- Devel - libvirt List Archives

[libvirt] [PATCH] bhyve: fix interface type handling for argv2xml
by Roman Bogorodskiy 19 Jan '17

19 Jan '17

When generating a domain XML from native command (i.e. via the connectDomainXMLFromNative call), we should use interface type 'bridge' rather than 'ethernet' because we only support bridges at this point. As we don't have bridge name explicitly specified on the command line, just use 'virbr0' as a default. --- src/bhyve/bhyve_parse_command.c | 9 ++++++--- tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net.xml | 6 ++++-- tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net4.xml | 3 ++- 3 files changed, 12 insertions(+), 6 deletions(-) diff --git a/src/bhyve/bhyve_parse_command.c b/src/bhyve/bhyve_parse_command.c index 619004298..5b8712388 100644 --- a/src/bhyve/bhyve_parse_command.c +++ b/src/bhyve/bhyve_parse_command.c @@ -507,9 +507,12 @@ bhyveParsePCINet(virDomainDefPtr def, if (VIR_ALLOC(net) < 0) goto cleanup; - /* Let's just assume it is VIR_DOMAIN_NET_TYPE_ETHERNET, it could also be - * a bridge, but this is the most generic option. */ - net->type = VIR_DOMAIN_NET_TYPE_ETHERNET; + /* As we only support interface type='bridge' and cannot + * guess the actual bridge name from the command line, + * try to come up with some reasonable defaults */ + net->type = VIR_DOMAIN_NET_TYPE_BRIDGE; + if (VIR_STRDUP(net->data.bridge.brname, "virbr0") < 0) + goto error; net->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI; net->info.addr.pci.slot = pcislot; diff --git a/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net.xml b/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net.xml index 09cc79b92..5895c8c53 100644 --- a/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net.xml +++ b/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net.xml @@ -12,13 +12,15 @@ <on_reboot>destroy</on_reboot> <on_crash>destroy</on_crash> <devices> - <interface type='ethernet'> + <interface type='bridge'> <mac address='52:54:00:00:00:00'/> + <source bridge='virbr0'/> <target dev='tap0'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x0'/> </interface> - <interface type='ethernet'> + <interface type='bridge'> <mac address='fe:ed:ad:ea:df:15'/> + <source bridge='virbr0'/> <target dev='tap1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/> </interface> diff --git a/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net4.xml b/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net4.xml index e1bda46a1..5f1972080 100644 --- a/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net4.xml +++ b/tests/bhyveargv2xmldata/bhyveargv2xml-virtio-net4.xml @@ -12,8 +12,9 @@ <on_reboot>destroy</on_reboot> <on_crash>destroy</on_crash> <devices> - <interface type='ethernet'> + <interface type='bridge'> <mac address='00:00:00:00:00:00'/> + <source bridge='virbr0'/> <target dev='tap1'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x1'/> </interface> -- 2.11.0

2 2

[libvirt] The libvirt Release Notes Game, v3.0.0 edition
by Andrea Bolognani 19 Jan '17

19 Jan '17

Hello everyone, and welcome to the libvirt Release Notes Game, v3.0.0 edition! Here's the deal: libvirt v3.0.0 is right around the corner, and so it's about time we get the release notes into shape. Most of the work has thankfully already been done during the development cycle, by bundling release notes entries with the code that implements the actual changes; however, some changes have managed to fly under the radar so far. Not documenting them properly in the release notes would be a disservice to our users, and would undersell what is once again a terrific release - I'd venture as far as calling it the best one of 2017! I've gone through all the commits between v2.5.0 and now, and highlighted the changes that look like they should be part of the release notes. Look for your name in the list below, and spend a few minutes to craft a beautiful release notes entry for each of the highlighted changes, then send the patch to the list as usual. Feel free to CC: me for a guaranteed[1] swift review! Please note that, unlike you, I don't have intimate knowledge of each and every line of code you've ever contributed to libvirt, and as such might have misjudged the weight and scope of the commits listed below, but that's the whole point: the only person worthy of documenting a change in the release notes is the author of the change himself! That said, if you feel that any of the changes below should not end up in the release notes feel free to raise an objection :) Happy documenting! ^^ * Cédric Bosdonnat - a30b08b libxl: define a per-domain logger. - 340bb6b libxl: add QED disk format support * Collin L. Walling, Jason J. Herne - 79d7201 s390: Cpu driver support for update and compare ..d47db7b qemu: command: Support new cpu feature argument syntax * Daniel P. Berrange - c500701 Add domain event for metadata changes ..dc2bfdc Update remote_protocol-structs for new events - 44f79a0 lxc: ensure libvirt_lxc and qemu-nbd move into systemd machine slice * intrigeri - de79efd AppArmor policy: support merged-/usr. ..a73e703 AppArmor: allow QEMU to set_process_name. * John Ferlan - 2b13361 nodedev: Add the ability to create vHBA by parent wwnn/wwpn or fabric_wwn * Laine Stump - 9838cad conf: use struct instead of int for each slot in virDomainPCIAddressBus ..5949b53 conf: eliminate virDomainPCIAddressReleaseSlot() in favor of ...Addr() * Lin Ma - 2922cd9 cpu: Add support for more AVX512 Intel features ..c80e6b9 cpu: Add support for pku and ospke Intel features for Memory Protection Keys * Marc Hartmayer - c344d4b conf: simplify functions virDomainSCSIDriveAddressIsUsedBy*() ..36d9965 tests: add test cases for address conflicts * Maxim Nestratov - ef5c8bb qemu: Fix pit timer tick policy=delay ..245d9ba tests: Add "no-kvm-pit-device" testcase * Mehdi Abaakouk - 013df87 Gathering vhostuser interface stats with ovs * Michal Privoznik - f55afd8 qemu: Create hugepage path on per domain basis * Nikolay Shirokovskiy - 61a0026 qemu: Fix xml dump of autogenerated websocket * Pavel Glushchak - 5bafa1d vz: set PVMT_DONT_CREATE_DISK migration flag ..b1f916a vz: added VIR_MIGRATE_NON_SHARED_INC migration flag support * Peter Krempa - b469853 qemu: blockjob: Fix locking of block copy/active block commit - f61e406 qemu: snapshot: Properly handle image locking - a4ed5b4 qemu: Don't try to find compression program for "raw" memory images - 9e93055 qemu: block copy: Forbid block copy to relative paths * Pino Toscano - 408a1ce rpc: libssh: allow a NULL known_hosts file ..1a5de3f remote: do not check for an existing config dir [1] Or you money back! -- Andrea Bolognani / Red Hat / Virtualization

4 4

[libvirt] [PATCH for 3.0.x] Disable use of namespaces by default
by Daniel P. Berrange 19 Jan '17

19 Jan '17

When namespaces are enabled there is currently breakage when using disk hotplug and when using AppArmor Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- I'm suggesting this for 3.0.x branch - we'll leave them enabled in master on the basis that we'll actually fix the real bugs there. src/qemu/qemu_conf.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c index 6613d59..aa05b46 100644 --- a/src/qemu/qemu_conf.c +++ b/src/qemu/qemu_conf.c @@ -317,13 +317,6 @@ virQEMUDriverConfigPtr virQEMUDriverConfigNew(bool privileged) if (!(cfg->namespaces = virBitmapNew(QEMU_DOMAIN_NS_LAST))) goto error; -#if defined(__linux__) - if (privileged && - virProcessNamespaceAvailable(VIR_PROCESS_NAMESPACE_MNT) == 0 && - virBitmapSetBit(cfg->namespaces, QEMU_DOMAIN_NS_MOUNT) < 0) - goto error; -#endif /* defined(__linux__) */ - #ifdef DEFAULT_LOADER_NVRAM if (virFirmwareParseList(DEFAULT_LOADER_NVRAM, &cfg->firmwares, -- 2.9.3

3 3

[libvirt] [PATCH] storage: avoid use of undefined GLUSTER_CLI variable
by Daniel P. Berrange 19 Jan '17

19 Jan '17

Previous commit tried to change configure logic such that the GLUSTER_CLI parameter would always be set: commit 9e97c8c0f0f3921d06bac2b92cd094a41373f748 Author: Peter Krempa <pkrempa(a)redhat.com> Date: Mon Jan 9 15:56:12 2017 +0100 storage: gluster: Remove build-time dependency on the 'gluster' cli tool This missed the fact that the AC_PATH_PROG call was itself inside an 'if' conditional that would not be called in with_storage_gluster was false. As a result, GLUSTER_CLI was still conditionally defined. Just kill the GLUSTER_CLI parameter and AC_PATH_PROG call entirely and pass a bare "gluster" string to virFindFileInPath instead. Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com> --- Technically this is a build-break fix for FreeBSD, but I want to get a review to double-check before pushing this. configure.ac | 9 --------- src/storage/storage_util.c | 2 +- 2 files changed, 1 insertion(+), 10 deletions(-) diff --git a/configure.ac b/configure.ac index a217fc1..7efaddb 100644 --- a/configure.ac +++ b/configure.ac @@ -603,15 +603,6 @@ LIBVIRT_STORAGE_CHECK_SHEEPDOG LIBVIRT_STORAGE_CHECK_GLUSTER LIBVIRT_STORAGE_CHECK_ZFS -if test "$with_storage_fs" = "yes" || - test "$with_storage_gluster" = "yes"; then - AC_PATH_PROG([GLUSTER_CLI], [gluster], [gluster], [$LIBVIRT_SBIN_PATH]) - if test "x$GLUSTER_CLI" != "x"; then - AC_DEFINE_UNQUOTED([GLUSTER_CLI], ["$GLUSTER_CLI"], - [Location or name of the gluster command line tool]) - fi -fi - with_storage=no for backend in dir fs lvm iscsi scsi mpath rbd disk; do if eval test \$with_storage_$backend = yes; then diff --git a/src/storage/storage_util.c b/src/storage/storage_util.c index 895c623..e16c1a4 100644 --- a/src/storage/storage_util.c +++ b/src/storage/storage_util.c @@ -2437,7 +2437,7 @@ virStorageBackendFindGlusterPoolSources(const char *host, int ret = -1; - if (!(glusterpath = virFindFileInPath(GLUSTER_CLI))) { + if (!(glusterpath = virFindFileInPath("gluster"))) { if (report) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("'gluster' command line tool not found")); -- 2.9.3

2 1

[libvirt] [PATCH] Fix segfault in GetIOThreadInfo
by Leonid Podolny 19 Jan '17

19 Jan '17

The pointers calculation included a typo which caused segfaults. Signed-off-by: Leonid Podolny <leonid(a)podolny.net> --- domain.go | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/domain.go b/domain.go index 515138a..74061c2 100644 --- a/domain.go +++ b/domain.go @@ -3887,7 +3887,7 @@ func (d *Domain) GetIOThreadInfo(flags DomainModificationImpact) ([]DomainIOThre info := make([]DomainIOThreadInfo, int(ret)) for i := 0; i < int(ret); i++ { - cinfo := (*(**C.virDomainIOThreadInfo)(unsafe.Pointer(uintptr(unsafe.Pointer(cinfolist)) + (unsafe.Sizeof(*cinfolist) + uintptr(i))))) + cinfo := (*(**C.virDomainIOThreadInfo)(unsafe.Pointer(uintptr(unsafe.Pointer(cinfolist)) + (unsafe.Sizeof(*cinfolist) * uintptr(i))))) ncpus := int(cinfo.cpumaplen * 8) cpumap := make([]bool, ncpus) -- 2.11.0

2 1

[libvirt] [PATCH go] Fix segfault in GetIOThreadInfo
by Leonid Podolny 19 Jan '17

19 Jan '17

2 1

[libvirt] [PATCH] qemu: Reset hostModelInfo in virQEMUCapsReset
by Jiri Denemark 19 Jan '17

19 Jan '17

Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/qemu/qemu_capabilities.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 89e9dd471..1e1b53b22 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -3752,6 +3752,9 @@ virQEMUCapsReset(virQEMUCapsPtr qemuCaps) VIR_FREE(qemuCaps->gicCapabilities); qemuCaps->ngicCapabilities = 0; + qemuMonitorCPUModelInfoFree(qemuCaps->hostCPUModelInfo); + qemuCaps->hostCPUModelInfo = NULL; + virCPUDefFree(qemuCaps->hostCPUModel); qemuCaps->hostCPUModel = NULL; } -- 2.11.0

3 2

[libvirt] [PATCH 0/6] storage: Cleanups of the storage backends
by Peter Krempa 19 Jan '17

19 Jan '17

I'm working on splitting of the storage backend codebase into loadable modules so that ceratin libraries (ceph, gluster) can be installed only when the functionality is required. This is a dump of cleanup patches, but they tend to get merge conflicts pretty often so I'm sending them prior to finishing the driver module stuff. Peter Krempa (6): storage: Remove common code from specific driver backend storage: Split utility functions from storage_backend.(ch) storage: fs: Compile file backends even if filesystem support is disabled storage: scsi: Fix build if SCSI backend is disabled but iSCSI is enabled storage: scsi: Remove private constants from header storage: sheepdog: Split out functions required for tests po/POTFILES.in | 1 + src/Makefile.am | 6 +- src/storage/storage_backend.c | 2807 +--------------------- src/storage/storage_backend.h | 124 - src/storage/storage_backend_disk.c | 153 +- src/storage/storage_backend_disk.h | 3 - src/storage/storage_backend_fs.c | 4 +- src/storage/storage_backend_gluster.c | 1 + src/storage/storage_backend_iscsi.c | 2 +- src/storage/storage_backend_logical.c | 1 + src/storage/storage_backend_mpath.c | 1 + src/storage/storage_backend_rbd.c | 1 + src/storage/storage_backend_scsi.c | 452 +--- src/storage/storage_backend_scsi.h | 8 - src/storage/storage_backend_sheepdog.c | 2 + src/storage/storage_backend_sheepdog.h | 5 - src/storage/storage_backend_sheepdog_priv.h | 29 + src/storage/storage_backend_zfs.c | 1 + src/storage/storage_driver.c | 1 + src/storage/storage_util.c | 3364 +++++++++++++++++++++++++++ src/storage/storage_util.h | 151 ++ tests/storagebackendsheepdogtest.c | 2 +- tests/storagevolxml2argvtest.c | 2 +- 23 files changed, 3602 insertions(+), 3519 deletions(-) create mode 100644 src/storage/storage_backend_sheepdog_priv.h create mode 100644 src/storage/storage_util.c create mode 100644 src/storage/storage_util.h -- 2.11.0

2 18

[libvirt] [PATCH V2] tests: fix compilation of shunloadtest
by Jim Fehlig 19 Jan '17

19 Jan '17

While local builds succeed fine, a build worker building in a chroot environment is encountering the following error with libvirt 3.0.0 release candidates [ 162s] shunloadtest.o: In function `main': [ 162s] /home/abuild/rpmbuild/BUILD/libvirt-3.0.0/tests/shunloadtest.c:110: undefined reference to `dlopen' [ 162s] /home/abuild/rpmbuild/BUILD/libvirt-3.0.0/tests/shunloadtest.c:114: undefined reference to `dlsym' [ 162s] /home/abuild/rpmbuild/BUILD/libvirt-3.0.0/tests/shunloadtest.c:133: undefined reference to `dlclose' [ 162s] /home/abuild/rpmbuild/BUILD/libvirt-3.0.0/tests/shunloadtest.c:111: undefined reference to `dlerror' [ 162s] /home/abuild/rpmbuild/BUILD/libvirt-3.0.0/tests/shunloadtest.c:115: undefined reference to `dlerror' [ 162s] /home/abuild/rpmbuild/BUILD/libvirt-3.0.0/tests/shunloadtest.c:116: undefined reference to `dlclose' Fix by appending DLOPEN_LIBS to shunloadtest_LDADD. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- V2: Use DLOPEN_LIBS instead of LDADDS tests/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/Makefile.am b/tests/Makefile.am index c7d4748..bcc2d1e 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -1348,7 +1348,7 @@ libshunload_la_LDFLAGS = $(MOCKLIBS_LDFLAGS) shunloadtest_SOURCES = \ shunloadtest.c -shunloadtest_LDADD = $(LIB_PTHREAD) +shunloadtest_LDADD = $(LIB_PTHREAD) $(DLOPEN_LIBS) shunloadtest_DEPENDENCIES = libshunload.la sysinfotest_SOURCES = \ -- 2.9.2

2 2

[libvirt] [PATCH] nss: Remove RES_USE_INET6 usage
by Andrea Bolognani 18 Jan '17

18 Jan '17

The recent deprecation in glibc (commit b76e065991ec) means the module will fail to build entirely: nss/libvirt_nss.c: In function '_nss_libvirt_gethostbyname_r': nss/libvirt_nss.c:363:13: error: RES_USE_INET6 is deprecated [-Werror] int af = ((_res.options & RES_USE_INET6) ? AF_INET6 : AF_INET); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~ This resolver option was removed shortly after being introduced, and application using it are already broken anyway. --- tools/nss/libvirt_nss.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/tools/nss/libvirt_nss.c b/tools/nss/libvirt_nss.c index b69e62c..9904026 100644 --- a/tools/nss/libvirt_nss.c +++ b/tools/nss/libvirt_nss.c @@ -360,9 +360,7 @@ NSS_NAME(gethostbyname)(const char *name, struct hostent *result, char *buffer, size_t buflen, int *errnop, int *herrnop) { - int af = ((_res.options & RES_USE_INET6) ? AF_INET6 : AF_INET); - - return NSS_NAME(gethostbyname3)(name, af, result, buffer, buflen, + return NSS_NAME(gethostbyname3)(name, AF_INET, result, buffer, buflen, errnop, herrnop, NULL, NULL); } -- 2.7.4

2 1