- Devel - libvirt List Archives

[libvirt] libvirt starts QEMU with bad CPU based on its XML
by Doug Goldstein 28 Mar '12

28 Mar '12

My XML configs for my VM machines have it set to AMD Opteron G3s, but libvirt is creating QEMU with a core2duo. This is not correct at all. I've included the setup and config of one machine that does this. But another machine that does this matches danpb's AMD Quad Core Dual NUMA system he posted on the mailing list recently. That machine is using a stock qemu-0.15.1 from upstream's git and libvirt git a few commits shy of the 0.9.11 rc1. The machine below is using libvirt 0.9.10 and qemu-kvm-1.0. Here's the setup: The host is: processor : 63 vendor_id : AuthenticAMD cpu family : 21 model : 1 model name : AMD Opteron(TM) Processor 6272 stepping : 2 microcode : 0x6000613 cpu MHz : 2099.875 cache size : 2048 KB physical id : 2 siblings : 16 core id : 7 cpu cores : 8 apicid : 79 initial apicid : 79 fpu : yes fpu_exception : yes cpuid level : 13 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid amd_dcm aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 popcnt aes xsave avx lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 nodeid_msr topoext perfctr_core arat cpb npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold bogomips : 4200.08 TLB size : 1536 4K pages clflush size : 64 cache_alignment : 64 address sizes : 48 bits physical, 48 bits virtual power management: ts ttp tm 100mhzsteps hwpstate [9] The guest is: processor : 1 vendor_id : AuthenticAMD cpu family : 6 model : 15 model name : Intel(R) Core(TM)2 Duo CPU T7700 @ 2.40GHz stepping : 11 cpu MHz : 2100.397 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt pdpe1gb lm pni cx16 popcnt lahf_lm cmp_legacy svm cr8legacy abm sse4a misalignsse 3dnowprefetch bogomips : 4200.01 The QEMU command line contains the following: -cpu core2duo,+wdt,+skinit,+osvw,+3dnowprefetch,+misalignsse,+sse4a,+abm,+cr8legacy,+extapic,+svm,+cmp_legacy,+lahf_lm,+rdtscp,+pdpe1gb,+fxsr_opt,+mmxext,+aes,+popcnt,+sse4.2,+sse4.1,+cx16,+ht While libvirt's XML contains: <cpu match='exact'> <model>Opteron_G3</model> <vendor>AMD</vendor> <feature policy='require' name='aes'/> <feature policy='require' name='skinit'/> <feature policy='require' name='vme'/> <feature policy='require' name='mmxext'/> <feature policy='require' name='fxsr_opt'/> <feature policy='require' name='cr8legacy'/> <feature policy='require' name='ht'/> <feature policy='require' name='3dnowprefetch'/> <feature policy='require' name='ssse3'/> <feature policy='require' name='wdt'/> <feature policy='require' name='extapic'/> <feature policy='require' name='pdpe1gb'/> <feature policy='require' name='osvw'/> <feature policy='require' name='sse4.1'/> <feature policy='require' name='cmp_legacy'/> <feature policy='require' name='sse4.2'/> </cpu> -- Doug Goldstein

2 1

[libvirt] [PATCH] virsh: plug memory leaks on failure path
by Alex Jia 28 Mar '12

28 Mar '12

Leaks are introduced in commit 1cf0e3d and fe383bb. * tools/virsh.c (cmdSnapshotList): fix memory leaks. * How to reproduce? % virsh snapshot-list foo --parent --roots error: --parent and --roots are mutually exclusive error: Failed to disconnect from the hypervisor, 1 leaked reference(s) % virsh snapshot-list foo --parent --tree error: --parent and --tree are mutually exclusive error: Failed to disconnect from the hypervisor, 1 leaked reference(s) % virsh snapshot-list foo --roots --tree error: --roots and --tree are mutually exclusive error: Failed to disconnect from the hypervisor, 1 leaked reference(s) ...... Signed-off-by: Alex Jia <ajia(a)redhat.com> --- tools/virsh.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5009b6b..f5c3b60 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -16357,11 +16357,13 @@ cmdSnapshotList(vshControl *ctl, const vshCmd *cmd) if (vshCommandOptBool(cmd, "roots")) { vshError(ctl, "%s", _("--parent and --roots are mutually exclusive")); + virDomainFree(dom); return false; } if (tree) { vshError(ctl, "%s", _("--parent and --tree are mutually exclusive")); + virDomainFree(dom); return false; } parent_filter = 1; @@ -16369,11 +16371,13 @@ cmdSnapshotList(vshControl *ctl, const vshCmd *cmd) if (tree) { vshError(ctl, "%s", _("--roots and --tree are mutually exclusive")); + virDomainFree(dom); return false; } if (from) { vshError(ctl, "%s", _("--roots and --from are mutually exclusive")); + virDomainFree(dom); } flags |= VIR_DOMAIN_SNAPSHOT_LIST_ROOTS; } @@ -16381,6 +16385,7 @@ cmdSnapshotList(vshControl *ctl, const vshCmd *cmd) if (tree) { vshError(ctl, "%s", _("--leaves and --tree are mutually exclusive")); + virDomainFree(dom); return false; } flags |= VIR_DOMAIN_SNAPSHOT_LIST_LEAVES; -- 1.7.1

2 2

[libvirt] [PATCH] virsh: add error return value
by Alex Jia 27 Mar '12

27 Mar '12

* tools/virsh.c (cmdSnapshotList): Avoiding to continue to execute subsequent codes, the programming should return false when meeting a error. For details, please see the following link: RHBZ: https://bugzilla.redhat.com/show_bug.cgi?id=807545 Signed-off-by: Alex Jia <ajia(a)redhat.com> --- tools/virsh.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/tools/virsh.c b/tools/virsh.c index 5009b6b..7e74744 100644 --- a/tools/virsh.c +++ b/tools/virsh.c @@ -16374,6 +16374,7 @@ cmdSnapshotList(vshControl *ctl, const vshCmd *cmd) if (from) { vshError(ctl, "%s", _("--roots and --from are mutually exclusive")); + return false; } flags |= VIR_DOMAIN_SNAPSHOT_LIST_ROOTS; } -- 1.7.1

2 2

[libvirt] Coverity automatic detection
by Alex Jia 27 Mar '12

27 Mar '12

This email is automatically generated. The test result is based on the following git commit: 851117b Enable all warnings permanently & default to -Werror for GIT builds Analysis summary report: ------------------------ Files analyzed : 306 Total LoC input to cov-analyze : 362576 Functions analyzed : 8807 Paths analyzed : 1075505 Defect occurrences found : 105 Total 7 ATOMICITY 15 CHECKED_RETURN 14 DEADCODE 1 EVALUATION_ORDER 9 FORWARD_NULL 13 LOCK 3 NEGATIVE_RETURNS 1 NO_EFFECT 3 NULL_RETURNS 1 OVERRUN_STATIC 16 RESOURCE_LEAK 1 RETURN_LOCAL 12 REVERSE_INULL 1 SIZEOF_MISMATCH 6 UNINIT 2 UNUSED_VALUE Exceeded path limit of 5000 paths in 0.55% of functions (normally up to 5% of functions encounter this limitation) For details, please see attachment. Regards, Alex

1 0

[libvirt] [PATCH] snapshot: fix virsh docs
by Eric Blake 27 Mar '12

27 Mar '12

Commit d42a2ff forgot to touch up virsh documentation, and commit 4e9953a mis-spelled the option name. * tools/virsh.pod (snapshot-create, snapshot-create-as): Fix typo and match recent change in flag meaning. --- tools/virsh.pod | 6 +++--- 1 files changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/virsh.pod b/tools/virsh.pod index 390bcec..76d2ea4 100644 --- a/tools/virsh.pod +++ b/tools/virsh.pod @@ -2347,7 +2347,7 @@ metadata again). If I<--reuse-external> is specified, and the snapshot XML requests an external snapshot with a destination of an existing file, then the -existing file is truncated and reused; otherwise, a snapshot is refused +destination must exist, and is reused; otherwise, a snapshot is refused to avoid losing contents of the existing files. If I<--quiesce> is specified, libvirt will try to use guest agent @@ -2367,7 +2367,7 @@ metadata is silently lost when the domain quits running (whether by command such as B<destroy> or by internal guest action). =item B<snapshot-create-as> I<domain> {[I<--print-xml>] -| [I<--no-metadata>] [I<--halt>] [I<--reuse-existing>]} [I<name>] +| [I<--no-metadata>] [I<--halt>] [I<--reuse-external>]} [I<name>] [I<description>] [I<--disk-only> [I<--quiesce>] [I<--atomic>] [[I<--diskspec>] B<diskspec>]...] @@ -2394,7 +2394,7 @@ results in the following XML: If I<--reuse-external> is specified, and the domain XML or I<diskspec> option requests an external snapshot with a destination of an existing -file, then the existing file is truncated and reused; otherwise, a +file, then the destination must exist, and is reused; otherwise, a snapshot is refused to avoid losing contents of the existing files. If I<--quiesce> is specified, libvirt will try to use guest agent -- 1.7.7.6

2 2

[libvirt] Turning off libvirtd mdns by default
by Stef Walter 27 Mar '12

27 Mar '12

In the GNOME UI we'd like to make use of Avahi discovery and name resolution "out of the box". A typical use case is for discovery of printers that are advertised using MDNS. This should work even on potentially 'hostile' networks such as a wireless access point in a print shop or airport. It should work without user configuration. https://fedoraproject.org/wiki/Desktop/Whiteboards/AvahiDefault In order to turn on Avahi by default, and make it work by default, we'd like to make it possible to use Avahi without advertising any information to the network by default. Advertising information to the network (even the host name) without the user's configuration or consent is a privacy issue. libvirtd advertises itself via MDNS on the network by default. I understand that MDNS discovery of libvirtd is really handy in many cases. However since one has to configure network access in libvirtd anyway -- none of the access methods work "out of the box" to my understanding -- I'd like to suggest turning off libvirtd's MDNS publishing by default. As part of setting up libvirtd for network access, the user would turn on mdns_adv. I hope that makes sense. Let me know if I've gotten something wrong. Would you accept a patch to do this? Or would you suggest that we try and do this downstream in the Fedora/RHEL packages instead? Cheers, Stef

5 10

[libvirt] [PATCH] qemu: Make migration fail when port profile association fails on the dst host
by Christian Benvenuti (benve) 27 Mar '12

27 Mar '12

In the current V3 migration protocol, Libvirt does not check the result of the function qemuMigrationVPAssociatePortProfiles This means that it is possible for a migration to complete successfully even when the VM loses network connectivity on the destination host. With this change libvirt aborts the migration (during the "finish" step) when the above function fails, that is to say when at least one of the port profile associations fails. Signed-off by: Christian Benvenuti <benve(a)cisco.com> --- src/qemu/qemu_migration.c | 24 ++++++++++++++++++++---- 1 files changed, 16 insertions(+), 4 deletions(-) diff --git a/src/qemu/qemu_migration.c b/src/qemu/qemu_migration.c index 802785f..38fbf83 100644 --- a/src/qemu/qemu_migration.c +++ b/src/qemu/qemu_migration.c @@ -2716,7 +2716,7 @@ qemuMigrationPerform(struct qemud_driver *driver, } } -static void +static int qemuMigrationVPAssociatePortProfiles(virDomainDefPtr def) { int i; int last_good_net = -1; @@ -2731,13 +2731,17 @@ qemuMigrationVPAssociatePortProfiles(virDomainDefPtr def) { virDomainNetGetActualDirectDev(net), -1, def->uuid, - VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH, false) < 0) + VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH, + false) < 0) { + VIR_ERROR("Port profile Associate failed for %s", net->ifname); goto err_exit; + } + VIR_DEBUG("Port profile Associate succeeded for %s", net->ifname); } last_good_net = i; } - return; + return 0; err_exit: for (i = 0; i < last_good_net; i++) { @@ -2751,6 +2755,7 @@ err_exit: VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH)); } } + return -1; } @@ -2805,7 +2810,18 @@ qemuMigrationFinish(struct qemud_driver *driver, goto endjob; } - qemuMigrationVPAssociatePortProfiles(vm->def); + if (qemuMigrationVPAssociatePortProfiles(vm->def) < 0) { + qemuProcessStop(driver, vm, 1, VIR_DOMAIN_SHUTOFF_FAILED); + virDomainAuditStop(vm, "failed"); + event = virDomainEventNewFromObj(vm, + VIR_DOMAIN_EVENT_STOPPED, + VIR_DOMAIN_EVENT_STOPPED_FAILED); + goto endjob; + } if (flags & VIR_MIGRATE_PERSIST_DEST) { virDomainDefPtr vmdef; -- 1.7.1

2 1

[libvirt] [PATCH] Enable all warnings permanently & default to -Werror for GIT builds
by Daniel P. Berrange 27 Mar '12

27 Mar '12

From: "Daniel P. Berrange" <berrange(a)redhat.com> Given that we auto-detect whether each -Wxxxx flag is supported by GCC, and we are warning-free and use automake silent rules, there is no compelling reason to allow compile warnings to be disabled. Replace the --enable-compile-warnings flag with a simpler --enable-werror flag, which defaults to 'yes' if building from GIT, or 'no' if building from tar.gz This helps ensure that everyone writing patches for libvirt will take care to fix their warning problems before submitting for review * autobuild.sh: Force -Werror * configure.ac: Update for LIBVIRT_COMPILE_WARNINGS macro change * m4/virt-compile-warnings.m4: Permanently enable all warnings, auto-enable Werror for GIT builds --- autobuild.sh | 4 +- configure.ac | 4 +- m4/virt-compile-warnings.m4 | 258 +++++++++++++++++++++---------------------- 3 files changed, 130 insertions(+), 136 deletions(-) diff --git a/autobuild.sh b/autobuild.sh index 1c289b4..7e8f645 100755 --- a/autobuild.sh +++ b/autobuild.sh @@ -20,7 +20,7 @@ cd build ../autogen.sh --prefix="$AUTOBUILD_INSTALL_ROOT" \ --enable-test-coverage \ --disable-nls \ - --enable-compile-warnings=error + --enable-werror # If the MAKEFLAGS envvar does not yet include a -j option, # add -jN where N depends on the number of processors. @@ -74,7 +74,7 @@ if [ -x /usr/bin/i686-pc-mingw32-gcc ]; then --build=$(uname -m)-pc-linux \ --host=i686-pc-mingw32 \ --prefix="$AUTOBUILD_INSTALL_ROOT/i686-pc-mingw32/sys-root/mingw" \ - --enable-compile-warnings=error \ + --enable-werror \ --without-libvirtd \ --without-python diff --git a/configure.ac b/configure.ac index 740129c..32cc8d0 100644 --- a/configure.ac +++ b/configure.ac @@ -106,7 +106,7 @@ VERSION_SCRIPT_FLAGS=-Wl,--version-script= VERSION_SCRIPT_FLAGS="-Wl,-M -Wl," AC_MSG_RESULT([$VERSION_SCRIPT_FLAGS]) -LIBVIRT_COMPILE_WARNINGS([maximum]) +LIBVIRT_COMPILE_WARNINGS AC_MSG_CHECKING([for CPUID instruction]) AC_COMPILE_IFELSE([AC_LANG_PROGRAM( @@ -2820,7 +2820,7 @@ AC_MSG_NOTICE([]) AC_MSG_NOTICE([Miscellaneous]) AC_MSG_NOTICE([]) AC_MSG_NOTICE([ Debug: $enable_debug]) -AC_MSG_NOTICE([ Warnings: $enable_compile_warnings]) +AC_MSG_NOTICE([ Use -Werror: $set_werror]) AC_MSG_NOTICE([Warning Flags: $WARN_CFLAGS]) AC_MSG_NOTICE([ Readline: $lv_use_readline]) AC_MSG_NOTICE([ Python: $with_python]) diff --git a/m4/virt-compile-warnings.m4 b/m4/virt-compile-warnings.m4 index 3a428c3..5527bff 100644 --- a/m4/virt-compile-warnings.m4 +++ b/m4/virt-compile-warnings.m4 @@ -7,139 +7,133 @@ AC_DEFUN([LIBVIRT_COMPILE_WARNINGS],[ dnl More compiler warnings dnl ****************************** - AC_ARG_ENABLE(compile-warnings, - [AC_HELP_STRING([--enable-compile-warnings=@<:@no/yes/error@:>@], - [Turn on compiler warnings])],, - [enable_compile_warnings="m4_default([$1],[yes])"]) - - case "$enable_compile_warnings" in - no) - try_compiler_flags="" - ;; - yes|minimum|maximum|error) - - # List of warnings that are not relevant / wanted - - # Don't care about C++ compiler compat - dontwarn="$dontwarn -Wc++-compat" - dontwarn="$dontwarn -Wabi" - dontwarn="$dontwarn -Wdeprecated" - # Don't care about ancient C standard compat - dontwarn="$dontwarn -Wtraditional" - # Don't care about ancient C standard compat - dontwarn="$dontwarn -Wtraditional-conversion" - # Ignore warnings in /usr/include - dontwarn="$dontwarn -Wsystem-headers" - # Happy for compiler to add struct padding - dontwarn="$dontwarn -Wpadded" - # GCC very confused with -O2 - dontwarn="$dontwarn -Wunreachable-code" - # Too many to deal with - dontwarn="$dontwarn -Wconversion" - # Too many to deal with - dontwarn="$dontwarn -Wsign-conversion" - # GNULIB gettext.h violates - dontwarn="$dontwarn -Wvla" - # Many GNULIB header violations - dontwarn="$dontwarn -Wundef" - # Need to allow bad cast for execve() - dontwarn="$dontwarn -Wcast-qual" - # We need to use long long in many places - dontwarn="$dontwarn -Wlong-long" - # We allow manual list of all enum cases without default: - dontwarn="$dontwarn -Wswitch-default" - # We allow optional default: instead of listing all enum values - dontwarn="$dontwarn -Wswitch-enum" - # Not a problem since we don't use -fstrict-overflow - dontwarn="$dontwarn -Wstrict-overflow" - # Not a problem since we don't use -funsafe-loop-optimizations - dontwarn="$dontwarn -Wunsafe-loop-optimizations" - # Things like virAsprintf mean we can't use this - dontwarn="$dontwarn -Wformat-nonliteral" - - # We might fundamentally need some of these disabled forever, but - # ideally we'd turn many of them on - dontwarn="$dontwarn -Wfloat-equal" - dontwarn="$dontwarn -Wdeclaration-after-statement" - dontwarn="$dontwarn -Wcast-qual" - dontwarn="$dontwarn -Wconversion" - dontwarn="$dontwarn -Wsign-conversion" - dontwarn="$dontwarn -Wpacked" - dontwarn="$dontwarn -Wunused-macros" - dontwarn="$dontwarn -Woverlength-strings" - dontwarn="$dontwarn -Wstack-protector" - - # Get all possible GCC warnings - gl_MANYWARN_ALL_GCC([maybewarn]) - - # Remove the ones we don't want, blacklisted earlier - gl_MANYWARN_COMPLEMENT([wantwarn], [$maybewarn], [$dontwarn]) - - # Check for $CC support of each warning - for w in $wantwarn; do - gl_WARN_ADD([$w]) - done - - # GNULIB uses '-W' (aka -Wextra) which includes a bunch of stuff. - # Unfortunately, this means you can't simply use '-Wsign-compare' - # with gl_MANYWARN_COMPLEMENT - # So we have -W enabled, and then have to explicitly turn off... - gl_WARN_ADD([-Wno-sign-compare]) - - # GNULIB expects this to be part of -Wc++-compat, but we turn - # that one off, so we need to manually enable this again - gl_WARN_ADD([-Wjump-misses-init]) - - # GNULIB turns on -Wformat=2 which implies -Wformat-nonliteral, - # so we need to manually re-exclude it. - gl_WARN_ADD([-Wno-format-nonliteral]) - - # This should be < 256 really. Currently we're down to 4096, - # but using 1024 bytes sized buffers (mostly for virStrerror) - # stops us from going down further - gl_WARN_ADD([-Wframe-larger-than=4096]) - dnl gl_WARN_ADD([-Wframe-larger-than=256]) - - # Silence certain warnings in gnulib, and use improved glibc headers - AC_DEFINE([lint], [1], - [Define to 1 if the compiler is checking for lint.]) - AC_DEFINE([_FORTIFY_SOURCE], [2], - [enable compile-time and run-time bounds-checking, and some warnings]) - - # Extra special flags - dnl -fstack-protector stuff passes gl_WARN_ADD with gcc - dnl on Mingw32, but fails when actually used - case $host in - *-*-linux*) - dnl Fedora only uses -fstack-protector, but doesn't seem to - dnl be great overhead in adding -fstack-protector-all instead - dnl gl_WARN_ADD([-fstack-protector]) - gl_WARN_ADD([-fstack-protector-all]) - gl_WARN_ADD([--param=ssp-buffer-size=4]) - ;; - esac - gl_WARN_ADD([-fexceptions]) - gl_WARN_ADD([-fasynchronous-unwind-tables]) - gl_WARN_ADD([-fdiagnostics-show-option]) - gl_WARN_ADD([-funit-at-a-time]) - - # Need -fipa-pure-const in order to make -Wsuggest-attribute=pure - # fire even without -O. - gl_WARN_ADD([-fipa-pure-const]) - # We should eventually enable this, but right now there are at - # least 75 functions triggering warnings. - gl_WARN_ADD([-Wno-suggest-attribute=pure]) - gl_WARN_ADD([-Wno-suggest-attribute=const]) - - if test "$enable_compile_warnings" = "error" - then - gl_WARN_ADD([-Werror]) - fi - ;; - *) - AC_MSG_ERROR(Unknown argument '$enable_compile_warnings' to --enable-compile-warnings) - ;; + AC_ARG_ENABLE([werror], + AS_HELP_STRING([--enable-werror], [Use -Werror (if supported)]), + [set_werror="$enableval"], + [if test -d $srcdir/.git; then + is_git_version=true + set_werror=yes + else + set_werror=no + fi]) + + # List of warnings that are not relevant / wanted + + # Don't care about C++ compiler compat + dontwarn="$dontwarn -Wc++-compat" + dontwarn="$dontwarn -Wabi" + dontwarn="$dontwarn -Wdeprecated" + # Don't care about ancient C standard compat + dontwarn="$dontwarn -Wtraditional" + # Don't care about ancient C standard compat + dontwarn="$dontwarn -Wtraditional-conversion" + # Ignore warnings in /usr/include + dontwarn="$dontwarn -Wsystem-headers" + # Happy for compiler to add struct padding + dontwarn="$dontwarn -Wpadded" + # GCC very confused with -O2 + dontwarn="$dontwarn -Wunreachable-code" + # Too many to deal with + dontwarn="$dontwarn -Wconversion" + # Too many to deal with + dontwarn="$dontwarn -Wsign-conversion" + # GNULIB gettext.h violates + dontwarn="$dontwarn -Wvla" + # Many GNULIB header violations + dontwarn="$dontwarn -Wundef" + # Need to allow bad cast for execve() + dontwarn="$dontwarn -Wcast-qual" + # We need to use long long in many places + dontwarn="$dontwarn -Wlong-long" + # We allow manual list of all enum cases without default: + dontwarn="$dontwarn -Wswitch-default" + # We allow optional default: instead of listing all enum values + dontwarn="$dontwarn -Wswitch-enum" + # Not a problem since we don't use -fstrict-overflow + dontwarn="$dontwarn -Wstrict-overflow" + # Not a problem since we don't use -funsafe-loop-optimizations + dontwarn="$dontwarn -Wunsafe-loop-optimizations" + # Things like virAsprintf mean we can't use this + dontwarn="$dontwarn -Wformat-nonliteral" + + # We might fundamentally need some of these disabled forever, but + # ideally we'd turn many of them on + dontwarn="$dontwarn -Wfloat-equal" + dontwarn="$dontwarn -Wdeclaration-after-statement" + dontwarn="$dontwarn -Wcast-qual" + dontwarn="$dontwarn -Wconversion" + dontwarn="$dontwarn -Wsign-conversion" + dontwarn="$dontwarn -Wpacked" + dontwarn="$dontwarn -Wunused-macros" + dontwarn="$dontwarn -Woverlength-strings" + dontwarn="$dontwarn -Wstack-protector" + + # Get all possible GCC warnings + gl_MANYWARN_ALL_GCC([maybewarn]) + + # Remove the ones we don't want, blacklisted earlier + gl_MANYWARN_COMPLEMENT([wantwarn], [$maybewarn], [$dontwarn]) + + # Check for $CC support of each warning + for w in $wantwarn; do + gl_WARN_ADD([$w]) + done + + # GNULIB uses '-W' (aka -Wextra) which includes a bunch of stuff. + # Unfortunately, this means you can't simply use '-Wsign-compare' + # with gl_MANYWARN_COMPLEMENT + # So we have -W enabled, and then have to explicitly turn off... + gl_WARN_ADD([-Wno-sign-compare]) + + # GNULIB expects this to be part of -Wc++-compat, but we turn + # that one off, so we need to manually enable this again + gl_WARN_ADD([-Wjump-misses-init]) + + # GNULIB turns on -Wformat=2 which implies -Wformat-nonliteral, + # so we need to manually re-exclude it. + gl_WARN_ADD([-Wno-format-nonliteral]) + + # This should be < 256 really. Currently we're down to 4096, + # but using 1024 bytes sized buffers (mostly for virStrerror) + # stops us from going down further + gl_WARN_ADD([-Wframe-larger-than=4096]) + dnl gl_WARN_ADD([-Wframe-larger-than=256]) + + # Silence certain warnings in gnulib, and use improved glibc headers + AC_DEFINE([lint], [1], + [Define to 1 if the compiler is checking for lint.]) + AC_DEFINE([_FORTIFY_SOURCE], [2], + [enable compile-time and run-time bounds-checking, and some warnings]) + + # Extra special flags + dnl -fstack-protector stuff passes gl_WARN_ADD with gcc + dnl on Mingw32, but fails when actually used + case $host in + *-*-linux*) + dnl Fedora only uses -fstack-protector, but doesn't seem to + dnl be great overhead in adding -fstack-protector-all instead + dnl gl_WARN_ADD([-fstack-protector]) + gl_WARN_ADD([-fstack-protector-all]) + gl_WARN_ADD([--param=ssp-buffer-size=4]) + ;; esac + gl_WARN_ADD([-fexceptions]) + gl_WARN_ADD([-fasynchronous-unwind-tables]) + gl_WARN_ADD([-fdiagnostics-show-option]) + gl_WARN_ADD([-funit-at-a-time]) + + # Need -fipa-pure-const in order to make -Wsuggest-attribute=pure + # fire even without -O. + gl_WARN_ADD([-fipa-pure-const]) + # We should eventually enable this, but right now there are at + # least 75 functions triggering warnings. + gl_WARN_ADD([-Wno-suggest-attribute=pure]) + gl_WARN_ADD([-Wno-suggest-attribute=const]) + + if test "$set_werror" = "yes" + then + gl_WARN_ADD([-Werror]) + fi WARN_LDFLAGS=$WARN_CFLAGS AC_SUBST([WARN_CFLAGS]) -- 1.7.7.6

2 1

[libvirt] [PATCH] qemu_monitor: Don't output snapshot format argument if type is NULL
by Peter Krempa 27 Mar '12

27 Mar '12

If the snapshot format type string was NULL, the JSON framework created an invalid JSON string. --- The other option would be to fix qemuMonitorJSONMakeCommandRaw that string arguments with a NULL argument would suppress outputing the complete option, but I'm afraid of breaking something. Background: http://www.redhat.com/archives/libvir-list/2012-March/msg01198.html src/qemu/qemu_monitor_json.c | 27 +++++++++++++++++++-------- 1 files changed, 19 insertions(+), 8 deletions(-) diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c index eeeb6a6..6fce58a 100644 --- a/src/qemu/qemu_monitor_json.c +++ b/src/qemu/qemu_monitor_json.c @@ -3112,14 +3112,25 @@ qemuMonitorJSONDiskSnapshot(qemuMonitorPtr mon, virJSONValuePtr actions, virJSONValuePtr cmd; virJSONValuePtr reply = NULL; - cmd = qemuMonitorJSONMakeCommandRaw(actions != NULL, - "blockdev-snapshot-sync", - "s:device", device, - "s:snapshot-file", file, - "s:format", format, - reuse ? "s:mode" : NULL, - reuse ? "existing" : NULL, - NULL); + if (format) { + cmd = qemuMonitorJSONMakeCommandRaw(actions != NULL, + "blockdev-snapshot-sync", + "s:device", device, + "s:snapshot-file", file, + "s:format", format, + reuse ? "s:mode" : NULL, + reuse ? "existing" : NULL, + NULL); + } else { + cmd = qemuMonitorJSONMakeCommandRaw(actions != NULL, + "blockdev-snapshot-sync", + "s:device", device, + "s:snapshot-file", file, + reuse ? "s:mode" : NULL, + reuse ? "existing" : NULL, + NULL); + } + if (!cmd) return -1; -- 1.7.3.4

2 4

[libvirt] [PATCH] Enable build of test suite programs by default for GIT checkouts
by Daniel P. Berrange 27 Mar '12

27 Mar '12

From: "Daniel P. Berrange" <berrange(a)redhat.com> Add a new flag '--with-test-suite' to configure to control whether the test suite binaries are built by default. ie built with a plain 'make', as opposed to delayed until 'make check' For builds from tar.gz tests will not be built by default. For builds from GIT, tests with be on by default, to try and ensure that patch developers don't accidentally break the test suites without noticing. * configure.ac: Add --with-test-suite * tests/Makefile.am: Use noinst_PROGRAMS instead of check_PROGRAMS if building tests by default. Consolidate setting of TESTS and {noinst,check}_PROGRAMS to avoid duplication --- configure.ac | 19 ++++++++ tests/Makefile.am | 119 +++++++++++++--------------------------------------- 2 files changed, 49 insertions(+), 89 deletions(-) diff --git a/configure.ac b/configure.ac index c1f9f45..740129c 100644 --- a/configure.ac +++ b/configure.ac @@ -2162,6 +2162,25 @@ AM_CONDITIONAL([WITH_PYTHON], [test "$with_python" = "yes"]) AC_SUBST([PYTHON_VERSION]) AC_SUBST([PYTHON_INCLUDES]) +AC_ARG_ENABLE([with-test-suite], + AC_HELP_STRING([--with-test-suite], [build test suite by default @<:@default=check@:>@]), + [case "${withval}" in + yes|no|check) ;; + *) AC_MSG_ERROR([bad value ${withval} for tests option]) ;; + esac], + [withval=check]) + +AC_MSG_CHECKING([Whether to build test suite by default]) +if test "$withval" = "check" ; then + if test -d $srcdir/.git ; then + withval=yes + else + withval=no + fi +fi +AC_MSG_RESULT([$withval]) +AM_CONDITIONAL([WITH_TESTS], [test "$withval" = "yes"]) + AC_ARG_ENABLE([test-coverage], AC_HELP_STRING([--enable-test-coverage], [turn on code coverage instrumentation @<:@default=no@:>@]), [case "${enableval}" in diff --git a/tests/Makefile.am b/tests/Makefile.am index 4a0686f..dd8bf4f 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -92,69 +92,68 @@ EXTRA_DIST = \ xml2vmxdata \ .valgrind.supp -check_PROGRAMS = virshtest conftest sockettest \ +test_helpers = commandhelper ssh conftest +test_programs = virshtest sockettest \ nodeinfotest virbuftest \ - commandtest commandhelper seclabeltest \ - virhashtest virnetmessagetest virnetsockettest ssh \ + commandtest seclabeltest \ + virhashtest virnetmessagetest virnetsockettest \ utiltest virnettlscontexttest shunloadtest \ virtimetest viruritest virkeyfiletest \ virauthconfigtest -check_LTLIBRARIES = libshunload.la - # This is a fake SSH we use from virnetsockettest ssh_SOURCES = ssh.c ssh_LDADD = $(COVERAGE_LDFLAGS) if WITH_XEN -check_PROGRAMS += xml2sexprtest sexpr2xmltest \ +test_programs += xml2sexprtest sexpr2xmltest \ xmconfigtest xencapstest statstest reconnect endif if WITH_QEMU -check_PROGRAMS += qemuxml2argvtest qemuxml2xmltest qemuxmlnstest \ +test_programs += qemuxml2argvtest qemuxml2xmltest qemuxmlnstest \ qemuargv2xmltest qemuhelptest domainsnapshotxml2xmltest \ qemumonitortest endif if WITH_LXC -check_PROGRAMS += lxcxml2xmltest +test_programs += lxcxml2xmltest endif if WITH_OPENVZ -check_PROGRAMS += openvzutilstest +test_programs += openvzutilstest endif if WITH_ESX -check_PROGRAMS += esxutilstest +test_programs += esxutilstest endif if WITH_VMX -check_PROGRAMS += vmx2xmltest xml2vmxtest +test_programs += vmx2xmltest xml2vmxtest endif if WITH_CIL -check_PROGRAMS += object-locking +test_programs += object-locking endif if HAVE_YAJL -check_PROGRAMS += jsontest +test_programs += jsontest endif -check_PROGRAMS += networkxml2xmltest +test_programs += networkxml2xmltest if WITH_NETWORK -check_PROGRAMS += networkxml2argvtest +test_programs += networkxml2argvtest endif -check_PROGRAMS += nwfilterxml2xmltest +test_programs += nwfilterxml2xmltest -check_PROGRAMS += storagevolxml2xmltest storagepoolxml2xmltest +test_programs += storagevolxml2xmltest storagepoolxml2xmltest -check_PROGRAMS += nodedevxml2xmltest +test_programs += nodedevxml2xmltest -check_PROGRAMS += interfacexml2xmltest +test_programs += interfacexml2xmltest -check_PROGRAMS += cputest +test_programs += cputest test_scripts = \ capabilityschematest \ @@ -185,6 +184,8 @@ test_scripts += \ virsh-optparse \ virsh-schedinfo \ virsh-synopsis + +test_programs += eventtest else EXTRA_DIST += \ test_conf.sh \ @@ -213,77 +214,16 @@ endif EXTRA_DIST += $(test_scripts) -TESTS = virshtest \ - nodeinfotest \ - virbuftest \ - sockettest \ - commandtest \ - seclabeltest \ - virhashtest \ - virnetmessagetest \ - virnetsockettest \ - virnettlscontexttest \ - virtimetest \ - viruritest \ - virkeyfiletest \ - virauthconfigtest \ - shunloadtest \ - utiltest \ - $(test_scripts) - -if HAVE_YAJL -TESTS += jsontest -endif - -if WITH_XEN -TESTS += xml2sexprtest \ - sexpr2xmltest \ - xmconfigtest \ - xencapstest \ - reconnect \ - statstest -endif - -if WITH_QEMU -TESTS += qemuxml2argvtest qemuxml2xmltest qemuxmlnstest qemuargv2xmltest \ - qemuhelptest domainsnapshotxml2xmltest nwfilterxml2xmltest \ - qemumonitortest -endif - -if WITH_LXC -TESTS += lxcxml2xmltest -endif - -if WITH_OPENVZ -TESTS += openvzutilstest -endif - -if WITH_ESX -TESTS += esxutilstest -endif - -if WITH_VMX -TESTS += vmx2xmltest xml2vmxtest -endif - -if WITH_LIBVIRTD -check_PROGRAMS += eventtest -TESTS += eventtest -endif - -TESTS += networkxml2xmltest - -if WITH_NETWORK -TESTS += networkxml2argvtest +if WITH_TESTS +noinst_PROGRAMS = $(test_programs) $(test_helpers) +noinst_LTLIBRARIES = libshunload.la +else +check_PROGRAMS = $(test_programs) $(test_helpers) +check_LTLIBRARIES = libshunload.la endif -TESTS += storagevolxml2xmltest storagepoolxml2xmltest - -TESTS += nodedevxml2xmltest - -TESTS += interfacexml2xmltest - -TESTS += cputest +TESTS = $(test_programs) \ + $(test_scripts) # NB, automake < 1.10 does not provide the real # abs_top_{src/build}dir or builddir variables, so don't rely @@ -306,6 +246,7 @@ TESTS_ENVIRONMENT = \ LC_ALL=C \ $(VG) + valgrind: $(MAKE) check VG="valgrind --quiet --leak-check=full --suppressions=$(srcdir)/.valgrind.supp" -- 1.7.7.6

2 1