The virConnectGetDomainCapabilities API accepts either a binary path
to the emulator, or desired guest arch. If guest arch is not given,
then the host arch is assumed.

In the case where the binary is not given, the code tried to find the
emulator binary in the existing list of cached emulator capabilities.
This is not valid since we switched to lazy population of the cache in:

  commit 3dd91af0
  Author: Daniel P. Berrangé <berrange@redhat.com>
  Date:   Mon Dec 2 13:04:26 2019 +0000

    qemu: stop creating capabilities at driver startup

As a result of this change, if there are no persistent guests defined
using the requested guest architecture, virConnectGetDomainCapabilities
will fail to find an emulator binary.

The solution is to stop relying on the cached capabilities to find the
binary and instead use the same logic we use to pick default a binary
per arch when populating capabilities.
Tested-by: NBoris Fiuczynski <fiuczy@linux.ibm.com>
Tested-by: NRichard W.M. Jones <rjones@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The "ps2" bus is only available on certain machines like x86. On
machines like s390x, we should refuse to add a device to this bus
instead of silently ignoring it.

Looking at the QEMU sources, PS/2 is only available if the QEMU binary
has the "i8042" device, so let's check for that and only allow "ps2"
devices if this QEMU device is available, or if we're on x86 anyway
(so we don't have to fake the QEMU_CAPS_DEVICE_I8042 capability in
all the tests that use <input ... bus='ps2'/> in their xml data).
Reported-by: NSebastian Mitterle <smitterl@redhat.com>
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1763191Signed-off-by: NThomas Huth <thuth@redhat.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

LXC driver is not able to retrieve IP addresses from domains. This
function was not implemented yet. It can be done using DHCP lease and
ARP table. Different from QEMU, LXC does not have an agent to fetch
this info, but other sources can be used.
Signed-off-by: NJulio Faracco <jcfaracco@gmail.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

QEMU driver has two functions: qemuGetDHCPInterfaces() and
qemuARPGetInterfaces() that are being used inside only one single
function. They can be turned into generic functions that other drivers
can use. This commit move both from QEMU driver tree to domain conf
tree.
Signed-off-by: NJulio Faracco <jcfaracco@gmail.com>
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

Simplify function logic by using g_autofree to free local variables so
that we can remove some goto statements that are used for cleanup.

Introduce a g_autoptr cleanup function for virNodeDeviceDef.
Signed-off-by: NJonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

Since commit <60d9ad6f> we require
GnuTLS and since commit <ac0d21c7>
we can actually drop the usage of WITH_GNUTLS.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

If virHostdevManagerGetDefault in qemuhotplugtest fails it works
for quite a while to later segfault when accessing
mgr->activePCIHostdevs.

Report the error details and break on a failed init to see the
real issue right away.
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
Signed-off-by: NChristian Ehrhardt <christian.ehrhardt@canonical.com>

The libxl driver already tries to call shutdown inhibit callback in the
right places, but only if it's set. That last part was missing,
resulting in premature shutdown when running libvirtd
--timeout=...
Signed-off-by: NMarek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Reviewed-by: NJim Fehlig <jfehlig@suse.com>

You normally want to run the locally compiled copy of virsh.  Trying
to run the installed version with the locally compiled library is a
recipe for problems with missing symbols and so on.  By adding tools
to the path we can ensure that (eg) the libguestfs test suite will use
compatible copies of the library and virsh.
Signed-off-by: NRichard W.M. Jones <rjones@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

This has been used in libguestfs and libnbd for quite a while as it
makes the ./run script easier to read and write.

See also:
http://stackoverflow.com/a/9631350Signed-off-by: NRichard W.M. Jones <rjones@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

This function grabs an agent job but ends a monitor job.
End the agent job instead.

https://bugzilla.redhat.com/show_bug.cgi?id=1792723Signed-off-by: NJán Tomko <jtomko@redhat.com>
Reported-by: NDan Zheng <dzheng@redhat.com>
Fixes: e005c95f

This is only 2 simple typo fixes for wrong documentation wording.
Signed-off-by: NJulio Faracco <jcfaracco@gmail.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>
Signed-off-by: NJán Tomko <jtomko@redhat.com>

Running bootstrap and autoreconf from autogen.sh produced different
files in build-aux directory.  The reason is that gnulib usually have
newer version of these files and overwrites them after the autoreconf
step.

In order to fix it remove the --install and --force options, in addition
introduce --verbose option in order to reflect what bootstrap is doing.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

The existence of AUTHORS file is required for GNU projects but since
commit <8bfb36db> we do not require
these files to exist.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

We already ignore most of these files and the .gitignore files as well.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Preparation for having bootstrap result in git.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

We are in process of removing gnulib and adopting meson as our build
system.  In order to help with the transition let's drop gnulib tests.

This will also help with the fact that before we will be able to drop
gnulib completely we will store output of bootstrap in git.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

It is pulled in by tests and used by our build system as well.
Make an explicit dependency on threadlib.  This can be later removed
by using GLib GThread.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

We already use this function and so far we've been lucky that the same
check is done by gnulib.  This will change once we will drop gnulib and
also make it obvious that we have to do the same check in Meson as well.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

The function virSecretGetSecretString calls into secret driver and is
used from other hypervisors drivers and as such makes more sense in
util.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Reported at build time by lintian:
manpage-section-mismatch usr/share/man/man8/virt-sanlock-cleanup.8.gz:3 8 != 1

And indeed the rst file says 1 while the makefile say 8:
 if WITH_SANLOCK
   manpages8_rst += manpages/virt-sanlock-cleanup.rst
 else ! WITH_SANLOCK

8 "System administration commands and daemons" seems to match, so fix
the rst file to match.
Signed-off-by: NChristian Ehrhardt <christian.ehrhardt@canonical.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>

Remove many imports of sys/ioctl.h which are redundant,
and conditionalize remaining usage that needs to compile
on Windows platforms.

The previous change to remove the "nonblocking" gnulib
module indirectly caused the loss of the "ioctl" gnulib
module that we did not explicitly list in bootstrap.conf
despite relying on.

Rather than re-introduce the "ioctl" module this patch
makes it redundant.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

This fixes a build bug introduced by

  commit fbf27730
  Author: Daniel P. Berrangé <berrange@redhat.com>
  Date:   Mon Dec 16 11:16:51 2019 +0000

    conf: add support for specifying CPU "dies" parameter
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Add few test cases for nbd+unix style URIs with few corner cases.

The NBD URI syntax is documented at
https://github.com/NetworkBlockDevice/nbd/blob/master/doc/uri.mdSigned-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

When parsing legacy NBD backing file strings such as
'nbd:unix:/tmp/sock:exportname=/' we'd fail to set the transport to
VIR_STORAGE_NET_HOST_TRANS_UNIX. This started to be a problem once we
actually started to generate config of the backing store on the command
line with -blockdev as the JSON code would try to format it as TCP and
fail with:

 internal error: argument key 'host' must not have null value

Set the type properly and add a test.

This bug was found by the libguestfs test suite in:

https://bugzilla.redhat.com/show_bug.cgi?id=1791614Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reported-by: NMing Xie <mxie@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Tested-by: NRichard W.M. Jones <rjones@redhat.com>

 * send, recv: we use write & read for sockets so don't
   need these portability wrappers
 * ioctl, fcntl, fcntl-h: any usage of these is conditionally
   compiled and excludes Windows
 * ttyname_r: this exists in all supported platforms that
   we require now
 * environ: the tests explicitly declare this global variable
 * intprops: the code has been converted / simplified
 * nonblocking: we have a custom impl now to work with our
   own sockets wrappers
 * openpty: custom checks in configure.ac cope with portability
 * accept, bind, connect, getpeername, getsockname, listen,
   setsockopt, socket: code needing Windows portability uses
   our wrapper functions
 * close: avoids abort when passed invalid FD on Windows.
   Our VIR_FORCE_CLOSE wrapper avoids calling close(-1)
   and it is reasonable to abort in other scenarios in
   the RPC client
 * physmem: the gnulib code has been partially imported
 * warnings, manywarnings: copy the files directly into
   our local m4 dir
 * verify: replaced by G_STATIC_ASSERT
 * pthread_sigmask: none of the fixed portability problems
   affect libvirt's usage on current supported platforms
 * termios: the header is now conditionally included only
   when needed
 * time_r: replaced with GDateTime APIs
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

gmtime_r/localtime_r are mostly used in combination with
strftime to format timestamps in libvirt. This can all
be replaced with GDateTime resulting in simpler code
that is also more portable.

There is some boundary condition problem in parsing POSIX
timezone offsets in GLib which tickles our test suite.
The test suite is hacked to avoid the problem. The upsteam
GLib bug report is

  https://gitlab.gnome.org/GNOME/glib/issues/1999Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The GNULIB termios module ensures termios.h exists (but
is none the less empty) when building for Windows. We
already exclude usage of the functions that would exist
in a real termios.h, so having an empty termios.h is
not especially useful.

It is simpler to just put all use of termios.h related
functions behind a "#ifndef WIN32" conditional.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

G_STATIC_ASSERT() is a drop-in functional equivalent of
the GNULIB verify() macro.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Libvirt's original atomic ops impls were largely copied
from GLib's code at the time. The only API difference
was that libvirt's virAtomicIntInc() would return a
value, but g_atomic_int_inc was void. We thus use
g_atomic_int_add(v, 1) instead, though this means
virAtomicIntInc() now returns the original value,
instead of the new value.

This rewrites libvirt's impl in terms of g_atomic_int*
as a short term conversion. The key motivation was to
quickly eliminate use of GNULIB's verify_expr() macro
which is not a direct match for G_STATIC_ASSERT_EXPR.
Long term all the callers should be updated to use
g_atomic_int* directly.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

We don't need all the platforms gnulib deals with, so
this is a cut down version of GNULIB's physmem.c
code. This also allows us to integrate libvirt's
error reporting functions closer to the error cause.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Convert to use socket wrappers. Aside from the header file
include change, this requires changing close -> closesocket
since our portability isn't trying to replace the close
function.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Windows sockets take a SOCKET HANDLE object instead of a
file descriptor. Wrap them in the same way that gnulib
does so that they use C runtime file descriptors.

While we could in theory use GSocket, it is hard to get
the exact same semantics libvirt has for its current
socket usage. Wrapping the Winsock2 APIs is thus the
easiest approach in the short term.

In changing the socke wrappers we need to re-implement
the nonblocking function too, since the GNULIB impl
expects to be used with the GNULIB sockets wrappers.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

All UNIX platforms we care about have openpty() in the libutil
library. Use of pty.h must also be made conditional, excluding
Win32.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Some syntax check rules validate usage of headers provided
by gnulib. We want to validate these only against the gnulib
modules we've chosen to use, not all modules, since we're
trying to eliminate them.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Some UNIX platforms don't declare 'environ' in their
header files. We can unconditionally declare it ourselves
to avoid this problem.

There is no need to do this in the aa-helper code
since that is Linux only code.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The GLib g_size_checked_mul() function is not quite the
same signature, and gives compiler warnings due to not
correctly casting from gsize to guint64/32. Implementing
a replacement for INT_MULTIPLY_OVERFLOW is easy enough
to do ourselves.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Introduce a vastly simpler VIR_INT64_STR_BUFLEN constant
which is large enough for all cases where we currently
use INT_BUFSIZE_BOUND. This eliminates most use of the
gnulib intprops.h header.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>