- Devel - libvirt List Archives

[libvirt] [PATCH] conf: Enforce scheduler name when parsing XML
by Peter Krempa 23 Jun '15

23 Jun '15

We require the scheduler name attribute in the schemas but the code would actually be fine when it was omitted. Make it mandatory. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1234729 --- src/conf/domain_conf.c | 49 ++++++++++++++++++++++++++----------------------- 1 file changed, 26 insertions(+), 23 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index e592adf..183e66c 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -14153,32 +14153,35 @@ virDomainThreadSchedParse(xmlNodePtr node, } VIR_FREE(tmp); - tmp = virXMLPropString(node, "scheduler"); - if (tmp) { - if ((pol = virProcessSchedPolicyTypeFromString(tmp)) <= 0) { - virReportError(VIR_ERR_INTERNAL_ERROR, - _("Invalid scheduler attribute: '%s'"), - tmp); + if (!(tmp = virXMLPropString(node, "scheduler"))) { + virReportError(VIR_ERR_INTERNAL_ERROR, "%s", + _("Missing scheduler attribute")); + goto error; + } + + if ((pol = virProcessSchedPolicyTypeFromString(tmp)) <= 0) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("Invalid scheduler attribute: '%s'"), + tmp); + goto error; + } + sp->policy = pol; + + VIR_FREE(tmp); + if (sp->policy == VIR_PROC_POLICY_FIFO || + sp->policy == VIR_PROC_POLICY_RR) { + tmp = virXMLPropString(node, "priority"); + if (!tmp) { + virReportError(VIR_ERR_XML_ERROR, "%s", + _("Missing scheduler priority")); goto error; } - sp->policy = pol; - - VIR_FREE(tmp); - if (sp->policy == VIR_PROC_POLICY_FIFO || - sp->policy == VIR_PROC_POLICY_RR) { - tmp = virXMLPropString(node, "priority"); - if (!tmp) { - virReportError(VIR_ERR_XML_ERROR, "%s", - _("Missing scheduler priority")); - goto error; - } - if (virStrToLong_i(tmp, NULL, 10, &sp->priority) < 0) { - virReportError(VIR_ERR_XML_ERROR, "%s", - _("Invalid value for element priority")); - goto error; - } - VIR_FREE(tmp); + if (virStrToLong_i(tmp, NULL, 10, &sp->priority) < 0) { + virReportError(VIR_ERR_XML_ERROR, "%s", + _("Invalid value for element priority")); + goto error; } + VIR_FREE(tmp); } return 0; -- 2.4.1

2 2

[libvirt] [PATCHv3 00/10] new API virDomainBlockSetWriteThreshold
by Eric Blake 23 Jun '15

23 Jun '15

v2 was here: https://www.redhat.com/archives/libvir-list/2015-June/msg00591.html This series depends on my yajl/json cleanups: https://www.redhat.com/archives/libvir-list/2015-June/msg01098.html and can also be found here: git clone git://repo.or.cz/libvirt/ericb.git threshold http://repo.or.cz/w/libvirt/ericb.git/shortlog/refs/heads/threshold Since then, I've rebased to master, addressed Peter's findings on v2, and tested that things actually work when combined with a single pending patch on qemu (I'll reply to this mail with the list id of that patch, where that patch in turn points to this series as justification). Here's a transcript of how I tested: I modified my 'f20' domain to include: <devices> <emulator>/home/eblake/qemu/x86_64-softmmu/qemu-system-x86_64</emulator> <disk type='block' device='disk'> <driver name='qemu' type='qcow2'/> <source dev='/dev/sda6'/> <target dev='vda' bus='virtio'/> </disk> then created a raw partition wrapper around the usual disk image: # qemu-img create -f qcow2 -o backing_file=/var/lib/libvirt/images/f20.img,backing_fmt=qcow2 /dev/sda6 Formatting '/dev/sda6', fmt=qcow2 size=12884901888 backing_file='/var/lib/libvirt/images/f20.img' backing_fmt='qcow2' encryption=off cluster_size=65536 lazy_refcounts=off refcount_bits=16 I then started the domain paused (as long as no guest I/O happens, the threshold cannot be exceeded, which gives me time to register a threshold and set up an event handler): # virsh start --paused f20 Domain f20 started # ./run tools/virsh -k0 domblkthreshold f20 vda 10000 threshold of vda set to 10000 bytes I was then able to play with domstats to see different thresholds as I registered them (such as playing with the parts-per-million or even with virsh 'percentage sugar'): # virsh -k0 domstats f20 --block | grep -C1 allo block.0.write-threshold=10000 block.0.allocation=0 block.0.capacity=12884901888 In another window, I set up my event loop: $ ./run tools/virsh -k0 event f20 --loop --all then in the first window, I resumed things: # virsh resume f20 and got this message in the second window, after several seconds had elapsed (the first few seconds in the guest is under control of Grub, which doesn't write to disk; when it transitioned over to the Linux boot, disk activity started happening and easily causes the qcow2 wrapper to exceed the threshold I set earlier): $ ./run tools/virsh -k0 -c qemu:///system event --loop --all event 'lifecycle' for domain f20: Resumed Unpaused event 'lifecycle' for domain f20: Resumed Unpaused event 'write-threshold' for domain f20 disk vda (/dev/sda6): threshold 10000 exceeded by 186608 bytes Sure enough, the registration had cleared, and the qcow2 wrapper had grown (it grew more than once in the process of booting, so by the time I re-queried, allocation had grown even further). # virsh -k0 domstats f20 --block | grep -C1 allo block.0.write-threshold=0 block.0.allocation=13434368 block.0.capacity=12884901888 I'd really like to see this series in 1.3.0 to commit to the API; there are still some further improvements we can make (in particular 'vda[1]' support, and also improving qemu event handling to re-grab a job lock if the cache of node names is not populated after a libvirtd restart), but as those improvements won't affect ABI, they can be deferred if they don't make it into 1.3.0. [I'm headed on vacation, so I will be really slow to respond in the next 3 weeks; someone else may need to push this if it is approved for 1.3.0] Eric Blake (10): threshold: new API virDomainBlockSetWriteThreshold threshold: expose new API in virsh threshold: new event object for tracking write threshold threshold: wire up threshold event in RPC threshold: add qemu capability bits threshold: learn a node name for a given qcow2 disk threshold: track an allocation node name for a storage source threshold: scrape threshold data from QMP threshold: add threshold event handling in qemu threshold: add write threshold setting in qemu daemon/remote.c | 51 +++ include/libvirt/libvirt-domain.h | 53 +++ src/conf/domain_event.c | 101 ++++- src/conf/domain_event.h | 16 +- src/driver-hypervisor.h | 7 + src/libvirt-domain.c | 98 +++++ src/libvirt_private.syms | 2 + src/libvirt_public.syms | 1 + src/qemu/qemu_capabilities.c | 12 +- src/qemu/qemu_capabilities.h | 4 +- src/qemu/qemu_domain.c | 105 ++++++ src/qemu/qemu_domain.h | 10 +- src/qemu/qemu_driver.c | 108 +++++- src/qemu/qemu_migration.c | 2 +- src/qemu/qemu_monitor.c | 79 +++- src/qemu/qemu_monitor.h | 28 ++ src/qemu/qemu_monitor_json.c | 305 ++++++++++++++- src/qemu/qemu_monitor_json.h | 11 + src/qemu/qemu_process.c | 44 +++ src/remote/remote_driver.c | 35 ++ src/remote/remote_protocol.x | 30 +- src/remote_protocol-structs | 16 + src/util/virstoragefile.c | 4 +- src/util/virstoragefile.h | 3 +- tests/qemucapabilitiesdata/caps_2.1.1-1.replies | 470 ++++++++++++------------ tests/qemumonitortest.c | 13 +- tools/virsh-domain-monitor.c | 108 ++++++ tools/virsh-domain.c | 24 ++ tools/virsh.pod | 26 ++ 29 files changed, 1508 insertions(+), 258 deletions(-) -- 2.4.3

2 16

[libvirt] [PATCH 0/2] Fix a couple issues found with disk backend
by John Ferlan 23 Jun '15

23 Jun '15

Validation of https://bugzilla.redhat.com/show_bug.cgi?id=1181062 raised a couple of questions about XML format and the defaults - these patches address those questions. John Ferlan (2): docs: Adjust Disk storage rng storage: Force setting of disk format type docs/schemas/storagepool.rng | 2 +- src/storage/storage_backend_disk.c | 5 +++-- 2 files changed, 4 insertions(+), 3 deletions(-) -- 2.1.0

2 7

[libvirt] [PATCH 0/2] Support for new watchdog model diag288 and action inject-nmi
by Boris Fiuczynski 23 Jun '15

23 Jun '15

This patche set provides support for the new watchdog model "diag288" including the new watchdog action "inject-nmi". Boris Fiuczynski (2): Support for the new watchdog action inject-nmi Support for the new watchdog model diag288 docs/formatdomain.html.in | 5 +++++ docs/schemas/domaincommon.rng | 2 ++ include/libvirt/libvirt-domain.h | 1 + src/conf/domain_conf.c | 6 ++++-- src/conf/domain_conf.h | 2 ++ src/qemu/qemu_command.c | 6 +++--- src/qemu/qemu_monitor_json.c | 2 +- tools/virsh-domain.c | 3 ++- 8 files changed, 20 insertions(+), 7 deletions(-) -- 2.3.0

2 3

[libvirt] lxc: setsid() usage
by Richard Weinberger 23 Jun '15

23 Jun '15

Hi! Why is libvirt-lxc issuing a setsid() in lxcContainerSetupFDs()? To me it seems like a hack to have a controlling TTY if PID 1 is /bin/bash. If one runs a sysv init style distro (like Debian) in libvirt-lxc the setsid() has a major downside, when getty spawns a login shell on /dev/tty1 it cannot become the controlling tty. Hence, if one presses ctrl-c in such a session, the container will reboot. Interestingly it does not happen when a systemd distro is used. Maybe because systemd completely closes and reopens the TTY? Thanks, //richard

2 5

[libvirt] [PATCH v2] nodeinfo: fix to parse present cpus rather than possible cpus
by Kothapally Madhu Pavan 23 Jun '15

23 Jun '15

Currently we are parsing all the possible cpus to get the nodeinfo. This fix will perform a check for present cpus before parsing. Signed-off-by: Kothapally Madhu Pavan <kmp(a)linux.vnet.ibm.com> --- src/nodeinfo.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/src/nodeinfo.c b/src/nodeinfo.c index 2fafe2d..0134aba 100644 --- a/src/nodeinfo.c +++ b/src/nodeinfo.c @@ -57,6 +57,7 @@ #define VIR_FROM_THIS VIR_FROM_NONE VIR_LOG_INIT("nodeinfo"); +virBitmapPtr nodeGetPresentCPUBitmap(void); #if defined(__FreeBSD__) || defined(__APPLE__) static int @@ -418,6 +419,7 @@ virNodeParseNode(const char *node, int processors = 0; DIR *cpudir = NULL; struct dirent *cpudirent = NULL; + virBitmapPtr present_cpumap = NULL; int sock_max = 0; cpu_set_t sock_map; int sock; @@ -438,12 +440,18 @@ virNodeParseNode(const char *node, goto cleanup; } + present_cpumap = nodeGetPresentCPUBitmap(); + /* enumerate sockets in the node */ CPU_ZERO(&sock_map); while ((direrr = virDirRead(cpudir, &cpudirent, node)) > 0) { if (sscanf(cpudirent->d_name, "cpu%u", &cpu) != 1) continue; + if (present_cpumap) + if (!(virBitmapIsBitSet(present_cpumap, cpu))) + continue; + if ((online = virNodeGetCpuValue(node, cpu, "online", 1)) < 0) goto cleanup; @@ -477,6 +485,10 @@ virNodeParseNode(const char *node, if (sscanf(cpudirent->d_name, "cpu%u", &cpu) != 1) continue; + if (present_cpumap) + if (!(virBitmapIsBitSet(present_cpumap, cpu))) + continue; + if ((online = virNodeGetCpuValue(node, cpu, "online", 1)) < 0) goto cleanup; @@ -537,6 +549,7 @@ virNodeParseNode(const char *node, ret = -1; } VIR_FREE(core_maps); + virBitmapFree(present_cpumap); return ret; }

2 1

[libvirt] [PATCH 0/3] yajl cleanups
by Eric Blake 23 Jun '15

23 Jun '15

Fixes 'make check' on RHEL 6, which I recently broke, then adds further improvements to the JSON parser. I'm tempted to push patch 1/3 as a build-breaker, but since the other two need review, I'll hold off and we can do all three as a series. Eric Blake (3): json: cope with older yajl semantics json: reject javascript comments json: reject trailing garbage src/util/virjson.c | 45 ++++++++++++++++++++++++++++++++++++++------- tests/jsontest.c | 5 ++++- 2 files changed, 42 insertions(+), 8 deletions(-) -- 2.4.3

3 7

[libvirt] bisected: Unable to get index for interface...
by Thomas Voegtle 23 Jun '15

23 Jun '15

Hi, On my own build linux system without systemd, I have following problem, when I try to start a qemu VM, libvirtd.log says: 2015-06-23 08:29:25.498+0000: 889: error : virNetDevGetIndex:842 : Unable to get index for interface vm0.0: No such device 2015-06-23 08:29:25.551+0000: 886: error : virNetDevGetIndex:842 : Unable to get index for interface vm0.0: No such device 2015-06-23 08:29:25.611+0000: 888: error : virNetDevGetIndex:842 : Unable to get index for interface vm0.0: No such device ... virsh dumpxml vm: ... <interface type='ethernet'> <mac address='52:54:00:cd:15:51'/> <target dev='vm0.0'/> <model type='virtio'/> <boot order='3'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x0f' function='0x0'/> </interface> ... I had no problems with 1.2.12, so I started a git bisect and I came down to this: # first bad commit: [4bbe1029f2fb6cd1c102794951a944c62fdbd0e6] qemu: fix ifindex array reported to systemd When I revert that commit on top of 1.2.13.1, problem is gone. I tried today's version (d10a5f58c75), problem is still in there. Hope this is useful to you? Thomas

1 0

[libvirt] [PATCH] admin: Fix mingw build by reordering includes
by Martin Kletzander 23 Jun '15

23 Jun '15

By trying to lead the way of clean includes, I sorted the lines alphabetically and that is a problem for mingw builds with gnulib. As 'configmake.h' defines DATADIR and 'datatypes.h' transitively includes 'winsock.h' that uses 'DATADIR' as a name for a struct, it's enough to reorder those. Even though this might be worked around in gnulib later on, this fixes the build for now. Signed-off-by: Martin Kletzander <mkletzan(a)redhat.com> --- src/libvirt-admin.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/libvirt-admin.c b/src/libvirt-admin.c index 11b6fe3238b8..b3fd0b3a8cf2 100644 --- a/src/libvirt-admin.c +++ b/src/libvirt-admin.c @@ -25,8 +25,9 @@ #include <rpc/rpc.h> #include "internal.h" -#include "configmake.h" #include "datatypes.h" +#include "configmake.h" + #include "viralloc.h" #include "virlog.h" #include "virnetclient.h" -- 2.4.4

2 1

[libvirt] [PATCH] docs: Properly mark acl.html dependencies
by Michal Privoznik 23 Jun '15

23 Jun '15

The acl.html file includes aclperms.htmlinc which is generated. However, there's a typo in the Makefile which makes make fail to see the dependencies correctly. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- docs/Makefile.am | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/Makefile.am b/docs/Makefile.am index b7b49cb..b807a0b 100644 --- a/docs/Makefile.am +++ b/docs/Makefile.am @@ -163,7 +163,7 @@ EXTRA_DIST= \ sitemap.html.in aclperms.htmlinc \ todo.pl hvsupport.pl todo.cfg-example -acl.html:: $(srcdir)/aclperms.htmlinc +acl.html: $(srcdir)/aclperms.htmlinc $(srcdir)/aclperms.htmlinc: $(top_srcdir)/src/access/viraccessperm.h \ $(srcdir)/genaclperms.pl Makefile.am -- 2.3.6

2 1