OpenVZ creates per-container kernel threads and they live in
container's cgroups. CRIU should skip these threads, it's impossiable to
get any state of them.
Signed-off-by: Andrew Vagin <avagin@virtuozzo.org>
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This patch(set) is inspired by similar from Andrey Vagin sent
sime time earlier.

The major idea is to artificially fail criu dump or restore at
specific places and let zdtm tests check whether failed dump
or restore resulted in anything bad.

This particular patch introduces the ability to tell criu "fail
at X point". Each point is specified with a integer constant
and with the next patches there will appear places over the
code checking for specific fail code being set and failing.

Two points are introduced -- early on dump, right after loading
the parasite and right after creation of the root task.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This option cannot be tested using classic zdtm tests as it implies
some data created before restore and passed through criu restore down
to the restored process (descriptor in our case).

So add inhfd_test class that creates such.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

One for checkpoint another for restore.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

We don't check the return value here, so let's return void.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

As of
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=fb7589a1
tun support is merged, so let's check it in ms as well.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

As of
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=5442e9fbd
this feature is in mainline, so let's check for it in ms.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

As of
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=49d063cb3
the mnt_id field is in mainline, so let's check for it in --ms.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

The /dev/rtc device major value may change depending on the
architecture, and kernel release. To ensure the test rtc is working
all the time, we no longer belong on hard coded major value but peek
it from /dev/rtc itself.
Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Since crtools is no more linked with object build for the
parasite/restorer blob, it no more needs the builtin memory services.
Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Tree ways are possible -- always, never and only for tests that failed. The
latter one is the default.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This class is essentially libcriu in python(yet for now it
has only some basic functions such as check/dump/restore).
It is needed to make life for python users even more easier,
i.e. hiding some nasty connection stuff. It is also using
criu swrk(COMM_BIN) communication method, instead of an old
system service, as we no longer recommend it.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

1. As pointed out by Coverity (CID 114629), mnt_ns_fd is closed,
but then the function calls try_clean_remaps(mnt_ns_fd)
which tries to close the file descriptor which is already closed.

To address this, let's use safe_close() which sets closed fd to -1.
As it also checks its argument, there's no need for explicit check
so let's remove "if" check before close().

2. As Pavel pointed out, "calling the whole try_clean_remaps()
is not required once we've passed the cleanup_mnt_ns() point".
This could be addressed by introducing yet another label, but
it's cleaner to just use a flag variable.

Note that since the second issue is being addressed, the first one
goes away, but let's keep the fix for it anyway, it might help in
the future.
Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

When running tests, there are a lot of errors like this:

 (00.000053) Error (sysctl.c:367): Can't open sysctl net/ipv4/tcp_rmem: No such file or directory

As Andrew explained, tests are running in a netns that lacks
this sysctl. Use CTL_FLAGS_OPTIONAL flag to hide the error message.

Since with this flag sysctl_op() will most probably return 0,
add an additional check for vect[0] == 0 to detect that the file
was not read and show an appropriate warning, i.e.
"TCP mem sysctls are not available. Using defaults."
Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

I spent some time trying to figure out why vect array
is declared as [2][3], also why vect[0] is not used.
The answer is commit 4bca68ba that removed code using it.

Let's save a few bytes, and make the code more readable.
Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

As our tests serve as an example for people on how to preferably
use criu, we need to also switch to swrk.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Not sure if it is okay, but this patch breaks backward compatibility,
as we now discourage people from using criu as a system service by default.
But it is better to be done rather sooner than later, considering criu service
is not widely used.

On a patch side, we just need to daemonize swrk if self-dump is requested.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

As we've deleted them in previous two patches.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

We don't need it without systemd configs.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

As it is not recomended to run criu as a root system service, lets
remove service file to stop encouraging people to do it.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This patch change the build chain to not use pie objects in the crtools
executable.

This done by building the shared source files twice:
 1. for parasite/restorer as '<file>-pie-build.o'
 2. for crtools as '<file>.o'
Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Reviewed-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

The kernel prior 4.3 is exporting FS_EVENT_ON_CHILD
bit via procfs fdinfo interface. This bit is kernel's
internal and should not be passed in inotify_add_watch
call. Thus simply filter it out when obtain from old
images for backward compatibility reason.

More details here https://lkml.org/lkml/2015/9/21/680Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

open_mountpoint helper is called when mount namespace are
already restored so we have to use local paths.

This is a continuation to patch a082f27b -- we keep mount
point id in fsnotify, thus we should use the namespace's
path for open.

https://jira.sw.ru/browse/PSBM-39957Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

If criu restore failed, criu should wait all processes because they
hold files, namespaces and other stuff that caller might want to
have released (in our case it was ploop device).

Here we do this only for cases when processes are restored in a pid
namespace. We'd like to do the same for non-ns case, but there's
no simple way to wait for a bunch of unconnected processes.

Another good side effect is that "Restoring FAILED." will be printed
at the end of the log (now after we kill init tasks still have time
to do smth and write log messages).

Cc: Nikita Spiridonov <nspiridonov@odin.com>
Reported-by: Nikita Spiridonov <nspiridonov@odin.com>
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

We need to use SIG_SETMASK instead of SIG_BLOCK.

SIG_SETMASK
      The set of blocked signals is set to the argument set.
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Otherwise it's impossible to understand from which namespace
this path should be opened.
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

For slave mount we need to make sure that all childrent
mountpoints of the master peer is mounted before we can
proceed otherwise there gonna be duplicated mountpoints
after the restore.

https://jira.sw.ru/browse/PSBM-39957Signed-off-by: Andrew Vagin <avagin@odin.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This flag means that there's lib/ directory where a plugin is.
Currently rtc test is such. The wait_pid_die timeout is risen
since this test dies for more than 6 seconds :)
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Sometimes tests fail before generating .out file, so print its
contents only if it exists.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

I constantly have binfmt_misc off on my set ups :) Looks like this fs is
far from being default for systems, so to make the test work (we plan to
rewrite it anyway) let's make this fs being optional.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>