Commit Graph

8542 Commits

Author SHA1 Message Date
Jiri Denemark
5b2d0bbc4c qemu: Add support for job phase
Asynchronous jobs may take long time to finish and may consist of
several phases which we need to now about to help with recovery/rollback
after libvirtd restarts.
2011-07-13 01:53:32 +02:00
Jiri Denemark
e6704af1fc qemu: Recover from interrupted jobs
Detect and react on situations when libvirtd was restarted or killed
when a job was active.
2011-07-13 01:53:32 +02:00
Jiri Denemark
ff340a84b8 qemu: Save job type in domain status XML
If libvirtd is restarted when a job is running, the new libvirtd process
needs to know about that to be able to recover and rollback the
operation.
2011-07-13 01:53:32 +02:00
Jiri Denemark
361842881e qemu: Allow all query commands to be run during long jobs
Query commands are safe to be called during long running jobs (such as
migration). This patch makes them all work without the need to
special-case every single one of them.

The patch introduces new job.asyncCond condition and associated
job.asyncJob which are dedicated to asynchronous (from qemu monitor
point of view) jobs that can take arbitrarily long time to finish while
qemu monitor is still usable for other commands.

The existing job.active (and job.cond condition) is used all other
synchronous jobs (including the commands run during async job).

Locking schema is changed to use these two conditions. While asyncJob is
active, only allowed set of synchronous jobs is allowed (the set can be
different according to a particular asyncJob) so any method that
communicates to qemu monitor needs to check if it is allowed to be
executed during current asyncJob (if any). Once the check passes, the
method needs to normally acquire job.cond to ensure no other command is
running. Since domain object lock is released during that time, asyncJob
could have been started in the meantime so the method needs to recheck
the first condition. Then, normal jobs set job.active and asynchronous
jobs set job.asyncJob and optionally change the list of allowed job
groups.

Since asynchronous jobs only set job.asyncJob, other allowed commands
can still be run when domain object is unlocked (when communicating to
remote libvirtd or sleeping). To protect its own internal synchronous
commands, the asynchronous job needs to start a special nested job
before entering qemu monitor. The nested job doesn't check asyncJob, it
only acquires job.cond and sets job.active to block other jobs.
2011-07-13 01:53:21 +02:00
Jiri Denemark
24f717ac22 qemu: Consolidate {Enter,Exit}Monitor{,WithDriver}
EnterMonitor and ExitMonitor methods are very similar to their
*WithDriver variants; consolidate them into EnterMonitorInternal and
ExitMonitorInternal to avoid (mainly future) code duplication.
2011-07-12 21:17:28 +02:00
Jiri Denemark
cec1d280ad qemu: Consolidate BeginJob{,WithDriver} into a single method
This avoids code duplication and also avoids relying on good luck that
ignore_value(virDomainObjUnref(obj)) doesn't set errno.
2011-07-12 21:17:28 +02:00
Jiri Denemark
90a422f071 qemu: Separate job related data into a new object 2011-07-12 21:17:28 +02:00
Eric Blake
0583825f3c maint: avoid incremental 'make syntax-check' failure
Incrementally running 'make syntax-check' on a tree previously
built after commit 62dee6f but before 44036460 fails sc_po_check
(because the generated qemu_dispatch.h gained translatable strings).
This is a followup to commit addaa537 for that scenario.

* cfg.mk (sc_po_check): Add another prereq.
($(srcdir)/daemon/qemu_dispatch.h): Add rule.
2011-07-12 11:16:44 -06:00
Daniel P. Berrange
52312385c0 Skip bugs which are CLOSED in todo list
* docs/todo.pl: Skip closed bugs
2011-07-12 17:10:33 +01:00
Daniel P. Berrange
65d68f62f1 Add missing cleanup for transient guests in UML driver
The UML inotify handler would kill off guests when certain
conditions arise, but it forgot to remove transient guests
from the list of domains

* src/uml/uml_driver.c: Cleanup transient guests
2011-07-12 17:05:30 +01:00
Daniel P. Berrange
f0c7103db5 Add auditing to the UML driver
* src/uml/uml_driver.c: Add audit hooks
2011-07-12 17:05:30 +01:00
Daniel P. Berrange
7d612c3059 Add auditing to the LXC driver
* src/lxc/lxc_driver.c: Add auditing hooks
2011-07-12 17:05:30 +01:00
Daniel P. Berrange
839f47b57f Add virtualization type in audit messages
Since a host can run several different virtualization types at
the same time, audit messages should allow domains to be identified.
Add a 'virt={qemu,kvm,uml,lxc,...}' key to domain audit messages

* src/conf/domain_audit.c: Identify virt type of guest
2011-07-12 17:05:30 +01:00
Daniel P. Berrange
e6e90c8d70 Add auditing of filesystems
When passing through filesystems from the host to a guest, the
host filesystem passed must be audited

* src/conf/domain_audit.{c,h}: Add virDomainAuditFS
2011-07-12 17:05:30 +01:00
Daniel P. Berrange
b43070ebfc Move qemu_audit.h helpers into shared code
The LXC and UML drivers can both make use of auditing. Move
the qemu_audit.{c,h} files to src/conf/domain_audit.{c,h}

* src/conf/domain_audit.c: Rename from src/qemu/qemu_audit.c
* src/conf/domain_audit.h: Rename from src/qemu/qemu_audit.h
* src/Makefile.am: Remove qemu_audit.{c,h}, add domain_audit.{c,h}
* src/qemu/qemu_audit.h, src/qemu/qemu_cgroup.c,
  src/qemu/qemu_command.c, src/qemu/qemu_driver.c,
  src/qemu/qemu_hotplug.c, src/qemu/qemu_migration.c,
  src/qemu/qemu_process.c: Update for changed audit API names
2011-07-12 17:05:25 +01:00
Matthias Bolte
38ea116eba esx: Initialize all instances of esxVMX_Data to zero
To avoid using or freeing uninitialized memory or pointers.
2011-07-12 17:08:36 +02:00
Matthias Bolte
b253452942 rpc: Fix compile error due to potentially unused parameter
connectDBus is only used if HAVE_DBUS is set. Therefore mark
it as potentially unused.
2011-07-12 17:06:16 +02:00
Jim Fehlig
b8adfcc60c Fix build when using polkit0
V2: Remove policy kit references from virNetServer and use DBus APIs
    directly, if available.
2011-07-12 08:44:55 -06:00
Daniel P. Berrange
109efd7941 Implement code to attach to external QEMU instances.
Given a PID, the QEMU driver reads /proc/$PID/cmdline and
/proc/$PID/environ to get the configuration. This is fed
into the ARGV->XML convertor to build an XML configuration
for the process.

/proc/$PID/exe is resolved to identify the full command
binary path

After checking for name/uuid uniqueness, an attempt is
made to connect to the monitor socket. If successful
then 'info status' and 'info kvm' are issued to determine
whether the CPUs are running and if KVM is enabled.

* src/qemu/qemu_driver.c: Implement virDomainQemuAttach
* src/qemu/qemu_process.h, src/qemu/qemu_process.c: Add
  qemuProcessAttach to connect to the monitor of an
  existing QEMU process
2011-07-12 15:39:04 +01:00
Daniel P. Berrange
7760eaa050 Add monitor API for checking whether KVM is enabled
When attaching to an external QEMU process, it is neccessary
to check if the process is using KVM or not. This can be done
using a monitor command

* src/qemu/qemu_monitor.c, src/qemu/qemu_monitor.h,
  src/qemu/qemu_monitor_json.c, src/qemu/qemu_monitor_json.h,
  src/qemu/qemu_monitor_text.c, src/qemu/qemu_monitor_text.h: Add
  API for checking if KVM is enabled
2011-07-12 15:39:04 +01:00
Daniel P. Berrange
80a4ee4695 Add a method for extracting QEMU argv from /proc
To enable attaching to externally launched QEMU, we need
to be able to reverse engineer a guest XML config based
on the argv for a PID in /proc

* src/qemu/qemu_command.c, src/qemu/qemu_command.h: Add
  qemuParseCommandLinePid which extracts QEMU config from
  argv in /proc, given a PID number
2011-07-12 15:39:04 +01:00
Daniel P. Berrange
ebbae359ff Change extract pidfile & monitor config from QEMU command line
When converting QEMU argv into a virDomainDefPtr, also extract
the pidfile, monitor character device config and the monitor
mode.

* src/qemu/qemu_command.c, src/qemu/qemu_command.h: Extract
  pidfile & monitor config from QEMU argv
* src/qemu/qemu_driver.c, tests/qemuargv2xmltest.c: Add extra
  params when calling qemuParseCommandLineString
2011-07-12 15:39:04 +01:00
Daniel P. Berrange
accfe952eb Keep pidfile path in qemuDomainObjPrivate struct
Avoid re-formatting the pidfile path everytime we need it. Create
it once when starting the guest, and preserve it until the guest
is shutdown.

* src/libvirt_private.syms, src/util/util.c,
  src/util/util.h: Add virFileReadPidPath
* src/qemu/qemu_domain.h: Add pidfile field
* src/qemu/qemu_process.c: Store pidfile path in qemuDomainObjPrivate
2011-07-12 15:39:03 +01:00
Daniel P. Berrange
eb7be6a606 Add a new 'virsh qemu-attach' command.
This command allows libvirt to attach to an existing QEMU
instance.

 $ qemu-kvm -cdrom ~/demo.iso \
     -monitor unix:/tmp/demo,server,nowait \
     -name foo \
     -uuid cece4f9f-dff0-575d-0e8e-01fe380f12ea  &
 $ QEMUPID=$!
 $ virsh qemu-attach $QEMUPID
2011-07-12 15:39:03 +01:00
Daniel P. Berrange
4403646001 Define remote wire protocol & impls for virDomainQemuAttach
This tweaks the RPC generator to cope with some naming
conventions used for the QEMU specific APIs

* daemon/remote.c: Server side dispatcher
* src/remote/remote_driver.c: Client side dispatcher
* src/remote/qemu_protocol.x: Wire protocol definition
* src/rpc/gendispatch.pl: Use '$structprefix' in method
  names, fix QEMU flags and fix dispatcher method names
2011-07-12 15:39:03 +01:00
Daniel P. Berrange
639f841346 Define a QEMU specific API to attach to a running QEMU process
Introduce a new API in libvirt-qemu.so

 virDomainPtr virDomainQemuAttach(virConnectPtr domain,
                                  unsigned long long pid,
                                  unsigned int flags);

This allows libvirtd to attach to an existing, externally
launched QEMU process. This is useful for QEMU developers who
prefer to launch QEMU themselves for debugging/devel reasons,
but still want the benefit of libvirt based tools like
virt-top, virt-viewer, etc

* include/libvirt/libvirt-qemu.h: Define virDomainQemuAttach
* src/driver.h, src/libvirt-qemu.c, src/libvirt_qemu.syms:
  Driver glue for virDomainQemuAttach
2011-07-12 15:39:03 +01:00
Alex Jia
9c5b190017 qemu: update configuration file
* src/qemu/qemu.conf: Add blkio controller into qemu.conf.
2011-07-12 07:24:43 -06:00
Eric Blake
f532bfa297 qemu: avoid fd leak on core dump failure
* src/qemu/qemu_driver.c (doCoreDump): Guarantee fd is closed.
2011-07-12 07:24:43 -06:00
Eric Blake
2ceb35e1cd maint: rename virtaudit to match file contents
* src/util/virtaudit.[ch]: Rename...
* src/util/viraudit.[ch]: ...to match virAudit* API.
* src/Makefile.am (UTIL_SOURCES): Reflect rename.
* daemon/libvirtd.c: Likewise.
* po/POTFILES.in: Likewise.
* src/libvirt_private.syms: Likewise.
* src/qemu/qemu_audit.c: Likewise.
2011-07-12 07:24:43 -06:00
Oskari Saarenmaa
9a0e6a8fae remote/ssh: support for no_verify.
Set StrictHostKeyChecking=no to auto-accept new ssh host keys if the
no_verify extra parameter was specified.  This won't disable host key
checking for already known hosts.  Includes a test and documentation.
2011-07-12 15:09:05 +02:00
KAMEZAWA Hiroyuki
2851d37855 Fix virsh inject-nmi man page
* tools/virsh.pod: move the description which was misplaced
2011-07-12 17:40:16 +08:00
Michal Privoznik
d50bb45b1a domain_conf: Free temporary variable
* src/conf/domain_conf.c: caller must free returned value of
  virXPathString in useserial
2011-07-12 17:25:36 +08:00
Osier Yang
6ddb83efcc esx: Fix a potential crash
Initialize ptr data.datastorePathWithoutFileName as NULL, otherwise
it might cause crash when trying to free it in cleanup.
2011-07-12 17:23:35 +08:00
Eric Blake
62dee6fa48 build: also check qemu_protocol for on-the-wire stability
Since we are going to add some libvirt-qemu.so entry points in
0.9.4, we might as well start checking for RPC stability, just
as for libvirt.so.

* src/Makefile.am (PROTOCOL_STRUCTS): New variable.
(remote_protocol-structs): Rename...
(%_protocol-structs): ...and make more generic.
* src/qemu_protocol-structs: New file.
2011-07-11 17:48:32 -06:00
Eric Blake
df94811f71 build: avoid requiring -lm
log2() is heavy when ffs() can do the same thing.  But ffs()
requires gnulib support for mingw.

This patch solves this linker error on Fedora 14.

/usr/bin/ld: libvirt_lxc-domain_conf.o: undefined reference to symbol 'log2@@GLIBC_2.2.5'
/usr/bin/ld: note: 'log2@@GLIBC_2.2.5' is defined in DSO /lib64/libm.so.6 so try adding it to the linker command line
/lib64/libm.so.6: could not read symbols: Invalid operation
collect2: ld returned 1 exit status

* .gnulib: Update to latest, for ffs.
* bootstrap.conf (gnulib_modules): Import ffs.
* src/conf/domain_conf.c (virDomainDefParseXML): Use ffs instead
of log2.
Reported by Dave Allan.
2011-07-11 17:32:11 -06:00
Jim Fehlig
00d3c5a603 Remove code no longer used after commit df0b57a9 2011-07-11 14:38:03 -06:00
Matthias Bolte
aa14709a47 Add domain type checking
The drivers were accepting domain configs without checking if those
were actually meant for them. For example the LXC driver happily
accepts configs with type QEMU.

Add a check for the expected domain types to the virDomainDefParse*
functions.
2011-07-11 19:38:51 +02:00
Alex Jia
7518ad753f remote: Fix memory leak
Detected in valgrind run:

==9184== 1 bytes in 1 blocks are definitely lost in loss record 1 of 19
==9184==    at 0x4A04A28: calloc (vg_replace_malloc.c:467)
==9184==    by 0x3073715F78: xdr_array (xdr_array.c:97)
==9184==    by 0x4CF97C9: xdr_remote_domain_get_security_label_ret (remote_protocol.c:1696)
==9184==    by 0x4D08741: virNetMessageDecodePayload (virnetmessage.c:286)
==9184==    by 0x4D00F78: virNetClientProgramCall (virnetclientprogram.c:318)
==9184==    by 0x4CE3887: call (remote_driver.c:3933)
==9184==    by 0x4CF71C6: remoteDomainGetSecurityLabel (remote_driver.c:1580)
==9184==    by 0x4CCA480: virDomainGetSecurityLabel (libvirt.c:7340)
==9184==    by 0x41993A: cmdDominfo (virsh.c:2414)
==9184==    by 0x411E92: vshCommandRun (virsh.c:12730)
==9184==    by 0x4211ED: main (virsh.c:14076)
==9184==
==9184== 2 bytes in 1 blocks are definitely lost in loss record 2 of 19
==9184==    at 0x4A04A28: calloc (vg_replace_malloc.c:467)
==9184==    by 0x3073715F78: xdr_array (xdr_array.c:97)
==9184==    by 0x4CF974F: xdr_remote_node_get_security_model_ret (remote_protocol.c:1713)
==9184==    by 0x4D08741: virNetMessageDecodePayload (virnetmessage.c:286)
==9184==    by 0x4D00F78: virNetClientProgramCall (virnetclientprogram.c:318)
==9184==    by 0x4CE3887: call (remote_driver.c:3933)
==9184==    by 0x4CF6F96: remoteNodeGetSecurityModel (remote_driver.c:1648)
==9184==    by 0x4CBF799: virNodeGetSecurityModel (libvirt.c:7382)
==9184==    by 0x4197D7: cmdDominfo (virsh.c:2394)
==9184==    by 0x411E92: vshCommandRun (virsh.c:12730)
==9184==    by 0x4211ED: main (virsh.c:14076)
==9184==
==9184== 8 bytes in 1 blocks are definitely lost in loss record 3 of 19
==9184==    at 0x4A04A28: calloc (vg_replace_malloc.c:467)
==9184==    by 0x3073715F78: xdr_array (xdr_array.c:97)
==9184==    by 0x4CF9729: xdr_remote_node_get_security_model_ret (remote_protocol.c:1710)
==9184==    by 0x4D08741: virNetMessageDecodePayload (virnetmessage.c:286)
==9184==    by 0x4D00F78: virNetClientProgramCall (virnetclientprogram.c:318)
==9184==    by 0x4CE3887: call (remote_driver.c:3933)
==9184==    by 0x4CF6F96: remoteNodeGetSecurityModel (remote_driver.c:1648)
==9184==    by 0x4CBF799: virNodeGetSecurityModel (libvirt.c:7382)
==9184==    by 0x4197D7: cmdDominfo (virsh.c:2394)
==9184==    by 0x411E92: vshCommandRun (virsh.c:12730)
==9184==    by 0x4211ED: main (virsh.c:14076)
==9184==
==9184== LEAK SUMMARY:
==9184==    definitely lost: 11 bytes in 3 blocks

* src/remote/remote_driver.c: Avoid leak on remoteDomainGetSecurityLabel
  and remoteNodeGetSecurityModel.
2011-07-11 09:50:07 -06:00
Eric Blake
9693e29395 tests: simplify formatting
The shell version would output 40 extra spaces for a test with
a multiple of 40 sub-tests, and the C version can use the same
printf optimization for avoiding a loop over single space output
as the shell version.

* tests/testutils.c (virtTestMain): Avoid loop for alignment.
* tests/test-lib.sh: Fix formatting when counter is multiple of 40.
2011-07-11 09:21:37 -06:00
Jim Fehlig
eb3143154e Do not drop kernel cmdline for xen pv domains
Kernel cmdline args can be passed to xen pv domains even when a
bootloader is specified.  The current config-to-sxpr mapping
ignores cmdline when bootloader is present.

Since the xend sub-driver is used with many xen toolstack versions,
this patch takes conservative approach of adding an else block to
existing !def->os.bootloader, and only appends sxpr if def->os.cmdline
is non-NULL.

V2: Fix existing testcase broken by this patch and add new testcases
2011-07-11 09:11:15 -06:00
Wen Congyang
8a8b45b889 kill vm if saving config failed in v3 protocol
If virDomainSaveConfig() failed, we will return NULL to source,
and the vm is still available to restart during confirm() step in
v3 protocol. So we should kill it off in qemuMigrationFinish().

In v2 protocol, we should not set vm to NULL, because we hold
a reference of vm and should unrefernce it.
2011-07-11 20:53:35 +08:00
Wen Congyang
586765fb65 RPC: fix argument's name 2011-07-11 20:53:29 +08:00
Wen Congyang
ecde731c72 fix typo error 2011-07-11 20:53:21 +08:00
Michal Privoznik
874e65aa15 bios: Add support for SGA
This patch creates new <bios> element which, at this time has only the
attribute useserial='yes|no'. This attribute allow users to use
Serial Graphics Adapter and see BIOS messages from the very first moment
domain boots up. Therefore, users can choose boot medium, set PXE, etc.
2011-07-11 11:47:14 +02:00
Daniel Veillard
920ffe1b0a Fix rpm build with sanlock and without QEmu
The qemu-sanlock.conf file is not installed in this case
2011-07-11 15:57:01 +08:00
Matthias Bolte
1428029738 vbox: Fix logic in storage driver open function
If the main driver is the vbox driver, then the open function
has to return an error if the private data is invalid.
2011-07-09 15:52:14 +02:00
Matthias Bolte
c3ab6b2b53 tests: Improve output of tests that decide to skip at runtime
Don't print OK/FAIL for tests that decide to be skipped after
calling virtTestMain. Delay printing of the indentation before
the first test until we know that the test didn't decide to be
skipped.

Also make the reconnect test use VIRT_TEST_MAIN.
2011-07-09 15:47:57 +02:00
Matthias Bolte
b1c9bf27cb tests: Fix compressed test output padding logic
The current logic tries to count from 1 to 40 and ignores paddings
of 0 and 1 to 40. This doesn't work for counter + 1 mod 40 == 0
like here for counter value 159

TEST: virsh-all
      ........................................ 40
      ........................................ 80
      ........................................ 120
      ....................................... 159 OK
PASS: virsh-all

Also seq isn't portable. Therefore, calculate the correct padding
length directly and use printf to output it at once.
2011-07-09 15:40:23 +02:00
Matthias Bolte
4540f8d227 tests: Use EXIT_AM_SKIP instead of 77 directly 2011-07-09 10:14:38 +02:00
Matthias Bolte
6ca8d68d70 rpc: Fix whitespace problem in generated code
Add missing line break and fix indention level.

Reported by Cole Robinson.
2011-07-09 00:10:04 +02:00