- Devel - libvirt List Archives

[libvirt] [PATCH v1 00/16] Start to modularize src/Makefile.am
by Daniel P. Berrangé 23 Feb '18

23 Feb '18

The src/Makefile.am is quite large and has quite poor grouping of rules / variables / etc. This makes it increasingly hard to identify all the rules relating to a particular area of code. Traditionally one might create Makefile.am's in each sub-directory and let make recurse into each. Recursive make is quite a bad idea in general though because it harms parallelization and means that make does not have a full view of dependencies. This series thus takes a different approach to modularization which is to make use of the "include" statement to pull in makefile fragments from subdirectories. automake fully expands all "include" statements when generating the Makefile.in, so we still end up with a single monolithic file for the eventual build. Aside from ensuring make still has a full view of dependencies, this also means that all variables are still in a global namespace. In doing this split, I've taken the opportunity to santize the variable declarations into a consistent style. Any variable that is assigned more than one value now uses line continuations with exactly one value per line and a trailing $(NULL). It would be nice to enforce this style with a syntax-check rule but I've not figured this out yet. This series only moves the virt drivers. So there's obviously a further series to follow behind this to finish the job for other drivers. Daniel P. Berrangé (16): make: split UML driver build rules into uml/Makefile.inc.am make: split PHyp driver build rules into phyp/Makefile.inc.am make: split test driver build rules into test/Makefile.inc.am make: split ESX driver build rules into esx/Makefile.inc.am make: split hyperv driver build rules into hyperv/Makefile.inc.am make: split vmware driver build rules into vmware/Makefile.inc.am make: split vbox driver build rules into vbox/Makefile.inc.am make: split openvz driver build rules into openvz/Makefile.inc.am make: split qemu driver build rules into qemu/Makefile.inc.am make: split bhyve driver build rules into bhyve/Makefile.inc.am make: split xenconfig driver build rules into xenconfig/Makefile.inc.am make: split libxl driver build rules into libxl/Makefile.inc.am make: split xen driver build rules into xen/Makefile.inc.am make: split xenapi driver build rules into xenapi/Makefile.inc.am make: split vz driver build rules into vz/Makefile.inc.am make: split lxc driver build rules into lxc/Makefile.inc.am src/Makefile.am | 872 +++--------------------------------------- src/bhyve/Makefile.inc.am | 73 ++++ src/esx/Makefile.inc.am | 90 +++++ src/hyperv/Makefile.inc.am | 59 +++ src/libxl/Makefile.inc.am | 104 +++++ src/lxc/Makefile.inc.am | 207 ++++++++++ src/openvz/Makefile.inc.am | 28 ++ src/phyp/Makefile.inc.am | 19 + src/qemu/Makefile.inc.am | 148 +++++++ src/test/Makefile.inc.am | 26 ++ src/uml/Makefile.inc.am | 46 +++ src/vbox/Makefile.inc.am | 76 ++++ src/vmware/Makefile.inc.am | 27 ++ src/vz/Makefile.inc.am | 38 ++ src/xen/Makefile.inc.am | 67 ++++ src/xenapi/Makefile.inc.am | 28 ++ src/xenconfig/Makefile.inc.am | 28 ++ 17 files changed, 1122 insertions(+), 814 deletions(-) create mode 100644 src/bhyve/Makefile.inc.am create mode 100644 src/esx/Makefile.inc.am create mode 100644 src/hyperv/Makefile.inc.am create mode 100644 src/libxl/Makefile.inc.am create mode 100644 src/lxc/Makefile.inc.am create mode 100644 src/openvz/Makefile.inc.am create mode 100644 src/phyp/Makefile.inc.am create mode 100644 src/qemu/Makefile.inc.am create mode 100644 src/test/Makefile.inc.am create mode 100644 src/uml/Makefile.inc.am create mode 100644 src/vbox/Makefile.inc.am create mode 100644 src/vmware/Makefile.inc.am create mode 100644 src/vz/Makefile.inc.am create mode 100644 src/xen/Makefile.inc.am create mode 100644 src/xenapi/Makefile.inc.am create mode 100644 src/xenconfig/Makefile.inc.am -- 2.14.3

2 17

[libvirt] [PATCH V2] libxl: round memory values to next 1MiB increment
by Jim Fehlig 23 Feb '18

23 Feb '18

libxl requires the memory sizes to be rounded to 1MiB increments. Attempting to start a domain that violates this requirement will fail with the marginally helpful error 2018-02-22 01:55:32.921+0000: xc: panic: xc_dom_boot.c:141: xc_dom_boot_mem_init: can't allocate low memory for domain: Out of memory 2018-02-22 01:55:32.921+0000: libxl: libxl_dom.c:671:libxl__build_dom: xc_dom_boot_mem_init failed: No such file or directory Round the maximum and current memory values to the next 1MiB increment when generating the libxl_domain_config object. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- V2: Set rounded memory values in virDomainDef object so they are correctly reflected in the live config. src/libxl/libxl_conf.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 57cb4c0ad..d074edc9c 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -367,6 +367,10 @@ libxlMakeDomBuildInfo(virDomainDefPtr def, } b_info->sched_params.weight = 1000; + /* Xen requires the memory sizes to be rounded to 1MiB increments */ + virDomainDefSetMemoryTotal(def, + VIR_ROUND_UP(virDomainDefGetMemoryInitial(def), 1024)); + def->mem.cur_balloon = VIR_ROUND_UP(def->mem.cur_balloon, 1024); b_info->max_memkb = virDomainDefGetMemoryInitial(def); b_info->target_memkb = def->mem.cur_balloon; if (hvm) { -- 2.16.1

2 1

[libvirt] [PATCH] libvirt.spec.in: Don't distribute libvirtd.upstart
by Michal Privoznik 23 Feb '18

23 Feb '18

Firstly, for rpm we are building libvirt with --init-script=systemd or --init-script=redhat. So upstart is never enabled. And only due to a bug we installed libvirtd.upstart file. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- libvirt.spec.in | 1 - 1 file changed, 1 deletion(-) diff --git a/libvirt.spec.in b/libvirt.spec.in index 1eca4e39d..8f46f58b5 100644 --- a/libvirt.spec.in +++ b/libvirt.spec.in @@ -1825,7 +1825,6 @@ exit 0 %{_sysconfdir}/rc.d/init.d/virtlogd %{_sysconfdir}/rc.d/init.d/virtlockd %endif -%doc daemon/libvirtd.upstart %config(noreplace) %{_sysconfdir}/sysconfig/libvirtd %config(noreplace) %{_sysconfdir}/sysconfig/virtlogd %config(noreplace) %{_sysconfdir}/sysconfig/virtlockd -- 2.16.1

2 1

[libvirt] [PATCH] qemu_hotplug: Drop dead code in net update
by Jiri Denemark 23 Feb '18

23 Feb '18

vm->def->nets[changeidx] can never be NULL for changeidx returned by virDomainNetFindIdx. Signed-off-by: Jiri Denemark <jdenemar(a)redhat.com> --- src/qemu/qemu_hotplug.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c index 87cf578acc..e0a5300f08 100644 --- a/src/qemu/qemu_hotplug.c +++ b/src/qemu/qemu_hotplug.c @@ -3061,12 +3061,7 @@ qemuDomainChangeNet(virQEMUDriverPtr driver, if ((changeidx = virDomainNetFindIdx(vm->def, newdev)) < 0) goto cleanup; devslot = &vm->def->nets[changeidx]; - - if (!(olddev = *devslot)) { - virReportError(VIR_ERR_OPERATION_FAILED, "%s", - _("cannot find existing network device to modify")); - goto cleanup; - } + olddev = *devslot; oldType = virDomainNetGetActualType(olddev); if (oldType == VIR_DOMAIN_NET_TYPE_HOSTDEV) { -- 2.16.2

2 1

[libvirt] [PATCH] libxl: round memory values to next 1MiB increment
by Jim Fehlig 22 Feb '18

22 Feb '18

libxl requires the memory sizes to be rounded to 1MiB increments. Attempting to start a domain that violates this requirement will fail with the marginally helpful error 2018-02-22 01:55:32.921+0000: xc: panic: xc_dom_boot.c:141: xc_dom_boot_mem_init: can't allocate low memory for domain: Out of memory 2018-02-22 01:55:32.921+0000: libxl: libxl_dom.c:671:libxl__build_dom: xc_dom_boot_mem_init failed: No such file or directory Round the maximum and current memory values to the next 1MiB increment when generating the libxl_domain_config object. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- The memory values could also be adjusted in the post parse callback, which might be better since the virDomainDef object would then reflect the actual values given to libxl. In general, I often find myself confused about the dividing line between adjusting such things in the post parse callback vs domain creation. src/libxl/libxl_conf.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c index 57cb4c0ad..02bbf4e28 100644 --- a/src/libxl/libxl_conf.c +++ b/src/libxl/libxl_conf.c @@ -367,8 +367,8 @@ libxlMakeDomBuildInfo(virDomainDefPtr def, } b_info->sched_params.weight = 1000; - b_info->max_memkb = virDomainDefGetMemoryInitial(def); - b_info->target_memkb = def->mem.cur_balloon; + b_info->max_memkb = VIR_ROUND_UP(virDomainDefGetMemoryInitial(def), 1024); + b_info->target_memkb = VIR_ROUND_UP(def->mem.cur_balloon, 1024); if (hvm) { char bootorder[VIR_DOMAIN_BOOT_LAST + 1]; -- 2.16.1

2 2

[libvirt] [PATCH] make: fix another VPATH bug impacting install of sysconf files
by Daniel P. Berrangé 22 Feb '18

22 Feb '18

Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- Pushed as build fix src/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Makefile.am b/src/Makefile.am index 2166e17dbe..8ceeda5756 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -2833,7 +2833,7 @@ install-sysconfig: for f in $(SYSCONF_FILES:%.sysconf=%) ; \ do \ tgt=`basename $$f`; \ - $(INSTALL_SCRIPT) $$f.sysconf $(DESTDIR)$(sysconfdir)/sysconfig/$$tgt; \ + $(INSTALL_SCRIPT) $(srcdir)/$$f.sysconf $(DESTDIR)$(sysconfdir)/sysconfig/$$tgt; \ done uninstall-sysconfig: -- 2.14.3

1 0

[libvirt] RFC: should we mandate use of VPATH builds ?
by Daniel P. Berrangé 22 Feb '18

22 Feb '18

One of the things we (well I) constantly screw up is VPATH builds. I've always tended to just build in the source tree, and as a result I'll often miss changes which break VPATH builds. Now our Jenkins CI catches this because it does VPATH builds only, so we don't bit-rot for very long. Unfortunately Travis CI does source tree builds only so I don't notice my mistakes when I test branches on Travis before pushing. I'll fix Travis to do VPATH builds too, but.... I notice that most modern build systems will refuse to ever do in source tree builds. eg Ninja/Meson/Perl/Python/Go/etc all put build artifacts in a separate directory from source. We could if we wanted to simply put a check right at the start of the configure.ac script to reject any attempt at doing a source tree build to force VPATH builds. This would mean we only ever have to care about one execution path instead of two execution paths. We could even make it such that autogen.sh will actually create a subdir "build" and change into it automatically before running configure, to help devs do the "right thing" by default. Essentially we're talking the difference between Today: - Source tree build ./autogen.sh && make - VPATH build mkdir build && cd build && ../autogen.sh && make Possible future - Source tree build No. - VPATH build ./autogen.sh && cd build && make Any opinions ? Do people really like source tree builds, or are they (like me) just doing them because its the easy/lazy option ? Regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|

5 4

[libvirt] [PATCH] src: Clean *.logrotate, .aug and libvirtd.policy files on 'make clean'
by Michal Privoznik 22 Feb '18

22 Feb '18

Before 3f055b5997c we were doing that. However after the commit we are leaving the files behind. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- Pushed under build breaker rule. src/Makefile.am | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/Makefile.am b/src/Makefile.am index 41a3cb262..2166e17db 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -1376,7 +1376,7 @@ augeastest_DATA += test_libvirtd.aug conf_DATA += remote/libvirtd.conf -CLEANFILES += tets_libvirtd.aug +CLEANFILES += test_libvirtd.aug if WITH_POLKIT if WITH_POLKIT0 @@ -1389,6 +1389,7 @@ endif ! WITH_POLKIT0 endif WITH_POLKIT BUILT_SOURCES += libvirtd.policy +CLEANFILES += libvirtd.policy libvirtd.policy: remote/libvirtd.policy.in $(top_builddir)/config.status $(AM_V_GEN) sed \ @@ -2868,6 +2869,7 @@ if WITH_LIBVIRTD LOGROTATE_FILES := $(LOGROTATE_FILES_IN:remote/%.in=%) BUILT_SOURCES += $(LOGROTATE_FILES) +CLEANFILES += $(LOGROTATE_FILES) %.logrotate: remote/%.logrotate.in $(AM_V_GEN)sed \ -- 2.16.1

2 3

[libvirt] [PATCH] make: fix VPATH build for libvirtd augeas check
by Daniel P. Berrangé 22 Feb '18

22 Feb '18

Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com> --- src/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Makefile.am b/src/Makefile.am index b8e2cbf139..41a3cb2628 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -2297,7 +2297,7 @@ check-augeas-virtlogd: test_virtlogd.aug if WITH_LIBVIRTD test_libvirtd.aug: remote/test_libvirtd.aug.in \ remote/libvirtd.conf $(AUG_GENTEST) - $(AM_V_GEN)$(AUG_GENTEST) remote/libvirtd.conf $< $@ + $(AM_V_GEN)$(AUG_GENTEST) $(srcdir)/remote/libvirtd.conf $< $@ check-augeas-libvirtd: test_libvirtd.aug $(AM_V_GEN)if test -x '$(AUGPARSE)'; then \ -- 2.14.3

2 1

[libvirt] [PATCH] virLogGetOutputs: remove unnecessary braces
by Ján Tomko 22 Feb '18

22 Feb '18

Commit 9275def reduced the if block to one line without removing the braces. Signed-off-by: Ján Tomko <jtomko(a)redhat.com> --- I pushed this yesterday as a build fix, but forgot to send an e-mail. src/util/virlog.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/util/virlog.c b/src/util/virlog.c index c03f4fb14..dd927f0ba 100644 --- a/src/util/virlog.c +++ b/src/util/virlog.c @@ -1219,9 +1219,8 @@ virLogGetOutputs(void) } } - if (virBufferError(&outputbuf)) { + if (virBufferError(&outputbuf)) goto error; - } virLogUnlock(); return virBufferContentAndReset(&outputbuf); -- 2.13.6

1 0