Currently each task subtracts number of zombies from
task_entries->nr_threads without locks, so if two tasks will do this
operation concurrently, the result may be unpredictable.
https://github.com/xemul/criu/issues/13
Cc: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Andrew Vagin <avagin@openvz.org>
Acked-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Users shouldn't use it directly anyway, they should
use criu_set* fucntions for that.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In case if socket's cwd lays on nested mount point
we might resolve its path a bit incorrectly
(mount_resolve_path helper should not obtain
paths with leading dot).
Thus send a path without leading dot for correct
name resolving.
Also add some error messages.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In this mode libcriu will execute criu binary in swrk
mode, so users are not always obliged to run criu service
daemon.
Signed-off-by: Ruslan Kuprieiev <rkuprieiev@cloudlinux.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use CRTOOLSVERSION instead of GITID to write criu.pc version. With
GITID, criu.pc had version '0' when built from tarballs.
Signed-off-by: Cédric Bosdonnat <cbosdonnat@suse.com>
Acked-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Make sure we don't pass any trash value here, because
the kernel does copy it explicitly. We allocate the
memory for frame as zero filled but stack segment
is special and zero is not acceptable (we've had
a discussion on LKML if we need a special handling
for zero ss but end up that new kernels need new CRIU
version, upon which all agreed). Finally in
commit 296bbf7e3 I managed to hit exactly
this problem :)
Reported-by: Andrey Wagin <avagin@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Andrey Wagin <avagin@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Also remove the cast of a pointer-to-void variable to the type
it already is.
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
* Added test for dumping\restoring of unnamed unix sockets.
Also test added to test/Makefile. Use make run for launch.
* Cleanup env after call run.sh in test/libcriu.
Signed-off-by: Artem Kuzmitskiy <artem.kuzmitskiy@lge.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
* Added functionality for dumping unnamed unix sockets.
When we call CRIU with dump option, for unnamed socket we
should pass it inode into --ext-unix-sk. Details about this problem
described in http://criu.org/External_UNIX_socket#What_to_do_with_socketpair.28.29-s.3F.
Usage example:
criu dump -D images -o dump.log -v4 --ext-unix-sk=4529709 -t 13506
* fix typo error in log output
Signed-off-by: Artem Kuzmitskiy <artem.kuzmitskiy@lge.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Unix sockets may be created with non-absolute (relative) path
(when kernel creates one it always use AT_FDCWD for name resolving),
So when we collect sockets we see them as having names without leading
slash.
In common cases for such sockets application doesn't change own
working directory after that but this is not always the true.
So we need to invent some name resolver. The good candidate is
IRMAP cache but after a number of testings I found that it might
slow down performance very dramatically. Thus we need some more
intelligent way here.
For a while, for common applications such as postfix, fetching
dumpee working directory and root is enough. So here what we do
- when socket get collected from diag interface we remember
its relative name parameters (device and inode) but postprone
name resolving to not bring perf penalty until really needed
- when we meet a socket to dump with relative name assigned we
try to use $cwd/name and $root/name for this socket to check
if it has been created in those directories. On success we
simply remember the directory in image and when restore such
socket call for chdir helper to change working dir and generate
relative name
v2:
- Use new unlink_stale to remove sockets we're to restore
- Use *at() helpers once we're changed working dir in bind_unix_sk
- Add more debug ouput
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It gonna be extende to support relative names.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In the case where there were multiple clients for a dgram socket, we were
restoring the queue for each client. Instead, we should pick one client and
she should restore the queue while the rest skip it.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The old and new address parameters passed to the mremap system
call must be page size aligned. On AArch64, the page size can
only be correctly determined at run time. This fixes the following
errors for CRIU on AArch64 kernels with CONFIG_ARM64_64K_PAGES=y.
call mremap(0x3ffb7d50000, 8192, 8192, MAYMOVE | FIXED, 0x2a000)
Error (rst-malloc.c:201): Can't mremap rst mem: Invalid argument
call mremap(0x3ffb7d90000, 8192, 8192, MAYMOVE | FIXED, 0x32000)
Error (rst-malloc.c:201): Can't mremap rst mem: Invalid argument
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This fixes the following error for CRIU on AArch64 kernels with
CONFIG_ARM64_64K_PAGES=y.
Error (cr-restore.c:2828): Can't mmap section for restore code
This occurred because the address being requested (0x16000 in
one case) was not page aligned.
Also change the capitalization of the pie_size() macro to make it
clear that the value is not necessarily a build-time constant.
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We use native cpuid, so this one is no longer used.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Add checkpoint and restore of the Altivec and VSX registers.
Currently we rely on the return value of ptrace to detect if the CPU is
supporting these features or not. In the future, we should rely on the
AT_HWCAP vector and check feature at restart time.
Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use hex output for the whole module with proper 0x prefix.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This will help to support compat mode in the future.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This patch adds capability to using libcriu from C++ code.
Signed-off-by: Artem Kuzmitskiy <artem.kuzmitskiy@lge.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When a TASK_HELPER would exit just before a zombie, sometimes the signal
would get coalesced, and we would miss the zombie exit, causing us to block
forever waiting for the zombie to complete. Let's use an entirely different
strategy for waiting on zombies: explicitly wait on them with waitid, and
use WNOWAIT to prevent their data from actually being reaped.
v2: don't decrement nr_{tasks,threads} in the loop
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Acked-by: Andrew Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We'll use this in the next patch for collecting the zombies without
actually waiting on them.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This saves some keystrokes and is equivalent to decode --pretty.
Signed-off-by: Tycho Andersen <tycho.andersen@canonical.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In criu 1.6 if no --manage-cgroups option been specified
we still restore default (known) properties. But in commit
c7d646afb373 we've enhanced its semantics occasionally break
backward compatibility: if no --manage-cgroups passed at all
it's assumed that one asks to not touch cgroups at all on
restore. To restore old behaviour setup "soft" mode by
default.
Reported-by: Andrew Vagin <avagin@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The kernel doesnt allow to call lseek on /dev/kmsg
if it has been opened with O_WRONLY mode so our
restore procedure fails.
Thus if we meet a file which fits the condition
above -- set it's position to predefined value
which tells criu to not call lseek on restore.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
mountinfo contains mangled paths. space, tab and back slash were
replaced with usual octal escape, so we need to replace these charecter
back.
v2: declare cure_path as static
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
For example:
mount -t tmpfs "" test
v2: don't leak memory
Reported-by: Ross Boucher <boucher@gmail.com>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
