Commit Graph

15262 Commits

Author SHA1 Message Date
Eric Blake
f06d55da7b maint: reset error on entrance to public API
We document that calling any public API wipes out all prior
libvirt errors in the same thread; but weren't obeying this
style in a few functions.

There are a couple of nested uses of virConnectRef (in lxc
and qemu reboot paths), but they should not be affected by
this change in semantics since there should not be any
previous error getting nuked (a later patch will clean up
the nested calls, along with abuse of virConnectClose on
cleanup paths which DOES nuke errors).

* src/libvirt.c (virGetVersion, virConnectRef, virDomainRef)
(virDomainGetSecurityLabel, virDomainGetSecurityLabelList)
(virDomainSetMetadata, virDomainGetMetadata)
(virNodeGetSecurityModel, virNetworkRef, virInterfaceRef)
(virStoragePoolRef, virStorageVolRef, virNodeDeviceGetName)
(virNodeDeviceRef, virSecretRef, virStreamRef, virNWFilterRef)
(virDomainSnapshotRef): Reset error on entrance.
(do_open): Drop redundant error reset.
* src/libvirt-qemu.c (virDomainQemuAgentCommand): Likewise.
* src/libvirt-lxc.c (virDomainLxcEnterNamespace)
(virDomainLxcEnterSecurityLabel): Likewise.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-03 09:38:31 -07:00
Eric Blake
ca0ea2a981 maint: improve error condition style in public API
While auditing error messages in libvirt.c, I found a couple
instances that had not been converted to modern error styles,
and a few places that failed to dispatch the error through
the known-good connection.

* src/libvirt.c (virDomainPinEmulator, virDomainGetDiskErrors)
(virDomainSendKey, virDomainGetSecurityLabelList)
(virDomainGetEmulatorPinInfo): Use typical error reporting.
(virConnectGetCPUModelNames, virConnectRegisterCloseCallback)
(virConnectUnregisterCloseCallback, virDomainGetUUID): Report
error through connection.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-03 07:15:23 -07:00
Eric Blake
5e3e211dd5 maint: split long lines in invalidArg errors
Style only.  In particular, the message on "flags 'affect live'
and 'affect config'" being mutually exclusive was already split
in some instances.

* src/libvirt.c: Wrap some long error messages to fit in 80 columns.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-03 07:11:46 -07:00
Eric Blake
ff65843480 maint: move debug statements first in public API
Most of our public APIs emit a debug log on entry, prior to anything
else.  There were a few exceptions where obvious failures were not
logged, so fix those.  When moving a debug earlier, this patch also
makes sure to avoid any NULL dereference during the log (the APIs
are supposed to gracefully fail if the user passes NULL for the object).

However, do NOT use VIR_DEBUG prior to virInitialize, since setting
up the error reporting can change where VIR_DEBUG output would be
routed.  Instead add documentation to virGlobalInit, virInitialize,
and virGetVersion that better explains initialization.

* src/libvirt.c (virGetVersion, virConnectRef, virDomainRef)
(virNetworkRef, virInterfaceRef, virStoragePoolRef)
(virStorageVolRef, virNodeDeviceRef, virSecretRef, virStreamRef)
(virNWFilterRef, virDomainSnapshotRef): Debug on function entry.
* src/libvirt-lxc.c (virDomainLxcEnterNamespace)
(virDomainLxcEnterSecurityLabel): Likewise.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-03 06:15:04 -07:00
Eric Blake
13f8372007 maint: improve debug of libvirt-{qemu,lxc} apis
I noticed that the virDomainQemuMonitorCommand debug output wasn't
telling me the name of the domain it was working on.  While it was
easy enough to determine which pointer matches the domain based on
other log messages, it is nicer to be consistent.

* src/util/viruuid.h (VIR_UUID_DEBUG): Moved here from...
* src/libvirt.c (VIR_UUID_DEBUG): ...here.
(VIR_ARG15, VIR_HAS_COMMA, VIR_DOMAIN_DEBUG_EXPAND)
(VIR_DOMAIN_DEBUG_PASTE, VIR_DOMAIN_DEBUG_0, VIR_DOMAIN_DEBUG_1)
(VIR_DOMAIN_DEBUG_2, VIR_DOMAIN_DEBUG): Move...
* src/datatypes.h: ...here.
* src/libvirt-qemu.c (virDomainQemuMonitorCommand)
(virDomainQemuAgentCommand): Better debug messages.
* src/libvirt-lxc.c (virDomainLxcOpenNamespace): Likewise.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-02 22:17:11 -07:00
Eric Blake
80aed9f87f maint: consistent formatting in libvirt.c
Preliminary cleanups to make search-and-replace easier in later
patches.  Many of these were done by grepping for (multiline)
pattern violations, then bundled all into one patch.

* src/libvirt.c: Uniform two spaces between functions, return
type and open brace on separate line, avoid blank lines around
open brace, label in column 1, drop redundant (), consistent
indentation for function headers split across lines.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-02 12:06:03 -07:00
Eric Blake
c8f95fff63 event: improve public API docs
Since libvirt 0.9.3, the entire virevent.c file has been a public
API, so improve the documentation in this file.  Also, fix a
potential core dump - it could only be triggered by bogus use of
the API and would only affect the caller (not libvirtd), but we
might as well be nice.

* src/libvirt.c (virConnectSetKeepAlive)
(virConnectDomainEventRegister, virConnectDomainEventRegisterAny)
(virConnectNetworkEventRegisterAny): Document event loop requirement.
* src/util/virevent.c (virEventAddHandle, virEventRemoveHandle)
(virEventAddTimeout, virEventRemoveTimeout): Likewise.
(virEventUpdateHandle, virEventUpdateTimeout): Likewise, and avoid
core dump if caller didn't register handler.
(virEventRunDefaultImpl): Expand example, and set up code block in
html docs.
(virEventRegisterImpl, virEventRegisterDefaultImpl): Document more
on the use of the event loop.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-02 07:29:54 -07:00
Eric Blake
cfd62c1f61 event: don't overwrite registration error message
Prior to this patch, an attempt to register an event without an
event loop started results in the vague:

libvirt: Remote Driver error : adding cb to list

Now it gives the much nicer:

libvirt:  error : internal error: could not initialize domain event timer

This also avoids hiding other reasonable error messages, such as
attempts to register a duplicate callback or OOM errors.

* src/remote/remote_driver.c (remoteConnectNetworkEventRegisterAny)
(remoteConnectDomainEventRegister)
(remoteConnectDomainEventRegisterAny): Preserve more detailed error.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-02 05:54:53 -07:00
Eric Blake
53e3f9e46e tests: fix syntax in previous patch
'make syntax-check' wasn't happy with commit fb00420.

* tests/virnetsockettest.c (checkProtocols): Typo fix.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-01 09:58:27 -07:00
Richard W.M. Jones
1063a0ff81 aarch64: Fix sysinfo test.
On AArch64 the kernel prints one "processor" (lower case 'p') line per
core.  As this was missing from the test data, virSysinfo was not
parsing any processors at all.

Fix the test data so the test now passes.
2014-01-01 16:34:13 +00:00
Richard W.M. Jones
9c225efc36 tests: sysinfotest: Actual/expected were swapped in error message. 2014-01-01 16:34:13 +00:00
Richard W.M. Jones
fb004207fa tests: virnetsockettest: Print the error if getifaddrs fails. 2014-01-01 16:34:13 +00:00
Richard W.M. Jones
fe881ae086 aarch64: Disable -fstack-protector.
I'm using gcc-4.8.2-7.fc21.aarch64 which does not appear to support
this option.
2014-01-01 14:21:21 +00:00
Eric Blake
a01cf01fff maint: update to latest gnulib
'make syntax-check' wants a newer gnulib for 2014 copyright.

Also, a couple of fixes for bootstrap issues reported on IRC:
- on some older glibc systems, ./configure could deadlock due to
a glibc malloc bug
- on FreeBSD systems, a broken autom4te coupled with gettext
0.18.3 prevents bootstrap; we can't work around it, but can at
least inform the user why they have a problem

And as always, portability fixes in other modules, some of which
are used by libvirt.

* .gnulib: Update to latest, in part for bootstrap improvements,
and for the new year.
* bootstrap: Resync to gnulib.
* gnulib/local/m4/ssize_t.m4.diff: Regenerate.

Signed-off-by: Eric Blake <eblake@redhat.com>
2014-01-01 06:02:47 -07:00
Eric Blake
169700970d docs: return paragraph must be last
Commit eb70ceb tried to create a code block for
libvirt-libvirt.html#virConnectGetType, but failed to note
that our doc generator treats everything after "Returns" as
part of the return description rather than looking for
paragraph and code layout.  Fix some other API that also had
generic details crammed into the return type paragraph.

* src/libvirt.c (virConnectOpen, virConnectOpenReadOnly)
(virConnectOpenAuth, virConnectListAllDomains): Fit doc pattern.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-31 23:21:00 -07:00
Laine Stump
3337a98a5e domain: don't try to interpret <driver> as virtio config for hostdev interfaces
This resolves:

  https://bugzilla.redhat.com/show_bug.cgi?id=1046337

The <driver> name attribute of an interface is interpreted in two
different ways depending on the <interface> type - if the interface is
type='hostdev', then the driver name describes which backend to use
for the hostdev device assignment (vfio or kvm), but if the interface
is any emulated type *and* the model type is "virtio", then the driver
name can be "vhost" or "qemu", telling which backend qemu should use
to communicate with the emulated device.

The problem comes when someone has defined a an interface like this
(which is accepted by the parser as long as no <driver name='xxx'/> is
specified):

    <interface type='hostdev'>
       ...
       <model type='virtio'/>
       ...
    </interface>

As libvirt storing this definition in the domain's status, the driver
name is automatically filled in with the backend that was
automatically decided by libvirt, so it stores this in the status:

    <interface type='hostdev'>
       ...
       <driver name='vfio'/>
       ...
       <model type='virtio'/>
       ...
    </interface>

This isn't noticed until the next time libvirtd is restarted - as it
is reading the status of all domains, it encounters the above
interface definition, logs an error:

  internal error: Unknown interface <driver name='vfio'> has been specified

and fails to reload the domain status, so the domain is marked as
inactive.

The solution is to stop the parser from interpreting <driver>
attributes as if the device was an emulated virtio device, when it is
actually a hostdev.

(Although the bug has existed since vfio support was added, it has
just recently become more apparent because libvirt previously didn't
automatically set the driver name for hostdev interfaces in the domain
status to vfio/kvm as it does since commit f094aa, first appearing in
v1.1.4.)
2013-12-31 10:56:21 +02:00
Eric Blake
eb70ceba8a docs: fix layout of code snippets
Similar to commit 52dbeac, we should indent code snippets in
other places to ensure they appear correctly in html.  See
http://libvirt.org/html/libvirt-libvirt.html#virNodeGetCPUStats
for an example improved by this patch.  Also fix some missing
semicolons in the examples.

* src/libvirt.c: Indent code samples in comments.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-28 11:28:04 -07:00
Eric Blake
9d30e078be maint: fix typo in error message
* src/util/virerror.h (virReportInvalidZeroArg): Fix double space.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-27 14:30:25 -07:00
Michal Privoznik
72ffbd1bf3 virkeycode: Allow ANSI_A
https://bugzilla.redhat.com/show_bug.cgi?id=1044806

Currently, sending the ANSI_A keycode from os_x codepage doesn't work as
it has a special value of 0x0. Our internal code handles that no
different to other not defined keycodes. Hence, in order to allow it we
must change all the undefined keycodes from 0 to -1 and adapt some code
too.

  # virsh send-key guestname --codeset os_x ANSI_A
  error: invalid keycode: 'ANSI_A'

  # virsh send-key guestname --codeset os_x ANSI_B
  # virsh send-key guestname --codeset os_x ANSI_C

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-12-24 17:36:47 +01:00
Michal Privoznik
aa4619337c lxcDomainShutdownFlags: Cleanup @flags usage
Currently, the @flags usage is a bit unclear at first sight to say the
least. There's no need for such unclear code especially when we can
borrow the working code from qemuDomainShutdownFlags().

In addition, this fixes one bug too. If user requested both
VIR_DOMAIN_SHUTDOWN_INITCTL and VIR_DOMAIN_SHUTDOWN_SIGNAL at the same
time, he is basically saying: 'Use the force Luke! If initctl fails try
sending a signal.' But with the current code we don't do that. If
initctl fails for some reason (e.g. inability to write to /dev/initctl)
we don't try sending any signal but fail immediately. To make things
worse, making a domain shutdown with bare _SIGNAL was working by blind
chance of a @rc variable being placed at correct place on the stack so
its initial value was zero.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-12-24 17:36:47 +01:00
Michal Privoznik
5bd7ac029e interface: Take interface status into account when starting and destroying
https://bugzilla.redhat.com/show_bug.cgi?id=956994

Currently, it is possible to start an interface that is already running:

 # virsh iface-start eth2
 Interface eth2 started

 # echo $?
 0

 # virsh iface-start eth2
 Interface eth2 started

 # echo $?
 0

 # virsh iface-start eth2
 Interface eth2 started

 # echo $?
 0

Same applies for destroying a dead interface. We should not allow such
state transitions.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-12-24 17:20:00 +01:00
Michal Privoznik
50f5468c96 interface: Introduce netcfInterfaceObjIsActive
This function barely wraps ncf_if_status() and error handling code.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-12-24 17:20:00 +01:00
Michal Privoznik
f7fee15ae8 event-test: Unregister close callback
When registering a close callback, the connection refcount is increased
as the connection object is passed to the callback and hence we must
prevent deleting it too soon. However, when closing the connection, the
connection object is just unrefed. So whenever a connection with a close
callback is closed, we end up with the connection object which has
exactly one reference. Leaving the code as-is doesn't mean the end of
the world as we know it, but why give a bad example?

==14531== 288 bytes in 1 blocks are still reachable in loss record 695 of 762
==14531==    at 0x4C2BDE4: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==14531==    by 0x4E9FE09: virAllocVar (viralloc.c:558)
==14531==    by 0x4EDBE45: virObjectNew (virobject.c:190)
==14531==    by 0x4F71AAC: virGetConnect (datatypes.c:116)
==14531==    by 0x4F78511: do_open (libvirt.c:1136)
==14531==    by 0x4F7B3AC: virConnectOpenAuth (libvirt.c:1481)
==14531==    by 0x4011D2: main (event-test.c:499)

(and other leaks tied to virGetConnect())

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-12-24 17:18:34 +01:00
Michal Privoznik
e26cae6b85 examples: Resurrect domsuspend example
This partially reverts 5eb4b04211 and 62774afb6b.

Rewrite the domsuspend example from scratch. This time do it right.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2013-12-24 17:17:35 +01:00
Dario Faggioli
9375ba2a6e libxl: correctly handle affinity reset in virDomainPinVcpu[Flags]
By actually removing the <vcpupin> element (from within the
<cputune> section) from the XML, rather than jus update it with
a fully set vcpu affinity mask.

Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>
Cc: Jim Fehlig <jfehlig@suse.com>
Cc: Ian Jackson <Ian.Jackson@eu.citrix.com>
2013-12-23 22:43:32 -07:00
Dario Faggioli
520c3fbd1b libxl: implement virDomainPinVcpuFlags
And use it to implement libxlDomainPinVcpu(), similarly to what
happens in the QEMU driver. This way, it is possible to both
query and change the vcpu affinity of a persistent but not
running domain.

In face, before this patch, we have:
 # virsh list --all
  Id    Name                           State
 ----------------------------------------------------
  5     debian_32                      running
  -     fedora20_64                    shut off
 # virsh vcpupin fedora20_64 0 2-4 --current
 error: this function is not supported by the connection driver: virDomainPinVcpuFlags

After (same situation as above):
 # virsh vcpupin  fedora20_64 0 2-4 --current
 # virsh vcpupin  fedora20_64 0
 VCPU: CPU Affinity
 ----------------------------------
    0: 2-4

Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>
Cc: Jim Fehlig <jfehlig@suse.com>
Cc: Ian Jackson <Ian.Jackson@eu.citrix.com>
2013-12-23 22:33:30 -07:00
Dario Faggioli
2682d0d522 libxl: implement virDomainGetVcpuPinInfo
So that it is possible to query vcpu related information of
a persistent but not running domain, like it is for the QEMU
driver.

In fact, before this patch, we have:
 # virsh list --all
  Id    Name                           State
 ----------------------------------------------------
  5     debian_32                      running
  -     fedora20_64                    shut off
 # virsh vcpuinfo fedora20_64
 error: this function is not supported by the connection driver: virDomainGetVcpuPinInfo

After (same situation as above, i.e., fedora20_64 not running):
 # virsh vcpuinfo fedora20_64
 VCPU:           0
 CPU:            N/A
 State:          N/A
 CPU time        N/A
 CPU Affinity:   yyyyyyyy

 VCPU:           1
 CPU:            N/A
 State:          N/A
 CPU time        N/A
 CPU Affinity:   yyyyyyyy

Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>
Cc: Jim Fehlig <jfehlig@suse.com>
Cc: Ian Jackson <Ian.Jackson@eu.citrix.com>
2013-12-23 22:17:03 -07:00
Eric Blake
7cc3a7189c virt-login-shell: clean up usage
I noticed a few odd things in 'virt-login-shell --help' output.

* tools/virt-login-shell.c (usage): At most one option accepted,
drop trailing colon.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-23 21:53:48 -07:00
Cédric Bosdonnat
78af457eef virnettlscontexttest fails with GNUTLS 3.0.28
On openSUSE 12.x with GNUTLS 3.0.28, virnettlscontexttest fails. It has
been reported to work from GNUTLS 3.1.11 on Fedora 19. Changed the
constraints on gnutls to 3.1+ for unit test cacert4req.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-23 13:16:07 -07:00
Laine Stump
8700b87800 qemu: re-add hostdev interfaces to hostdev array on libvirtd restart
This resolves:

   https://bugzilla.redhat.com/show_bug.cgi?id=1045002

If a domain has an <interface type='hostdev'> or an <interface
type='network'> where the network itself is a pool of hostdev devices,
then libvirt will internally keep that device on both the interface
list *and* the hostdev list for the domain. One of the places this
comes in handy is when a new device is being added and libvirt wants
to find a unique "alias" name for it - it just scans through the
hostdev array and makes sure it picks a name that doesn't match the
alias of any device in that array.

However, when libvirtd was restarted, if there was an <interface
type='network'> with the network being a hostdev pool, the device
would not be added to the reconstructed internal hostdev array, so its
alias would not be found during a scan of the hostdev array, thus
attempts to add a new hostdev (or <interface type='hostdev'> or
<interface type='network'>) would result in a message like this:

internal error: unable to execute QEMU command 'device_add':
                Duplicate ID 'hostdev0' for device

This patch simply fixes the existing code in the domain XML parser
that fixes up the hostdev array in the case of <interface
type='hostdev'> to do the same thing in the case of <interface
type='network'> with a hostdev network.

This bug has existed since the very first addition of hostdev networks
to libvirt (0.10.0).
2013-12-23 13:16:57 +02:00
Laine Stump
c0f511ee2b qemu: avoid duplicate security label restore on hostdev attach failure
This eliminates the misleading error message that was being logged
when a vfio hostdev hotplug failed:

  error: unable to set user and group to '107:107' on '/dev/vfio/22':
         No such file or directory

as documented in:

  https://bugzilla.redhat.com/show_bug.cgi?id=1035490

Commit ee414b5d (pushed as a fix for Bug 1016511 and part of Bug
1025108) replaced the single call to
virSecurityManagerSetHostdevLabel() in qemuDomainAttachHostDevice()
with individual calls to that same function in each
device-type-specific attach function (for PCI, USB, and SCSI). It also
added a corresponding call to virSecurityManagerRestoreHostdevLabel()
in the error handling of the device-type-specific functions, but
forgot to remove the common call to that from
qemuDomainAttachHostDevice() - this resulted in a duplicate call to
virSecurityManagerRestoreHostdevLabel(), with the second occurrence
being after (e.g.) a PCI device has already been re-attached to the
host driver, thus destroying some of the device nodes / links that we
then attempted to re-label (e.f. /dev/vfio/22) and generating an error
log that obscured the original error.
2013-12-23 13:16:57 +02:00
Laine Stump
6d867f72f4 qemu: properly set MaxMemLock when hotplugging with VFIO
This resolves:

  https://bugzilla.redhat.com/show_bug.cgi?id=1035490

virProcessSetMaxMemLock() (which is a wrapper over prlimit(3)) expects
the memory size in bytes, but libvirt's domain definition (which was
being used by qemuDomainAttachHostPciDevice()) stores all memory
tuning parameters in KiB. This was being accounted for when setting
MaxMemLock at domain startup time (so cold-plugged devices would
work), but not for hotplug.

This patch simplifies the few lines that call
virProcessSetMemMaxLock(), and multiply the amount * 1024 so that
we're locking the correct amount of memory.

What remains a mystery to me is why hot-plug of a managed='no' device
would succeed (at least on my system) while managed='yes' would
fail. I guess in one case the memory was coincidentally already
resident and in the other it wasn't.
2013-12-23 13:16:57 +02:00
John Ferlan
d53c57b806 PanicCheckABIStability: Need to check for existence
Commit id '4313fead' added a call to virDomainPanicCheckABIStability()
which did not check whether the panic device existed before making a call
to virDomainDeviceInfoCheckABIStability() which ended up segfaulting:

Thread 1 (Thread 0x7f5332837700 (LWP 10964)):
    (src=<optimized out>, dst=<optimized out>)
    at conf/domain_conf.c:13007
    (dst=<optimized out>, src=<optimized out>)
    at conf/domain_conf.c:13712
    (src=<optimized out>, dst=<optimized out>)
    at conf/domain_conf.c:14056
    (domain=domain@entry=0x7f53000057c0, vm=vm@entry=0x7f53000036d0,
     defptr=defptr@entry=0x7f5332836978, snap=snap@entry=0x7f5332836970,
     update_current=update_current@entry=0x7f5332836962, flags=flags@entry=1)
    at conf/snapshot_conf.c:1230
    (domain=0x7f53000057c0, xmlDesc=<optimized out>, flags=1)
    at qemu/qemu_driver.c:12719
    (domain=domain@entry=0x7f53000057c0, xmlDesc=0x7f53000081d0
     "<domainsnapshot>\n  <name>snap2</name>\n
     <description>new-desc</description>\n  <state>running</state>\n
     <parent>\n    <name>snap1</name>\n  </parent>\n
     <creationTime>1387487268</creationTime>\n  <memory s"..., flags=1)
    at libvirt.c:19695
...

(gdb) up 3
(gdb) print *other->def->dom
$2 = {virtType = 2, id = -1, ..
...
  rng = 0x0, panic = 0x0, namespaceData = 0x0,...
...
(gdb) print *def->dom
$3 = {virtType = 2, id = -1, ...
...
  rng = 0x0, panic = 0x0, namespaceData = 0x0,...
...
(gdb)

Also seen using following sequence:

virsh save $dom $file
virsh save-image-edit $file
  add (or remove) a <panic/> line
  <devices>
  ...
    <panic>
      <address type='isa' iobase='0x505'/>
    </panic>
  ...
  </devices>
2013-12-21 09:49:58 -05:00
Bamvor Jian Zhang
b03eba1376 libxl: fix segfault when domain create fail
there is a segfault in libxl logging in libxl_ctx_free when domain
create fail. because the log output handler vmessage is freed by
xtl_logger_destroy before libxl_ctx_free in virDomainObjListRemove.
move xtl_logger_destroy after libxl_ctx_free could fix this bug.

Signed-off-by: Bamvor Jian Zhang <bjzhang@suse.com>
2013-12-20 11:49:24 -07:00
Dario Faggioli
f9ee91d355 libxl: avoid crashing if calling `virsh numatune' on inactive domain
by, in libxlDomainGetNumaParameters(), calling libxl_bitmap_init() as soon as
possible, which avoids getting to 'cleanup:', where libxl_bitmap_dispose()
happens, without having initialized the nodemap, and hence crashing after some
invalid free()-s:

 # ./daemon/libvirtd -v
 *** Error in `/home/xen/libvirt.git/daemon/.libs/lt-libvirtd': munmap_chunk(): invalid pointer: 0x00007fdd42592666 ***
 ======= Backtrace: =========
 /lib64/libc.so.6(+0x7bbe7)[0x7fdd3f767be7]
 /lib64/libxenlight.so.4.3(libxl_bitmap_dispose+0xd)[0x7fdd2c88c045]
 /home/xen/libvirt.git/daemon/.libs/../../src/.libs/libvirt_driver_libxl.so(+0x12d26)[0x7fdd2caccd26]
 /home/xen/libvirt.git/src/.libs/libvirt.so.0(virDomainGetNumaParameters+0x15c)[0x7fdd4247898c]
 /home/xen/libvirt.git/daemon/.libs/lt-libvirtd(+0x1d9a2)[0x7fdd42ecc9a2]
 /home/xen/libvirt.git/src/.libs/libvirt.so.0(virNetServerProgramDispatch+0x3da)[0x7fdd424e9eaa]
 /home/xen/libvirt.git/src/.libs/libvirt.so.0(+0x1a6f38)[0x7fdd424e3f38]
 /home/xen/libvirt.git/src/.libs/libvirt.so.0(+0xa81e5)[0x7fdd423e51e5]
 /home/xen/libvirt.git/src/.libs/libvirt.so.0(+0xa783e)[0x7fdd423e483e]
 /lib64/libpthread.so.0(+0x7c53)[0x7fdd3febbc53]
 /lib64/libc.so.6(clone+0x6d)[0x7fdd3f7e1dbd]

Signed-off-by: Dario Faggili <dario.faggioli@citrix.com>
Cc: Jim Fehlig <jfehlig@suse.com>
Cc: Ian Jackson <Ian.Jackson@eu.citrix.com>
2013-12-20 11:11:50 -07:00
Martin Kletzander
9faf3f2950 Fix crash in lxcDomainSetMemoryParameters
The function doesn't check whether the request is made for active or
inactive domain.  Thus when the domain is not running it still tries
accessing non-existing cgroups (priv->cgroup, which is NULL).

I re-made the function in order for it to work the same way it's qemu
counterpart does.

Reproducer:
 1) Define an LXC domain
 2) Do 'virsh memtune <domain> --hard-limit 133T'

Backtrace:
 Thread 6 (Thread 0x7fffec8c0700 (LWP 26826)):
 #0  0x00007ffff70edcc4 in virCgroupPathOfController (group=0x0, controller=3,
     key=0x7ffff75734bd "memory.limit_in_bytes", path=0x7fffec8bf718) at util/vircgroup.c:1764
 #1  0x00007ffff70e9206 in virCgroupSetValueStr (group=0x0, controller=3,
     key=0x7ffff75734bd "memory.limit_in_bytes", value=0x7fffe409f360 "1073741824")
     at util/vircgroup.c:669
 #2  0x00007ffff70e98b4 in virCgroupSetValueU64 (group=0x0, controller=3,
     key=0x7ffff75734bd "memory.limit_in_bytes", value=1073741824) at util/vircgroup.c:740
 #3  0x00007ffff70ee518 in virCgroupSetMemory (group=0x0, kb=1048576) at util/vircgroup.c:1904
 #4  0x00007ffff70ee675 in virCgroupSetMemoryHardLimit (group=0x0, kb=1048576)
     at util/vircgroup.c:1944
 #5  0x00005555557d54c8 in lxcDomainSetMemoryParameters (dom=0x7fffe40cc420,
     params=0x7fffe409f100, nparams=1, flags=0) at lxc/lxc_driver.c:774
 #6  0x00007ffff72c20f9 in virDomainSetMemoryParameters (domain=0x7fffe40cc420,
     params=0x7fffe409f100, nparams=1, flags=0) at libvirt.c:4051
 #7  0x000055555561365f in remoteDispatchDomainSetMemoryParameters (server=0x555555eb7e00,
     client=0x555555ec4b10, msg=0x555555eb94e0, rerr=0x7fffec8bfb70, args=0x7fffe40b8510)
     at remote_dispatch.h:7621
 #8  0x00005555556133fd in remoteDispatchDomainSetMemoryParametersHelper (server=0x555555eb7e00,
     client=0x555555ec4b10, msg=0x555555eb94e0, rerr=0x7fffec8bfb70, args=0x7fffe40b8510,
     ret=0x7fffe40b84f0) at remote_dispatch.h:7591
 #9  0x00007ffff73b293f in virNetServerProgramDispatchCall (prog=0x555555ec3ae0,
     server=0x555555eb7e00, client=0x555555ec4b10, msg=0x555555eb94e0)
     at rpc/virnetserverprogram.c:435
 #10 0x00007ffff73b207f in virNetServerProgramDispatch (prog=0x555555ec3ae0,
     server=0x555555eb7e00, client=0x555555ec4b10, msg=0x555555eb94e0)
     at rpc/virnetserverprogram.c:305
 #11 0x00007ffff73a4d2c in virNetServerProcessMsg (srv=0x555555eb7e00, client=0x555555ec4b10,
     prog=0x555555ec3ae0, msg=0x555555eb94e0) at rpc/virnetserver.c:165
 #12 0x00007ffff73a4e8d in virNetServerHandleJob (jobOpaque=0x555555ec3e30, opaque=0x555555eb7e00)
     at rpc/virnetserver.c:186
 #13 0x00007ffff7187f3f in virThreadPoolWorker (opaque=0x555555eb7ac0) at util/virthreadpool.c:144
 #14 0x00007ffff718733a in virThreadHelper (data=0x555555eb7890) at util/virthreadpthread.c:161
 #15 0x00007ffff468ed89 in start_thread (arg=0x7fffec8c0700) at pthread_create.c:308
 #16 0x00007ffff3da26bd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2013-12-20 12:47:14 +00:00
Martin Kletzander
f8c1cb9021 CVE-2013-6436: fix crash in lxcDomainGetMemoryParameters
The function doesn't check whether the request is made for active or
inactive domain.  Thus when the domain is not running it still tries
accessing non-existing cgroups (priv->cgroup, which is NULL).

I re-made the function in order for it to work the same way it's qemu
counterpart does.

Reproducer:
 1) Define an LXC domain
 2) Do 'virsh memtune <domain>'

Backtrace:
 Thread 6 (Thread 0x7fffec8c0700 (LWP 13387)):
 #0  0x00007ffff70edcc4 in virCgroupPathOfController (group=0x0, controller=3,
     key=0x7ffff75734bd "memory.limit_in_bytes", path=0x7fffec8bf750) at util/vircgroup.c:1764
 #1  0x00007ffff70e958c in virCgroupGetValueStr (group=0x0, controller=3,
     key=0x7ffff75734bd "memory.limit_in_bytes", value=0x7fffec8bf7c0) at util/vircgroup.c:705
 #2  0x00007ffff70e9d29 in virCgroupGetValueU64 (group=0x0, controller=3,
     key=0x7ffff75734bd "memory.limit_in_bytes", value=0x7fffec8bf810) at util/vircgroup.c:804
 #3  0x00007ffff70ee706 in virCgroupGetMemoryHardLimit (group=0x0, kb=0x7fffec8bf8a8)
     at util/vircgroup.c:1962
 #4  0x00005555557d590f in lxcDomainGetMemoryParameters (dom=0x7fffd40024a0,
     params=0x7fffd40027a0, nparams=0x7fffec8bfa24, flags=0) at lxc/lxc_driver.c:826
 #5  0x00007ffff72c28d3 in virDomainGetMemoryParameters (domain=0x7fffd40024a0,
     params=0x7fffd40027a0, nparams=0x7fffec8bfa24, flags=0) at libvirt.c:4137
 #6  0x000055555563714d in remoteDispatchDomainGetMemoryParameters (server=0x555555eb7e00,
     client=0x555555ebaef0, msg=0x555555ebb3e0, rerr=0x7fffec8bfb70, args=0x7fffd40024e0,
     ret=0x7fffd4002420) at remote.c:1895
 #7  0x00005555556052c4 in remoteDispatchDomainGetMemoryParametersHelper (server=0x555555eb7e00,
     client=0x555555ebaef0, msg=0x555555ebb3e0, rerr=0x7fffec8bfb70, args=0x7fffd40024e0,
     ret=0x7fffd4002420) at remote_dispatch.h:4050
 #8  0x00007ffff73b293f in virNetServerProgramDispatchCall (prog=0x555555ec3ae0,
     server=0x555555eb7e00, client=0x555555ebaef0, msg=0x555555ebb3e0)
     at rpc/virnetserverprogram.c:435
 #9  0x00007ffff73b207f in virNetServerProgramDispatch (prog=0x555555ec3ae0,
     server=0x555555eb7e00, client=0x555555ebaef0, msg=0x555555ebb3e0)
     at rpc/virnetserverprogram.c:305
 #10 0x00007ffff73a4d2c in virNetServerProcessMsg (srv=0x555555eb7e00, client=0x555555ebaef0,
     prog=0x555555ec3ae0, msg=0x555555ebb3e0) at rpc/virnetserver.c:165
 #11 0x00007ffff73a4e8d in virNetServerHandleJob (jobOpaque=0x555555ebc7e0, opaque=0x555555eb7e00)
     at rpc/virnetserver.c:186
 #12 0x00007ffff7187f3f in virThreadPoolWorker (opaque=0x555555eb7ac0) at util/virthreadpool.c:144
 #13 0x00007ffff718733a in virThreadHelper (data=0x555555eb7890) at util/virthreadpthread.c:161
 #14 0x00007ffff468ed89 in start_thread (arg=0x7fffec8c0700) at pthread_create.c:308
 #15 0x00007ffff3da26bd in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:113

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2013-12-20 12:46:47 +00:00
Eric Blake
3e67714e48 docs: improve event-related documentation
While looking at event code, I noticed that the documentation was
trying to refer me to functions that don't exist.  Also fix some
typos and poor formatting.

* src/libvirt.c (virConnectDomainEventDeregister)
(virConnectDomainEventRegisterAny)
(virConnectDomainEventDeregisterAny)
(virConnectNetworkEventRegisterAny)
(virConnectNetworkEventDeregisterAny): Link to correct function.
* include/libvirt.h.in (VIR_DOMAIN_EVENT_CALLBACK)
(VIR_NETWORK_EVENT_CALLBACK): Likewise.
(virDomainEventID, virConnectDomainEventGenericCallback)
(virNetworkEventID, virConnectNetworkEventGenericCallback):
Improve docs.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-19 09:45:50 -07:00
Eric Blake
3a96cebceb storage: fix bogus target in gluster volume xml
Commit 6cd60b6 was flat out broken - it tried to print into the
wrong variable.  My testing was obviously too cursory (did the
name get a slash added?); valgrind would have caught the error.
Thankfully it didn't hit any release.

Reported by Peter Krempa.

* src/storage/storage_backend_gluster.c
(virStorageBackendGlusterRefreshVol): Fix bogus code.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-19 09:17:11 -07:00
Daniel P. Berrange
147a3d2487 Remove stray semicolon after pragma macros
The VIR_WARNINGS_NO_CAST_ALIGN / VIR_WARNINGS_RESET should
not have any trailing ';' since they are pragmas. The use
of a ';' results in an empty statement which confuses CIL.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-12-19 16:02:38 +00:00
Cédric Bosdonnat
5e397d9c0f Fix crash in virsystemdtest with dbus 1.7.6
D-bus introduced some changes in its locking code. Overriding the init
function skips the new locking init and thus crashes later in libvirt
test. Removing the function makes the test pass again.
2013-12-19 11:20:05 +00:00
Peter Krempa
f17c968827 storage: Avoid forward declaration of virStorageVolDelete
Move the code around so that the forward declaration isn't needed. Also
fix code style of the opening brace of the function by moving it to a
separate line.
2013-12-19 11:53:41 +01:00
Peter Krempa
b560946c19 storage: Add gluster pool filter and fix virsh pool listing
Recent addition of the gluster pool type omitted fixing the virsh and
virConnectListAllStoragePool filters. A typecast of the converting
function in virsh showed that also the sheepdog pool was omitted in the
command parser.

This patch adds gluster pool filtering support and fixes virsh to
properly convert all supported storage pool types. The added typecast
should avoid doing such mistakes in the future.
2013-12-19 11:01:50 +01:00
Peter Krempa
32487c86c6 snapshot: schema: Split out snapshot disk driver definition
Extract the definition to a new type to allow avoiding of duplication.
2013-12-19 10:20:56 +01:00
Daniel P. Berrange
c6803a7312 Remove redefinition of bool type when --enable-test-locking
Old versions of CIL did not understand the 'bool' data type,
but at least 1.7.3 does now cope. We can remove the old hack
which redefined bool and no longer compiles successfully.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-12-18 17:03:35 +00:00
Daniel P. Berrange
7d3a1c8bd1 Fix race leading to crash when setting up dbus watches
Currently the virDBusAddWatch does

  virEventAddHandle(fd, flags,
                    virDBusWatchCallback,
                    watch, NULL);
  dbus_watch_set_data(watch, info, virDBusWatchFree);

Unfortunately this is racy - since the event loop is in a
different thread, the virDBusWatchCallback method may be
run before we get to calling dbus_watch_set_data. We must
reverse the order of these calls

See https://bugzilla.redhat.com/show_bug.cgi?id=885445

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-12-18 16:59:02 +00:00
Dario Faggioli
284d2fc299 libxl: libxl_get_max_cpus returning a libxl error from 4.4 onward
Starting from commit 2e82c18c in Xen (will be included in Xen 4.4)
both libxl_get_max_cpus() and libxl_get_max_nodes() start returning
a proper libxl error code, in case of failure. This patch fixes
this in the libxl driver.

Note that, although it is now basically impossible for them to return
0, that would, theoretically, still be wrong. Also, checking that the
returned value is '<= 0' makes the code correct for both Xen 4.4 and
Xen 4.3 (and 4.2), and that is why we go for it (rather than
just '< 0').

Signed-off-by: Dario Faggioli <dario.faggioli@citrix.com>
Cc: Jim Fehlig <jfehlig@suse.com>
Cc: Ian Jackson <Ian.Jackson@eu.citrix.com>
Cc: Martin Kletzander <mkletzan@redhat.com>
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2013-12-18 17:08:41 +01:00
Peter Krempa
ef56cc43d1 virsh: domain: Fix undefine with storage of 'volume' disks
The undefine code that removes the storage along with the VM didn't take
into account the existence of 'volume' type disks. Add the functionality.
2013-12-18 16:43:07 +01:00
Peter Krempa
878b8fa429 virsh: man: Mention that volumes need to be in storage pool for undefine
https://bugzilla.redhat.com/show_bug.cgi?id=1044445

When undefining a VM with storage the man page doesn't explicitly
mention that the volumes need to be a part of the storage pool otherwise
it won't work.
2013-12-18 16:42:10 +01:00
Eric Blake
a21cfb0f48 qemu: ask for -enable-fips when FIPS is required
On a system that is enforcing FIPS, most libraries honor the
current mode by default.  Qemu, on the other hand, refused to
honor FIPS mode unless you add the '-enable-fips' command
line option; worse, this option is not discoverable via QMP,
and is only present on binaries built for Linux.  So, if we
detect FIPS mode, then we unconditionally ask for FIPS; either
qemu is new enough to have the option and then correctly
cripple insecure VNC passwords, or it is so old that we are
correctly avoiding a FIPS violation by preventing qemu from
starting.  Meanwhile, if we don't detect FIPS mode, then
omitting the argument is safe whether the qemu has the option
(but it would do nothing because FIPS is disabled) or whether
qemu lacks the option (including in the case where we are not
running on Linux).

The testsuite was a bit interesting: we don't want our test
to depend on whether it is being run in FIPS mode, so I had
to tweak things to set the capability bit outside of our
normal interaction with capability parsing.

This fixes https://bugzilla.redhat.com/show_bug.cgi?id=1035474

* src/qemu/qemu_capabilities.h (QEMU_CAPS_ENABLE_FIPS): New bit.
* src/qemu/qemu_capabilities.c (virQEMUCapsInitQMP): Conditionally
set capability according to detection of FIPS mode.
* src/qemu/qemu_command.c (qemuBuildCommandLine): Use it.
* tests/qemucapabilitiestest.c (testQemuCaps): Conditionally set
capability to test expected output.
* tests/qemucapabilitiesdata/caps_1.2.2-1.caps: Update list.
* tests/qemucapabilitiesdata/caps_1.6.0-1.caps: Likewise.

Signed-off-by: Eric Blake <eblake@redhat.com>
2013-12-18 07:05:29 -07:00