This commit is separate due to unusual paradigm compared to the
most source files.

Within whole vircgroup.c we 'return -errno', e.g. 'return -ENOMEM'.
However, in this specific function virCgroupAddTaskStrController
we weren't returning -ENOMEM but -1 despite fact that later in
the function we are returning one of errno values indeed.

Commit '18b14012' refactored the Xen code resulting in a Coverity
warning about possible NULL reference if the path where the XM driver
takes puts the def on it's list.  Moved/duplicated the virGetDomain()
call to pacify the possible NULL deref.

Only /dev/random and /dev/hwrng are accepted, list them explicitly.

Since f03dcc5d we use [::] as the listening address both on qemu
command line in -incoming and in nbd-server-start QMP command.
However the latter requires just :: without the braces.

Don't free the stream on error if we've successfully added it
to the hash table, since it will be freed by virChrdevHashEntryFree
callback.

Preserve the error message before calling virStreamFree, since it
resets the error.

Introduced by 47161382, crashing since 69218922.

Reported by Sergey Fionov on libvir-list.

Previously, we were freeing verr fields instead of rerr which we've
allocated just a line above.

In bf1fe848 I've introduced 'newName' variable to substitute the old
'const char *name' as previously we had an ugly code there:

  name = strdup(name);

However, some parts of the function were not updated, so they were still
calling VIR_FREE(name) instead of VIR_FREE(newName).

There is possibility to jump to 'cleanup' label without tapfd variable
being initialized. In the label, VIR_FORCE_CLOSE(tapfd) is called which
can have fatal consequences.

The same issue as (already fixed) in virDomainCreate -
REMOTE_PROC_DOMAIN_CREATE_WITH_FLAGS doesn't return new domain ID, only
-1 on error or 0 on success.

Besides this one fix it is more general problem - local domain object
ID can desynchronize with the real one, for example in case of another
client creates/destroys domain in the meantime. Perhaps virDomainGetID
should be called remotely (with all performance implications...)? Or
some event-based notification used?
Signed-off-by: NMarek Marczykowski <marmarek@invisiblethingslab.com>

Commit 16251193 forgot to include "virstring.h". Pushed under
build-breaker rule.

Don't print 'OPTION' if there's no options. Just behaves as DESCRIPTION
does.
This mostly affects 'interface' command group.
Signed-off-by: NZhang Xiaohe <zhangxh@cn.fujitsu.com>
Reported-by: NLi Yang <liyang.fnst@cn.fujitsu.com>
Signed-off-by: NEric Blake <eblake@redhat.com>

For future work we need _virDomainEventGraphicsAddress and
_virDomainEventGraphicsSubjectIdentity members to be char * not const
char *. We are strdup()-ing them anyway, so they should have been char *
anyway (from const correctness POV). However, we don't want users to
change passed values, so we need to make the callback's argument const.

Although this is an API change (not ABI though), real callers won't be
impacted. Why?
 1. these callback members are read-only, so it is less likely that
someone is trying to assign into the struct members.
 2. The only way to register a virConnectDomainEventGraphicsCallback is
to cast it through a call to virConnectDomainEventRegisterAny.  That is,
even if the user's callback function leaves out the const, we never use
the typedef as the direct type of any API parameter.  Since they are
already casting their function pointer into a munged type before
registering it, their code will continue to compile.

In my previous patches I enabled the IFF_MULTI_QUEUE flag every
time the user requested multiqueue TAP device. However, this
works only at runtime. During build time the flag may be
undeclared.

In order to learn libvirt multiqueue several things must be done:

1) The '/dev/net/tun' device needs to be opened multiple times with
IFF_MULTI_QUEUE flag passed to ioctl(fd, TUNSETIFF, &ifr);

2) Similarly, '/dev/vhost-net' must be opened as many times as in 1)
in order to keep 1:1 ratio recommended by qemu and kernel folks.

3) The command line construction code needs to switch from 'fd=X' to
'fds=X:Y:...:Z' and from 'vhostfd=X' to 'vhostfds=X:Y:...:Z'.

4) The monitor handling code needs to learn to pass multiple FDs.

Currently, we have one huge function to construct qemu command line.
This is very ineffective esp. if there's a fault somewhere.

This attribute is going to represent number of queues for
multique vhost network interface. This commit implements XML
extension part of the feature and add one test as well. For now,
we can only do xml2xml test as qemu command line generation code
is not adapted yet.

example: qemu ${otherargs} \
             -vnc 127.0.0.1:0,share=allow-exclusive

 -vnc :5900,share=allow-exclusive
allows clients to ask for exclusive access which is
implemented by dropping other connections Connecting
multiple clients in parallel requires all clients asking
for a shared session (vncviewer: -shared switch)

 -vnc :5900,share=force-shared
disables exclusive client access.  Useful for shared
desktop sessions, where you don't want someone forgetting
specify -shared disconnect everybody else.

 -vnc :5900,share=ignore
completely ignores the shared flag and allows everybody
connect unconditionally

QEMU_CAPS_VNC_SHARE_POLICY (qemu >= 1.1)

In my review of 31532cab I missed the fact that VIR_STRDUP
now returns 1 on success, and 0 if the source was NULL.

(This still doesn't add proper OOM error handling.)

Only a few cases are allowed:

1) The expression is empty for "for" loop, E.g.

  for (i = 0; ; i++)

2) An empty statement

  while (write(statuswrite, &status, 1) == -1 &&
         errno == EINTR)
      ; /* empty */

3) ";" is inside double-quote, I.e, as part of const string. E.g.

  vshPrint(ctl, "a ; b ; cd;\n");

The "for" loop in src/rpc/virnettlscontext.c is the special case,
1) applies for it, so change it together in this patch.

The more common habit is to add the comment after the statements.

Introduced by commit 7ac2c4fe, pushed under build-breaker rule.

virConnectListAllInterfaces should support to list all of
interfaces when the value of flags is 0. The behaviour is
consistent with other virConnectListAll* APIs

https://bugzilla.redhat.com/show_bug.cgi?id=965169 documents a
problem starting domains when cgroups are enabled; I was able
to reliably reproduce the race about 5% of the time when I added
hooks to domain startup by 3 seconds (as that seemed to be about
the length of time that qemu created and then closed a temporary
thread, probably related to aio handling of initially opening
a disk image).  The problem has existed since we introduced
virCgroupMoveTask in commit 91028296 (v0.10.0).

There are some inherent TOCTTOU races when moving tasks between
kernel cgroups, precisely because threads can be created or
completed in the window between when we read a thread id from the
source and when we write to the destination.  As the goal of
virCgroupMoveTask is merely to move ALL tasks into the new
cgroup, it is sufficient to iterate until no more threads are
being created in the old group, and ignoring any threads that
die before we can move them.

It would be nicer to start the threads in the right cgroup to
begin with, but by default, all child threads are created in
the same cgroup as their parent, and we don't want vcpu child
threads in the emulator cgroup, so I don't see any good way
of avoiding the move.  It would also be nice if the kernel were
to implement something like rename() as a way to atomically move
a group of threads from one cgroup to another, instead of forcing
a window where we have to read and parse the source, then format
and write back into the destination.

* src/util/vircgroup.c (virCgroupAddTaskStrController): Ignore
ESRCH, because a thread ended between read and write attempts.
(virCgroupMoveTask): Loop until all threads have moved.
Signed-off-by: NEric Blake <eblake@redhat.com>