Laine Stump [Fri, 10 Dec 2010 18:54:48 +0000 (13:54 -0500)]
Convert dhcpStartDhcpDaemon from virRun to virCommand
This is pretty straightforward - even though dnsmasq gets daemonized
and uses a pid file, those things are both handled by the dnsmasq
binary itself. And libvirt doesn't need any of the output of the
dnsmasq command either, so we just setup the args and call
virRun(). Mainly it was just a (mostly) mechanical job of replacing
the APPEND_ARG() macro (and some other *printfs()) with
virCommandAddArg*().
Hu Tao [Fri, 10 Dec 2010 05:52:11 +0000 (13:52 +0800)]
Update documentation of watchdog dump option and add test data for it
The xml watchdog dump option is converted to qemu watchdog pause arg
but it is not reasonable to convert it back from qemu watchdog pause
arg since there already is a xml watchdog pause option, so a test for
the dump option to convert it from arg to xml is not added.
Josh Durgin [Tue, 7 Dec 2010 19:56:34 +0000 (11:56 -0800)]
qemu: Add RBD support and some network disk fixes
Changes common to all network disks:
-Make source name optional in the domain schema, since NBD doesn't use it
-Add a hostName type to the domain schema, and use it instead of genericName, which doesn't include .
-Don't leak host names or ports
-Set the source protocol in qemuParseCommandline
Jiri Denemark [Thu, 9 Dec 2010 10:18:32 +0000 (11:18 +0100)]
qemu: Distinguish between domain shutdown and crash
When we get an EOF event on monitor connection, it may be a result of
either crash or graceful shutdown. QEMU which supports async events
(i.e., we are talking to it using JSON monitor) emits SHUTDOWN event on
graceful shutdown. In case we don't get this event by the time monitor
connection is closed, we assume the associated domain crashed.
Ryan Harper [Wed, 8 Dec 2010 21:30:12 +0000 (14:30 -0700)]
qemu: call drive_del in DetachPciDiskDevice
Currently libvirt doesn't confirm whether the guest has responded to the
disk removal request. In some cases this can leave the guest with
continued access to the device while the mgmt layer believes that it has
been removed. With a recent qemu monitor command[1] we can
deterministically revoke a guests access to the disk (on the QEMU side)
to ensure no futher access is permitted.
This patch adds support for the drive_del() command and introduces it
in the disk removal paths. If the guest is running in a QEMU without this
command we currently explicitly check for unknown command/CommandNotFound
and log the issue.
If QEMU supports the command we issue the drive_del command after we attempt
to remove the device. The guest may respond and remove the block device
before we get to attempt to call drive_del. In that case, we explicitly check
for 'Device not found' from the monitor indicating that the target drive
was auto-deleted upon guest responds to the device removal notification.
Ryan Harper [Fri, 22 Oct 2010 14:14:22 +0000 (09:14 -0500)]
qemu: call drive_unplug in DetachPciDiskDevice
Currently libvirt doesn't confirm whether the guest has responded to the
disk removal request. In some cases this can leave the guest with
continued access to the device while the mgmt layer believes that it has
been removed. With a recent qemu monitor command[1] we can
deterministically revoke a guests access to the disk (on the QEMU side)
to ensure no futher access is permitted.
This patch adds support for the drive_unplug() command and introduces it
in the disk removal paths. There is some discussion to be had about how
to handle the case where the guest is running in a QEMU without this
command (and the fact that we currently don't have a way of detecting
what monitor commands are available).
Changes since v2:
- use VIR_ERROR to report when unplug command not found
Changes since v1:
- return > 0 when command isn't present, < 0 on command failure
- detect when drive_unplug command isn't present and log error
instead of failing entire command
Osier Yang [Sat, 4 Dec 2010 15:34:01 +0000 (23:34 +0800)]
qemu: Only build devstr when needs (attach PCI controller)
- qemudDomainAttachPciControllerDevice: Don't build "devstr"
if "-device" of qemu is not available, as "devstr" will only
be used by "qemuMonitorAddDevice", which depends on "-device"
argument of qemu is supported.
Jiri Denemark [Wed, 8 Dec 2010 13:45:00 +0000 (14:45 +0100)]
qemu: Use -vga none only if it is supported
Commit febc591683cf51e4551f8bcf3ce279a776056e1c introduced -vga none in
case no video card is included in domain XML. However, old qemu
versions do not support this and such domain cannot be successfully
started.
This fixes a misleading error message saying the libnl package
needs to be installed, when it's really the libnl-devel package
needing to be installed.
Eric Blake [Wed, 8 Dec 2010 00:31:14 +0000 (17:31 -0700)]
maint: update to latest gnulib
* .gnulib: Update to latest, for at least a stdint.h fix
* src/storage/storage_driver.c (storageVolumeZeroSparseFile)
(storageWipeExtent): Use better type, although it still triggers
spurious -Wformat warning on MacOS's gcc.
Eric Blake [Fri, 3 Dec 2010 00:29:34 +0000 (17:29 -0700)]
openvz: convert popen to virCommand
popen must be matched with pclose (not fclose), or it will leak
resources. Furthermore, it is a lousy interface when it comes to
signal handling. We're much better off using our decent command
wrapper. Note that virCommand guarantees that VIR_FREE(outbuf) is
both required and safe to call, whether virCommandRun succeeded or
failed.
* src/openvz/openvz_conf.c (openvzLoadDomains, openvzGetVEID):
Replace popen with virCommand usage.
Eric Blake [Fri, 3 Dec 2010 21:14:16 +0000 (14:14 -0700)]
command: improve behavior on no output
Guarantee that outbuf/errbuf are allocated on success, even if to the
empty string. Caller always has to free the result, and empty output
check requires checking if *outbuf=='\0'. Makes the API easier to use
safely. Failure is best effort allocation (some paths, like
out-of-memory, cannot allocate a buffer, but most do), so caller must
free buffer on failure.
* docs/internals/command.html.in: Update documentation.
* src/util/command.c (virCommandSetOutputBuffer)
(virCommandSetErrorBuffer, virCommandProcessIO) Guarantee empty
string on no output.
* tests/commandtest.c (test17): New test.
Eric Blake [Mon, 6 Dec 2010 23:36:34 +0000 (16:36 -0700)]
command: enforce fd vs. buffer considerations
* docs/internals/command.html.in: Better documentation of buffer
vs. fd considerations.
* src/util/command.c (virCommandRunAsync): Reject raw execution
with string io.
(virCommandRun): Reject execution with user-specified fds not
visiting a regular file.
Eric Blake [Fri, 3 Dec 2010 20:16:31 +0000 (13:16 -0700)]
uuid: require smbios uuid and domain uuid to match
* src/conf/domain_conf.c (virDomainDefParseXML): Prefer sysinfo
uuid over generating one, and if both uuids are present, require
them to be identical.
* src/qemu/qemu_conf.c (qemuBuildSmbiosSystemStr): Allow skipping
the uuid.
(qemudBuildCommandLine): Adjust caller; <smbios mode=host/> must
not use host uuid in place of guest uuid.
Eric Blake [Wed, 1 Dec 2010 23:50:59 +0000 (16:50 -0700)]
qemu: avoid adding "" in smbios arguments
The log lists things like -smbios type=1,vendor="Red Hat", which
is great for shell parsing, but not so great when you realize that
execve() then passes those literal "" on as part of the command
line argument, such that qemu sets SMBIOS with extra literal quotes.
The eventual addition of virCommand is needed before we have the API
to shell-quote a string representation of a command line, so that the
log can still be pasted into a shell, but without inserting extra
bytes into the execve() arguments.
* src/qemu/qemu_conf.c (qemuBuildSmbiosBiosStr)
(qemuBuildSmbiosSystemStr): Qemu doesn't like quotes around uuid
arguments, and the remaining quotes are passed literally to
smbios, making <smbios mode='host'/> inaccurate. Removing the
quotes makes the log harder to parse, but that can be fixed later
with virCommand improvements.
* tests/qemuxml2argvdata/qemuxml2argv-smbios.args: 'Fix' test; it
will need fixing again once virCommand learns how to shell-quote a
potential command line.
Laurent Léonard [Sun, 5 Dec 2010 17:53:26 +0000 (18:53 +0100)]
syntax error "Bad fd number" when stopping libvirt-guests
When libvirt-guests is being stopped, I get the following message:
$Running guests on default URI: test-vm
$Suspending guests on default URI...
$Suspending test-vm: /etc/init.d/libvirt-guests: 340: Syntax error: Bad fd
number
Eric Blake [Sat, 4 Dec 2010 21:33:23 +0000 (14:33 -0700)]
threads: add virThreadID for debugging use
* src/util/threads.h (virThreadID): New prototype.
* src/util/threads-pthread.c (virThreadID): New function.
* src/util/threads-win32.c (virThreadID): Likewise.
* src/libvirt_private.syms (threads.h): Export it.
* daemon/event.c (virEventInterruptLocked): Use it to avoid
warning on BSD systems.
Osier Yang [Mon, 6 Dec 2010 09:41:10 +0000 (17:41 +0800)]
util: Fix bug which will cause libvirtd crash
"virCommandRun": if "cmd->outbuf" or "cmd->errbuf" is NULL,
libvirtd will be crashed when trying to start a qemu domain
(which invokes "virCommandRun"), it caused by we try to use
"*cmd->outbuf" and "*cmd->errbuf" regardless of cmd->outbuf
or cmd->errbuf is NULL.
Jiri Denemark [Fri, 3 Dec 2010 09:48:31 +0000 (10:48 +0100)]
qemu: Fix a possible deadlock in p2p migration
Two more calls to remote libvirtd have to be surrounded by
qemuDomainObjEnterRemoteWithDriver() and
qemuDomainObjExitRemoteWithDriver() to prevent possible deadlock between
two communicating libvirt daemons.
Jiri Denemark [Fri, 3 Dec 2010 08:31:48 +0000 (09:31 +0100)]
Change return value of VIR_DRV_SUPPORTS_FEATURE to bool
virDrvSupportsFeature API is allowed to return -1 on error while all but
one uses of VIR_DRV_SUPPORTS_FEATURE only check for (non)zero return
value. Let's make this macro return zero on error, which is what
everyone expects anyway.
Osier Yang [Fri, 3 Dec 2010 07:34:23 +0000 (15:34 +0800)]
virsh: Remove redundant optional option for cmdHelp
Remove the optional option "group", as cmdHelp should accepts
only one option ("virsh help" supports both command and command
group now, and user rarely uses the options, so it doesn't matter
much for it being longer, :-)
Create file in virFileWriteStr() if it doesn't exist
This patch adds a mode_t parameter to virFileWriteStr().
If mode is different from 0, virFileWriteStr() will try
to create the file if it doesn't exist.
Osier Yang [Thu, 2 Dec 2010 00:25:17 +0000 (08:25 +0800)]
virsh: move two commands from domain group to storage pool group
* tools/virsh.c (find-storage-pool-sources-as and find-storage-pool-sources
should't be in command group "Domain Management", move them to group
"Storage Pool".
Eric Blake [Mon, 22 Nov 2010 23:09:13 +0000 (16:09 -0700)]
qemu: convert to virCommand
* src/qemu/qemu_conf.c (qemudExtractVersionInfo): Check for file
before executing it here, rather than in callers.
(qemudBuildCommandLine): Rewrite with virCommand.
* src/qemu/qemu_conf.h (qemudBuildCommandLine): Update signature.
* src/qemu/qemu_driver.c (qemuAssignPCIAddresses)
(qemudStartVMDaemon, qemuDomainXMLToNative): Adjust callers.
Port hooks and iptables code to new command execution APIs
This proof of concept shows how two existing uses of virExec
and virRun can be ported to the new virCommand APIs, and how
much simpler the code becomes
This introduces a new set of APIs in src/util/command.h
to use for invoking commands. This is intended to replace
all current usage of virRun and virExec variants, with a
more flexible and less error prone API.
* src/util/command.c: New file.
* src/util/command.h: New header.
* src/Makefile.am (UTIL_SOURCES): Build it.
* src/libvirt_private.syms: Export symbols internally.
* tests/commandtest.c: New test.
* tests/Makefile.am (check_PROGRAMS): Run it.
* tests/commandhelper.c: Auxiliary program.
* tests/commanddata/test2.log - test15.log: New expected outputs.
* cfg.mk (useless_free_options): Add virCommandFree.
(msg_gen_function): Add virCommandError.
* po/POTFILES.in: New translation.
* .x-sc_avoid_write: Add exemption.
* tests/.gitignore: Ignore new built file.
tests: Support for faking emulator in qemuxml2argv
This patch allows for using custom scripts instead of /usr/bin/qemu
emulator in domain XML. To do so, one would specify relative path to the
custom script in <emulator/>. The path needs to be relative to
qemuxml2argvdata directory and it will be transparently made absolute in
runtime. The expected command line needs to contain the exact relative
path as was used in domain XML.
The problem is RelaxNG schema for domain XML only allows for absolute
path within <emulator/>. To workaround it, an extra '/' must be added at
the beginning of the path. That is, instead of "./qemu.sh" or
"../emulator/qemu.sh" one would use "/./qemu.sh" or
"/../emulator/qemu.sh". The extra slash is removed before further
processing. I don't like this workaround, it's very ugly but it's the
best option I was able to come up with. Relaxing domain XML schema is
not an option IMO.
Osier Yang [Thu, 2 Dec 2010 01:19:30 +0000 (09:19 +0800)]
virsh: Remove using phy as default disk driver in cmdAttachDisk
* tools/virsh.c (virsh shouldn't use 'phy' as the disk driver if
user doesn't specify "--driver", it causes bugs, as not all of
hypervisor driver supports 'phy', and actually hypervisor should
known the correct default disk driver and subdriver, so remove it)
Osier Yang [Wed, 1 Dec 2010 12:35:52 +0000 (20:35 +0800)]
qemu: Use macro for max and min vnc port instead of number
* src/qemu/qemu_driver.c (though MACROS QEMU_VNC_PORT_MAX, and
QEMU_VNC_PORT_MIN are defined at the beginning, numbers (65535, 5900)
are still used, replace them)
The arguments passed to the thread function must be allocated on
the heap, rather than the stack, since it is possible for the
spawning thread to continue before the new thread runs at all.
In such a case, it is possible that the area of stack where the
thread args were stored is overwritten.
* src/util/threads-pthread.c, src/util/threads-win32.c: Allocate
thread arguments on the heap
Osier Yang [Wed, 1 Dec 2010 12:24:58 +0000 (20:24 +0800)]
virsh: update help for "virsh help help"
As virsh help supports both command and command group now,
update "cmdHelp" to print consite help, (this patch is
increment of "7829052757953023b0826e0293ffe18ed4ab89e9").
And also remove redundant empty line in "vshUsage".
Hu Tao [Tue, 30 Nov 2010 07:12:54 +0000 (15:12 +0800)]
Fall back to QEMUD_SAVE_FORMAT_RAW if compression method fails.
When dumping a domain, it's reasonable to save dump-file in raw format
if dump format is misconfigured or the corresponding compress program
is not available rather then fail dumping.
Osier Yang [Tue, 30 Nov 2010 06:37:04 +0000 (14:37 +0800)]
virsh: Categorize commands into groups for virsh help
Change the virsh help out. The new output of "virsh help" and
"virsh --help" will be like:
Secret (help keyword 'secret'):
secret-define define or modify a secret from an XML file
secret-dumpxml secret attributes in XML
secret-set-value set a secret value
secret-get-value Output a secret value
secret-undefine undefine a secret
secret-list list secrets
Snapshot (help keyword 'snapshot'):
snapshot-create Create a snapshot
snapshot-current Get the current snapshot
snapshot-delete Delete a domain snapshot
snapshot-dumpxml Dump XML for a domain snapshot
snapshot-list List snapshots for a domain
snapshot-revert Revert a domain to a snapshot
Also support output help information of specified command group, e.g.
% ./tools/virsh help "Network Filter"
Network Filter (help keyword 'filter'):
nwfilter-define define or update a network filter from an XML file
nwfilter-undefine undefine a network filter
nwfilter-dumpxml network filter information in XML
nwfilter-list list network filters
nwfilter-edit edit XML configuration for a network filter
Each group has a help keyword, e.g.
% ./tools/virsh help filter
Network Filter (help keyword 'filter'):
nwfilter-define define or update a network filter from an XML file
nwfilter-undefine undefine a network filter
nwfilter-dumpxml network filter information in XML
nwfilter-list list network filters
nwfilter-edit edit XML configuration for a network filter
* tools/virsh.c:
- introduce new struct "vshCmdGrp" and macros to define the groups.
- split previous array "commands" into small arrays which are orgnized
by group
- changed some functions, e.g. "vshCmdDefSearch"
- Added new functions, e.g. "vshCmdGrpSearch"
- commands of each group are in "alphabetical order" now.
- command groups are in "alphabetical order" now.
- the commands are categorized with reference of
http://wiki.libvirt.org/page/VirshHelpV2 (by Justin)
- the modifications doesn't affect tests
Stefan Berger [Tue, 30 Nov 2010 17:50:54 +0000 (12:50 -0500)]
802.1Qbg: use pre-associate state at beginning of inc. migr
This patch introduces the usage of the pre-associate state of the IEEE 802.1Qbg standard on incoming VM migration on the target host. It is in response to bugzilla entry 632750.
For being able to differentiate the exact reason as to why a macvtap device is being created, either due to a VM creation or an incoming VM migration, I needed to pass that reason as a parameter from wherever qemudStartVMDaemon is being called in order to determine whether to send an ASSOCIATE (VM creation) or a PRE-ASSOCIATE (incoming VM migration) towards lldpad.
I am also fixing a problem with the virsh domainxml-to-native call on the way.
Gerhard successfully tested the patch with a recent blade network 802.1Qbg-compliant switch.
The patch should not have any side-effects on the 802.1Qbh support in libvirt, but Roopa (cc'ed) may want to verify this.
Guido Günther [Sun, 28 Nov 2010 21:52:44 +0000 (22:52 +0100)]
OpenVZ: take veid from vmdef->name when defining new domains
We currently use the next free veid although there's one given in the
domain xml. This currently breaks defining new domains since vmdef->name
and veid don't match leading to the following error later on:
error: Failed to define domain from 110.xml
error: internal error Could not set UUID
Since silently ignoring vmdef->name is not nice respect it instead. We
avoid veid collisions in the upper levels already.
and makes virRaiseErrorFull() log errors at debug priority
when called from inside libvirtd. This stops libvirtd from
polluting it's own log with client errors at error priority
that'll be reported and logged on the client side anyway.
Wen Congyang [Thu, 25 Nov 2010 08:38:32 +0000 (16:38 +0800)]
correct the arguments of migrate_speed
When we set migrate_speed by json, we receive the following
error message:
libvirtError: internal error unable to execute QEMU command
'migrate_set_speed': Invalid parameter type, expected: number
The reason is that: the arguments of migrate_set_speed
by json is json number, not json string.
Jiri Denemark [Wed, 24 Nov 2010 10:25:19 +0000 (11:25 +0100)]
Fix host CPU counting on unusual NUMA topologies
The nodeinfo structure includes
nodes : the number of NUMA cell, 1 for uniform mem access
sockets : number of CPU socket per node
cores : number of core per socket
threads : number of threads per core
which does not work well for NUMA topologies where each node does not
consist of integral number of CPU sockets.
We also have VIR_NODEINFO_MAXCPUS macro in public libvirt.h which
computes maximum number of CPUs as (nodes * sockets * cores * threads).
As a result, we can't just change sockets to report total number of
sockets instead of sockets per node. This would probably be the easiest
since I doubt anyone is using the field directly. But because of the
macro, some apps might be using sockets indirectly.
This patch leaves sockets to be the number of CPU sockets per node (and
fixes qemu driver to comply with this) on machines where sockets can be
divided by nodes. If we can't divide sockets by nodes, we behave as if
there was just one NUMA node containing all sockets. Apps interested in
NUMA should consult capabilities XML, which is what they probably do
anyway.
This way, the only case in which apps that care about NUMA may break is
on machines with funky NUMA topology. And there is a chance libvirt
wasn't able to start any guests on those machines anyway (although it
depends on the topology, total number of CPUs and kernel version).
Nothing changes at all for apps that don't care about NUMA.