This is required to support checkpoint and restore of timers
that notify via file descriptors on ARM and AArch64.
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Check that files opened before switching to new mount namespace
remain in it after restore. Right now this is not so :( Andrey is
fixing the issue.
Christopher, can you check whether the ns_child's call to system()
works in your minimal set-ups (it launches cat and awk). If not,
then I should rewrite this routine in pure C.
The first version was written by Pavel (xemul@).
v2: don't use test_init_ns
don't call awk and cat
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
v2: add a comment before mntns_get_root_by_mnt_id(-1);
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Here is a problem with ghost files. Links are created on restore, but
they can't be created on any mount point, because a mount point can be
non-root bind-mount of another one. So we need to find the root mount
and create all links there.
v2: clean up
v3: add optimization for the case when both links on the same mount
point.
v4: don't look up mount points by mnt_id in a second time.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
On restore we add a temporary root to a mount point path. It's convinient
for restoring mount namespaces, but real paths are used for restoring
link-remap files.
v2: replace the offset field on a char * field
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The m->bind will be used to get the "lowest" mount. A mount is
called "lowest" in the chain of bind-mounts when its m->bind is
NULL. I.e. -- it's the mount of the root dir of an FS, all the
others are bind-mounted from it.
The propagate_siblings() is called for each new mount and sets
the bind field for others. A part of these other mounts can be
already mounted.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This modifies the x86 VDSO code to work on AArch64.
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
While it duplicates hundreds of lines of code, this is the
short term strategy Cyrill and I have agreed to for supporting
VDSOs across multiple architectures [1]. With better
understanding of where things differ per-architecture, or even
improved consolidation in the kernel, we can hopefully move to
a more shared implementation in the future.
1. http://lists.openvz.org/pipermail/criu/2014-August/015218.html
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Christopher Covington <cov@codeaurora.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
As it was reported, some systems don't use /etc/passwd.
On such systems getpwuid fails with undefined errno(see getpwuid(3))
not allowing criu to restrict ids with user additional groups.
Luckily, on such systems criu is run as root, so we can
just skip obtaining additional groups, as they don't matter
for root.
Reported-by: Christopher Covington <cov@codeaurora.org>
Signed-off-by: Ruslan Kuprieiev <kupruser@gmail.com>
Tested-by: Christopher Covington <cov@codeaurora.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This test creates 2 users to check how secure is using criu with setuid bit set.
Signed-off-by: Ruslan Kuprieiev <kupruser@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Currently check_ids is used to check uids only. So lets rename it
and drop crid argument, as we can use cr_uid directly.
Signed-off-by: Ruslan Kuprieiev <kupruser@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Currently, we only check if process gids match primary gid of user.
But process and user have additional groups too. So lets:
1) check that process rgid,egid and sgid are in the user's grouplist.
2) on restore check that user has all groups from the images.
Signed-off-by: Ruslan Kuprieiev <kupruser@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Otherwise we might take a look on last vma from previous
cycle not running vdso analisys at all.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Occasionally I ruined the ability to do a in-place remap for
pre 3.16 kernels. Bring it back.
CID 1230182: Logically dead code (DEADCODE)
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
No need for second if() statement, merge everything
in previous one.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When performing "make clean" rpc_pb2.pyc should be deleted as well.
Signed-off-by: Ruslan Kuprieiev <kupruser@gmail.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
openat() is used to open files, so absolute pathes can't be used
v2: change the comment
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Here is a bug now. A path for devices and directories should be
generated with a specified root.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
For that mnt namespaces should be dumped after files.
v2: rework enumeration of namespaces in dump_mnt_namespaces()
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Currently collect_shared is executed for each mount namespace separately,
so we miss the dependencies between namespaces.
For example, we have two namespaces, the first one contains a tmpfs mount
and the second one contains a non-root bind-mount of this tmpfs. Without
this patch this example can't be dumped.
On restore mnt_build_tree() is called for all namespaces at once, thus
this bug doesn't exist there.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In worst case we might be tryin to proxify vdso zone
when vvar only is remmaped but vdso itself is not yet
(left and right zones shifting). Thus vdso_proxify
will complain that vdso is not yet mapped refusing
to restore.
Thus wait until everything is re-mapped then call
for proxification helper.
Reported: Mr. Jenkins
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Andrew Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In latest linux-next the vdso zone is placed _after_ vvar
zone so eventually we need to handle any combination of
the following cases
- no vvar zone
- vvar before vdso
- vvar after vdso
Here we address all them.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Since we might have a several vDSO zones lets hide
handling in arch-specific routines.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In linux kernel 3.17 most probably the vvar and vdso zones will
be in reverse order, ie vvar first and vdso later so do extended
test for these VMAs coming in one bundle.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
- xfree works well with nil argument no need for additional tests.
- no need for @ret variable, we either success returning 0 explicitly,
either fail with explicit -1
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
