- Devel - libvirt List Archives

[libvirt] [PATCH] Free slicename in virSystemdCreateMachine
by Ján Tomko 17 Sep '13

17 Sep '13

1,003 bytes in 1 blocks are definitely lost in loss record 599 of 635 ==404== by 0x50728A7: virBufferAddChar (virbuffer.c:185) ==404== by 0x50BC466: virSystemdEscapeName (virsystemd.c:67) ==404== by 0x50BC6B2: virSystemdMakeSliceName (virsystemd.c:108) ==404== by 0x50BC870: virSystemdCreateMachine (virsystemd.c:169) ==404== by 0x5078267: virCgroupNewMachine (vircgroup.c:1498) --- src/util/virsystemd.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/util/virsystemd.c b/src/util/virsystemd.c index 7674cc6..e72b7f0 100644 --- a/src/util/virsystemd.c +++ b/src/util/virsystemd.c @@ -239,5 +239,6 @@ cleanup: VIR_FREE(username); VIR_FREE(creatorname); VIR_FREE(machinename); + VIR_FREE(slicename); return ret; } -- 1.8.1.5

2 2

[libvirt] [PATCH] Always free network and graphics cookies
by Ján Tomko 17 Sep '13

17 Sep '13

qemuMigrationEatCookie has flags to control if these should be parsed, but it does not fill mig->flags. These cookies might get leaked if these flags are not set by qemuMigrationBakeCookie. 42 (32 direct, 10 indirect) bytes in 1 blocks are definitely lost in loss record 361 of 662 ==123== by 0x1BA33FCA: qemuMigrationEatCookie (qemu_migration.c:678) ==123== by 0x1BA34A1E: qemuMigrationRun (qemu_migration.c:3108) ==123== by 0x1BA3622B: doNativeMigrate (qemu_migration.c:3343) ==123== by 0x1BA3B408: qemuMigrationPerform (qemu_migration.c:4138) --- src/qemu/qemu_migration.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index 69a9013..d7b89fc 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -202,11 +202,8 @@ static void qemuMigrationCookieFree(qemuMigrationCookiePtr mig) if (!mig) return; - if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) - qemuMigrationCookieGraphicsFree(mig->graphics); - - if (mig->flags & QEMU_MIGRATION_COOKIE_NETWORK) - qemuMigrationCookieNetworkFree(mig->network); + qemuMigrationCookieGraphicsFree(mig->graphics); + qemuMigrationCookieNetworkFree(mig->network); VIR_FREE(mig->localHostname); VIR_FREE(mig->remoteHostname); -- 1.8.1.5

2 2

[libvirt] [PATCHv2] build: make autobuild require rpm build deps
by Eric Blake 17 Sep '13

17 Sep '13

I spent far too long on a new machine trying to figure out why ./autobuild.sh failed due to an rpm build failure (complaining that libvirt_parthelper was supposed to be packaged but was not built), and finally traced it to a missing parted-devel installation. I learned that --nodeps is in place for autobuilder setups, but for developers, removing it would make rpmbuild error out much sooner for a less cryptic failure. * autobuild.sh: Conditionally drop --nodeps from rpmbuild lines. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- v2: make --nodeps conditional based on $AUTOBUILD_COUNTER; suggested by Dan Berrange. autobuild.sh | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/autobuild.sh b/autobuild.sh index e5aa35c..3109b49 100755 --- a/autobuild.sh +++ b/autobuild.sh @@ -8,6 +8,13 @@ set -v test -n "$1" && RESULTS=$1 || RESULTS=results.log : ${AUTOBUILD_INSTALL_ROOT=$HOME/builder} +# If run under the autobuilder, we must use --nodeps with rpmbuild; +# but this can lead to odd error diagnosis for normal development. +nodeps= +if test "${AUTOBUILD_COUNTER+set}"; then + nodeps=--nodeps +fi + test -f Makefile && make -k distclean || : rm -rf coverage @@ -60,7 +67,7 @@ else fi if test -f /usr/bin/rpmbuild ; then - rpmbuild --nodeps \ + rpmbuild $nodeps \ --define "extra_release $EXTRA_RELEASE" \ --define "_sourcedir `pwd`" \ -ba --clean libvirt.spec @@ -111,7 +118,7 @@ fi if test -x /usr/bin/i686-w64-mingw32-gcc && test -x /usr/bin/x86_64-w64-mingw32-gcc ; then if test -f /usr/bin/rpmbuild ; then - rpmbuild --nodeps \ + rpmbuild $nodeps \ --define "extra_release $EXTRA_RELEASE" \ --define "_sourcedir `pwd`" \ -ba --clean mingw-libvirt.spec -- 1.8.3.1

2 2

Re: [libvirt] rpmbuild libvirt 1.1.2 problem
by jsjshaowentao＠21cn.com 17 Sep '13

17 Sep '13

1 0

[libvirt] [PATCHv2] cgroup: Move [qemu|lxc]GetCpuBWStatus to vicgroup.c and refactor it
by Peter Krempa 17 Sep '13

17 Sep '13

The function existed in two identical instances in lxc and qemu. Move it to vircgroup.c and simplify it. Refactor the callers too. --- Notes: Version 2: - added a fallback implementation that doesn't spam logs with errors src/libvirt_private.syms | 1 + src/lxc/lxc_driver.c | 52 +++++------------------------------------------- src/qemu/qemu_driver.c | 51 +++++------------------------------------------ src/util/vircgroup.c | 38 +++++++++++++++++++++++++++++++++++ src/util/vircgroup.h | 2 ++ 5 files changed, 51 insertions(+), 93 deletions(-) diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms index 35f0f1b..82e3d6f 100644 --- a/src/libvirt_private.syms +++ b/src/libvirt_private.syms @@ -1217,6 +1217,7 @@ virCgroupSetMemory; virCgroupSetMemoryHardLimit; virCgroupSetMemorySoftLimit; virCgroupSetMemSwapHardLimit; +virCgroupSupportsCpuBW; # util/virclosecallbacks.h diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c index b587c22..87ced95 100644 --- a/src/lxc/lxc_driver.c +++ b/src/lxc/lxc_driver.c @@ -1574,45 +1574,10 @@ static char *lxcConnectGetHostname(virConnectPtr conn) } - -/* - * check whether the host supports CFS bandwidth - * - * Return 1 when CFS bandwidth is supported, 0 when CFS bandwidth is not - * supported, -1 on error. - */ -static int lxcGetCpuBWStatus(virCgroupPtr cgroup) -{ - char *cfs_period_path = NULL; - int ret = -1; - - if (!cgroup) - return 0; - - if (virCgroupPathOfController(cgroup, VIR_CGROUP_CONTROLLER_CPU, - "cpu.cfs_period_us", &cfs_period_path) < 0) { - VIR_INFO("cannot get the path of cgroup CPU controller"); - ret = 0; - goto cleanup; - } - - if (access(cfs_period_path, F_OK) < 0) { - ret = 0; - } else { - ret = 1; - } - -cleanup: - VIR_FREE(cfs_period_path); - return ret; -} - - static char *lxcDomainGetSchedulerType(virDomainPtr dom, int *nparams) { char *ret = NULL; - int rc; virDomainObjPtr vm; virLXCDomainObjPrivatePtr priv; @@ -1639,13 +1604,10 @@ static char *lxcDomainGetSchedulerType(virDomainPtr dom, } if (nparams) { - rc = lxcGetCpuBWStatus(priv->cgroup); - if (rc < 0) - goto cleanup; - else if (rc == 0) - *nparams = 1; - else + if (virCgroupSupportsCpuBW(priv->cgroup)) *nparams = 3; + else + *nparams = 1; } ignore_value(VIR_STRDUP(ret, "posix")); @@ -1872,12 +1834,8 @@ lxcDomainGetSchedulerParametersFlags(virDomainPtr dom, if (virDomainGetSchedulerParametersFlagsEnsureACL(dom->conn, vm->def) < 0) goto cleanup; - if (*nparams > 1) { - rc = lxcGetCpuBWStatus(priv->cgroup); - if (rc < 0) - goto cleanup; - cpu_bw_status = !!rc; - } + if (*nparams > 1) + cpu_bw_status = virCgroupSupportsCpuBW(priv->cgroup); if (!(caps = virLXCDriverGetCapabilities(driver, false))) goto cleanup; diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index ae1948f..0caeb08 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -7325,44 +7325,10 @@ cleanup: } -/* - * check whether the host supports CFS bandwidth - * - * Return 1 when CFS bandwidth is supported, 0 when CFS bandwidth is not - * supported, -1 on error. - */ -static int qemuGetCpuBWStatus(virCgroupPtr cgroup) -{ - char *cfs_period_path = NULL; - int ret = -1; - - if (!cgroup) - return 0; - - if (virCgroupPathOfController(cgroup, VIR_CGROUP_CONTROLLER_CPU, - "cpu.cfs_period_us", &cfs_period_path) < 0) { - VIR_INFO("cannot get the path of cgroup CPU controller"); - ret = 0; - goto cleanup; - } - - if (access(cfs_period_path, F_OK) < 0) { - ret = 0; - } else { - ret = 1; - } - -cleanup: - VIR_FREE(cfs_period_path); - return ret; -} - - static char *qemuDomainGetSchedulerType(virDomainPtr dom, int *nparams) { char *ret = NULL; - int rc; virDomainObjPtr vm = NULL; qemuDomainObjPrivatePtr priv; @@ -7389,13 +7355,10 @@ static char *qemuDomainGetSchedulerType(virDomainPtr dom, } if (nparams) { - rc = qemuGetCpuBWStatus(priv->cgroup); - if (rc < 0) - goto cleanup; - else if (rc == 0) - *nparams = 1; - else + if (virCgroupSupportsCpuBW(priv->cgroup)) *nparams = 5; + else + *nparams = 1; } ignore_value(VIR_STRDUP(ret, "posix")); @@ -8728,12 +8691,8 @@ qemuDomainGetSchedulerParametersFlags(virDomainPtr dom, if (virDomainGetSchedulerParametersFlagsEnsureACL(dom->conn, vm->def) < 0) goto cleanup; - if (*nparams > 1) { - rc = qemuGetCpuBWStatus(priv->cgroup); - if (rc < 0) - goto cleanup; - cpu_bw_status = !!rc; - } + if (*nparams > 1) + cpu_bw_status = virCgroupSupportsCpuBW(priv->cgroup); if (!(caps = virQEMUDriverGetCapabilities(driver, false))) goto cleanup; diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c index a615f28..e99caf5 100644 --- a/src/util/vircgroup.c +++ b/src/util/vircgroup.c @@ -3061,6 +3061,36 @@ cleanup: } +/** + * virCgroupSupportsCpuBW(): + * Check whether the host supports CFS bandwidth. + * + * Return true when CFS bandwidth is supported, + * false when CFS bandwidth is not supported. + */ +bool +virCgroupSupportsCpuBW(virCgroupPtr cgroup) +{ + char *path = NULL; + int ret = false; + + if (!cgroup) + return false; + + if (virCgroupPathOfController(cgroup, VIR_CGROUP_CONTROLLER_CPU, + "cpu.cfs_period_us", &path) < 0) { + virResetLastError(); + goto cleanup; + } + + ret = virFileExists(path); + +cleanup: + VIR_FREE(path); + return ret; +} + + #else /* !VIR_CGROUP_SUPPORTED */ bool @@ -3646,4 +3676,12 @@ virCgroupIsolateMount(virCgroupPtr group ATTRIBUTE_UNUSED, return -1; } + +bool +virCgroupSupportsCpuBW(virCgroupPtr cgroup ATTRIBUTE_UNUSED) +{ + VIR_DEBUG("Control groups not supported on this platform"); + return false; +} + #endif /* !VIR_CGROUP_SUPPORTED */ diff --git a/src/util/vircgroup.h b/src/util/vircgroup.h index 7bb4b2a..835eb30 100644 --- a/src/util/vircgroup.h +++ b/src/util/vircgroup.h @@ -207,4 +207,6 @@ int virCgroupIsolateMount(virCgroupPtr group, const char *oldroot, const char *mountopts); +bool virCgroupSupportsCpuBW(virCgroupPtr cgroup); + #endif /* __VIR_CGROUP_H__ */ -- 1.8.3.2

2 2

[libvirt] [PATCH] build: fix build with latest rawhide kernel headers
by Eric Blake 17 Sep '13

17 Sep '13

Bother those kernel developers. In the latest rawhide, kernel and glibc have now been unified so that <netinet/in.h> and <linux/in6.h> no longer clash; but <linux/if_bridge.h> is still not self-contained. Because of the latest header change, the build is failing with: checking for linux/param.h... no configure: error: You must install kernel-headers in order to compile libvirt with QEMU or LXC support with details: In file included from conftest.c:561:0: /usr/include/linux/in6.h:71:18: error: field 'flr_dst' has incomplete type struct in6_addr flr_dst; We need a workaround to avoid our workaround :) * configure.ac (KERNEL_HEADERS_WORKAROUND): New test. * src/util/virnetdevbridge.c (includes): Use it. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- This qualifies as a build-breaker; but last time I pushed a patch in this area of code without review, it got reverted, so I'll wait for the review. I've tested on RHEL 5, RHEL 6, Fedora 19 (all of which have glibc and kernel differences, although not the same error messages) and rawhide (where glibc and kernel finally agree, but where the kernel headers _still_ assume that the caller already declared ipv6 structs). configure.ac | 38 ++++++++++++++++++++++++++++---------- src/util/virnetdevbridge.c | 24 ++++++++++++++---------- 2 files changed, 42 insertions(+), 20 deletions(-) diff --git a/configure.ac b/configure.ac index 8426863..8acccdd 100644 --- a/configure.ac +++ b/configure.ac @@ -1051,18 +1051,36 @@ dnl check for kernel headers required by src/bridge.c dnl if test "$with_linux" = "yes"; then if test "$with_qemu" = "yes" || test "$with_lxc" = "yes" ; then + # Various kernel versions have headers that are not self-standing, but + # yet are incompatible with the corresponding glibc headers. In order + # to guarantee compilation across a wide range of versions (from RHEL 5 + # to rawhide), we first have to probe whether glibc and kernel can be + # used in tandem; and if not, provide workarounds that ensure that + # ABI-compatible IPv6 types are present for use by the kernel headers. + # These probes mirror the usage in virnetdevbridge.c + AC_CACHE_CHECK([whether kernel and glibc headers are compatible], + [lv_cv_kernel_glibc_compatible], + [AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[ + #include <netinet/in.h> + #include <linux/in6.h> + ]])], + [lv_cv_kernel_glibc_compatible=yes], + [lv_cv_kernel_glibc_compatible=no])]) + if test "x$lv_cv_kernel_glibc_compatible" != xyes; then + AC_DEFINE([KERNEL_HEADERS_WORKAROUND], [1], + [define to 1 if Linux kernel headers require a workaround to avoid + compilation errors when mixed with glibc headers]) + fi AC_CHECK_HEADERS([linux/param.h linux/sockios.h linux/if_bridge.h linux/if_tun.h],, [AC_MSG_ERROR([You must install kernel-headers in order to compile libvirt with QEMU or LXC support])], - [[/* The kernel folks broke their headers when used with particular - * glibc versions; although the structs are ABI compatible, the - * C type system doesn't like struct redefinitions. We work around - * the problem here in the same manner as in virnetdevbridge.c. */ - #include <netinet/in.h> - #define in6_addr in6_addr_ - #define sockaddr_in6 sockaddr_in6_ - #define ipv6_mreq ipv6_mreq_ - #define in6addr_any in6addr_any_ - #define in6addr_loopback in6addr_loopback_ + [[#include <netinet/in.h> + #if KERNEL_HEADERS_WORKAROUND + # define in6_addr in6_addr_ + # define sockaddr_in6 sockaddr_in6_ + # define ipv6_mreq ipv6_mreq_ + # define in6addr_any in6addr_any_ + # define in6addr_loopback in6addr_loopback_ + #endif #include <linux/in6.h> ]]) fi diff --git a/src/util/virnetdevbridge.c b/src/util/virnetdevbridge.c index e4daa27..b78ab07 100644 --- a/src/util/virnetdevbridge.c +++ b/src/util/virnetdevbridge.c @@ -39,22 +39,26 @@ #ifdef __linux__ # include <linux/sockios.h> # include <linux/param.h> /* HZ */ +# if KERNEL_HEADERS_WORKAROUND /* Depending on the version of kernel vs. glibc, there may be a collision * between <net/in.h> and kernel IPv6 structures. The different types * are ABI compatible, but choke the C type system; work around it by * using temporary redefinitions. */ -# define in6_addr in6_addr_ -# define sockaddr_in6 sockaddr_in6_ -# define ipv6_mreq ipv6_mreq_ -# define in6addr_any in6addr_any_ -# define in6addr_loopback in6addr_loopback_ +# define in6_addr in6_addr_ +# define sockaddr_in6 sockaddr_in6_ +# define ipv6_mreq ipv6_mreq_ +# define in6addr_any in6addr_any_ +# define in6addr_loopback in6addr_loopback_ +# endif # include <linux/in6.h> # include <linux/if_bridge.h> /* SYSFS_BRIDGE_ATTR */ -# undef in6_addr -# undef sockaddr_in6 -# undef ipv6_mreq -# undef in6addr_any -# undef in6addr_loopback +# if KERNEL_HEADERS_WORKAROUND +# undef in6_addr +# undef sockaddr_in6 +# undef ipv6_mreq +# undef in6addr_any +# undef in6addr_loopback +# endif # define JIFFIES_TO_MS(j) (((j)*1000)/HZ) # define MS_TO_JIFFIES(ms) (((ms)*HZ)/1000) -- 1.8.3.1

2 2

[libvirt] [PATCH] build: make autobuild require rpm build deps
by Eric Blake 17 Sep '13

17 Sep '13

I spent far too long on a new machine trying to figure out why ./autobuild.sh failed during the rpm build failure (complaining that libvirt_parthelper was supposed to be packaged but was not built), and finally traced it to a missing parted-devel installation. I don't know why we have --nodeps in place, but removing it would make rpmbuild error out much sooner, and with a much less cryptic failure case. * autobuild.sh: Drop --nodeps from rpmbuild lines. Signed-off-by: Eric Blake <eblake(a)redhat.com> --- I'm okay ditching this patch if someone can explain why the --nodeps line should be there. autobuild.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/autobuild.sh b/autobuild.sh index e5aa35c..0ab5f5e 100755 --- a/autobuild.sh +++ b/autobuild.sh @@ -60,7 +60,7 @@ else fi if test -f /usr/bin/rpmbuild ; then - rpmbuild --nodeps \ + rpmbuild \ --define "extra_release $EXTRA_RELEASE" \ --define "_sourcedir `pwd`" \ -ba --clean libvirt.spec @@ -111,7 +111,7 @@ fi if test -x /usr/bin/i686-w64-mingw32-gcc && test -x /usr/bin/x86_64-w64-mingw32-gcc ; then if test -f /usr/bin/rpmbuild ; then - rpmbuild --nodeps \ + rpmbuild \ --define "extra_release $EXTRA_RELEASE" \ --define "_sourcedir `pwd`" \ -ba --clean mingw-libvirt.spec -- 1.8.3.1

2 2

[libvirt] [PATCHv3 RESEND 0/5] Handling of undefine and redefine snapshots with VirtualBox 4.2
by Manuel VIVES 16 Sep '13

16 Sep '13

Hi, This is a serie of patches in order to support undefining and redefining snapshots with VirtualBox 4.2. The serie of patches is rather big, and adds among other things some utility functions unrelated to VirtualBox in patches 1 & 2. The code review could be done in several parts: e.g. patches 1 & 2 separately to validate the utility functions. The VirtualBox API provides only high level operations to manipulate snapshots, so it not possible to support flags like VIR_DOMAIN_SNAPSHOT_CREATE_REDEFINE and VIR_DOMAIN_SNAPSHOT_DELETE_METADATA_ONLY with only API calls. Following an IRC talk with Eric Blake, the decision was taken to emulate these behaviours by manipulating directly the .vbox XML files. The first two patches are some util methods for handling uuid and strings that will be used after. The third patch brings more details in the snapshot XML returned by libvirt. We will need those modifications in order to redefine the snapshots. The fourth patch brings the support of the VIR_DOMAIN_SNAPSHOT_CREATE_REDEFINE and VIR_DOMAIN_SNAPSHOT_CREATE_CURRENT flags in virDomainSnapshotCreateXML. The fifth and last patch brings the support of the VIR_DOMAIN_SNAPSHOT_DELETE_METADATA_ONLY flag in virDomainSnapshotDelete. The patches are only for Virtualbox 4.2 Regards, Manuel VIVES V3: * Use of STREQ_NULLABLE instead of STREQ in one case * Fix the method for finding uuids according to Ján Tomko review V2: * Fix a licence problem with the method for string replacement Manuel VIVES (5): viruuid.h/c: Util method for finding uuid patterns in some strings virstring.h/c: Util method for making some find and replace in strings vbox_tmpl.c: Better XML description for snapshots vbox_tmpl.c: Patch for redefining snapshots vbox_tmpl.c: Add methods for undefining snapshots po/POTFILES.in | 1 + src/conf/domain_conf.c | 2 +- src/libvirt_private.syms | 2 + src/util/virstring.c | 48 ++ src/util/virstring.h | 2 + src/util/viruuid.c | 81 ++ src/util/viruuid.h | 1 + src/vbox/vbox_tmpl.c | 1846 +++++++++++++++++++++++++++++++++++++++++++--- 8 files changed, 1871 insertions(+), 112 deletions(-) -- 1.7.10.4

1 5

[libvirt] One question about ACS for pci passthrough
by Li Zhang 16 Sep '13

16 Sep '13

Hi all, In libvirt, it doesn't allow non-ACS switch to assign devices to guests. Is this required for X86 to support PCI passthrough? On my power machine, ACS capability can't be found. I think HW doesn't support it. I don't know whether this is a special situation, and I have to set relaxed_acs_check = 1. Thanks. --Li

2 1

[libvirt] [PATCH 0/3] Clean up usage of access(PATH, F_OK)
by Peter Krempa 16 Sep '13

16 Sep '13

Peter Krempa (3): util: Declare that virFileExists shall honor errno cgroup: Move [qemu|lxc]GetCpuBWStatus to vicgroup.c and refactor it cleanup: Kill usage of access(PATH, F_OK) in favor of virFileExists() src/libvirt_private.syms | 1 + src/lxc/lxc_driver.c | 52 ++++------------------------------- src/openvz/openvz_conf.c | 2 +- src/parallels/parallels_storage.c | 2 +- src/qemu/qemu_capabilities.c | 4 +-- src/qemu/qemu_cgroup.c | 6 ++-- src/qemu/qemu_driver.c | 51 ++++------------------------------ src/qemu/qemu_process.c | 2 +- src/storage/storage_backend_fs.c | 2 +- src/storage/storage_backend_logical.c | 2 +- src/storage/storage_backend_mpath.c | 8 +----- src/storage/storage_backend_scsi.c | 3 +- src/util/vircgroup.c | 31 ++++++++++++++++++++- src/util/vircgroup.h | 2 ++ src/util/virfile.c | 8 +++++- src/util/virpci.c | 2 +- src/util/virscsi.c | 2 +- src/util/virutil.c | 6 ++-- 18 files changed, 67 insertions(+), 119 deletions(-) -- 1.8.3.2

3 8