0) return NULL; asm("mfpvr %0" - : "=r" (data->ppc.pvr)); + : "=r" (cpuData->data.ppc.pvr)); - return data; + return cpuData; } #endif -- 1.8.1.4

3 5

[libvirt] Use flock() instead of fcntl()
by David Weber 29 Jul '13

29 Jul '13

Hi, we are interested in using virtlockd on an OCFS2 shared filesystem. We are now facing the problem that virtlockd uses fcntl() locks which aren't supported by OCFS2 with the o2cb cluster stack and we want to avoid using indirect leases. OCFS2 instead supports flock() which is quite similar to fcntl(). I attached a patch which makes libvirt use flock() *instead* of fcntl() and it seems to work. NFS on the contrast only supports fcntl() so it should be configurable which lock type to use. I'm not very experienced with locking, so would such a patch be acceptable or do you see possible problems with it? Cheers, David >From b823a9a9bd60a870d64341c4273c42d4eeba8d9b Mon Sep 17 00:00:00 2001 From: David Weber <wb(a)munzinger.de> Date: Thu, 25 Jul 2013 08:20:20 +0000 Subject: [PATCH] Use flock() instead of fcntl() --- src/util/virfile.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/util/virfile.c b/src/util/virfile.c index 8f0eec3..e243c26 100644 --- a/src/util/virfile.c +++ b/src/util/virfile.c int virFileLock(int fd, bool shared, off_t start, off_t len) .l_len = len, }; - if (fcntl(fd, F_SETLK, &fl) < 0) + if (flock(fd, LOCK_EX | LOCK_NB) < 0) return -errno; return 0; int virFileUnlock(int fd, off_t start, off_t len) .l_len = len, }; - if (fcntl(fd, F_SETLK, &fl) < 0) + if (flock(fd, LOCK_UN) < 0) return -errno; return 0; -- 1.8.1.5

3 10

[libvirt] [PATCH] build: fix VPATH 'make check'
by Eric Blake 29 Jul '13

29 Jul '13

A VPATH build 'make check' was failing with: GEN check-driverimpls Can't open ../../src/../../src/lxc/lxc_monitor_protocol.h: No such file or directory at ../../src/check-driverimpls.pl line 29, <> line 27153. Can't open ../../src/../../src/lxc/lxc_monitor_protocol.c: No such file or directory at ../../src/check-driverimpls.pl line 29, <> line 27153. ... GEN check-aclrules cannot read ../../src/../../src/remote/remote_protocol.x at ../../src/check-aclrules.pl line 128. because $(srcdir) was being prepended to file names that already included it. * src/Makefile.am (check-driverimpls): Don't add srcdir twice. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- Pushing under the build-breaker rule. src/Makefile.am | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/src/Makefile.am b/src/Makefile.am index 84372cb..62e427e 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -499,11 +499,13 @@ STATEFUL_DRIVER_SOURCE_FILES = \ check-driverimpls: $(AM_V_GEN)$(PERL) $(srcdir)/check-driverimpls.pl \ $(filter /%,$(DRIVER_SOURCE_FILES)) \ - $(addprefix $(srcdir)/,$(filter-out /%,$(DRIVER_SOURCE_FILES))) + $(filter $(srcdir)/%,$(DRIVER_SOURCE_FILES)) \ + $(addprefix $(srcdir)/,$(filter-out $(srcdir)/%, \ + $(filter-out /%,$(DRIVER_SOURCE_FILES)))) check-aclrules: $(AM_V_GEN)$(PERL) $(srcdir)/check-aclrules.pl \ - $(addprefix $(srcdir)/,$(filter-out /%,$(REMOTE_PROTOCOL))) \ + $(REMOTE_PROTOCOL) \ $(addprefix $(srcdir)/,$(filter-out /%,$(STATEFUL_DRIVER_SOURCE_FILES))) EXTRA_DIST += check-driverimpls.pl check-aclrules.pl -- 1.8.3.1

1 0

[libvirt] [PATCH] python: Drop TODO
by Cole Robinson 29 Jul '13

29 Jul '13

File hasn't been really touched for 7 years. And with recent rawhide changes it contributed to an RPM build failure. Let's drop it. This also removes installation of a libvirt-python doc dir, so drop handling of it from the RPM spec. --- libvirt.spec.in | 7 ------- python/Makefile.am | 7 ------- python/TODO | 6 ------ 3 files changed, 20 deletions(-) delete mode 100644 python/TODO diff --git a/libvirt.spec.in b/libvirt.spec.in index e0e0004..81101e8 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1440,12 +1440,6 @@ rm -f $RPM_BUILD_ROOT%{_datadir}/augeas/lenses/libvirtd_lxc.aug rm -f $RPM_BUILD_ROOT%{_datadir}/augeas/lenses/tests/test_libvirtd_lxc.aug %endif -%if ! %{with_python} -rm -rf $RPM_BUILD_ROOT%{_datadir}/doc/libvirt-python-%{version} -%else -rm -rf $RPM_BUILD_ROOT%{_datadir}/doc/libvirt-python-%{version}/examples -%endif - %if ! %{with_qemu} rm -rf $RPM_BUILD_ROOT%{_sysconfdir}/libvirt/qemu.conf rm -rf $RPM_BUILD_ROOT%{_sysconfdir}/logrotate.d/libvirtd.qemu @@ -2060,7 +2054,6 @@ fi %{_libdir}/python*/site-packages/libvirt_lxc.py* %{_libdir}/python*/site-packages/libvirtmod* %doc python/tests/*.py -%doc python/TODO %doc examples/python %doc examples/domain-events/events-python %endif diff --git a/python/Makefile.am b/python/Makefile.am index 9e957d1..7eb42c6 100644 --- a/python/Makefile.am +++ b/python/Makefile.am @@ -31,10 +31,6 @@ INCLUDES = \ AM_CFLAGS = $(WARN_CFLAGS) -DOCS_DIR = $(datadir)/doc/libvirt-python-$(VERSION) - -DOCS = ${srcdir}/TODO - CLASSES_EXTRA = \ libvirt-override-virConnect.py \ libvirt-override-virDomain.py \ @@ -152,9 +148,6 @@ install-data-local: $(INSTALL) -m 0644 libvirt.py $(DESTDIR)$(pyexecdir) $(INSTALL) -m 0644 libvirt_lxc.py $(DESTDIR)$(pyexecdir) $(INSTALL) -m 0644 libvirt_qemu.py $(DESTDIR)$(pyexecdir) - $(mkinstalldirs) $(DESTDIR)$(DOCS_DIR) - @(for doc in $(DOCS) ; \ - do $(INSTALL) -m 0644 $$doc $(DESTDIR)$(DOCS_DIR) ; done) uninstall-local: rm -f $(DESTDIR)$(pyexecdir)/libvirt.py diff --git a/python/TODO b/python/TODO deleted file mode 100644 index 4627482..0000000 --- a/python/TODO +++ /dev/null @@ -1,6 +0,0 @@ -- Need to complete, add custom wrapper function for those - which could not be handled fully automatically -- Check the names generated, iD is bad, fix the name generation routine -- add examples, web page and python based test suite - -Daniel Veillard -- 1.8.3.1

2 2

[libvirt] [PATCH 0/3] qemu: libvirt RDMA live migration support
by mrhines＠linux.vnet.ibm.com 29 Jul '13

29 Jul '13

From: "Michael R. Hines" <mrhines(a)us.ibm.com> QEMU has in tree now planned for 1.6 support for RDMA-based live migration. Changes to libvirt: 1. QEMU has a new 'setup' phase in their state machine. 2. Expose the 'x-rdma' migration protocol URI. 3. Expose the 'x-rdma-pin-all' capability for pre-registration of memory. Michael R. Hines (3): qemu: handle new 'setup' migration state qemu: RDMA migration support using 'x-rdma' URI qemu: memory pre-pinning support for RDMA migration include/libvirt/libvirt.h.in | 3 + src/qemu/qemu_capabilities.c | 7 +++ src/qemu/qemu_capabilities.h | 4 ++ src/qemu/qemu_command.c | 8 +++ src/qemu/qemu_migration.c | 131 ++++++++++++++++++++++++++++++++++++------ src/qemu/qemu_migration.h | 3 +- src/qemu/qemu_monitor.c | 7 ++- src/qemu/qemu_monitor.h | 13 +++++ src/qemu/qemu_monitor_json.c | 18 ++++++ tools/virsh-domain.c | 7 +++ 10 files changed, 178 insertions(+), 23 deletions(-) -- 1.7.10.4

5 15

[libvirt] [PATCH 0/2] Rework client connection handling
by Michal Privoznik 29 Jul '13

29 Jul '13

There are few cases where users don't want to raise 'max_client', but are doing many concurrent connection and don't want them to fail too. However, we are currently accept()-ing the incoming request even though we have reached the limit. If that's the case, error is reported and connection is thrown away. Gross. What about leaving the requests we know we can't handle yet in the listen() queue and accepting only those we know we can handle? For more info see: https://bugzilla.redhat.com/show_bug.cgi?id=981729 Michal Privoznik (2): RPC: Don't accept client if it would overcommit max_clients Introduce max_queued_clients daemon/libvirtd-config.c | 1 + daemon/libvirtd-config.h | 1 + daemon/libvirtd.aug | 1 + daemon/libvirtd.c | 4 ++++ daemon/libvirtd.conf | 6 ++++++ src/locking/lock_daemon.c | 2 +- src/lxc/lxc_controller.c | 1 + src/rpc/virnetserver.c | 40 ++++++++++++++++++++++++++++++++++++++++ src/rpc/virnetserverservice.c | 15 +++++++++++++-- src/rpc/virnetserverservice.h | 6 ++++++ 10 files changed, 74 insertions(+), 3 deletions(-) -- 1.8.1.5

3 11

[libvirt] Entering freeze for libvirt-1.1.1
by Daniel Veillard 29 Jul '13

29 Jul '13

As planned I tagged the release candidate 1 for libvirt 1.1.1 in git, I also made tarball and rpms available on the ftp: ftp://libvirt.org/libvirt/ There is still some patches not completely ready after initial review it is probably safe to push the simplest ones, but not big sets now. Bug fixes are okay until final release obviously ! The release seems to work fine in my local testing (my fedora 19 crashed when shutdown a window guest, but that sounds like a kernel problem). Goal is to release next tuesday, please give it a try, and report problems :-) thanks ! Daniel -- Daniel Veillard | Open Source and Standards, Red Hat veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

4 5

[libvirt] [RFC] Image Fleecing for Libvirt (BZ 955734, 905125)
by Richard W.M. Jones 29 Jul '13

29 Jul '13

On Mon, Jul 15, 2013 at 05:57:12PM +0800, Fam Zheng wrote: > Hi all, > > QEMU-KVM BZ 955734, and libvirt BZ 905125 are about feature "Read-only > point-in-time throwaway snapshot". The development is ongoing on > upstream, which implements the core functionality by QMP command > drive-backup. I want to demonstrate the HMP/QMP commands here for image > fleecing tasks (again) and make sure this interface looks ready and > satisfying from Libvirt point of view. > > We get cheap point-in-time snapshot, and export it through built in NBD > server, by commands described below: > > 1. qemu-img create -f qcow2 -o backing_file=RUNNING-VM.img BACKUP.qcow2 > > (although the backing_file option is not honoured in the next step > because we *override* backing file with an existing > BlockDriverState, giving it here does no harm and also makes sure > the created image is of right size.) > > 2. (HMP) drive_add backing=ide0-hd0,file=BACKUP.qcow2,id=target0,if=none > > (where ide0-hd0 is the running BlockDriverState name for > RUNNING-VM.img) > > 3. (QMP) drive-backup device=ide0-hd0 mode=drive sync=none target=target0 > > (NewImageMode 'drive' means target is looked up as a device id, sync > mode 'none' means don't copy any data except copy-on-write the > point in time snapshot data) > > 4. (QMP) nbd-server-add device=target0 > > When image fleecing done: > > 1. (QMP) block-job-complete device=ide0-hd0 > > 2. (HMP) drive_del target0 > > 3. rm BACKUP.qcow2 > > Note: HMP drive_add/drive_del has no counterpart in QMP now but a new > command blockdev-add to do similar things is WIP, which can be an > alternative in QMP flavor. > > Any comments are welcome! > > -- > Best regards, > Fam Zheng -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Fedora Windows cross-compiler. Compile Windows programs, test, and build Windows installers. Over 100 libraries supported. http://fedoraproject.org/wiki/MinGW

5 11

[libvirt] [PATCH] caps: use -device for primary video when qemu >=1.6
by Guannan Ren 29 Jul '13

29 Jul '13

https://bugzilla.redhat.com/show_bug.cgi?id=981094 The commit 0ad9025ef introduce qemu flag QEMU_CAPS_DEVICE_VIDEO_PRIMARY for using -device VGA, -device cirrus-vga, -device vmware-svga and -device qxl-vga. In use, for -device qxl-vga, mouse doesn't display in guest window like the desciption in above bug. This patch try to use -device for primary video when qemu >=1.6 which is safe. --- src/qemu/qemu_capabilities.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 5dc3c9e..08406b8 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -1189,8 +1189,6 @@ virQEMUCapsComputeCmdFlags(const char *help, virQEMUCapsSet(qemuCaps, QEMU_CAPS_VNC_SHARE_POLICY); } - if (version >= 1002000) - virQEMUCapsSet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY); return 0; } @@ -2424,7 +2422,6 @@ virQEMUCapsInitQMPBasic(virQEMUCapsPtr qemuCaps) virQEMUCapsSet(qemuCaps, QEMU_CAPS_MACHINE_OPT); virQEMUCapsSet(qemuCaps, QEMU_CAPS_DUMP_GUEST_CORE); virQEMUCapsSet(qemuCaps, QEMU_CAPS_VNC_SHARE_POLICY); - virQEMUCapsSet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY); } /* Capabilities that are architecture depending @@ -2597,6 +2594,9 @@ virQEMUCapsInitQMP(virQEMUCapsPtr qemuCaps, if (qemuCaps->version >= 1003001) virQEMUCapsSet(qemuCaps, QEMU_CAPS_VNC_WEBSOCKET); + if (qemuCaps->version >= 1006000) + virQEMUCapsSet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY); + if (virQEMUCapsProbeQMPCommands(qemuCaps, mon) < 0) goto cleanup; if (virQEMUCapsProbeQMPEvents(qemuCaps, mon) < 0) -- 1.8.3.1

3 5

Re: [libvirt] Entering freeze for libvirt-1.1.1
by Daniel Veillard 29 Jul '13

29 Jul '13

On Fri, Jul 26, 2013 at 06:01:23AM -0400, Alex Jia wrote: > Hello Daniel, > I gave a basic try for libvirt-1.1.1-rc1 on RHEL6. > > [root@202 libvirt-1.1.1]# make > > Notes, sometimes, I can meet the following errors, it's not reproducible each time, > maybe, it's a network issue. > > <slice> > > I/O error : Attempt to load network entity http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd > formatdomain.html.in:2: warning: failed to load external entity "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" > C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" > ^ > formatdomain.html.in:4582: parser error : Entity 'mdash' not defined > <li> 'i6300esb' — the recommended device, you need to install xhtml1-dtds for local validation of generated HTML [...] > TEST: virnetsockettest > ........!!!.!!! 15 FAIL weird ... > Notes, works well with PATH environment setting, but it still is failed for > running 'make -C tests valgrind' due to previous 'make check' fails, whether > we my bypass/fix this for 'make check'. BTW, I also checked libvirt upstream > with "make -C tests valgrind", it's fine and without memory leaks. > > > [root@202 libvirt-1.1.1]# make syntax-check > GEN HACKING > GEN bracket-spacing-check > ./build-aux/vc-list-files: Failed to determine type of version control used in /home/ajia/Workspace/rc1/libvirt-1.1.1 > Can't open perl script "./build-aux/bracket-spacing.pl": No such file or directory > maint.mk: incorrect whitespace, see HACKING for rules > make: *** [bracket-spacing-check] Error 1 > > [root@202 libvirt-1.1.1]# ls ./build-aux/bracket-spacing.pl > ls: cannot access ./build-aux/bracket-spacing.pl: No such file or directory > > Notes, it's okay on libvirt upstream and can find ./build-aux/bracket-spacing.pl file. > > > To run coverity on libvirt-1.1.1-rc1 and attach report as an attachment. thanks ! Daniel -- Daniel Veillard | Open Source and Standards, Red Hat veillard(a)redhat.com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | virtualization library http://libvirt.org/

3 2

[libvirt] [PATCH] build: avoid uninitialized use warning
by Eric Blake 26 Jul '13

26 Jul '13

Otherwise, with new enough gcc compiling at -O2, the build fails with: ../../src/conf/domain_conf.c: In function ‘virDomainDeviceDefPostParse’: ../../src/conf/domain_conf.c:2821:29: error: ‘cnt’ may be used uninitialized in this function [-Werror=maybe-uninitialized] for (i = 0; i < *cnt; i++) { ^ ../../src/conf/domain_conf.c:2795:20: note: ‘cnt’ was declared here size_t i, *cnt; ^ ../../src/conf/domain_conf.c:2794:30: error: ‘arrPtr’ may be used uninitialized in this function [-Werror=maybe-uninitialized] virDomainChrDefPtr **arrPtr; ^ * src/conf/domain_conf.c (virDomainChrGetDomainPtrs): Always assign into output parameters. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- Pushing under the build-breaker rule. I actually found this while running ./autobuild.sh; but there, we run libtool in both static and shared modes, and libtool, in its "infinite wisdom", assumes that if the static build succeeds, then the shared build can have compiler output redirected to /dev/null, which meant I was stuck debugging this output: CC libvirt_driver_test_la-test_driver.lo CC libvirt_driver_remote_la-remote_driver.lo CC libvirt_driver_remote_la-remote_protocol.lo make[3]: *** [libvirt_conf_la-domain_conf.lo] Error 1 make[3]: *** Waiting for unfinished jobs.... thanks for nothing, libtool. src/conf/domain_conf.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 0e74039..a86be8c 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -10221,6 +10221,8 @@ virDomainChrGetDomainPtrs(virDomainDefPtr vmdef, break; case VIR_DOMAIN_CHR_DEVICE_TYPE_LAST: + *arrPtr = NULL; + *cntPtr = NULL; break; } } -- 1.8.3.1

1 0

[libvirt] [PATCH] Fix probing of legacy Xen driver to not leave URI set
by Daniel P. Berrange 26 Jul '13

26 Jul '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> When the legacy Xen driver probes with a NULL URI, and finds itself running on Xen, it will set conn->uri. A little bit later though it checks to see if libxl support exists, and if so declines the driver. This leaves the conn->uri set to 'xen:///', so if libxl also declines it, it prevents probing of the QEMU driver. Once a driver has set the conn->uri, it must *never* decline an open request. So we must move the libxl check earlier Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/xen/xen_driver.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-) diff --git a/src/xen/xen_driver.c b/src/xen/xen_driver.c index 39334b7..4ae38d3 100644 --- a/src/xen/xen_driver.c +++ b/src/xen/xen_driver.c @@ -365,6 +365,13 @@ xenUnifiedConnectOpen(virConnectPtr conn, virConnectAuthPtr auth, unsigned int f if (!xenUnifiedProbe()) return VIR_DRV_OPEN_DECLINED; +#ifdef WITH_LIBXL + /* Decline xen:// URI if xend is not running and libxenlight + * driver is potentially available. */ + if (!xenUnifiedXendProbe()) + return VIR_DRV_OPEN_DECLINED; +#endif + if (!(conn->uri = virURIParse("xen:///"))) return VIR_DRV_OPEN_ERROR; } else { @@ -374,6 +381,12 @@ xenUnifiedConnectOpen(virConnectPtr conn, virConnectAuthPtr auth, unsigned int f STRCASENEQ(conn->uri->scheme, "http")) return VIR_DRV_OPEN_DECLINED; +#ifdef WITH_LIBXL + /* Decline xen:// URI if xend is not running and libxenlight + * driver is potentially available. */ + if (!xenUnifiedXendProbe()) + return VIR_DRV_OPEN_DECLINED; +#endif /* Return an error if the path isn't '' or '/' */ if (conn->uri->path && @@ -395,13 +408,6 @@ xenUnifiedConnectOpen(virConnectPtr conn, virConnectAuthPtr auth, unsigned int f } } -#ifdef WITH_LIBXL - /* Decline xen:// URI if xend is not running and libxenlight - * driver is potentially available. */ - if (!xenUnifiedXendProbe()) - return VIR_DRV_OPEN_DECLINED; -#endif - /* We now know the URI is definitely for this driver, so beyond * here, don't return DECLINED, always use ERROR */ -- 1.8.3.1

2 1

[libvirt] [PATCH] Set default partition in libvirtd instead of libvirt_lxc
by Daniel P. Berrange 26 Jul '13

26 Jul '13

From: "Daniel P. Berrange" <berrange(a)redhat.com> By setting the default partition in libvirt_lxc it is not visible when querying the live XML. Move setting of the default partition into libvirtd virLXCProcessStart Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- src/lxc/lxc_cgroup.c | 14 -------------- src/lxc/lxc_process.c | 14 ++++++++++++++ 2 files changed, 14 insertions(+), 14 deletions(-) diff --git a/src/lxc/lxc_cgroup.c b/src/lxc/lxc_cgroup.c index af91b04..0b0ca02 100644 --- a/src/lxc/lxc_cgroup.c +++ b/src/lxc/lxc_cgroup.c @@ -433,20 +433,6 @@ virCgroupPtr virLXCCgroupCreate(virDomainDefPtr def) { virCgroupPtr cgroup = NULL; - if (!def->resource) { - virDomainResourceDefPtr res; - - if (VIR_ALLOC(res) < 0) - goto cleanup; - - if (VIR_STRDUP(res->partition, "/machine") < 0) { - VIR_FREE(res); - goto cleanup; - } - - def->resource = res; - } - if (def->resource->partition[0] != '/') { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, _("Resource partition '%s' must start with '/'"), diff --git a/src/lxc/lxc_process.c b/src/lxc/lxc_process.c index 1a5686f..247e516 100644 --- a/src/lxc/lxc_process.c +++ b/src/lxc/lxc_process.c @@ -1007,6 +1007,20 @@ int virLXCProcessStart(virConnectPtr conn, return -1; } + if (!vm->def->resource) { + virDomainResourceDefPtr res; + + if (VIR_ALLOC(res) < 0) + goto cleanup; + + if (VIR_STRDUP(res->partition, "/machine") < 0) { + VIR_FREE(res); + goto cleanup; + } + + vm->def->resource = res; + } + if (virAsprintf(&logfile, "%s/%s.log", cfg->logDir, vm->def->name) < 0) return -1; -- 1.8.1.4

2 1

[libvirt] [PATCH v2 0/2] Split driver StateAutoStart from StateInitialization
by John Ferlan 26 Jul '13

26 Jul '13

This is an update to change the 'int' AutoStart function to 'void' as was discussed in the initial review: https://www.redhat.com/archives/libvir-list/2013-July/msg01674.html John Ferlan (2): Separate out StateAutoStart from StateInitialize virStateDriver - Separate AutoStart from Initialize src/driver.h | 4 ++++ src/libvirt.c | 14 +++++++++++++- src/libxl/libxl_driver.c | 16 +++++++++++++--- src/lxc/lxc_driver.c | 16 ++++++++++++++-- src/network/bridge_driver.c | 18 +++++++++++++++++- src/qemu/qemu_driver.c | 17 +++++++++++++++-- src/storage/storage_driver.c | 18 +++++++++++++++++- src/uml/uml_driver.c | 17 +++++++++++++++-- 8 files changed, 108 insertions(+), 12 deletions(-) -- 1.8.1.4

1 3

[libvirt] [PATCH 0/7] Probe QEMU binary for host CPU and use it for computations
by Jiri Denemark 26 Jul '13

26 Jul '13

Since QEMU and kvm may filter some host CPU features or add efficiently emulated features, asking QEMU binary for host CPU data provides better results when we later use the data for building guest CPUs. Jiri Denemark (7): cpu: Add support for loading and storing CPU data cpu: Export few x86-specific APIs x86: Ignore CPUID functions greater than 10 qemu: Add monitor APIs to fetch CPUID data from QEMU qemu: Make QMP probing process reusable qemu: Probe QEMU binary for host CPU qemu: Use host CPU from QEMU for computations src/cpu/cpu.c | 41 ++++ src/cpu/cpu.h | 13 ++ src/cpu/cpu_x86.c | 161 +++++++++++--- src/cpu/cpu_x86.h | 10 + src/cpu/cpu_x86_data.h | 1 + src/libvirt_private.syms | 9 + src/qemu/qemu_capabilities.c | 234 ++++++++++++++------- src/qemu/qemu_capabilities.h | 2 + src/qemu/qemu_command.c | 32 ++- src/qemu/qemu_domain.c | 21 +- src/qemu/qemu_monitor.c | 21 ++ src/qemu/qemu_monitor.h | 3 + src/qemu/qemu_monitor_json.c | 162 ++++++++++++++ src/qemu/qemu_monitor_json.h | 6 + tests/Makefile.am | 1 + .../qemumonitorjson-getcpu-empty.data | 2 + .../qemumonitorjson-getcpu-empty.json | 46 ++++ .../qemumonitorjson-getcpu-filtered.data | 4 + .../qemumonitorjson-getcpu-filtered.json | 46 ++++ .../qemumonitorjson-getcpu-full.data | 4 + .../qemumonitorjson-getcpu-full.json | 46 ++++ .../qemumonitorjson-getcpu-host.data | 5 + .../qemumonitorjson-getcpu-host.json | 45 ++++ tests/qemumonitorjsontest.c | 74 +++++++ 24 files changed, 881 insertions(+), 108 deletions(-) create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-empty.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-empty.json create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-filtered.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-filtered.json create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-full.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-full.json create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-host.data create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-getcpu-host.json -- 1.8.3.2

5 32

[libvirt] [PATCH 0/4] fix two issues about diskchain
by Guannan Ren 26 Jul '13

26 Jul '13

1, Report an backing missing error on guest boot if its disks have broken backing file chain 2, List volume even if its diskchain is broken Guannan Ren(4) [PATCH 1/4] qemu: refactor qemuDomainCheckDiskPresence for only disk [PATCH 2/4] qemu: report error if disk backing files doesn't exist [PATCH 3/4] qemu: check presence of each disk in chain [PATCH 4/4] storage: list volumes even if its diskchain is broken src/qemu/qemu_domain.c | 108 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------------------------------------------- src/qemu/qemu_process.c | 7 ------- src/storage/storage_backend_fs.c | 4 +++- src/util/virstoragefile.c | 23 +++++++++++++++-------- tests/virstoragetest.c | 16 ++++++++-------- 5 files changed, 90 insertions(+), 68 deletions(-)

2 8

[libvirt] [PATCH] build: avoid -lgcrypt with newer gnutls
by Eric Blake 26 Jul '13

26 Jul '13

https://bugzilla.redhat.com/show_bug.cgi?id=951637 Newer gnutls uses nettle, rather than gcrypt, which is a lot nicer regarding initialization. Yet we were unconditionally initializing gcrypt even when gnutls wouldn't be using it, and having two crypto libraries linked into libvirt.so is pointless. The ldd probe in configure borrows from our libnl-1 vs. libnl-3 code. * configure.ac (WITH_GNUTLS): Probe whether to add -lgcrypt, and define a witness WITH_GNUTLS_GCRYPT. * src/libvirt.c (virTLSMutexInit, virTLSMutexDestroy) (virTLSMutexLock, virTLSMutexUnlock, virTLSThreadImpl) (virGlobalInit): Honor the witness. * libvirt.spec.in (BuildRequires): Make gcrypt usage conditional, no longer needed in Fedora 19. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- Tested with 'ldd src/.libs/libvirt.so | grep -E "(gcry|net|tls)"': - on RHEL 6.4 and Fedora 18, pre- and post-patch remain unchanged (use of just libgnutls/libgcrypt) - on Fedora 19, pre-patch linked against libgnutls, libgcrypt, and libnettle, post-patch linked against just libgnutls and libnettle This should probably go in for 1.1.1, but it's not a build-breaker so it needs review. configure.ac | 27 +++++++++++++++++++++------ libvirt.spec.in | 2 ++ src/libvirt.c | 10 ++++++---- 3 files changed, 29 insertions(+), 10 deletions(-) diff --git a/configure.ac b/configure.ac index cc9942a..de209e2 100644 --- a/configure.ac +++ b/configure.ac @@ -1098,13 +1098,28 @@ if test "x$with_gnutls" != "xno"; then AC_MSG_ERROR([You must install the GnuTLS library in order to compile and run libvirt]) fi else - dnl Not all versions of gnutls include -lgcrypt, and so we add - dnl it explicitly for the calls to gcry_control/check_version - GNUTLS_LIBS="$GNUTLS_LIBS -lgcrypt" - dnl We're not using gcrypt deprecated features so define - dnl GCRYPT_NO_DEPRECATED to avoid deprecated warnings - GNUTLS_CFLAGS="$GNUTLS_CFLAGS -DGCRYPT_NO_DEPRECATED" + dnl If gnutls linked against -lgcrypt, then we must initialize gcrypt + dnl prior to using gnutls. Newer versions of gnutls use -lnettle, in + dnl which case we don't want to drag in gcrypt ourselves. + gnutls_ldd= + for dir in /usr/lib64 /usr/lib /usr/lib/*-linux-gnu*; do + if test -f $dir/libgnutls.so; then + gnutls_ldd=`(ldd $dir/libgnutls.so) 2>&1` + break + fi + done + case $gnutls_ldd in + '' | *libgcrypt.so*) + GNUTLS_LIBS="$GNUTLS_LIBS -lgcrypt" + dnl We're not using gcrypt deprecated features so define + dnl GCRYPT_NO_DEPRECATED to avoid deprecated warnings + GNUTLS_CFLAGS="$GNUTLS_CFLAGS -DGCRYPT_NO_DEPRECATED" + AC_DEFINE_UNQUOTED([WITH_GNUTLS_GCRYPT], 1, + [whether GNUTLS uses gcrypt]) + ;; + *) ;; # Assume no gcrypt usage + esac dnl gnutls 3.x moved some declarations to a new header AC_CHECK_HEADERS([gnutls/crypto.h], [], [], [[ diff --git a/libvirt.spec.in b/libvirt.spec.in index e0e0004..4320281 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -422,7 +422,9 @@ BuildRequires: readline-devel BuildRequires: ncurses-devel BuildRequires: gettext BuildRequires: libtasn1-devel +%if (0%{?rhel} && 0%{?rhel} < 7) || (0%{?fedora} && 0%{?fedora} < 19) BuildRequires: libgcrypt-devel +%endif BuildRequires: gnutls-devel BuildRequires: libattr-devel %if %{with_libvirtd} diff --git a/src/libvirt.c b/src/libvirt.c index 444c1c3..9775b97 100644 --- a/src/libvirt.c +++ b/src/libvirt.c @@ -55,7 +55,9 @@ #include "intprops.h" #include "virconf.h" #if WITH_GNUTLS -# include <gcrypt.h> +# if WITH_GNUTLS_GCRYPT +# include <gcrypt.h> +# endif # include "rpc/virnettlscontext.h" #endif #include "vircommand.h" @@ -270,7 +272,7 @@ winsock_init(void) #endif -#ifdef WITH_GNUTLS +#ifdef WITH_GNUTLS_GCRYPT static int virTLSMutexInit(void **priv) { virMutexPtr lock = NULL; @@ -323,7 +325,7 @@ static struct gcry_thread_cbs virTLSThreadImpl = { virTLSMutexUnlock, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL }; -#endif +#endif /* WITH_GNUTLS_GCRYPT */ /* Helper macros to implement VIR_DOMAIN_DEBUG using just C99. This * assumes you pass fewer than 15 arguments to VIR_DOMAIN_DEBUG, but @@ -407,7 +409,7 @@ virGlobalInit(void) virErrorInitialize() < 0) goto error; -#ifdef WITH_GNUTLS +#ifdef WITH_GNUTLS_GCRYPT /* * This sequence of API calls it copied exactly from * gnutls 2.12.23 source lib/gcrypt/init.c, with -- 1.8.3.1

2 3

[libvirt] [PATCH 0/2] Split driver StateAutoStart from StateInitialization
by John Ferlan 26 Jul '13

26 Jul '13

The post push review/comments for the chap authentication determined that trying to connect to qemu driver from within the storage auto start would not be successful, see the following and followups https://www.redhat.com/archives/libvir-list/2013-July/msg01409.html These patches will split the virStateInitialize() into two phases - the first being the bulk of the existing code and the second being just running the auto start functionality for each of the drivers that needs/support it. I realize it's probably too late for 1.1.1, but figured it'd be good to be ready when the barn door opens again. John Ferlan (2): Separate out StateAutoStart from StateInitialize virStateDriver - Separate AutoStart from Initialize src/driver.h | 4 ++++ src/libvirt.c | 23 ++++++++++++++++++++++- src/libxl/libxl_driver.c | 18 +++++++++++++++--- src/lxc/lxc_driver.c | 18 ++++++++++++++++-- src/network/bridge_driver.c | 20 +++++++++++++++++++- src/qemu/qemu_driver.c | 18 ++++++++++++++++-- src/storage/storage_driver.c | 19 ++++++++++++++++++- src/uml/uml_driver.c | 18 ++++++++++++++++-- 8 files changed, 126 insertions(+), 12 deletions(-) -- 1.8.1.4

4 9

[libvirt] [PATCH 0/4] Test logical volume cloning in storagevolxml2argvtest
by Ján Tomko 26 Jul '13

26 Jul '13

Ján Tomko (4): Use separate macros for failure/success in vol-to-argv test Move pool XML out of storagevolxml2argvdata Move volume XMLs out of storagevolxml2argvdata Add inputpool to storagevolxml2argvtest .../storagevolxml2argvdata/logical-from-qcow2.argv | 2 + tests/storagevolxml2argvdata/pool-dir.xml | 18 --- .../storagevolxml2argvdata/qcow2-from-logical.argv | 2 + tests/storagevolxml2argvdata/vol-file.xml | 20 ---- tests/storagevolxml2argvdata/vol-qcow2-1.1.xml | 32 ------ tests/storagevolxml2argvtest.c | 128 ++++++++++++++++----- .../vol-qcow2-0.10-lazy.xml | 0 tests/storagevolxml2xmlin/vol-qcow2-1.1.xml | 2 +- tests/storagevolxml2xmlin/vol-qcow2-lazy.xml | 2 +- .../vol-qcow2-nobacking.xml | 0 tests/storagevolxml2xmlin/vol-qcow2.xml | 2 +- .../vol-qcow2-0.10-lazy.xml} | 8 +- tests/storagevolxml2xmlout/vol-qcow2-1.1.xml | 2 +- tests/storagevolxml2xmlout/vol-qcow2-lazy.xml | 2 +- .../vol-qcow2-nobacking.xml} | 16 +-- tests/storagevolxml2xmlout/vol-qcow2.xml | 2 +- tests/storagevolxml2xmltest.c | 2 + 17 files changed, 120 insertions(+), 120 deletions(-) create mode 100644 tests/storagevolxml2argvdata/logical-from-qcow2.argv delete mode 100644 tests/storagevolxml2argvdata/pool-dir.xml create mode 100644 tests/storagevolxml2argvdata/qcow2-from-logical.argv delete mode 100644 tests/storagevolxml2argvdata/vol-file.xml delete mode 100644 tests/storagevolxml2argvdata/vol-qcow2-1.1.xml rename tests/{storagevolxml2argvdata => storagevolxml2xmlin}/vol-qcow2-0.10-lazy.xml (100%) rename tests/{storagevolxml2argvdata => storagevolxml2xmlin}/vol-qcow2-nobacking.xml (100%) rename tests/{storagevolxml2argvdata/vol-qcow2-lazy.xml => storagevolxml2xmlout/vol-qcow2-0.10-lazy.xml} (84%) rename tests/{storagevolxml2argvdata/vol-qcow2.xml => storagevolxml2xmlout/vol-qcow2-nobacking.xml} (55%) -- 1.8.1.5

3 13

[libvirt] [PATCH 0/4] Fix a pair of crashes in DNS XML parsing
by Ján Tomko 26 Jul '13

26 Jul '13

*** BLUAAARRRGHB HERE *** Ján Tomko (4): Set the number of elements to 0 in virNetwork*Clear Don't check validity of missing attributes in DNS SRV XML Remove double space in error messages Remove redundant free in virNetworkDNSHostDefParseXML src/conf/network_conf.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) -- 1.8.1.5

3 13

[libvirt] [PATCH] conf:Fix a copy paste error
by Alex Jia 26 Jul '13

26 Jul '13

Signed-off-by: Alex Jia <ajia(a)redhat.com> --- src/conf/domain_conf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 10cb7f6..0e74039 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -5153,7 +5153,7 @@ virDomainDiskDefParseXML(virDomainXMLOptionPtr xmlopt, xmlStrEqual(cur->name, BAD_CAST "product")) { product = (char *)xmlNodeGetContent(cur); - if (strlen(vendor) > PRODUCT_LEN) { + if (strlen(product) > PRODUCT_LEN) { virReportError(VIR_ERR_XML_ERROR, "%s", _("disk product is more than 16 characters")); goto error; -- 1.8.3.1

3 3

[libvirt] [PATCH] formatdomain.html.in: Correctly use <code/> in #elementQoS
by Michal Privoznik 26 Jul '13

26 Jul '13

Interestingly, we had <code>floor<code> ... </code>outbound</code> which results in much larger block of text to be written in code style that intended. --- docs/formatdomain.html.in | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in index 7601aaa..d2cee67 100644 --- a/docs/formatdomain.html.in +++ b/docs/formatdomain.html.in @@ -3539,12 +3539,12 @@ qemu-kvm -net nic,model=? /dev/null forward type of route, nat, or no forward at all). Moreover, the virtual network the interface is connected to is required to have at least inbound QoS set (<code>average</code> at least). Moreover, with - <code>floor<code> attribute users don't need to specify + <code>floor</code> attribute users don't need to specify <code>average</code>. However, <code>peak</code> and <code>burst</code> attributes still require <code>average</code>. Currently, linux kernel doesn't allow ingress qdiscs to have any classes therefore <code>floor</code> can be applied only on <code>inbound</code> and not - </code>outbound</code>. <span class="since">Since 1.0.1</span> + <code>outbound</code>. <span class="since">Since 1.0.1</span> </p> <h5><a name="elementVlanTag">Setting VLAN tag (on supported network types only)</a></h5> -- 1.8.1.5

2 3

[libvirt] [PATCH] Fix virsh snapshot-list error reporting
by Jim Fehlig 26 Jul '13

26 Jul '13

Noticed that the expected "not supported" error is dropped when invoking 'virsh snapshot-list dom' on a Xen installation running the libxl driver virsh snapshot-list test error: Invalid snapshot: virDomainSnapshotFree The error is overwritten by a call to virDomainSnapshotFree in cleanup code within cmdSnapshotList. Prevent overwritting the real error by not calling virDomainSnapshotFree with a NULL virDomainSnapshotPtr. --- This is one possible fix for the bug. The other would be to silently return in virDomainSnapshotFree on a NULL virDomainSnapshotPtr. tools/virsh-snapshot.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/tools/virsh-snapshot.c b/tools/virsh-snapshot.c index cfe8ee9..db9715b 100644 --- a/tools/virsh-snapshot.c +++ b/tools/virsh-snapshot.c @@ -1678,7 +1678,8 @@ cleanup: vshSnapshotListFree(snaplist); VIR_FREE(parent_snap); VIR_FREE(state); - virDomainSnapshotFree(start); + if (start) + virDomainSnapshotFree(start); xmlXPathFreeContext(ctxt); xmlFreeDoc(xml); VIR_FREE(doc); -- 1.8.1.4

2 2

[libvirt] [PATCH 0/7] backport of getGroupList to v0.10.2-maint
by Eric Blake 25 Jul '13

25 Jul '13

https://bugzilla.redhat.com/show_bug.cgi?id=964358 Since it was on Fedora 18 that I first noticed the deadlock possible when a child process calls getpwuid_r while the parent owned the lock in a different thread, I'm interested in backporting my recent work on virGetGroupList to v0.10.2-maint. However, I hit enough conflict resolution that I'd like a review of my backport decisions before pushing this to the stable branch. Daniel P. Berrange (1): Fix potential deadlock across fork() in QEMU driver Eric Blake (6): util: improve user lookup helper util: add virGetGroupList util: make virSetUIDGID async-signal-safe security: framework for driver PreFork handler security_dac: compute supplemental groups before fork security: fix deadlock with prefork configure.ac | 6 +- src/libvirt_private.syms | 3 + src/lxc/lxc_container.c | 2 +- src/qemu/qemu_process.c | 8 ++ src/security/security_dac.c | 56 ++++++++-- src/security/security_driver.h | 4 + src/security/security_manager.c | 30 +++++ src/security/security_manager.h | 3 + src/security/security_stack.c | 28 +++++ src/storage/storage_backend.c | 16 ++- src/util/util.c | 237 ++++++++++++++++++++++++---------------- src/util/util.h | 4 +- 12 files changed, 285 insertions(+), 112 deletions(-) -- 1.8.3.1

2 16