- Devel - libvirt List Archives

[libvirt] [PATCH] maint: avoid 'make syntax-check' from tarball
by Eric Blake 18 Feb '11

18 Feb '11

* .gnulib: update to latest gnulib for maint.mk fixes --- In testing the latest release candidate, I ran 'make syntax-check' out of habit, and it died miserably (even hanging waiting for input on stdin). So I fixed maint.mk upstream: http://thread.gmane.org/gmane.comp.lib.gnulib.bugs/25340 However, we're so close to the release that I don't know if we want this now, or if we just document that 0.8.8 is the last release where 'make syntax-check' on the tarball is expected to fail. * .gnulib ecb020f...aa0f5d7 (30): > maintainer-makefile: make syntax-check a no-op from tarballs > longlong: tune, particularly for common case of c99 > getloadavg: set errno > getloadavg: omit unused var > autoupdate > doc: update users.txt > Consistent macro naming for macros that use GCC __attribute__. > Don't interfere with a program's definition of __attribute__. > maint: correct a ChangeLog attribution > update from texinfo > mbrtowc: Add more tests for native Windows platforms. > mbrtowc: Work around native Windows bug. > mbsinit: Work around mingw bug. > mbsinit: Don't crash for a NULL argument. > Don't interfere with a program's definition of __attribute__. > Fix last ChangeLog entry. > stdlib: don't get in the way of non-GCC __attribute__ > quotearg test: Avoid test failure on mingw. > setlocale: Prefer gnulib's override over libintl's override. > stdlib: support non-GCC __attribute__ > wcsrtombs: Work around bug on native Windows. > mbsrtowcs: Work around bug on native Windows. > Avoid setlocale bugs in tests. > Fix ChangeLog entry. > setlocale: Workaround native Windows bug. > Remove unused test-setlocale1 argument. > Tests for module 'setlocale'. > New module 'setlocale'. > Prepare for locale dependent tests on mingw. > Prepare for locale dependent tests on mingw. .gnulib | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/.gnulib b/.gnulib index ecb020f..aa0f5d7 160000 --- a/.gnulib +++ b/.gnulib @@ -1 +1 @@ -Subproject commit ecb020f2b046790d76d05d377ca1e864e2bc9fdc +Subproject commit aa0f5d7586efe7044f6ca9e07be3f579ee0d5618 -- 1.7.4

2 3

[libvirt] [PATCH v3] virsh: freecell --all getting wrong NUMA nodes count
by Michal Privoznik 18 Feb '11

18 Feb '11

Virsh freecell --all was not only getting wrong NUMA nodes count, but even the NUMA nodes IDs. They doesn't have to be continuous, as I've found out during testing this. Therefore a modification of nodeGetCellsFreeMemory() error message. --- Hope it's good this time :) src/nodeinfo.c | 3 +- tools/virsh.c | 68 ++++++++++++++++++++++++++++++++++++++++++------------- 2 files changed, 54 insertions(+), 17 deletions(-) diff --git a/src/nodeinfo.c b/src/nodeinfo.c index 22d53e5..f4ea36e 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -468,7 +468,8 @@ nodeGetCellsFreeMemory(virConnectPtr conn ATTRIBUTE_UNUSED, long long mem; if (numa_node_size64(n, &mem) < 0) { nodeReportError(VIR_ERR_INTERNAL_ERROR, - "%s", _("Failed to query NUMA free memory")); + _("Failed to query NUMA free memory for node: %d"), + n); goto cleanup; } freeMems[numCells++] = mem; diff --git a/tools/virsh.c b/tools/virsh.c index 50d5e33..2e7cd72 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -2283,9 +2283,15 @@ cmdFreecell(vshControl *ctl, const vshCmd *cmd) int ret; int cell, cell_given; unsigned long long memory; - unsigned long long *nodes = NULL; + xmlNodePtr *nodes = NULL; + unsigned long nodes_cnt; + unsigned long *nodes_id = NULL; + unsigned long long *nodes_free = NULL; int all_given; - virNodeInfo info; + int i; + char *cap_xml = NULL; + xmlDocPtr xml = NULL; + xmlXPathContextPtr ctxt = NULL; if (!vshConnectionUsability(ctl, ctl->conn)) @@ -2301,32 +2307,56 @@ cmdFreecell(vshControl *ctl, const vshCmd *cmd) } if (all_given) { - if (virNodeGetInfo(ctl->conn, &info) < 0) { - vshError(ctl, "%s", _("failed to get NUMA nodes count")); + cap_xml = virConnectGetCapabilities(ctl->conn); + if (!cap_xml) { + vshError(ctl, "%s", _("unable to get node capabilities")); goto cleanup; } - if (!info.nodes) { - vshError(ctl, "%s", _("no NUMA nodes present")); - goto cleanup; - } + xml = xmlReadDoc((const xmlChar *) cap_xml, "node.xml", NULL, + XML_PARSE_NOENT | XML_PARSE_NONET | + XML_PARSE_NOWARNING); - if (VIR_ALLOC_N(nodes, info.nodes) < 0) { - vshError(ctl, "%s", _("could not allocate memory")); + if (!xml) { + vshError(ctl, "%s", _("unable to get node capabilities")); goto cleanup; } - ret = virNodeGetCellsFreeMemory(ctl->conn, nodes, 0, info.nodes); - if (ret != info.nodes) { + ctxt = xmlXPathNewContext(xml); + nodes_cnt = virXPathNodeSet("/capabilities/host/topology/cells/cell", + ctxt, &nodes); + + if (nodes_cnt == -1) { vshError(ctl, "%s", _("could not get information about " - "all NUMA nodes")); + "NUMA topology")); goto cleanup; } + nodes_free = vshCalloc(ctl, nodes_cnt, sizeof(*nodes_free)); + nodes_id = vshCalloc(ctl, nodes_cnt, sizeof(*nodes_id)); + + for (i = 0; i < nodes_cnt; i++) { + unsigned long id; + char *val = virXMLPropString(nodes[i], "id"); + if (virStrToLong_ul(val, NULL, 10, &id)) { + vshError(ctl, "%s", _("conversion from string failed")); + goto cleanup; + } + VIR_FREE(val); + nodes_id[i]=id; + ret = virNodeGetCellsFreeMemory(ctl->conn, &(nodes_free[i]), id, 1); + if (ret != 1) { + vshError(ctl, _("failed to get free memory for NUMA node " + "nuber: %lu"), id); + goto cleanup; + } + } + memory = 0; - for (cell = 0; cell < info.nodes; cell++) { - vshPrint(ctl, "%5d: %10llu kB\n", cell, (nodes[cell]/1024)); - memory += nodes[cell]; + for (cell = 0; cell < nodes_cnt; cell++) { + vshPrint(ctl, "%5lu: %10llu kB\n", nodes_id[cell], + (nodes_free[cell]/1024)); + memory += nodes_free[cell]; } vshPrintExtra(ctl, "--------------------\n"); @@ -2351,7 +2381,13 @@ cmdFreecell(vshControl *ctl, const vshCmd *cmd) func_ret = TRUE; cleanup: + xmlXPathFreeContext(ctxt); + if (xml) + xmlFreeDoc(xml); VIR_FREE(nodes); + VIR_FREE(nodes_free); + VIR_FREE(nodes_id); + VIR_FREE(cap_xml); return func_ret; } -- 1.7.3.5

2 1

[libvirt] Reconfiguring doesn't trigger regeneration of libvirt.syms
by Matthias Bolte 18 Feb '11

18 Feb '11

Consider this sequence of commands $ ./configure ... $ make ... GEN libvirt.syms ... $ ./configure --with-driver-modules ... $ make ... libvirt.syms doesn't get regenerated but it should as it should contain virDriverLoadModule now. make clean after a configure run doesn't help, as make clean dosen't remove libvirt.syms. So either the libvirt.syms rule needs to depend on something that triggers a regeneration of libvirt.syms after a reconfiguration or make clean needs to remove libvirt.syms. I'm not sure about the correct solution here. Any suggestions? Matthias

2 3

[libvirt] [PATCH] build: speed up non-maintainer builds
by Eric Blake 18 Feb '11

18 Feb '11

* configure.ac (gl_ASSERT_NO_GNULIB_POSIXCHECK): Use to reduce time spent in configure. --- This skips out on the bulk of the 'declared without a macro' checks during configure time (they are only useful if you are going to run with CFLAGS=-DGNULIB_POSIXCHECK=1, but that's a maintainer action); it also results in a smaller config.h (no HAVE_RAW_DECL substitutions). Worth including in 0.8.8? configure.ac | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/configure.ac b/configure.ac index 3cd824a..7172b92 100644 --- a/configure.ac +++ b/configure.ac @@ -9,6 +9,11 @@ dnl Make automake keep quiet about wildcards & other GNUmake-isms AM_INIT_AUTOMAKE([-Wno-portability]) AM_MAINTAINER_MODE([enable]) +# Maintainer note - comment this line out if you plan to rerun +# GNULIB_POSIXCHECK testing to see if libvirt should be using more modules. +# Leave it uncommented for normal releases, for faster ./configure. +gl_ASSERT_NO_GNULIB_POSIXCHECK + # Use the silent-rules feature when possible. m4_ifndef([AM_SILENT_RULES], [m4_define([AM_SILENT_RULES],[])]) AM_SILENT_RULES([yes]) -- 1.7.4

2 2

[libvirt] Driver ESX
by Cédric Penas 18 Feb '11

18 Feb '11

Hi, I am currently working on a project where I must develop a function to clone domain for the driver ESX. I immediately saw that the function cloneVM_Task () from VMware API was not supported for standalone ESX. Does anyone know the best way to still implement this function or a workaround ? Why is it not already implemented ? Thank you in advance, Cédric

2 2

[libvirt] [PATCH 0/2] Add support for multiple serial ports into the Xen driver
by Michal Novotny 18 Feb '11

18 Feb '11

Hi, this is the patch to add support for multiple serial ports to the libvirt Xen driver. It support both old style (serial = "pty") and new style (serial = [ "/dev/ttyS0", "/dev/ttyS1" ]) definition and tests for xml2sexpr, sexpr2xml and xmconfig have been added as well. Written and tested on RHEL-5 Xen dom0 and working as designed but the Xen version have to have patch for RHBZ #614004 but this patch is for upstream version of libvirt. Also, this patch is addressing issue described in RHBZ #670789. Differences between v2 and v3: * Fixed handling of port number in virDomainChrDefParseTargetXML() * Fixed serial port handling if we have definition of device in non-zero port number * Added second test for first port undefined (i.e. port value > 0 for just one serial port) Michal Signed-off-by: Michal Novotny <minovotn(a)redhat.com> Michal Novotny (2): Fix virDomainChrDefParseTargetXML() to parse the target port if available Add support for multiple serial ports into the Xen driver src/conf/domain_conf.c | 7 +- src/xen/xend_internal.c | 86 ++++++++++-- src/xen/xm_internal.c | 137 ++++++++++++++++--- .../sexpr2xml-fv-serial-dev-2-ports.sexpr | 1 + .../sexpr2xml-fv-serial-dev-2-ports.xml | 53 ++++++++ .../sexpr2xml-fv-serial-dev-2nd-port.sexpr | 1 + .../sexpr2xml-fv-serial-dev-2nd-port.xml | 49 +++++++ tests/sexpr2xmltest.c | 2 + .../test-fullvirt-serial-dev-2-ports.cfg | 25 ++++ .../test-fullvirt-serial-dev-2-ports.xml | 55 ++++++++ .../test-fullvirt-serial-dev-2nd-port.cfg | 25 ++++ .../test-fullvirt-serial-dev-2nd-port.xml | 53 ++++++++ tests/xmconfigtest.c | 2 + .../xml2sexpr-fv-serial-dev-2-ports.sexpr | 1 + .../xml2sexpr-fv-serial-dev-2-ports.xml | 44 +++++++ .../xml2sexpr-fv-serial-dev-2nd-port.sexpr | 1 + .../xml2sexpr-fv-serial-dev-2nd-port.xml | 40 ++++++ tests/xml2sexprtest.c | 2 + 18 files changed, 546 insertions(+), 38 deletions(-) create mode 100644 tests/sexpr2xmldata/sexpr2xml-fv-serial-dev-2-ports.sexpr create mode 100644 tests/sexpr2xmldata/sexpr2xml-fv-serial-dev-2-ports.xml create mode 100644 tests/sexpr2xmldata/sexpr2xml-fv-serial-dev-2nd-port.sexpr create mode 100644 tests/sexpr2xmldata/sexpr2xml-fv-serial-dev-2nd-port.xml create mode 100644 tests/xmconfigdata/test-fullvirt-serial-dev-2-ports.cfg create mode 100644 tests/xmconfigdata/test-fullvirt-serial-dev-2-ports.xml create mode 100644 tests/xmconfigdata/test-fullvirt-serial-dev-2nd-port.cfg create mode 100644 tests/xmconfigdata/test-fullvirt-serial-dev-2nd-port.xml create mode 100644 tests/xml2sexprdata/xml2sexpr-fv-serial-dev-2-ports.sexpr create mode 100644 tests/xml2sexprdata/xml2sexpr-fv-serial-dev-2-ports.xml create mode 100644 tests/xml2sexprdata/xml2sexpr-fv-serial-dev-2nd-port.sexpr create mode 100644 tests/xml2sexprdata/xml2sexpr-fv-serial-dev-2nd-port.xml -- 1.7.3.2

1 3

[libvirt] [PATCH v2] Requires gettext for client package
by Osier Yang 18 Feb '11

18 Feb '11

libvirt-guests invokes functions in gettext.sh, so we need to require gettext package in spec file. Demo with the fix: % rpm -q gettext package gettext is not installed % rpm -ivh libvirt-client-0.8.8-1.fc14.x86_64.rpm error: Failed dependencies: gettext is needed by libvirt-client-0.8.8-1.fc14.x86_64 * libvirt.spec.in --- libvirt.spec.in | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index d4208e8..c4b4e7f 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -415,6 +415,7 @@ Requires: ncurses # So remote clients can access libvirt over SSH tunnel # (client invokes 'nc' against the UNIX socket on the server) Requires: nc +Requires: gettext %if %{with_sasl} Requires: cyrus-sasl # Not technically required, but makes 'out-of-box' config -- 1.7.4

4 6

[libvirt] cleanup patches
by Christophe Fergeau 18 Feb '11

18 Feb '11

Hi, Here are 2 cleanup patches for 2 smallish issues I found when looking at how virHash was used in libvirt. Christophe

3 5

[libvirt] [PATCH] check device-mapper when building with mpath or disk storage driver
by Wen Congyang 18 Feb '11

18 Feb '11

When I build libvirt without libvirtd, I receive the following errors: GEN virsh.1 CCLD virsh ../src/.libs/libvirt.so: undefined reference to `dm_is_dm_major' collect2: ld returned 1 exit status make[3]: *** [virsh] Error 1 My build step: # ./autogen.sh --without-libvirtd # make dist # rpmbuild --nodeps --define "_sourcedir `pwd`" --define "_without_libvirtd 1" -ba libvirt.spec This bug was caused by commit df1011ca. Signed-off-by: Wen Congyang <wency(a)cn.fujitsu.com> --- configure.ac | 47 ++++++++++++++++++++++++----------------------- src/util/util.c | 3 +++ 2 files changed, 27 insertions(+), 23 deletions(-) diff --git a/configure.ac b/configure.ac index 2bb6918..2f22c18 100644 --- a/configure.ac +++ b/configure.ac @@ -1697,29 +1697,6 @@ if test "$with_storage_mpath" = "check" && test "$with_linux" = "yes"; then fi AM_CONDITIONAL([WITH_STORAGE_MPATH], [test "$with_storage_mpath" = "yes"]) -if test "$with_storage_mpath" = "yes"; then - DEVMAPPER_CFLAGS= - DEVMAPPER_LIBS= - PKG_CHECK_MODULES([DEVMAPPER], [devmapper >= $DEVMAPPER_REQUIRED], [], [DEVMAPPER_FOUND=no]) - if test "$DEVMAPPER_FOUND" = "no"; then - # devmapper is missing pkg-config files in ubuntu, suse, etc - save_LIBS="$LIBS" - save_CFLAGS="$CFLAGS" - DEVMAPPER_FOUND=yes - AC_CHECK_HEADER([libdevmapper.h],,[DEVMAPPER_FOUND=no]) - AC_CHECK_LIB([devmapper], [dm_task_run],,[DEVMAPPER_FOUND=no]) - DEVMAPPER_LIBS="-ldevmapper" - LIBS="$save_LIBS" - CFLAGS="$save_CFLAGS" - fi - if test "$DEVMAPPER_FOUND" = "no" ; then - AC_MSG_ERROR([You must install device-mapper-devel/libdevmapper >= $DEVMAPPER_REQUIRED to compile libvirt]) - fi - -fi -AC_SUBST([DEVMAPPER_CFLAGS]) -AC_SUBST([DEVMAPPER_LIBS]) - LIBPARTED_CFLAGS= LIBPARTED_LIBS= if test "$with_storage_disk" = "yes" || @@ -1781,6 +1758,30 @@ AM_CONDITIONAL([WITH_STORAGE_DISK], [test "$with_storage_disk" = "yes"]) AC_SUBST([LIBPARTED_CFLAGS]) AC_SUBST([LIBPARTED_LIBS]) +if test "$with_storage_mpath" = "yes" || + test "$with_storage_disk" = "yes"; then + DEVMAPPER_CFLAGS= + DEVMAPPER_LIBS= + PKG_CHECK_MODULES([DEVMAPPER], [devmapper >= $DEVMAPPER_REQUIRED], [], [DEVMAPPER_FOUND=no]) + if test "$DEVMAPPER_FOUND" = "no"; then + # devmapper is missing pkg-config files in ubuntu, suse, etc + save_LIBS="$LIBS" + save_CFLAGS="$CFLAGS" + DEVMAPPER_FOUND=yes + AC_CHECK_HEADER([libdevmapper.h],,[DEVMAPPER_FOUND=no]) + AC_CHECK_LIB([devmapper], [dm_task_run],,[DEVMAPPER_FOUND=no]) + DEVMAPPER_LIBS="-ldevmapper" + LIBS="$save_LIBS" + CFLAGS="$save_CFLAGS" + fi + if test "$DEVMAPPER_FOUND" = "no" ; then + AC_MSG_ERROR([You must install device-mapper-devel/libdevmapper >= $DEVMAPPER_REQUIRED to compile libvirt]) + fi + +fi +AC_SUBST([DEVMAPPER_CFLAGS]) +AC_SUBST([DEVMAPPER_LIBS]) + dnl dnl check for libcurl (ESX/XenAPI) dnl diff --git a/src/util/util.c b/src/util/util.c index a7ce4b3..cd4ddf3 100644 --- a/src/util/util.c +++ b/src/util/util.c @@ -3125,6 +3125,8 @@ virTimestamp(void) return timestamp; } +#if defined WITH_STORAGE_MPATH || defined WITH_STORAGE_DISK +/* Currently only disk and mpath storage drivers use this function. */ bool virIsDevMapperDevice(const char *devname) { @@ -3137,3 +3139,4 @@ virIsDevMapperDevice(const char *devname) return false; } +#endif -- 1.7.1

1 0

[libvirt] Driver ESX
by Cédric Penas 18 Feb '11

18 Feb '11

Hi, I am currently working on a project where I must develop a function to clone domain for the driver ESX. I immediately saw that the function cloneVM_Task () from VMware API was not supported for standalone ESX. Does anyone know the best way to still implement this function or a workaround ? Thank you in advance, Cédric

1 0