Commit Graph

14636 Commits

Author SHA1 Message Date
Daniel P. Berrange
a50de5d827 Fix crash on OOM when parsing disk security label
If an OOM error occurs in virSecurityDeviceLabelDefParseXML the
cleanup code may free an uninitialized pointer, causing a crash

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-24 10:51:28 +01:00
lawrancejing
e9f5f2e749 conf: Fix virNetworkAssignDef's comment. 2013-09-24 17:02:19 +08:00
Daniel P. Berrange
6fdafe278d Add test case for virNetServerClient object identity code
Start a test case for the virNetServerClient object, which
initially checks the creation of a virIdentityPtr object.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-24 09:37:26 +01:00
Daniel P. Berrange
199679585b Add a virNetSocketNewConnectSockFD method
To allow creation of a virNetSocketPtr instance from a pre-opened
socketpair FD, add a virNetSocketNewConnectSockFD method.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-24 09:37:26 +01:00
Giuseppe Scrivano
de51dc9c9a python: add bindings for virConnectGetCPUModelNames
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 16:59:39 -06:00
Giuseppe Scrivano
ea45b23cfc virsh: add function to get the CPU models for an arch
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 16:42:10 -06:00
Giuseppe Scrivano
dda4548c7b virConnectGetCPUModelNames: add the support for the test protocol
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 16:40:30 -06:00
Giuseppe Scrivano
cbcecd7ab1 virConnectGetCPUModelNames: add the support for qemu
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 15:52:14 -06:00
Giuseppe Scrivano
fd69544965 virConnectGetCPUModelNames: implement the remote protocol
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 15:50:35 -06:00
Giuseppe Scrivano
36cc09420a cpu: add function to get the models for an arch
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 15:50:30 -06:00
Giuseppe Scrivano
f90857b32a libvirt: add new public API virConnectGetCPUModelNames
The new function virConnectGetCPUModelNames allows to retrieve the list
of CPU models known by the hypervisor for a specific architecture.

Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 15:41:50 -06:00
Eric Blake
b0d9207bc6 build: ensure 'make check' sees up-to-date config.h
Nehal J. Wani reported on IRC a rather interesting build failure:

In file included from util/virnetdevbridge.c:53:0:
/usr/include/linux/in6.h:30:8: error: redefinition of 'struct in6_addr'
 struct in6_addr {
        ^

I traced it to the fact that he ran 'git pull; make check' across
commit e62e0094.  What happened is that the configure changes
result in a new variable that was set to be defined on his system,
but config.h was not regenerated to contain the value of that
variable.  Running 'make' instead of 'make check' cleaned up the
problem.  A bit more investigation, and I see that in Makefile.am,
automake sticks rules that rebuild config.h as part of 'make all',
and that we also had a dependency 'check-local: all'; BUT the
rule for check-local is run only at the point when the top-level
directory is visited.  Automake documents that SUBDIRS should
contain an explicit '.' at the point the top-level should be
visited (defaulting to last, if it doesn't appear).  Sure enough,
with this patch, 'make check' now does the top-level 'all' rules,
which regenerates 'config.h' BEFORE compiling any code that might
depend on changed content of that file.

* Makefile.am (SUBDIRS): Put '.' first, not last.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-23 15:18:24 -06:00
Daniel P. Berrange
e4697b92ab Fix typo in identity code which is pre-requisite for CVE-2013-4311
The fix for CVE-2013-4311 had a pre-requisite enhancement
to the identity code

  commit db7a5688c0
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Thu Aug 22 16:00:01 2013 +0100

    Also store user & group ID values in virIdentity

This had a typo which caused the group ID to overwrite the
user ID string. This meant any checks using this would have
the wrong ID value. This only affected the ACL code, not the
initial polkit auth. It also leaked memory.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-23 14:07:55 -06:00
Jiri Denemark
0543e0d212 docs: Load libvirt_access*.xml from build dir
The xml files are generated in build directory and thus docs/newapi.xsl
was not able to find them in a VPATH build.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2013-09-23 14:16:45 +02:00
Chen Hanxiao
9a08e2cbc6 LXC: Check the existence of dir before resolving symlinks
If a dir does not exist, raise an immediate error in logs
rather than letting virFileResolveAllLinks fail, since this
gives better error reporting to the user.

Signed-off-by: Chen Hanxiao <chenhanxiao@cn.fujitsu.com>
2013-09-23 11:22:17 +01:00
Chen Hanxiao
0e618c6aff LXC: follow the unit style of /proc/meminfo
When FUSE is enabled, the LXC container is setup with
a custom /proc/meminfo file. This file uses "KB" as a
suffix, rather than "kB" which is the kernel's style.
Fix this inconsistency to avoid confusing apps.

Signed-off-by: Chen Hanxiao <chenhanxiao@cn.fujitsu.com>
2013-09-23 11:01:07 +01:00
Peter Krempa
5702c014b2 conf: Do better job when comparing features ABI compatibility
The ABI compatibility check for domain features didn't check the
expanded HyperV and APIC EOI values, thus possibly allowing change in
guest ABI.

Add the check and use typecasted switch statement to warn developers
when adding a new HyperV feature.
2013-09-22 19:24:38 +02:00
Daniel P. Berrange
49a5262d2f Fix potential use of uninitialized value in virDomainGetVcpuPinInfo
The virDomainGetVcpuPinInfo python wrapper had a potential use of
uninitialized values

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-20 17:22:39 +01:00
Martin Kletzander
484cc3217b qemu: Fix seamless SPICE migration
Since the wait is done during migration (still inside
QEMU_ASYNC_JOB_MIGRATION_OUT), the code should enter the monitor as such
in order to prohibit all other jobs from interfering in the meantime.
This patch fixes bug #1009886 in which qemuDomainGetBlockInfo was
waiting on the monitor condition and after GetSpiceMigrationStatus
mangled its internal data, the daemon crashed.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1009886
2013-09-20 17:11:10 +02:00
Doug Goldstein
7457cbe871 VMware: Make version parsing testable and add tests
This splits up the version parsing code into a callable API like QEMU
help/version string parsing so that we can test it as we need to add
additional patterns for newer versions/products.
2013-09-20 08:23:31 -05:00
Doug Goldstein
e7a1ce9d07 VMware: Store vmrun binary's path in the driver
Rather than looking up the path to vmrun each time we call it, look it
up once and save it. This sets up the ability for us to detect where the
path is on Mac OS X and not have to look it up each time we execute it.
2013-09-20 08:23:31 -05:00
Doug Goldstein
ea5805ed29 VMware: Convert driver type defines to enum
The VMware driver supports multiple backends for the VMware Player and
VMware Workstation, convert this logic into enum and use VIR_ENUM_IMPL()
to provide conversions to and from strings.
2013-09-20 08:23:31 -05:00
Ján Tomko
5b36ab900a Don't dereference NULL in qemumonitorjsontest
In case of an error, qemuMonitorTestNewSimple returns NULL.
Error out instead of dereferencing it.

Found by Coverity, reported by John Ferlan.
2013-09-20 13:46:13 +02:00
Daniel P. Berrange
83c24493df Add checking of dbus_message_iter_append_basic return value
Coverity complains that the test suite did not check the
return value of dbus_message_iter_append_basic() as we did
in most other places.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-20 12:38:24 +01:00
Laine Stump
30bb4c4b54 qemu: use "ide" as device name for implicit SATA controller on Q35
This resolves https://bugzilla.redhat.com/show_bug.cgi?id=1008903

The Q35 machinetype has an implicit SATA controller at 00:1F.2 which
isn't given the "expected" id of ahci0 by qemu when it's created. The
original suggested solution to this problem was to not specify any
controller for the disks that use the default controller and just
specify "unit=n" instead; qemu should then use the first IDE or SATA
controller for the disk.

Unfortunately, this "solution" is ignorant of the fact that in the
case of SATA disks, the "unit" attribute in the disk XML is actually
*not* being used for the unit, but is instead used to specify the
"bus" number; each SATA controller has 6 buses, and each bus only
allows a single unit. This makes it nonsensical to specify unit='n'
where n is anything other than 0. It also means that the only way to
connect more than a single device to the implicit SATA controller is
to explicitly give the bus names, which happen to be "ide.$n", where
$n can be replaced by the disk's "unit" number.
2013-09-20 07:03:23 -04:00
Jiri Denemark
604ae65744 daemon: Avoid dead code in polkit auth 2013-09-20 11:23:01 +02:00
Simone Gotti
fe64499dd1 virsh: add missing "async" option in opts_block_commit
After commit 8aecd35126 it'll detect
that a required option is not defined and it will assert and exit with:

virsh.c:1364: vshCommandOpt: Assertion `valid->name' failed.

Problem has been latent since commit ed23b106.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-19 12:14:03 -06:00
Michal Privoznik
2cf3a4e526 qemumonitorjsontest: Introduce DO_TEST_SIMPLE
This macro is there to test the simplest monitor functions we have,
those in the form of: int ( *func) (qemuMonitorPtr). So far, we have
seven such functions.
2013-09-19 12:31:59 +02:00
Michal Privoznik
4f1b6df127 qemumonitorjsontest: Test CPU state handling code
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-09-19 12:31:59 +02:00
Christophe Fergeau
1533322211 daemon: Remove more hardcoded paths from help output
A previous patch used existing #define for the various files in /etc/pki
instead of hardcoding them in the help output. However I missed that
remote_driver.h contains #define for more paths that are present
in the daemon help output.
This commit uses the existing constants for the path to the
configuration file and to the libvirt sockets.
2013-09-19 09:32:42 +02:00
Christophe Fergeau
c5cc41584b Fix LIBVIRTD_CONFIGURATION_FILE constant
It's not used anywhere, but should be pointing to
$sysconfdir/libvirt/libvirtd.conf, not $sysconfdir/libvirtd.conf
2013-09-19 09:31:57 +02:00
Jonathan Lebon
b8e4644d1a docs: fix virEventAddHandle return details
In commit 6d41cb8, the interface for virEventAddHandleFunc was changed.
This patch updates the documentation for virEventAddHandle to reflect
the new significance of the return value. Also, both functions now
mention -1 for failure.
2013-09-18 13:06:13 -06:00
Daniel P. Berrange
e7f400a110 Fix crash in remoteDispatchDomainMemoryStats (CVE-2013-4296)
The 'stats' variable was not initialized to NULL, so if some
early validation of the RPC call fails, it is possible to jump
to the 'cleanup' label and VIR_FREE an uninitialized pointer.
This is a security flaw, since the API can be called from a
readonly connection which can trigger the validation checks.

This was introduced in release v0.9.1 onwards by

  commit 158ba8730e
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Wed Apr 13 16:21:35 2011 +0100

    Merge all returns paths from dispatcher into single path

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-18 12:41:14 -06:00
Giuseppe Scrivano
ab42c22627 doc: fix XML for the RNG device example
Add a missing '/' to close the "source" element.

Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
2013-09-18 17:14:49 +02:00
Daniel P. Berrange
922b7fda77 Add support for using 3-arg pkcheck syntax for process (CVE-2013-4311)
With the existing pkcheck (pid, start time) tuple for identifying
the process, there is a race condition, where a process can make
a libvirt RPC call and in another thread exec a setuid application,
causing it to change to effective UID 0. This in turn causes polkit
to do its permission check based on the wrong UID.

To address this, libvirt must get the UID the caller had at time
of connect() (from SO_PEERCRED) and pass a (pid, start time, uid)
triple to the pkcheck program.

This fix requires that libvirt is re-built against a version of
polkit that has the fix for its CVE-2013-4288, so that libvirt
can see 'pkg-config --variable pkcheck_supports_uid polkit-gobject-1'

Signed-off-by: Colin Walters <walters@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-18 15:13:42 +01:00
Daniel P. Berrange
e65667c0c6 Ensure system identity includes process start time
The polkit access driver will want to use the process start
time field. This was already set for network identities, but
not for the system identity.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-18 15:04:27 +01:00
Daniel P. Berrange
db7a5688c0 Also store user & group ID values in virIdentity
Future improvements to the polkit code will require access to
the numeric user ID, not merely user name.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-18 15:04:27 +01:00
Christophe Fergeau
d0d013d9fb daemon: Don't hardcode pki paths in help output
There are constants for these paths in remote_driver.h so we can
use these rather than duplicating them in the help output.

Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-18 07:56:38 -06:00
Christophe Fergeau
521b78d0af daemon: Fix 'caert.pem' typo in privileged help output
The help message indicates that the CA certificate is
$sysconfdir/pki/CA/caert.pem while the actual path is
$sysconfdir/pki/CA/cacert.pem

Signed-off-by: Christophe Fergeau <cfergeau@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-18 07:56:25 -06:00
Jiri Denemark
13e9bad55a qemu: Avoid dangling job in qemuDomainSetBlockIoTune
virDomainSetBlockIoTuneEnsureACL was incorrectly called after we already
started a job. As a result of this, the job was not cleaned up when an
access driver had forbidden the action.
2013-09-18 10:37:48 +02:00
Tomas Meszaros
ffb7ccaa70 virsh: Add vshCompleter to each option
completer and completer_flags added to the _vshCmdOptDef
structure so it will be possible for completion generators to
conveniently call option completer functions with desired flags.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-17 17:47:33 -06:00
Diego Woitasen
22547b4c98 Add forwarder attribute to <dns/> element
Useful to set custom forwarders instead of using the contents of
/etc/resolv.conf. It helps me to setup dnsmasq as local nameserver to
resolve VM domain names from domain 0, when domain option is used.

Signed-off-by: Diego Woitasen <diego.woitasen@vhgroup.net>
Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-17 17:47:33 -06:00
Doug Goldstein
7ab7c9a2e9 VMX: Add a VMWare Fusion 5 configuration for tests
A user was having an issue with this specific VMWare Fusion config and
he gave me permission to add it as part of our test suite to further
expand our VMX test coverage. Unfortunately our VMX parser and
generator does not support many features contained within and just
silently ignores fields it does not understand so they had to
be removed out in the xml2vmx test. The original unmodified version
exists in the vmx2xml test.
2013-09-17 14:10:41 -05:00
Doug Goldstein
834aebcc2f VMX: Add support for 'auto detect' fileNames
VMWare Fusion 5 can set the CD-ROM's device name to be 'auto detect' when
using the physical drive via 'cdrom-raw' device type. VMWare will then
connect to first available host CD-ROM to the virtual machine upon start
up according to VMWare documentation. If no device is available, it
appears that the device will remain disconnected.

To better model this a CD-ROM that is marked as "auto detect" when in
the off state would be modeled as the following with this patch:
  <disk type='block' device='lun'>
    <source startupPolicy='optional'/>
    <target dev='hda' bus='ide'/>
    <address type='drive' controller='0' bus='0' target='0' unit='0'/>
  </disk>

Once the domain transitions to the powered on state, libvirt can
populate the remaining source data with what is connected, if anything.
However future power cycles, the domain may not always start with that
device attached.
2013-09-17 14:10:40 -05:00
Doug Goldstein
4b5652d0dc Allow <source> for type=block to have no dev
Currently the XML parser already allows the following syntax:
  <disk type='block' device='cdrom'>
    <source startupPolicy='optional'/>
    <target dev='hda' bus='ide'/>
    <address type='drive' controller='0' bus='0' target='0' unit='0'/>
  </disk>

But it if the dev value is NULL then it would not have the leading
"<source ", resulting in invalid XML.
2013-09-17 14:10:40 -05:00
Eric Blake
caf6589233 build: skip ld_preload tests on non-Linux systems
A cross build to mingw fails with:

  CC       virsystemdtest-virsystemdtest.o
../../tests/virsystemdtest.c: In function 'testCreateNoSystemd':
../../tests/virsystemdtest.c:97:9: error: implicit declaration of function 'unsetenv' [-Werror=implicit-function-declaration]
         unsetenv("FAIL_NO_SERVICE");
         ^
../../tests/virsystemdtest.c:97:9: error: nested extern declaration of 'unsetenv' [-Werror=nested-externs]

We could cop out and pull in the gnulib unsetenv module.  But when
you stop and think about it, this test requires LD_PRELOAD to work,
and systemd is a Linux-only concept anyways, both of which mean
the test could never work on mingw in the first place.  Simpler is
to just fix the test to behave like our other LD_PRELOAD tests.

* tests/virsystemdtest.c: Provide non-Linux implementation.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-17 11:11:25 -06:00
Eric Blake
0e1f012c1d build: hoist system-specific checks before library checks
Commit f92c7e3 fixed a regression for native builds, but introduced
a regression for cross-compilation builds; in particular,
./autobuild.sh on a Fedora system with mingw cross-compiler fails
with:

checking for qemu-kvm... /usr/bin/qemu-kvm
checking for yajl_parse_complete in -lyajl... no
checking for yajl_tree_parse in -lyajl... no
configure: error: You must install the libyajl library & headers to compile libvirt

Since we default $with_qemu to 'yes' rather than 'check', and then
flip that default based on platform-specific checks, those platform
specifics need to come prior to any library checks that depend on
the value of $with_qemu.

* configure.ac: Ensure system defaults are sane before checking
for things that make decisions based on system default.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-17 10:25:42 -06:00
Eric Blake
ceaa536d49 maint: ignore recently-added test
* .gitignore: Ignore metadatatest.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-09-17 10:24:41 -06:00
Peter Krempa
935c031730 tests: metadatatest: Quiesce errors on expected paths
Use the helper added in previous patch to quiesce errors from this test
that was spamming logs on normal test runs.
2013-09-17 16:46:53 +02:00
Peter Krempa
43f68a4f9e test: Refactor setting of dummy error handlers
Multiple tests need to register a function to quiesce errors from
libvirt when using a connection and doing negative tests. Each of those
tests had a static function to do so. This can be replaced by a utility
function that enables the errors when debug is enabled.

This patch adds virtTestQuiesceLibvirtErrors() and refactors test that
use private handlers.
2013-09-17 16:45:53 +02:00