- Devel - libvirt List Archives

[libvirt] [PATCH v3 1/2] remote: set neventCallbacks to zero at DEREG_CB
by xinhua.Cao 12 Nov '17

12 Nov '17

set neventCallbacks to zero after VIR_FREE(eventCallbacks) was called --- daemon/remote.c | 1 + 1 file changed, 1 insertion(+) diff --git a/daemon/remote.c b/daemon/remote.c index 07557e9..cbcb6e8 100644 --- a/daemon/remote.c +++ b/daemon/remote.c @@ -1686,6 +1686,7 @@ void remoteRelayConnectionClosedEvent(virConnectPtr conn ATTRIBUTE_UNUSED, int r VIR_WARN("unexpected %s event deregister failure", name); \ } \ VIR_FREE(eventCallbacks); \ + neventCallbacks = 0; \ } while (0); /* -- 2.8.3

2 3

[libvirt] bhyve: grub-bhyve: support overriding just --root flag
by Christian Schwarz 11 Nov '17

11 Nov '17

Hi, I was trying to get a GPT-formatted VM boot on FreeBSD using the bhyve driver and the grub-bhyve bootloader. Turns out that libvirt 3.9.0 hardcodes the boot partition to (hd0,msdos1) or allows overriding it completly using <bootloader_args>. I hacked together a patch that allows overring just the --root argument to grub-bhyve and updated the documentation: https://github.com/problame/libvirt/commit/5fd1265c05987d907d9f1d9913dbee83… Obviously, this does not meet quality standards and should not be merged as is, but maybe spawn some discussion (if anyone is actually using bhyve + libvirt). Cheers, Christian --- commit 5fd1265c05987d907d9f1d9913dbee832a227889 Author: Christian Schwarz <me(a)cschwarz.com> Date: Sat Nov 11 16:15:05 2017 +0100 bhyve: grub-bhyve: support overriding just the --root argument in domain config diff --git a/docs/drvbhyve.html.in b/docs/drvbhyve.html.in index 63260afae..2583bfa01 100644 --- a/docs/drvbhyve.html.in +++ b/docs/drvbhyve.html.in @@ -300,17 +300,26 @@ are omitted, libvirt will try and infer boot ordering from user-supplied <boot order='N'> configuration in the domain. Failing that, it will boot the first disk in the domain (either <code>cdrom</code>- or <code>disk</code>-type devices). If the disk type is <code>disk</code>, it will -attempt to boot from the first partition in the disk image.</p> +attempt to boot from the first partition in the disk image, assuming +an <code>msdos</code> partitioning scheme +(i.e. <code>grub-bhyve --root hd0,msdos1</code>). +You can override this behavior using <code>bootloader_args</code> or <code>bootloader_grub_root</code>. +</p> <pre> ... <bootloader>/usr/local/sbin/grub-bhyve</bootloader> + <bootloader_args>...</bootloader_args> + +<bootloader_grub_root>hd0,gpt1</bootloader_grub_root> ... </pre> -<p>Caveat: <code>bootloader_args</code> does not support any quoting. -Filenames, etc, must not have spaces or they will be tokenized incorrectly.</p> +<p>Caveats when using <code>bootloader_args</code>: it does not support any quoting. +Filenames, etc, must not have spaces or they will be tokenized incorrectly. +Additionally, you will have to maintain your own <code>--device-map</code> +file and keep it in sync with the domain XML.</p> <h3><a id="uefi">Using UEFI bootrom, VNC, and USB tablet</a></h3> diff --git a/src/bhyve/bhyve_command.c b/src/bhyve/bhyve_command.c index 55032ae1d..6cab6e516 100644 --- a/src/bhyve/bhyve_command.c +++ b/src/bhyve/bhyve_command.c @@ -774,15 +774,21 @@ virBhyveProcessBuildGrubbhyveCmd(virDomainDefPtr def, } virCommandAddArg(cmd, "--root"); - if (userdef != NULL) { - if (userdef->device == VIR_DOMAIN_DISK_DEVICE_CDROM) + if (def->os.bootloaderGrubRoot != NULL) { + virCommandAddArg(cmd, def->os.bootloaderGrubRoot); + } else { + + if (userdef != NULL) { + if (userdef->device == VIR_DOMAIN_DISK_DEVICE_CDROM) + virCommandAddArg(cmd, "cd"); + else + virCommandAddArg(cmd, "hd0,msdos1"); + } else if (cd != NULL) { virCommandAddArg(cmd, "cd"); - else + } else { virCommandAddArg(cmd, "hd0,msdos1"); - } else if (cd != NULL) { - virCommandAddArg(cmd, "cd"); - } else { - virCommandAddArg(cmd, "hd0,msdos1"); + } + } virCommandAddArg(cmd, "--device-map"); diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 7dfd7b54e..ecd1f71dd 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -18141,6 +18141,7 @@ virDomainDefParseXML(xmlDocPtr xml, def->os.bootloader = virXPathString("string(./bootloader)", ctxt); def->os.bootloaderArgs = virXPathString("string(./bootloader_args)", ctxt); + def->os.bootloaderGrubRoot = virXPathString("string(./bootloader_grub_root)", ctxt); tmp = virXPathString("string(./os/type[1])", ctxt); if (!tmp) { diff --git a/src/conf/domain_conf.h b/src/conf/domain_conf.h index e3f060b12..f969e9195 100644 --- a/src/conf/domain_conf.h +++ b/src/conf/domain_conf.h @@ -1884,6 +1884,7 @@ struct _virDomainOSDef { char *slic_table; virDomainLoaderDefPtr loader; char *bootloader; + char *bootloaderGrubRoot; char *bootloaderArgs; int smbios_mode;

1 0

[libvirt] [PATCH 00/11] Make auto completion better
by Michal Privoznik 10 Nov '17

10 Nov '17

After initial RFC [1] I had some time to work on this and here is the result. What's implemented? =================== Auto completion for both interactive and non-interactive virsh/virt-admin. Known limitations ================= Currently, just options completers work. I mean, to bring up list of domains you have to: virsh # start --domain <TAB><TAB> Doing bare: virsh # start <TAB><TAB> brings up list of --options. I am working on this meanwhile. But one can argue that having to use --option is not that much of a problem since we have good completion now. The other known limitation - and actually room for others to join and help - is missing completers. I mean, the last 3 patches give an example how to do that. But that is still very few. We need more completers. How does this work? =================== The basic idea is simple, when defining opts for command two new struct members can be set: {.name = "mac", .type = VSH_OT_STRING, + .completer = virshDomainInterfaceCompleter, + .completer_flags = VIRSH_DOMAIN_INTERFACE_COMPLETER_MAC, .help = N_("MAC address") }, @completer is the callback that is used when user wants to bring up list of possible strings. @completer_flags can then be used to refine return value of @completer. In this specific example, virshDomainInterfaceCompleter() brings up list of interfaces for given domain. It tries to return /interface/target/@dev but if not found (e.g. because domain is not running), /interface/mac/@address is returned otherwise. However, in one specific case we are interested only in MAC addresses (regardless of domain state) and thus VIRSH_DOMAIN_INTERFACE_COMPLETER_MAC flag is specified which alters @completer behaviour. For non-interactive virsh/virt-admin there's tools/bash-completion/vsh script which does no more than calling: $1 complete $USER_INPUT (where $1 is name of the binary user intents to run). How to test this? ================= Interactive completion should work out of the box. Just make sure you're connected. Completers don't connect! You certainly don't want ssh's 'Password:' prompt show on <TAB><TAB>, do you? Non-interactive completers do connect, but in order to avoid any password prompts, /dev/stdin is closed before anything else is done. In order to test it, just: libvirt.git $ source tools/bash-completion/vsh Now, bash completion should work: libvirt.git $ ./tools/virsh -c qemu:///system start --domain <TAB><TAB> Notes ===== This is that part of vsh that nobody wants to touch, but with these patches in we have the framework to just write small functions (as can be seen in examples). User would benefit from this! As usual, you can find all the patches on my github [2]. Michal 1: https://www.redhat.com/archives/libvir-list/2017-October/msg01374.html 2: https://github.com/zippy2/libvirt/tree/bash_autocomplete Michal Privoznik (11): vshCommandStringParse: Allow retrieving partial result vshCommandOpt: Relax check for valid options vsh: Add @silent to vshConnectionHook vsh: Fix vshCompleter signature vsh: Call vshCmdOptDef.completer properly vshCompleter: Pass partially parsed command vsh: Introduce complete command tools: Provide bash autompletion file virsh: Introduce virshDomainNameCompleter virsh: Introduce virshDomainInterfaceCompleter virt-admin: Introduce vshAdmServerCompleter configure.ac | 3 + libvirt.spec.in | 2 + m4/virt-bash-completion.m4 | 74 ++++++++++++++++ tools/Makefile.am | 40 ++++++++- tools/bash-completion/vsh | 73 ++++++++++++++++ tools/virsh-completer.c | 150 +++++++++++++++++++++++++++++++++ tools/virsh-completer.h | 41 +++++++++ tools/virsh-domain-monitor.c | 33 ++++---- tools/virsh-domain.c | 188 +++++++++++++++++++++-------------------- tools/virsh-snapshot.c | 24 +++--- tools/virsh.c | 72 +++++++++------- tools/virsh.h | 12 ++- tools/virt-admin-completer.c | 76 +++++++++++++++++ tools/virt-admin-completer.h | 33 ++++++++ tools/virt-admin.c | 62 ++++++++------ tools/vsh.c | 195 ++++++++++++++++++++++++++++++++----------- tools/vsh.h | 23 ++++- 17 files changed, 873 insertions(+), 228 deletions(-) create mode 100644 m4/virt-bash-completion.m4 create mode 100644 tools/bash-completion/vsh create mode 100644 tools/virsh-completer.c create mode 100644 tools/virsh-completer.h create mode 100644 tools/virt-admin-completer.c create mode 100644 tools/virt-admin-completer.h -- 2.13.6

3 31

[libvirt] [PATCH] tests: fix TTY check in virTestUseTerminalColors()
by Pino Toscano 10 Nov '17

10 Nov '17

Since colors would be used when writing to stdout, then check that stdout is a TTY, instead of stdin. This avoids the usage of terminal color codes when the output is directed to file. --- tests/testutils.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/testutils.c b/tests/testutils.c index 915cd3b79..9266f15e8 100644 --- a/tests/testutils.c +++ b/tests/testutils.c @@ -92,7 +92,7 @@ bool virTestOOMActive(void) static int virTestUseTerminalColors(void) { - return isatty(STDIN_FILENO); + return isatty(STDOUT_FILENO); } static unsigned int -- 2.13.6

2 1

[libvirt] [PATCH] numa: avoid failure in nodememstats on non-NUMA systems
by Viktor Mihajlovski 10 Nov '17

10 Nov '17

libvirt reports a fake NUMA topology in virConnectGetCapabilities even if built without numactl support. The fake NUMA topology consists of a single cell representing the host's cpu and memory resources. Currently this is the case for ARM and s390[x] RPM builds. A client iterating over NUMA cells obtained via virConnectGetCapabilities and invoking virNodeGetMemoryStats on them will see an internal failure "NUMA isn't available on this host". An example for such a client is VDSM. Since the intention seems to be that libvirt always reports at least a single cell it is necessary to return "fake" node memory statistics matching the previously reported fake cell in case NUMA isn't supported on the system. Signed-off-by: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com> --- src/util/virhostmem.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/src/util/virhostmem.c b/src/util/virhostmem.c index a9ba278..fa04a37 100644 --- a/src/util/virhostmem.c +++ b/src/util/virhostmem.c @@ -267,6 +267,14 @@ virHostMemGetStats(int cellNum ATTRIBUTE_UNUSED, FILE *meminfo; int max_node; + /* + * Even if built without numactl, libvirt claims + * to have a one-cells NUMA topology. In such a + * case return the statistics for the entire host. + */ + if (!virNumaIsAvailable() && cellNum == 0) + cellNum = VIR_NODE_MEMORY_STATS_ALL_CELLS; + if (cellNum == VIR_NODE_MEMORY_STATS_ALL_CELLS) { if (VIR_STRDUP(meminfo_path, MEMINFO_PATH) < 0) return -1; -- 1.9.1

3 4

[libvirt] [PATCH 0/2] vbox: Add support for 5.2.x
by Dawid Zamirski 09 Nov '17

09 Nov '17

With each minor VBOX release a new SDK header needs to be added and wired up to the driver. This is basically the same as last time [1]. The first patch will be sent as a compressed attachment because it's a large header file taken from the SDK that would not pass the mailing list size constraints. [1] https://www.redhat.com/archives/libvir-list/2016-November/msg00355.html Dawid Zamirski (2): vbox: Add vbox 5.2 CAPI header file. vbox: Add support for 5.2.x src/Makefile.am | 1 + src/vbox/vbox_CAPI_v5_2.h | 26870 ++++++++++++++++++++++++++++++++++++++++ src/vbox/vbox_V5_2.c | 13 + src/vbox/vbox_common.h | 2 + src/vbox/vbox_storage.c | 2 + src/vbox/vbox_tmpl.c | 2 + src/vbox/vbox_uniformed_api.h | 1 + 7 files changed, 26891 insertions(+) create mode 100644 src/vbox/vbox_CAPI_v5_2.h create mode 100644 src/vbox/vbox_V5_2.c -- 2.14.2

2 4

[libvirt] [PATCH 0/4] qemu: fix usage of disk and storage source private data
by Peter Krempa 09 Nov '17

09 Nov '17

There were a few cases where we'd use private data for a storage source which was not allocated. Fix it by checking whether it is allocated. Qemu command line parser allocated disk definition via VIR_ALLOC rather than the proper function. This resulted into the disk private data strucutre not being allocated. That one is required. Peter Krempa (4): qemu: parse: Allocate disk definition with private data qemu: Tolerate storage source private data being NULL qemu: domain: Don't allocate storage source private data if not needed qemu: process: Setup disk secrets when preparing disks src/qemu/qemu_command.c | 18 ++++++++++++++---- src/qemu/qemu_domain.c | 14 ++++++++------ src/qemu/qemu_hotplug.c | 17 ++++++++++++----- src/qemu/qemu_parse_command.c | 4 +--- src/qemu/qemu_process.c | 14 +++++++++----- 5 files changed, 44 insertions(+), 23 deletions(-) -- 2.14.3

2 6

[libvirt] [PATCH] conf: Properly parse <backingStore/>
by Peter Krempa 09 Nov '17

09 Nov '17

The terminator would not be parsed properly since the XPath selector was looking for an populated element, and also the code did not bother assigning the terminating virStorageSourcePtr to the backingStore property of the parent. Some tests would catch it if there wasn't bigger fallout from the change to backing store termination in a693fdba0111. Fix them properly now. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1509110 --- src/conf/domain_conf.c | 14 ++++++-------- .../qemuxml2xmlout-disk-active-commit.xml | 1 + .../qemuxml2xmlout-disk-backing-chains-active.xml | 3 +++ .../qemuxml2xmlout-disk-mirror-active.xml | 4 ++++ .../qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-old.xml | 4 ++++ .../qemuxml2xmlout-seclabel-static-labelskip.xml | 1 + 6 files changed, 19 insertions(+), 8 deletions(-) diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c index 7dfd7b54e..a1ca307c4 100644 --- a/src/conf/domain_conf.c +++ b/src/conf/domain_conf.c @@ -8545,23 +8545,21 @@ virDomainDiskBackingStoreParse(xmlXPathContextPtr ctxt, char *idx = NULL; int ret = -1; - if (!(ctxt->node = virXPathNode("./backingStore[*]", ctxt))) { + if (!(ctxt->node = virXPathNode("./backingStore", ctxt))) { ret = 0; goto cleanup; } - if (!(type = virXMLPropString(ctxt->node, "type"))) { - /* terminator does not have a type */ - if (VIR_ALLOC(backingStore) < 0) - goto cleanup; + if (VIR_ALLOC(backingStore) < 0) + goto cleanup; + /* terminator does not have a type */ + if (!(type = virXMLPropString(ctxt->node, "type"))) { + VIR_STEAL_PTR(src->backingStore, backingStore); ret = 0; goto cleanup; } - if (VIR_ALLOC(backingStore) < 0) - goto cleanup; - if (!(flags & VIR_DOMAIN_DEF_PARSE_INACTIVE) && (idx = virXMLPropString(ctxt->node, "index")) && virStrToLong_uip(idx, NULL, 10, &backingStore->id) < 0) { diff --git a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-active-commit.xml b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-active-commit.xml index 5766e4aea..cc26af109 100644 --- a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-active-commit.xml +++ b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-active-commit.xml @@ -20,6 +20,7 @@ <backingStore type='block' index='1'> <format type='raw'/> <source dev='/dev/HostVG/QEMUGuest1'/> + <backingStore/> </backingStore> <mirror type='block' job='active-commit'> <format type='raw'/> diff --git a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-backing-chains-active.xml b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-backing-chains-active.xml index 828defcc2..d1fd2442c 100644 --- a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-backing-chains-active.xml +++ b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-backing-chains-active.xml @@ -49,6 +49,7 @@ <backingStore type='file' index='6'> <format type='raw'/> <source file='/tmp/Fedora-17-x86_64-Live-KDE.iso'/> + <backingStore/> </backingStore> </backingStore> </backingStore> @@ -63,6 +64,7 @@ <source protocol='gluster' name='Volume1/Image'> <host name='example.org' port='6000'/> </source> + <backingStore/> <target dev='vdc' bus='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/> </disk> @@ -79,6 +81,7 @@ <backingStore type='file' index='1'> <format type='qcow2'/> <source file='/tmp/image.qcow'/> + <backingStore/> </backingStore> <target dev='vdd' bus='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/> diff --git a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-active.xml b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-active.xml index 252bde338..c1e8a33ec 100644 --- a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-active.xml +++ b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-active.xml @@ -16,6 +16,7 @@ <emulator>/usr/bin/qemu-system-i686</emulator> <disk type='block' device='disk'> <source dev='/dev/HostVG/QEMUGuest1'/> + <backingStore/> <mirror type='block' job='copy' ready='yes'> <source dev='/dev/HostVG/QEMUGuest1Copy'/> </mirror> @@ -24,12 +25,14 @@ </disk> <disk type='block' device='cdrom'> <source dev='/dev/HostVG/QEMUGuest2'/> + <backingStore/> <target dev='hdc' bus='ide'/> <readonly/> <address type='drive' controller='0' bus='1' target='0' unit='0'/> </disk> <disk type='file' device='disk'> <source file='/tmp/data.img'/> + <backingStore/> <mirror type='file' file='/tmp/copy.img' format='qcow2' job='copy'> <format type='qcow2'/> <source file='/tmp/copy.img'/> @@ -39,6 +42,7 @@ </disk> <disk type='file' device='disk'> <source file='/tmp/logs.img'/> + <backingStore/> <mirror type='file' file='/tmp/logcopy.img' format='qcow2' job='copy' ready='abort'> <format type='qcow2'/> <source file='/tmp/logcopy.img'/> diff --git a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-old.xml b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-old.xml index f4bd39a58..e390bc02f 100644 --- a/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-old.xml +++ b/tests/qemuxml2xmloutdata/qemuxml2xmlout-disk-mirror-old.xml @@ -16,6 +16,7 @@ <emulator>/usr/bin/qemu-system-i686</emulator> <disk type='block' device='disk'> <source dev='/dev/HostVG/QEMUGuest1'/> + <backingStore/> <mirror type='file' file='/dev/HostVG/QEMUGuest1Copy' job='copy' ready='yes'> <source file='/dev/HostVG/QEMUGuest1Copy'/> </mirror> @@ -24,12 +25,14 @@ </disk> <disk type='block' device='cdrom'> <source dev='/dev/HostVG/QEMUGuest2'/> + <backingStore/> <target dev='hdc' bus='ide'/> <readonly/> <address type='drive' controller='0' bus='1' target='0' unit='0'/> </disk> <disk type='file' device='disk'> <source file='/tmp/data.img'/> + <backingStore/> <mirror type='file' file='/tmp/copy.img' format='qcow2' job='copy'> <format type='qcow2'/> <source file='/tmp/copy.img'/> @@ -39,6 +42,7 @@ </disk> <disk type='file' device='disk'> <source file='/tmp/logs.img'/> + <backingStore/> <target dev='vdb' bus='virtio'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x04' function='0x0'/> </disk> diff --git a/tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-static-labelskip.xml b/tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-static-labelskip.xml index 91f573db7..d37b950cb 100644 --- a/tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-static-labelskip.xml +++ b/tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-static-labelskip.xml @@ -18,6 +18,7 @@ <source dev='/dev/HostVG/QEMUGuest1'> <seclabel model='selinux' labelskip='yes'/> </source> + <backingStore/> <target dev='hda' bus='ide'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </disk> -- 2.14.3

2 1

[libvirt] [PATCH 0/2] qemu: block: Fix issues pointed out by coverity
by Peter Krempa 09 Nov '17

09 Nov '17

Few unlikely memleaks and a more likely NULL deref. Peter Krempa (2): qemu: block: Break out early on invalid storage sources qemu: block: Don't leak server JSON object from protocol generators src/qemu/qemu_block.c | 76 ++++++++++++++++++++++++++++++++------------------- 1 file changed, 48 insertions(+), 28 deletions(-) -- 2.14.3

2 3

[libvirt] KVM Forum BoF session notes: QEMU configuration/command-line/QMP
by Eduardo Habkost 09 Nov '17

09 Nov '17

Hi, Below are the notes I collected during the BoF session about QEMU configuration, command-line and QMP. Sorry for taking so long to send them. All inaccuracies and mistakes below are my own fault. (I learned the hard way that trying to take notes while participating in the discussion at the same time is not a good idea.) -------------------------------- KVM Forum 2017 BoF session: configuration, command-line and QMP 2017-10-26 People: Jiri Denemark, Laine Stump, Peter Krempa, Markus Armbruster, Eduardo Habkost, Igor Mammedov, Kevin Wolf, Eduardo Otubo, Alistair Francis, Kashyap Chamarthy, Max Reitz ---------------- First topic: early QMP and -paused series by Igor ehabkost tried to explain the original problem and Igor's proposal[1]. Markus: startup time is important, some things can make it slow. Personally, not so much concerned about querying stuff. Just cache it! Markus, about starting a monitor earlier: "it just makes sense". One problem is: command-line processing ordering mess in QEMU. It's tempting but very scary. very easy to screw up. About command-line ordering, “I'd just go left-to-right”. ehabkost: is gradual change with NUMA stuff the right path? (Markus: "what's the right path?"; ehabkost: "I don't know"). If you are using CPU hotplug you still might have some stuff not appearing in the command-line, because it uses query-hotpluggable-cpus. pkrempa: if you're using cpu hotplug you don't care about startup time. NUMA is not different. Markus's objection to current -paused series: the need for another option/state. We basically have 3 states/phases: 1: between exec and monitor available 2: between monitor and machines start to runs (in that phase a device_add is/can be a cold plug) 3: then the machine actually runs, and all device_add is hotplug ehabkost: that's how I understand it, too. ehabkost: question for libvirt developers: is there an "expansion" operation where libvirt could fill-up missing data on the XML based on info queried from QEMU? (Answer seems to be yes) ---------------- Some comments from Igor and Peter about query-hotpluggable-cpus and the need to choose the right device type (sometimes TYPE_CPU is not appropriate for device_add, but only CPU cores, like on Power). ehabkost: that's probably another use case for query-device-slots: knowing that power machines can't device_add threads, but just cores. Tricky part: providing slot information at qemu-probing time (with -machine none), before any machine was built. ---------------- Some discussion about QEMU capability caching, why we need it, what we need to get rid of it (answer: early QMP?). Markus: would --query-FOO for some FOOs help? libvirt developers seem to prefer early QMP. ---------------- Now that libvirt queries QEMU for host CPUID capabilities, there are more interesting ways the cache may need to be invalidated. Markus: “there's a it of a combinatorial explosion, i.e. query results can depend on so much more than just the QEMU version and machine type. Makes invalidating the cache tricky, and possibly even finding something useful in it insufficiently likely to be worth the trouble.” ---------------- Some discussion about libvirt/qemu versions compatibility. Markus proposes that we could choose to make newer QEMU require a more recent libvirt version. OpenStack Nova re-evaluates every 6 months what next minimum libvirt and QEMU versions to use (via constants: NEXT_MIN_LIBVIRT_VERSION, NEXT_MIN_QEMU_VERSION) Markus: “The version decoupling we have is really nice, but I think an occasional dependency would be tolerable. Consider it when it saves lots of work.” Markus: “Also: drop support for old versions of QEMU. Anything before 1.5?” ---------------- References and notes: [1] References for “Early QMP”: * Igor’s proposal at: Subject: [Qemu-devel] [RFC 0/6] enable numa configuration before machine_init() from HMP/QMP https://www.mail-archive.com/qemu-devel@nongnu.org/msg488601.html * And what Eduardo has suggested at: http://www.linux-kvm.org/images/4/46/03x06A-Eduardo_HabkostMachine-type_Int… Comments from Markus: > Related: "Why /two/ config interfaces?" in my KVM Forum talk. The two > being CLI and QMP. > > We've long held the idea to support starting QEMU with a minimal command > line, then do all configuration via QMP. Not possible today, because > lots of things can only be done via CLI. In part because QMP becomes > available only after quite a few config boats have sailed. > > Making QMP available earlier during startup ("early QMP") would be a > stepping stone towards the "only QMP" goal. > > Related: rebasing CLI onto QAPI would make replicating CLI-only > interfaces in QMP easier. -- Eduardo

1 0

[libvirt] [PATCH] apparmor: add network netlink raw rule
by Cédric Bosdonnat 09 Nov '17

09 Nov '17

The rule 'network netlink raw' fixes these denials on libvirtd start: apparmor="DENIED" operation="create" profile="/usr/sbin/libvirtd" pid=12969 comm="libvirtd" family="netlink" sock_type="raw" protocol=0 requested_mask="create" denied_mask="create" --- examples/apparmor/usr.sbin.libvirtd | 1 + 1 file changed, 1 insertion(+) diff --git a/examples/apparmor/usr.sbin.libvirtd b/examples/apparmor/usr.sbin.libvirtd index 819068ffc..8ac5233cc 100644 --- a/examples/apparmor/usr.sbin.libvirtd +++ b/examples/apparmor/usr.sbin.libvirtd @@ -36,6 +36,7 @@ network inet6 dgram, network packet dgram, network packet raw, + network netlink raw, ptrace (trace) peer=unconfined, ptrace (trace) peer=/usr/sbin/libvirtd, -- 2.14.3

3 2

[libvirt] [PATCH v3 00/11] qemu: replace nested job with interruptible async job state
by Nikolay Shirokovskiy 09 Nov '17

09 Nov '17

Hello, everyone. This is successor to [1] "[PATCH v2 RFC 0/4] qemu: replace nested job with interruptible async job state". I tried to expand on series rationale, to make better splitting and to add comments. ABSTRACT Let's use interruptible async job state instead of nested jobs to control concurrent execution of async and regular jobs. RATIONALE I'm not quite happy with current implementation of concurrent execution async and regular jobs because it is not straigthforward. What is current implementation of jobs concurrency? While async job is running any compatible regular job is allowed to run concurrently. Practically it means that as soon as async job drops the domain lock another rpc thread can run in, find domain in the domain list and start a regular job. This is what happens when for example migration drops the lock before waiting for migration completion event or before sending request to destination side. But if we are going to send command to qemu monitor and accordingly drop domain lock this concurrency is not desired because qemu monitor can not handle simultaneus commands now. (This is probably not the only reason. I guess we don't want to think about what can happen if regular job will run concurrenly for example to migration at any place. Ability to run concurrently during disk mirror or state transfer which take the most time of migration looks enough). Thus nested jobs were introduced. Now if async job wants to send command to monitor it start special nested jobs beforehand thus effectively block concurrent regular jobs. To me it sounds like hacky way to implement the fact that async job allows concurrency only in limited places. Let's call them interruptible. So I suggest introduce interruptible state for async job. Before we are going to wait for migration completion for example we enter this state allowing regular jobs run concurrently. Later when we meet migration completion condition we wait for concurrent regular job to finish if any before proceed. This final step is required in this approach by definition - when async job is not interruptible no regular jobs can run concurrenly. On practice this protects us from sending qemu monitor command from async job while concurrent regular job is not finished its qemu monitor communication yet. Sounds like this final wait can have consequenses on migration. Like what if finishing concurrent job take too much time. In this case we get increased migration downtime. Well we will get this increase with current appoach too. Because right after migration completion condition is met we fetch migration stats and wait in enter monitor function anyway until concurrent job is finished. This also demontrates that current appoach isolates async and concurrent regular job loosely. It is possible that regular job starts then async job resumes and continues its execution until it requires qemu monitor and only then regular job finishes its execution. I guess it will be easy to reason about concurrency if when concurrent regular is started async job can not continue until regular job is finished. Check patch 4 for a demonstration that job isolation is good. TESTS To test concurrency I ran about 100 times a migration of non-empty domain with shared disk in a run with concurrent domstats for this domain in a tight loop. Looks good. MISSING PARTS - Enter/exit interruptible state at the end/begin of migration phases. Until this point is done we have a bit of degradation because domain won't response for example to queries between migration phases. - Update THREADS.txt I'm going to implement missing parts as soon as RFC has overall approvement. TODO 1 Drop driver argument from calls to qemuDomainObjEnterMonitor 2 Replace qemuDomainObjEnterMonitorAsync with qemuDomainObjEnterMonitor 3 Make qemuDomainObjExitMonitor void and drop its driver argument. [1] https://www.redhat.com/archives/libvir-list/2017-May/msg00018.html Nikolay Shirokovskiy (11): qemu: introduce routines for interruptible state qemu: check interruptible state instead of taking nested job qemu: enter interruptlible state where appropriate qemu: remove nested job usage from qemuProcessStop qemu: remove unused qemuDomainObjBeginNestedJob qemu: remove QEMU_JOB_ASYNC_NESTED qemu: remove liveness check from qemuDomainObjExitMonitor qemu: drop qemuDomainObjEnterMonitorInternal qemu: drop qemuDomainObjExitMonitorInternal conf: cleanup: drop virDomainObjWait qemu: rename qemuDomainNestedJobAllowed src/conf/domain_conf.c | 19 ---- src/conf/domain_conf.h | 1 - src/libvirt_private.syms | 1 - src/qemu/qemu_domain.c | 265 ++++++++++++++++++++++++++++++---------------- src/qemu/qemu_domain.h | 17 ++- src/qemu/qemu_driver.c | 2 +- src/qemu/qemu_migration.c | 18 ++-- src/qemu/qemu_process.c | 19 +--- 8 files changed, 195 insertions(+), 147 deletions(-) -- 1.8.3.1

1 11

[libvirt] [PATCH] virconf: properly set the end of content
by Jim Fehlig 09 Nov '17

09 Nov '17

There was a recent report of the xen-xl converter not handling config files missing an ending newline https://www.redhat.com/archives/libvir-list/2017-October/msg01353.html Commit 3cc2a9e0 fixed a similar problem when parsing content of a file but missed parsing in-memory content. But AFAICT, the better fix is to properly set the end of the content when initializing the virConfParserCtxt in virConfParse(). This commit reverts the part of 3cc2a9e0 that appends a newline to files missing it, and fixes setting the end of content when initializing virConfParserCtxt. A test is also added to check parsing in-memory content missing an ending newline. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/util/virconf.c | 13 ++-------- tests/virconftest.c | 69 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 71 insertions(+), 11 deletions(-) diff --git a/src/util/virconf.c b/src/util/virconf.c index 39c2bd917..a82a509ca 100644 --- a/src/util/virconf.c +++ b/src/util/virconf.c @@ -705,7 +705,7 @@ virConfParse(const char *filename, const char *content, int len, ctxt.filename = filename; ctxt.base = ctxt.cur = content; - ctxt.end = content + len - 1; + ctxt.end = content + len; ctxt.line = 1; ctxt.conf = virConfCreate(filename, flags); @@ -745,7 +745,7 @@ virConfReadFile(const char *filename, unsigned int flags) { char *content; int len; - virConfPtr conf = NULL; + virConfPtr conf; VIR_DEBUG("filename=%s", NULLSTR(filename)); @@ -757,17 +757,8 @@ virConfReadFile(const char *filename, unsigned int flags) if ((len = virFileReadAll(filename, MAX_CONFIG_FILE_SIZE, &content)) < 0) return NULL; - if (len && len < MAX_CONFIG_FILE_SIZE && content[len - 1] != '\n') { - VIR_DEBUG("appending newline to busted config file %s", filename); - if (VIR_REALLOC_N(content, len + 2) < 0) - goto cleanup; - content[len++] = '\n'; - content[len] = '\0'; - } - conf = virConfParse(filename, content, len, flags); - cleanup: VIR_FREE(content); return conf; diff --git a/tests/virconftest.c b/tests/virconftest.c index a8b18bae0..3cf0df3ac 100644 --- a/tests/virconftest.c +++ b/tests/virconftest.c @@ -77,6 +77,72 @@ static int testConfRoundTrip(const void *opaque) } +static int testConfMemoryNoNewline(const void *opaque ATTRIBUTE_UNUSED) +{ + const char *srcdata = \ + "ullong = '123456789'\n" \ + "string = 'foo'\n" \ + "uint = 12345"; + + virConfPtr conf = virConfReadString(srcdata, 0); + int ret = -1; + virConfValuePtr val; + unsigned long long llvalue; + char *str = NULL; + int uintvalue; + + if (!conf) + return -1; + + if (!(val = virConfGetValue(conf, "ullong"))) + goto cleanup; + + if (val->type != VIR_CONF_STRING) + goto cleanup; + + if (virStrToLong_ull(val->str, NULL, 10, &llvalue) < 0) + goto cleanup; + + if (llvalue != 123456789) { + fprintf(stderr, "Expected '123' got '%llu'\n", llvalue); + goto cleanup; + } + + if (virConfGetValueType(conf, "string") != + VIR_CONF_STRING) { + fprintf(stderr, "expected a string for 'string'\n"); + goto cleanup; + } + + if (virConfGetValueString(conf, "string", &str) < 0) + goto cleanup; + + if (STRNEQ_NULLABLE(str, "foo")) { + fprintf(stderr, "Expected 'foo' got '%s'\n", str); + goto cleanup; + } + + if (virConfGetValueType(conf, "uint") != VIR_CONF_ULLONG) { + fprintf(stderr, "expected an unsigned long for 'uint'\n"); + goto cleanup; + } + + if (virConfGetValueInt(conf, "uint", &uintvalue) < 0) + goto cleanup; + + if (uintvalue != 12345) { + fprintf(stderr, "Expected 12345 got %ud\n", uintvalue); + goto cleanup; + } + + ret = 0; + cleanup: + VIR_FREE(str); + virConfFree(conf); + return ret; +} + + static int testConfParseInt(const void *opaque ATTRIBUTE_UNUSED) { const char *srcdata = \ @@ -414,6 +480,9 @@ mymain(void) if (virTestRun("no-newline", testConfRoundTrip, "no-newline") < 0) ret = -1; + if (virTestRun("memory-no-newline", testConfMemoryNoNewline, NULL) < 0) + ret = -1; + if (virTestRun("int", testConfParseInt, NULL) < 0) ret = -1; -- 2.14.2

2 2

[libvirt] [PATCH] virconf: properly set the end of content
by Jim Fehlig 08 Nov '17

08 Nov '17

There was a recent report of the xen-xl converter not handling config files missing an ending newline https://www.redhat.com/archives/libvir-list/2017-October/msg01353.html Commit 3cc2a9e0 fixed a similar problem when parsing content of a file but missed parsing in-memory content. But AFAICT, the better fix is to properly set the end of the content when initializing the virConfParserCtxt in virConfParse(). This commit reverts the part of 3cc2a9e0 that appends a newline to files missing it, and fixes setting the end of content when initializing virConfParserCtxt. Signed-off-by: Jim Fehlig <jfehlig(a)suse.com> --- src/util/virconf.c | 13 ++----------- 1 file changed, 2 insertions(+), 11 deletions(-) diff --git a/src/util/virconf.c b/src/util/virconf.c index 39c2bd917..a82a509ca 100644 --- a/src/util/virconf.c +++ b/src/util/virconf.c @@ -705,7 +705,7 @@ virConfParse(const char *filename, const char *content, int len, ctxt.filename = filename; ctxt.base = ctxt.cur = content; - ctxt.end = content + len - 1; + ctxt.end = content + len; ctxt.line = 1; ctxt.conf = virConfCreate(filename, flags); @@ -745,7 +745,7 @@ virConfReadFile(const char *filename, unsigned int flags) { char *content; int len; - virConfPtr conf = NULL; + virConfPtr conf; VIR_DEBUG("filename=%s", NULLSTR(filename)); @@ -757,17 +757,8 @@ virConfReadFile(const char *filename, unsigned int flags) if ((len = virFileReadAll(filename, MAX_CONFIG_FILE_SIZE, &content)) < 0) return NULL; - if (len && len < MAX_CONFIG_FILE_SIZE && content[len - 1] != '\n') { - VIR_DEBUG("appending newline to busted config file %s", filename); - if (VIR_REALLOC_N(content, len + 2) < 0) - goto cleanup; - content[len++] = '\n'; - content[len] = '\0'; - } - conf = virConfParse(filename, content, len, flags); - cleanup: VIR_FREE(content); return conf; -- 2.14.2

2 2

[libvirt] [PATCH v3 0/5] nwfilter common object adjustments
by John Ferlan 08 Nov '17

08 Nov '17

v2: https://www.redhat.com/archives/libvir-list/2017-July/msg00673.html (and a few pings along the way) Don't think much survived from v2 - this is a fresh start anyway. Perhaps old patch 2 the same, but beyond that a different approach to remove recursive read/write locks and replace with using rwlock read/write where the write's are in very tight confines. I've run the changes through avocado with success. There were some really strange deadlocks along the way - even causing libvirtd to go defunct. There's a lot of strange ways to use/access the nwfilters. John Ferlan (5): nwfilter: Add update locking to Initialization nwfilter: Remove unnecessary UUID comparison bypass nwfilter: Convert _virNWFilterObj to use virObjectRWLockable nwfilter: Convert _virNWFilterObjList to use virObjectRWLockable nwfilter: Remove need for nwfilterDriverLock in some API's src/conf/virnwfilterobj.c | 555 +++++++++++++++++++++++---------- src/conf/virnwfilterobj.h | 11 +- src/libvirt_private.syms | 3 +- src/nwfilter/nwfilter_driver.c | 77 +++-- src/nwfilter/nwfilter_gentech_driver.c | 11 +- 5 files changed, 433 insertions(+), 224 deletions(-) -- 2.13.6

1 6

[libvirt] [PATCH 0/4] Couple more hotplug cleanups
by John Ferlan 08 Nov '17

08 Nov '17

Patches speak for themselves - they also resolve a couple of new Coverity whines. John Ferlan (4): conf,qemu: Check for NULL addrs in virDomainUSBAddressRelease conf,qemu: Check for NULL addrs in virDomainUSBAddressEnsure qemu: Remove unnecessary virtio disk detach info.alias check qemu: Remove unnecessary controller detach info.alias check src/conf/domain_addr.c | 5 ++++- src/conf/domain_addr.h | 4 ++-- src/qemu/qemu_domain_address.c | 7 ++----- src/qemu/qemu_hotplug.c | 43 ++++++++++++++---------------------------- 4 files changed, 22 insertions(+), 37 deletions(-) -- 2.13.6

2 7

Re: [libvirt] [PATCH 0/7] x86: Rework KVM-defaults compat code, enable kvm_pv_unhalt by default
by Eduardo Habkost 08 Nov '17

08 Nov '17

(CCing libvir-list) On Mon, Oct 09, 2017 at 05:47:44PM +0200, Paolo Bonzini wrote: > On 09/10/2017 17:15, Waiman Long wrote: > > On 10/09/2017 09:39 AM, Paolo Bonzini wrote: > >> On 06/10/2017 23:52, Eduardo Habkost wrote: > >>> This series enables kvm_pv_unhalt by default on pc-*-2.11 and > >>> newer. > >> > >> I've discussed PV spinlocks with some folks at Microsoft for a few weeks > >> now, and I'm not 100% sure about enabling kvm_pv_unhalt by default. > >> It's probably a good idea overall, but it does come with some caveats. > >> > >> The problem is that there were two different implementations of fair > >> spinlocks in Linux, ticket spinlocks and queued spinlocks. When > >> kvm_pv_unhalt is disabled, ticket spinlocks sucked badly indeed; queued > >> spinlocks however simply revert to unfair spinlocks, which loses the > >> fairness but has the best performance. See virt_spin_lock in > >> arch/x86/include/asm/qspinlock.h. > >> > >> Now, fair spinlocks are only really needed for large NUMA machines. > >> With a single NUMA node, for example, test-and-set spinlocks work well > >> enough; there's not _much_ need for fairness in practice, and the > >> paravirtualization does introduce some overhead. > >> > >> Therefore, the best performance would be achieved with kvm_pv_unhalt > >> disabled on small VMs, and enabled on large VMs spanning multiple host > >> NUMA nodes. > >> > >> Waiman, Davidlohr, do you have an opinion on this as well? > > > > I agree. Using unfair lock in a small VM is good for performance. The > > only problem I see is how do we define what is small. Now, even a > > machine with a single socket can have up to 28 cores, 56 threads. If a > > VM has up to 56 vCPUs, we may still want pvqspinlock to be used. > > Yes. Another possibility is to enable it when there is >1 NUMA node in > the guest. We generally don't do this kind of magic but higher layers > (oVirt/OpenStack) do. Can't the guest make this decision, instead of the host? > > It also depends on how worse performance is with PV qspinlocks, > especially since now there's also vcpu_is_preempted that has to be > thrown in the mix. A lot more testing is needed. > This is an interesting case for "-cpu host"/host-passthrough: usage of "-cpu host" has an implicit assumption that no feature will decrease performance just because we told the guest that it is available. Can't the guest choose the most reasonable option based on the information we provide to it, instead of having to hide information from the guest to get better performance? > > Is the kvm_pv_unhalt flag a global one for all VMs within a machine? Or > > can it be different for each VM? We may want to have this flag > > dynamically determined depending on the configuration of the VM. > > It's per-VM, so that's easy. -- Eduardo

3 8

[libvirt] [PATCH] qemu-ns: Detect /dev/* mount point duplicates even better
by Michal Privoznik 08 Nov '17

08 Nov '17

In 4f1570720218302 I've tried to make duplicates detection for nested /dev mount better. However, I've missed the obvious case when there are two same mount points. For instance if: # mount --bind /dev/blah /dev/blah # mount --bind /dev/blah /dev/blah Yeah, very unlikely but possible. Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com> --- src/qemu/qemu_domain.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c index 702463547..61d28337b 100644 --- a/src/qemu/qemu_domain.c +++ b/src/qemu/qemu_domain.c @@ -8352,7 +8352,7 @@ qemuDomainGetPreservedMounts(virQEMUDriverConfigPtr cfg, while (j < nmounts) { char *c = STRSKIP(mounts[j], mounts[i]); - if (c && *c == '/') { + if (c && (*c == '/' || *c == '\0')) { VIR_DEBUG("Dropping path %s because of %s", mounts[j], mounts[i]); VIR_DELETE_ELEMENT(mounts, j, nmounts); } else { -- 2.13.6

3 2

[libvirt] [PATCH] conf: Fix type for @liveStatus in virDomainObjListLoadAllConfigs
by Peter Krempa 08 Nov '17

08 Nov '17

Use bool instead of an int. --- src/bhyve/bhyve_driver.c | 4 ++-- src/conf/virdomainobjlist.c | 2 +- src/conf/virdomainobjlist.h | 2 +- src/libxl/libxl_driver.c | 6 +++--- src/lxc/lxc_driver.c | 6 +++--- src/qemu/qemu_driver.c | 6 +++--- src/uml/uml_driver.c | 4 ++-- 7 files changed, 15 insertions(+), 15 deletions(-) diff --git a/src/bhyve/bhyve_driver.c b/src/bhyve/bhyve_driver.c index 5c432b25e..dd6e8abc6 100644 --- a/src/bhyve/bhyve_driver.c +++ b/src/bhyve/bhyve_driver.c @@ -1294,7 +1294,7 @@ bhyveStateInitialize(bool privileged, if (virDomainObjListLoadAllConfigs(bhyve_driver->domains, BHYVE_STATE_DIR, - NULL, 1, + NULL, true, bhyve_driver->caps, bhyve_driver->xmlopt, NULL, NULL) < 0) @@ -1302,7 +1302,7 @@ bhyveStateInitialize(bool privileged, if (virDomainObjListLoadAllConfigs(bhyve_driver->domains, BHYVE_CONFIG_DIR, - BHYVE_AUTOSTART_DIR, 0, + BHYVE_AUTOSTART_DIR, false, bhyve_driver->caps, bhyve_driver->xmlopt, NULL, NULL) < 0) diff --git a/src/conf/virdomainobjlist.c b/src/conf/virdomainobjlist.c index b9f78c572..87a742b1e 100644 --- a/src/conf/virdomainobjlist.c +++ b/src/conf/virdomainobjlist.c @@ -559,7 +559,7 @@ int virDomainObjListLoadAllConfigs(virDomainObjListPtr doms, const char *configDir, const char *autostartDir, - int liveStatus, + bool liveStatus, virCapsPtr caps, virDomainXMLOptionPtr xmlopt, virDomainLoadConfigNotify notify, diff --git a/src/conf/virdomainobjlist.h b/src/conf/virdomainobjlist.h index 60220ca0d..bb186bde3 100644 --- a/src/conf/virdomainobjlist.h +++ b/src/conf/virdomainobjlist.h @@ -72,7 +72,7 @@ void virDomainObjListRemoveLocked(virDomainObjListPtr doms, int virDomainObjListLoadAllConfigs(virDomainObjListPtr doms, const char *configDir, const char *autostartDir, - int liveStatus, + bool liveStatus, virCapsPtr caps, virDomainXMLOptionPtr xmlopt, virDomainLoadConfigNotify notify, diff --git a/src/libxl/libxl_driver.c b/src/libxl/libxl_driver.c index 951937f14..40328a6cb 100644 --- a/src/libxl/libxl_driver.c +++ b/src/libxl/libxl_driver.c @@ -763,7 +763,7 @@ libxlStateInitialize(bool privileged, if (virDomainObjListLoadAllConfigs(libxl_driver->domains, cfg->stateDir, cfg->autostartDir, - 1, + true, cfg->caps, libxl_driver->xmlopt, NULL, NULL) < 0) @@ -775,7 +775,7 @@ libxlStateInitialize(bool privileged, if (virDomainObjListLoadAllConfigs(libxl_driver->domains, cfg->configDir, cfg->autostartDir, - 0, + false, cfg->caps, libxl_driver->xmlopt, NULL, NULL) < 0) @@ -815,7 +815,7 @@ libxlStateReload(void) virDomainObjListLoadAllConfigs(libxl_driver->domains, cfg->configDir, cfg->autostartDir, - 1, + true, cfg->caps, libxl_driver->xmlopt, NULL, libxl_driver); diff --git a/src/lxc/lxc_driver.c b/src/lxc/lxc_driver.c index 0069e5e92..b3447100f 100644 --- a/src/lxc/lxc_driver.c +++ b/src/lxc/lxc_driver.c @@ -1679,7 +1679,7 @@ static int lxcStateInitialize(bool privileged, /* Get all the running persistent or transient configs first */ if (virDomainObjListLoadAllConfigs(lxc_driver->domains, cfg->stateDir, - NULL, 1, + NULL, true, caps, lxc_driver->xmlopt, NULL, NULL) < 0) @@ -1690,7 +1690,7 @@ static int lxcStateInitialize(bool privileged, /* Then inactive persistent configs */ if (virDomainObjListLoadAllConfigs(lxc_driver->domains, cfg->configDir, - cfg->autostartDir, 0, + cfg->autostartDir, false, caps, lxc_driver->xmlopt, NULL, NULL) < 0) @@ -1755,7 +1755,7 @@ lxcStateReload(void) virDomainObjListLoadAllConfigs(lxc_driver->domains, cfg->configDir, - cfg->autostartDir, 0, + cfg->autostartDir, false, caps, lxc_driver->xmlopt, lxcNotifyLoadDomain, lxc_driver); diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c index 47f85b9bf..6132bc4a9 100644 --- a/src/qemu/qemu_driver.c +++ b/src/qemu/qemu_driver.c @@ -911,7 +911,7 @@ qemuStateInitialize(bool privileged, /* Get all the running persistent or transient configs first */ if (virDomainObjListLoadAllConfigs(qemu_driver->domains, cfg->stateDir, - NULL, 1, + NULL, true, qemu_driver->caps, qemu_driver->xmlopt, NULL, NULL) < 0) @@ -933,7 +933,7 @@ qemuStateInitialize(bool privileged, /* Then inactive persistent configs */ if (virDomainObjListLoadAllConfigs(qemu_driver->domains, cfg->configDir, - cfg->autostartDir, 0, + cfg->autostartDir, false, qemu_driver->caps, qemu_driver->xmlopt, NULL, NULL) < 0) @@ -1015,7 +1015,7 @@ qemuStateReload(void) cfg = virQEMUDriverGetConfig(qemu_driver); virDomainObjListLoadAllConfigs(qemu_driver->domains, cfg->configDir, - cfg->autostartDir, 0, + cfg->autostartDir, false, caps, qemu_driver->xmlopt, qemuNotifyLoadDomain, qemu_driver); cleanup: diff --git a/src/uml/uml_driver.c b/src/uml/uml_driver.c index 1846835cc..ab7fa7f27 100644 --- a/src/uml/uml_driver.c +++ b/src/uml/uml_driver.c @@ -566,7 +566,7 @@ umlStateInitialize(bool privileged, if (virDomainObjListLoadAllConfigs(uml_driver->domains, uml_driver->configDir, - uml_driver->autostartDir, 0, + uml_driver->autostartDir, false, uml_driver->caps, uml_driver->xmlopt, NULL, NULL) < 0) @@ -634,7 +634,7 @@ umlStateReload(void) umlDriverLock(uml_driver); virDomainObjListLoadAllConfigs(uml_driver->domains, uml_driver->configDir, - uml_driver->autostartDir, 0, + uml_driver->autostartDir, false, uml_driver->caps, uml_driver->xmlopt, umlNotifyLoadDomain, uml_driver); -- 2.14.3

2 1

[libvirt] [PATCH 00/12] qemu: Fully populate JSON formatters for the protocol layer (blockdev-add saga)
by Peter Krempa 08 Nov '17

08 Nov '17

This adds the missing formatters for JSON properties for the storage. John Ferlan (1): qemu: block: Add JSON props generator for iSCSI protocol This patch was stolen from the iSCSI saga and fixed, since the formatter did not format the port number into the portal string. Peter Krempa (11): qemu: block: Use proper type for servers for VxHS disks qemu: process: Split out useful parts from qemuBuildNetworkDriveURI storage: Don't store leading '/' in image name when splitting out volume storage: Store RBD image name as pool and image name qemu: block: Add JSON props generator for 'curl' based storage backends qemu: block: Add JSON props generator for NBD storage backing qemu: block: Add JSON props generator for RBD storage backing qemu: block: Add JSON props generator for sheepdog storage backing qemu: block: Add JSON props generator for ssh storage backing qemu: block: Add node-names to JSON backing storage strings tests: Add testing of storage backend JSON props formatter src/conf/domain_conf.c | 17 +- src/libxl/libxl_conf.c | 2 +- src/qemu/qemu_block.c | 403 ++++++++++++++++++++- src/qemu/qemu_block.h | 4 + src/qemu/qemu_command.c | 40 +- src/util/virstoragefile.c | 26 +- src/xenconfig/xen_xl.c | 2 +- tests/Makefile.am | 14 +- tests/domainsnapshotxml2xmlin/disk_snapshot.xml | 2 +- tests/domainsnapshotxml2xmlout/disk_snapshot.xml | 2 +- tests/qemublocktest.c | 189 ++++++++++ .../qemuxml2argv-disk-drive-network-gluster.args | 2 +- ...muxml2argv-disk-drive-network-tlsx509-vxhs.args | 8 +- .../qemuxml2argv-disk-drive-network-vxhs.args | 4 +- tests/virstoragetest.c | 3 +- 15 files changed, 640 insertions(+), 78 deletions(-) create mode 100644 tests/qemublocktest.c -- 2.14.3

3 23

[libvirt] [PATCH] util: storage: Fix parsing of IPv6 portal address for iSCSI
by Peter Krempa 08 Nov '17

08 Nov '17

Split on the last colon and avoid parsing port if the split remainder contains the closing square bracket, so that IPv6 addresses are interpreted correctly. --- src/util/virstoragefile.c | 3 ++- tests/virstoragetest.c | 20 ++++++++++++++++++++ 2 files changed, 22 insertions(+), 1 deletion(-) diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c index 3a2d2aa05..e7fcb1238 100644 --- a/src/util/virstoragefile.c +++ b/src/util/virstoragefile.c @@ -3008,7 +3008,8 @@ virStorageSourceParseBackingJSONiSCSI(virStorageSourcePtr src, if (VIR_STRDUP(src->hosts->name, portal) < 0) goto cleanup; - if ((port = strchr(src->hosts->name, ':'))) { + if ((port = strrchr(src->hosts->name, ':')) && + !strchr(port, ']')) { if (virStringParsePort(port + 1, &src->hosts->port) < 0) goto cleanup; diff --git a/tests/virstoragetest.c b/tests/virstoragetest.c index cfcd8a79c..52a685d91 100644 --- a/tests/virstoragetest.c +++ b/tests/virstoragetest.c @@ -1578,6 +1578,26 @@ mymain(void) "<source protocol='iscsi' name='iqn.2016-12.com.virttest:emulated-iscsi-noauth.target/6'>\n" " <host name='test.org' port='1234'/>\n" "</source>\n"); + TEST_BACKING_PARSE("json:{\"file\":{\"driver\":\"iscsi\"," + "\"transport\":\"tcp\"," + "\"portal\":\"[2001::0]:1234\"," + "\"target\":\"iqn.2016-12.com.virttest:emulated-iscsi-noauth.target\"," + "\"lun\":6" + "}" + "}", + "<source protocol='iscsi' name='iqn.2016-12.com.virttest:emulated-iscsi-noauth.target/6'>\n" + " <host name='[2001::0]' port='1234'/>\n" + "</source>\n"); + TEST_BACKING_PARSE("json:{\"file\":{\"driver\":\"iscsi\"," + "\"transport\":\"tcp\"," + "\"portal\":\"[2001::0]\"," + "\"target\":\"iqn.2016-12.com.virttest:emulated-iscsi-noauth.target\"," + "\"lun\":6" + "}" + "}", + "<source protocol='iscsi' name='iqn.2016-12.com.virttest:emulated-iscsi-noauth.target/6'>\n" + " <host name='[2001::0]' port='3260'/>\n" + "</source>\n"); TEST_BACKING_PARSE("json:{\"file\":{\"driver\":\"sheepdog\"," "\"vdi\":\"test\"," "\"server\":{ \"type\":\"inet\"," -- 2.14.3

2 1

[libvirt] [PATCH go-xml] Add support for host paravirt bootloader.
by Brandon Bergren 08 Nov '17

08 Nov '17

--- domain.go | 2 ++ 1 file changed, 2 insertions(+) diff --git a/domain.go b/domain.go index 2b749b2..f4ef35c 100644 --- a/domain.go +++ b/domain.go @@ -791,6 +791,8 @@ type Domain struct { VCPU *DomainVCPU `xml:"vcpu"` VCPUs *DomainVCPUs `xml:"vcpus"` CPUTune *DomainCPUTune `xml:"cputune"` + Bootloader string `xml:"bootloader,omitempty"` + BootloaderArgs string `xml:"bootloader_args,omitempty"` Resource *DomainResource `xml:"resource"` SysInfo *DomainSysInfo `xml:"sysinfo"` OS *DomainOS `xml:"os"` -- 2.14.1

2 3

[libvirt] [PATCH 0/4] tests: Add caps for QEMU 2.10.0 on non-x86
by Andrea Bolognani 08 Nov '17

08 Nov '17

Patches 1, 2 and 4 are too big for the list, so I'm sending a [REDACTED] version. The full series is available from https://github.com/andreabolognani/libvirt/ in the "caps" branch. Andrea Bolognani (4): tests: Add caps for QEMU 2.10.0 on aarch64 (GICv2) tests: Add caps for QEMU 2.10.0 on aarch64 (GICv3) tests: Rename ppc64le caps to ppc64 tests: Add caps for QEMU 2.10.0 on ppc64 ...qemu_2.6.0.ppc64le.xml => qemu_2.6.0.ppc64.xml} | 2 +- tests/domaincapstest.c | 2 +- ...e.replies => caps_2.10.0-gicv2.aarch64.replies} | 23145 ++++++++----------- .../caps_2.10.0-gicv2.aarch64.xml | 303 + ...e.replies => caps_2.10.0-gicv3.aarch64.replies} | 23145 ++++++++----------- .../caps_2.10.0-gicv3.aarch64.xml | 303 + ...0.ppc64le.replies => caps_2.10.0.ppc64.replies} | 4714 ++-- ...aps_2.9.0.ppc64le.xml => caps_2.10.0.ppc64.xml} | 31 +- ....0.ppc64le.replies => caps_2.6.0.ppc64.replies} | 0 ...caps_2.6.0.ppc64le.xml => caps_2.6.0.ppc64.xml} | 0 ....0.ppc64le.replies => caps_2.9.0.ppc64.replies} | 0 ...caps_2.9.0.ppc64le.xml => caps_2.9.0.ppc64.xml} | 0 tests/qemucapabilitiestest.c | 7 +- 13 files changed, 24043 insertions(+), 27609 deletions(-) rename tests/domaincapsschemadata/{qemu_2.6.0.ppc64le.xml => qemu_2.6.0.ppc64.xml} (98%) copy tests/qemucapabilitiesdata/{caps_2.9.0.ppc64le.replies => caps_2.10.0-gicv2.aarch64.replies} (69%) create mode 100644 tests/qemucapabilitiesdata/caps_2.10.0-gicv2.aarch64.xml copy tests/qemucapabilitiesdata/{caps_2.9.0.ppc64le.replies => caps_2.10.0-gicv3.aarch64.replies} (69%) create mode 100644 tests/qemucapabilitiesdata/caps_2.10.0-gicv3.aarch64.xml copy tests/qemucapabilitiesdata/{caps_2.9.0.ppc64le.replies => caps_2.10.0.ppc64.replies} (83%) copy tests/qemucapabilitiesdata/{caps_2.9.0.ppc64le.xml => caps_2.10.0.ppc64.xml} (99%) rename tests/qemucapabilitiesdata/{caps_2.6.0.ppc64le.replies => caps_2.6.0.ppc64.replies} (100%) rename tests/qemucapabilitiesdata/{caps_2.6.0.ppc64le.xml => caps_2.6.0.ppc64.xml} (100%) rename tests/qemucapabilitiesdata/{caps_2.9.0.ppc64le.replies => caps_2.9.0.ppc64.replies} (100%) rename tests/qemucapabilitiesdata/{caps_2.9.0.ppc64le.xml => caps_2.9.0.ppc64.xml} (100%) -- 2.13.6

2 9

[libvirt] [PATCH v3 0/5] Predictable file names for memory-backend-file
by Michal Privoznik 08 Nov '17

08 Nov '17

v3 of: https://www.redhat.com/archives/libvir-list/2017-October/msg01091.html diff to v2: - Pushed 1/4 and 2/4 from the original series - Split 3/4 into smaller patches Michal Privoznik (5): qemu: Set alias for memory cell in qemuBuildMemoryCellBackendStr qemu: Rename qemuProcessBuildDestroyHugepagesPath qemu: Destroy whole memory tree qemu: Use predictable file names for memory-backend-file news: Document predictable file names for memory-backend-file docs/news.xml | 11 ++ src/qemu/qemu_command.c | 9 +- src/qemu/qemu_conf.c | 55 ++++++++- src/qemu/qemu_conf.h | 9 +- src/qemu/qemu_driver.c | 17 +++ src/qemu/qemu_hotplug.c | 2 +- src/qemu/qemu_process.c | 137 +++++++++++++++------ src/qemu/qemu_process.h | 8 +- .../qemuxml2argv-cpu-numa-memshared.args | 6 +- .../qemuxml2argv-fd-memory-numa-topology.args | 3 +- .../qemuxml2argv-fd-memory-numa-topology2.args | 6 +- .../qemuxml2argv-fd-memory-numa-topology3.args | 9 +- .../qemuxml2argv-hugepages-memaccess2.args | 9 +- 13 files changed, 218 insertions(+), 63 deletions(-) -- 2.13.6

4 14

[libvirt] Libvirt config converter can't handle file not ending with new line
by Wei Liu 07 Nov '17

07 Nov '17

Hi Jim I discover a problem when using xen_xl converter. When the file in question doesn't end with a new line, I get the following error: error: configuration file syntax error: memory conf:53: expecting a value After digging a bit (but haven't read libvirt code), it appears that the file didn't end with a new line. I have worked around this, but I think this is an issue that should be fixed. Thanks, Wei.

3 4