This reverts commit e4b980c8.

When a binary links against a .a archive (as opposed to a shared library),
any symbols which are marked as 'weak' get silently dropped. As a result
when the binary later runs, those 'weak' functions have an address of
0x0 and thus crash when run.

This happened with virtlogd and virtlockd because they don't link to
libvirt.so, but instead just libvirt_util.a and libvirt_rpc.a. The
virRandomBits symbols was weak and so left out of the virtlogd &
virtlockd binaries, despite being required by virHashTable functions.

Various other binaries like libvirt_lxc, libvirt_iohelper, etc also
link directly to .a files instead of libvirt.so, so are potentially
at risk of dropping symbols leading to a later runtime crash.

This is normal linker behaviour because a weak symbol is not treated
as undefined, so nothing forces it to be pulled in from the .a You
have to force the linker to pull in weak symbols using -u$SYMNAME
which is not a practical approach.

This risk is silent bad linkage that affects runtime behaviour is
not acceptable for a fix that was merely trying to fix the test
suite. So stop using __weak__ again.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

This reverts commit b9473d8b.

Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

When libvirt starts a new QEMU domain, it replaces host-model CPUs with
the appropriate custom CPU definition. However, when reconnecting to a
domain started by older libvirt (< 2.3), the domain would still have a
host-model CPU in its active definition.

https://bugzilla.redhat.com/show_bug.cgi?id=1463957Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

qemuProcessReconnect will need to call additional functions which were
originally defined further in qemu_process.c.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Separated from qemuProcessUpdateAndVerifyCPU to handle updating of an
active guest CPU definition according to live data from QEMU.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

In addition to updating a guest CPU definition the function verifies
that all required features are provided to the guest. Let's make it
obvious by calling it qemuProcessUpdateAndVerifyCPU.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Separated from qemuProcessUpdateLiveGuestCPU. The function makes sure
a guest CPU provides all features required by a domain definition.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Separated from qemuProcessUpdateLiveGuestCPU. Its purpose is to fetch
guest CPU data from a running QEMU process. The data can later be used
to verify and update the active guest CPU definition.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

CPU features unknown to a hypervisor will not be present in dataDisabled
even though the features won't naturally be enabled because.
Thus any features we asked for which are not in dataEnabled should be
considered disabled.
Signed-off-by: NJiri Denemark <jdenemar@redhat.com>

When checking ABI stability between two domain definitions, we first
make migratable copies of them. However, we also asked for the guest CPU
to be updated, even though the updated CPU is supposed to be already
included in the original definitions. Moreover, if we do this on the
destination host during migration, we're potentially updating the
definition with according to an incompatible host CPU.

While updating the CPU when checking ABI stability doesn't make any
sense, it actually just worked because updating the CPU doesn't do
anything for custom CPUs (only host-model CPUs are affected) and we
updated both definitions in the same way.

Less then a year ago commit v2.3.0-rc1~42 stopped updating the CPU in
the definition we got internally and only the user supplied definition
was updated. However, the same commit started updating host-model CPUs
to custom CPUs which are not affected by the request to update the CPU.
So it still seemed to work right, unless a user upgraded libvirt 2.2.0
to a newer version while there were some domains with host-model CPUs
running on the host. Such domains couldn't be migrated with a user
supplied XML since libvirt would complain:

    Target CPU mode custom does not match source host-model

The fix is pretty straightforward, we just need to stop updating the CPU
when checking ABI stability.

https://bugzilla.redhat.com/show_bug.cgi?id=1463957Signed-off-by: NJiri Denemark <jdenemar@redhat.com>
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>

Currently the scan of the /proc/mounts file used to find cgroup mount
points doesn't take into account that mount points may hidden by other
mount points. For, example in certain Kubernetes environments the
/proc/mounts contains the following lines:

  cgroup /sys/fs/cgroup/net_prio,net_cls cgroup ...
  tmpfs /sys/fs/cgroup tmpfs ...
  cgroup /sys/fs/cgroup/net_cls,net_prio cgroup ...

In this particular environment the first mount point is hidden by the
second one. The correct mount point is the third one, but libvirt will
never process it because it only checks the first mount point for each
controller (net_cls in this case). So libvirt will try to use the first
mount point, which doesn't actually exist, and the complete detection
process will fail.

To avoid that issue this patch changes the virCgroupDetectMountsFromFile
function so that when there are duplicates it takes the information from
the last line in /proc/mounts. This requires removing the previous
explicit condition to skip duplicates, and adding code to free the
memory used by the processing of duplicated lines.

Related-To: https://bugzilla.redhat.com/1468214
Related-To: https://github.com/kubevirt/libvirt/issues/4Signed-off-by: NJuan Hernandez <jhernand@redhat.com>

After 426dc5eb qemuCaps and virDomainDefPtr are unused here,
remove it from the call stack
Reviewed-by: NAndrea Bolognani <abologna@redhat.com>
Signed-off-by: NCole Robinson <crobinso@redhat.com>

Obviously, old gcc-s ale sad when a variable shares the name with
a function. And we do have such variable (added in 4d8a914b):
@mount. Rename it to @mountpoint so that compiler's happy again.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

The way we create devices under /dev is highly linux specific.
For instance we do mknod(), mount(), umount(), etc. Some
platforms are even missing some of these functions. Then again,
as declared in qemuDomainNamespaceAvailable(): namespaces are
linux only. Therefore, to avoid obfuscating the code by trying to
make it compile on weird platforms, just provide a non-linux stub
for qemuDomainAttachDeviceMknodRecursive(). At the same time,
qemuDomainAttachDeviceMknodHelper() which actually calls the
non-existent functions is moved under ifdef __linux__ block since
its only caller is in that block too.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel P. Berrange <berrange@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1467826

Commit id 'b9b1aa63' was supposed to add logic to set the allocation
for sparse files when wr_highest_offset was zero; however, an unconditional
setting was done just prior. For block devices, this means allocation is
always returning 0 since 'actual-size' will be zero.

Remove the unconditional setting and add the note about it being possible
to still be zero for block devices. As soon as the guest starts writing to
the volume, the allocation value will then be obtainable from qemu via
the wr_highest_offset.

The API documents that it peeks into the VM disk. We can't do that
currently for non raw images so report an error.

Refactor the access to storage driver usage along with
qemuDomainStorageFileInit which ensures that we access the file with
correct DAC uid/gid.

Allow specifying offset to read an arbitrary position in the file. This
warrants a rename to virStorageFileRead.

The helper methods for actually accessing the storage objects don't
really belong to the main storage driver implementation file. Split them
out.

The Win32 platform will fail to link if you use weak symbols
because it is incompatible with exporting symbols in a DLL:

Cannot export virRandomGenerateWWN: symbol wrong type (2 vs 3)

We only need weak symbols for our test suite to do LD_PRELOAD
and this doesn't work on Win32, so we can just drop the hack
for Win32
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

If we exceed a fixed limit in RPC code we get a horrible message
like this, if the parameter type is a 'string', because we forgot
to initialize the error message type field:

  $ virsh snapshot-list ostack1
  error: too many remote undefineds: 1329 > 1024

It would also be useful to know which RPC call and field was
exceeded. So this patch makes us report:

  $ virsh snapshot-list ostack1
  error: too many remote undefineds: 1329 > 1024,
  in parameter 'names' for 'virDomainSnapshotListNames'
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

On domain startup, bind host or bind service can be omitted
and we will format a working command line.

Extend this to hotplug as well and specify the service to QEMU
even if the host is missing.

https://bugzilla.redhat.com/show_bug.cgi?id=1452441

We have a temporary pointer to the currently processed parameter.
Use it to save three bytes per use.

We assign the unsigned long value of the currently processed
parameter to a temporary value_ul variable. Use it consistently
in all cases.

Currently all mockable functions are annotated with the 'noinline'
attribute. This is insufficient to guarantee that a function can
be reliably mocked with an LD_PRELOAD. The C language spec allows
the compiler to assume there is only a single implementation of
each function. It can thus do things like propagating constant
return values into the caller at compile time, or creating
multiple specialized copies of the function body each optimized
for a different caller. To prevent these optimizations we must
also set the 'noclone' and 'weak' attributes.

This fixes the test suite when libvirt.so is built with CLang
with optimization enabled.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The TODO macro expands to an fprintf() call and is used in several
places in the Xen driver. Anything that wishes to print such debug
messages should use the logging macros. In this case though, all the
places in the Xen driver should have been raising a formal libvirt
error instead. Add proper error handling and delete the TODO macro
to prevent future misuse.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

The HOST_NAME_MAX, INET_ADDRSTRLEN and VIR_LOOPBACK_IPV4_ADDR
constants are only used by a handful of files, so are better
kept in virsocketaddr.h or the source file that uses them.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

We only ever test libvirt with GCC or CLang which provides a
GCC compatible compilation environment. Between them, these
compilers cover every important operating system platform,
even Windows.

Mandate their use to make it explicit that we don't care about
compilers like Microsoft VCC or other UNIX vendor C compilers.

GCC 4.4 was picked as the baseline, since RHEL-6 ships 4.4.7
and that lets us remove a large set of checks. There is a slight
issue that CLang reports itself as GCC 4.2, so we must also check
if __clang__ is defined. We could check a particular CLang version
too, but that would require someone to figure out a suitable min
version which is fun because OS-X reports totally different CLang
version numbers from CLang builds on Linux/BSD
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Back in this commit:

  commit b436a8ae
  Author: Fabian Freyer <fabian.freyer@physik.tu-berlin.de>
  Date:   Thu Jun 9 00:50:35 2016 +0000

    gnulib: add getopt module

config-post.h was modified to define __GNUC_PREREQ, but the
original definition was never removed from internal.h, and
that is now dead code since config.h is always the first file
included.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Currently, the only type of chardev that we create the backend
for in the namespace is type='dev'. This is not enough, other
backends might have files under /dev too. For instance channels
might have a unix socket under /dev (well, bind mounted under
/dev from a different place).
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1462060

Just like in the previous commit, when attaching a file based
device which has its source living under /dev (that is not a
device rather than a regular file), calling mknod() is no help.
We need to:

1) bind mount device to some temporary location
2) enter the namespace
3) move the mount point to desired place
4) umount it in the parent namespace from the temporary location

At the same time, the check in qemuDomainNamespaceSetupDisk makes
no longer sense. Therefore remove it.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1462060

When building a qemu namespace we might be dealing with bare
regular files. Files that live under /dev. For instance
/dev/my_awesome_disk:

  <disk type='file' device='disk'>
    <driver name='qemu' type='qcow2'/>
    <source file='/dev/my_awesome_disk'/>
    <target dev='vdc' bus='virtio'/>
  </disk>

  # qemu-img create -f qcow2 /dev/my_awesome_disk 10M

So far we were mknod()-ing them which is
obviously wrong. We need to touch the file and bind mount it to
the original:

1) touch /var/run/libvirt/qemu/fedora.dev/my_awesome_disk
2) mount --bind /dev/my_awesome_disk /var/run/libvirt/qemu/fedora.dev/my_awesome_disk

Later, when the new /dev is built and replaces original /dev the
file is going to live at expected location.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Currently, we silently assume that file we are creating in the
namespace is either a link or a device (character or block one).
This is not always the case. Therefore instead of doing something
wrong, claim about unsupported file type.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Currently, we silently assume that file we are creating in the
namespace is either a link or a device (character or block one).
This is not always the case. Therefore instead of doing something
wrong, claim about unsupported file type.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

This function is going to be used on other places, so
instead of copying code we can just call the function.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

https://bugzilla.redhat.com/show_bug.cgi?id=1459592

In 290a00e4 I've tried to fix the process of building a
qemu namespace when dealing with file mount points. What I
haven't realized then is that we might be dealing not with just
regular files but also special files (like sockets). Indeed, try
the following:

1) socat unix-listen:/tmp/soket stdio
2) touch /dev/socket
3) mount --bind /tmp/socket /dev/socket
4) virsh start anyDomain

Problem with my previous approach is that I wasn't creating the
temporary location (where mount points under /dev are moved) for
anything but directories and regular files.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

It comes very handy to have source path for chardevs. We already
have such function: virDomainAuditChardevPath() but it's static
and has name not suitable for exposing. Moreover, while exposing
it change its name slightly to virDomainChrSourceDefGetPath.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NJohn Ferlan <jferlan@redhat.com>

Now that the JSON deflattener is working sanely we can always attempt
the deflattening so that we can then parse the tree as expected.