All our supported Linux distros now have this header.
It has never existed on FreeBSD / macOS / Mingw.
Reviewed-by: NPavel Hrdina <phrdina@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

Signed-off-by: NDaniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: NMichal Privoznik <mprivozn@redhat.com>

We need to mock virCgroupV2DevicesAvailable() in order to remove any
dependency on kernel as BPF devices might not be available.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NDaniel Henrique Barboza <danielhb413@gmail.com>

'viralloc.h' does not provide any type or macro which would be necessary
in headers. Prevent leakage of the inclusion.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>

Quite a few of the tests have a need to mock the stat() / lstat()
functions and they are taking somewhat different & inconsistent
approaches none of which are actually fully correct. This is shown
by fact that 'make check' fails on 32-bit hosts. Investigation
revealed that the code was calling into the native C library impl,
not getting intercepted by our mocks.

The POSIX stat() function might resolve to any number of different
symbols in the C library.

The may be an additional stat64() function exposed by the headers
too.

On 64-bit hosts the stat & stat64 functions are identical, always
refering to the 64-bit ABI.

On 32-bit hosts they refer to the 32-bit & 64-bit ABIs respectively.

Libvirt uses _FILE_OFFSET_BITS=64 on 32-bit hosts, which causes the
C library to transparently rewrite stat() calls to be stat64() calls.
Libvirt will never see the 32-bit ABI from the traditional stat()
call. We cannot assume this rewriting is done using a macro. It might
be, but on GLibC it is done with a magic __asm__ statement to apply
the rewrite at link time instead of at preprocessing.

In GLibC there may be two additional functions exposed by the headers,
__xstat() and __xstat64(). When these exist, stat() and stat64() are
transparently rewritten to call __xstat() and __xstat64() respectively.
The former symbols will not actally exist in the library at all, only
the header. The leading "__" indicates the symbols are a private impl
detail of the C library that applications should not care about.
Unfortunately, because we are trying to mock replace the C library,
we need to know about this internal impl detail.

With all this in mind the list of functions we have to mock will depend
on several factors

 - If _FILE_OFFSET_BITS is set, then we are on a 32-bit host, and we
   only need to mock stat64 and __xstat64. The other stat / __xstat
   functions exist, but we'll never call them so they can be ignored
   for mocking.

 - If _FILE_OFFSET_BITS is not set, then we are on a 64-bit host and
   we should mock stat, stat64, __xstat & __xstat64. Either may be
   called by app code.

 - If __xstat & __xstat64 exist, then stat & stat64 will not exist
   as symbols in the library, so the latter should not be mocked.

The same all applies to lstat()

These rules are complex enough that we don't want to duplicate them
across every mock file, so this centralizes all the logic in a helper
file virmockstathelper.c that should be #included when needed. The
code merely need to provide a filename rewriting callback called
virMockStatRedirect(). Optionally VIR_MOCK_STAT_HOOK can be defined
as a macro if further processing is needed inline.
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

The mock fopen() function will abort if "/proc/mounts" is
requested with "r" permissions and VIR_CGROUP_MOCK_FILENAME
env var is not set.

Unfortunately this is triggering by the libselinux library
constructor when it tries to read /proc/mounts to find out
if selinuxfs is mounted in an unusual place.

This, however, only affects libselinux in Debian as that
opens with "r", while in Fedora / RHEL it opens "re" and
thus luckily never triggered the abort(), instead getting
an EACCESS.
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

In many files there are header comments that contain an Author:
statement, supposedly reflecting who originally wrote the code.
In a large collaborative project like libvirt, any non-trivial
file will have been modified by a large number of different
contributors. IOW, the Author: comments are quickly out of date,
omitting people who have made significant contribitions.

In some places Author: lines have been added despite the person
merely being responsible for creating the file by moving existing
code out of another file. IOW, the Author: lines give an incorrect
record of authorship.

With this all in mind, the comments are useless as a means to identify
who to talk to about code in a particular file. Contributors will always
be better off using 'git log' and 'git blame' if they need to  find the
author of a particular bit of code.

This commit thus deletes all Author: comments from the source and adds
a rule to prevent them reappearing.

The Copyright headers are similarly misleading and inaccurate, however,
we cannot delete these as they have legal meaning, despite being largely
inaccurate. In addition only the copyright holder is permitted to change
their respective copyright statement.
Reviewed-by: NErik Skultety <eskultet@redhat.com>
Signed-off-by: NDaniel P. Berrangé <berrange@redhat.com>

We need to create the cgroup v2 sysfs the same way as we do for
cgroup v1.

This introduces new VIR_CGROUP_MOCK_MODE env variable which will
configure which cgroup mode each test requires.  There are three
different modes:

    - legacy: only cgroup v1 is available and it's the default mode
    - hybrid: both cgroup v1 and cgroup v2 are available and have some
        controllers
    - unified: only cgroup v2 is available
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

Remove the trailing '/' from prefix.  This change is required in order
to introduce tests for unified cgroups.  They are usually mounted in
'/sys/fs/cgroup'.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

Move all the cgroup data into separate files out of vircgroupmock.c
and rework the fopen function to load data from files.  This will
make it easier to add more test cases.
Reviewed-by: NFabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

Reviewed-by: NFabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: NJán Tomko <jtomko@redhat.com>
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

Coverity noted that each of the fmemopen called used the strlen value
in order to allocate space, but that neglected space for terminating
null string. So just add 1 to the strlen.
Signed-off-by: NJohn Ferlan <jferlan@redhat.com>
Reviewed-by: NErik Skultety <eskultet@redhat.com>

All of the ones being removed are pulled in by internal.h. The only
exception is sanlock which expects the application to include <stdint.h>
before sanlock's headers, because sanlock prototypes use fixed width
int, but they don't include stdint.h themselves, so we have to leave
that one in place.
Signed-off-by: NErik Skultety <eskultet@redhat.com>
Acked-by: NMichal Privoznik <mprivozn@redhat.com>

It doesn't really make sense for us to have stdlib.h and string.h but
not stdio.h in the internal.h header.
Signed-off-by: NErik Skultety <eskultet@redhat.com>
Acked-by: NMichal Privoznik <mprivozn@redhat.com>

Clang complains about it:

  error: second argument to 'va_arg' is of promotable type
  'mode_t' (aka 'unsigned short'); this va_arg has undefined
  behavior because arguments will be promoted to 'int'
  [-Werror,-Wvarargs]

    mode = va_arg(ap, mode_t);
                      ^~~~~~

Work around the issue by passing int to va_arg() and casting
its return value to mode_t afterwards.
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>
Reviewed-by: NDaniel P. Berrangé <berrange@redhat.com>

Right-aligning backslashes when defining macros or using complex
commands in Makefiles looks cute, but as soon as any changes is
required to the code you end up with either distractingly broken
alignment or unnecessarily big diffs where most of the changes
are just pushing all backslashes a few characters to one side.

Generated using

  $ git grep -El '[[:blank:]][[:blank:]]\\$' | \
    grep -E '*\.([chx]|am|mk)$$' | \
    while read f; do \
      sed -Ei 's/[[:blank:]]*[[:blank:]]\\$/ \\/g' "$f"; \
    done
Signed-off-by: NAndrea Bolognani <abologna@redhat.com>

Found by running valgrind for these tests.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

It wasn't as great idea as I thought. Thing around stat() are
more complicated than that. Therefore we need to revert
86d1705a plus drop use of the macro as introduced in
later patches.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

There is some magic going on when it comes to stat() or lstat().
Basically, stat() can either be a regular function, an inline
function that calls __xstat(_STAT_VER, ...) or a macro that does
the same as the inline func. Don't ask why is that, just read the
documentation in sys/stat.h and make sure you have a bucket next
to you. Anyway, currently there will not be both stat and __xstat
symbols at the same time, as one of them gets overwritten to the
other one during compilation. But this is not true anymore once
we start chaining our mocking libraries. Therefore we need a
wrapper that calls desired function from glibc.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Instead of introducing our own wrapper for dlsym()
we can use the one provided by virmock.h.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

So in glibc-2.23 sys/sysmacros.h is no longer included from sys/types.h
and we don't build because of the usage of major/minor/makedev macros.
Autoconf already has AC_HEADER_MAJOR macro that check where exactly
these functions/macros are defined, so let's use that.
Signed-off-by: NMartin Kletzander <mkletzan@redhat.com>

There's been a report on the upstream list [1] describing we
access /sys/devices/system/cpu/present directly on the host from
within our test suite. This may end up in unpredictable results
as no all linux systems are required to have that file. Mock
access to the file.

libvirt.git/tests $ ../run strace vircgrouptest
...
access("/sys/devices/system/cpu/present", F_OK) = 0
...
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Instead of fakesysfsdir, which is very generic, use fakesysfspcidir and
fakesysfscgroupdir. This makes it explicit what part of the fake sysfs
filesystem they're referring to, and also leaves open the possibility of
handling files in two unrelated parts of the fake sysfs filesystem.

No functional changes.

The old name is no longer accurate, since now we're using its value as
the root of the fake filesystem.

No functional changes.

We might need to mock files living outside SYSFS_PREFIX later on,
so it's better to treat the temporary directory we are passed via
the environment as the root of the fake filesystem and create
SYSFS_PREFIX inside it.

The environment variable name will be changed to reflect the new use
we're making of it in a later commit.

Per-cpu stats are only shown for present CPUs in the cgroups,
but we were only parsing the largest CPU number from
/sys/devices/system/cpu/present and looking for stats even for
non-present CPUs.
This resulted in:
internal error: cpuacct parse error

Most of our code base uses space after comma but not before;
fix the remaining uses before adding a syntax check.

* tests/sysinfotest.c: Consistently use commas.
* tests/viratomictest.c: Likewise.
* tests/vircgroupmock.c: Likewise.
* tools/virsh-domain.c: Likewise.
* tools/virsh-volume.c: Likewise.
Signed-off-by: NEric Blake <eblake@redhat.com>

Currently, if access(path, mode) is invoked, we check if @path has this
special prefix SYSFS_PREFIX. If it does, we modify the path a bit and
call realaccess. If it doesn't we act just like a wrapper and call
realaccess directly. However, we are mocking fopen() as well. And as one
can clearly see there, fopen("/proc/cgroups") will succeed. Hence, we
have an error in our mocked access(): We need to check whether @path is
not equal to /proc/cgroups as it may not exists on real system we're
running however we definitely know how to fopen() it.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>

Debian systems may run the 'systemd-logind' daemon, which causes the
/sys/fs/cgroup/systemd  mount to be setup, but no other cgroup
controllers are created. While the LXC driver considers cgroups to
be mandatory, the QEMU driver is supposed to accept them as optional.

We detect whether they are present by looking in /proc/mounts for
any mounts of type 'cgroups', but this is not sufficient. We need to
skip any named mounts (as seen by a name=XXX string in the mount
options), so that we only detect actual resource controllers.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Some users in Ubuntu/Debian seem to have a setup where all the
cgroup controllers are mounted on /sys/fs/cgroup rather than
any /sys/fs/cgroup/<controller> name. In the loop which detects
which controllers are present for a mount point we were modifying
'mnt_dir' field in the 'struct mntent' var, but not always restoring
the original value. This caused detection to break in the all-in-one
mount setup.

Fix that logic bug and add test case coverage for this mount
setup.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Clang does not like it when you pass a static variable to an
inline function

 vircgroupmock.c:462:22: error: static variable 'fakesysfsdir' is
  used in an inline function with external linkage [-Werror,-Wstatic-in-inline]

Just make the var non-static to avoid this
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

If a user cgroup name begins with "cgroup.", "_" or with any of
the controllers from /proc/cgroups followed by a dot, then they
need to be prefixed with a single underscore. eg if there is
an object "cpu.service", then this would end up as "_cpu.service"
in the cgroup filesystem tree, however, "waldo.service" would
stay "waldo.service", at least as long as nobody comes up with
a cgroup controller called "waldo".

Since we require a '.XXXX' suffix on all partitions, there is
no scope for clashing with the kernel 'tasks' and 'release_agent'
files.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

If a cgroup controller is co-mounted with another, eg

   /sys/fs/cgroup/cpu,cpuacct

Then it is a requirement that there exist symlinks at

   /sys/fs/cgroup/cpu
   /sys/fs/cgroup/cpuacct

pointing to the real mount point. Add support to virCgroupPtr
to detect and track these symlinks
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Some aspects of the cgroups setup / detection code are quite subtle
and easy to break. It would greatly benefit from unit testing, but
this is difficult because the test suite won't have privileges to
play around with cgroups. The solution is to use monkey patching
via LD_PRELOAD to override the fopen, open, mkdir, access functions
to redirect access of cgroups files to some magic stubs in the
test suite.

Using this we provide custom content for the /proc/cgroup and
/proc/self/mounts files which report a fixed cgroup setup. We
then override open/mkdir/access so that access to the cgroups
filesystem gets redirected into files in a temporary directory
tree in the test suite build dir.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>