A pid namespace is created if a pid of the first task is 1.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When we dump pid ns, dump_pstree is called after dump_tasks
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
If we try to dump a process from another pid ns,
we have not another way to get its pid.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Add struct pid and use it everywhere. This struct contains
two fields: pid and real_pid.
real_pid is a pid outside of the target pid namespace.
pid is in the target pid namespace
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Otherwise its value might be taken from heap (it's xmalloc'ed).
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
- align members
- comment with empty line
No func changes.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's static function near the file_desc_ops,
no need for a long name.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Checkpoint and restore do use fdtype_ops
and file_desc_ops intensively so reorder
pipes functions, this allow us to make
them static and drop additional declarations.
No func changes.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
There is no need to make it xmalloc'ed, it's
global for all processes dumped.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
xemul: Why we haven't done this earlier? Because the pe entry contained the
amount of bytes in the pipe and this value had to be calculated before we
wrote this entry into the image. Now the pipe data is stored in another img
file and we can happily split pe dump and pde dump into two calls.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Will need it for fifo data dumping. At moment
the helper is pushed out of compilation procedure
for easier review, and will be real used in next
patch.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
And make it dep. on PIPE_DEF_BUFFERS, this points to
a limit from kernel side.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
sys_exit() takes down only the currect thread. exit_group() should
be used to take down all threads.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Looks-good-to: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
If address already in use then need increase port.
The problem occured because the socket_listen.c и socket_aio.c
can try to use the same address.
Signed-off-by: Victor Konyashkin <vkonyashkin@parallels.com>
Acked-by: Andrew Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
That is what it is -- a hash for file
descriptors based on their ID.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
* Create fifo and two descriptors rw and ro
* Write data and close rw descriptor
* Suspend/resume
* Read data from ro descriptor
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use them explicitly. This fixes problem
| LINK crtools
| /usr/lib/librt.so: could not read symbols: File in wrong format
| collect2: ld returned 1 exit status
| make[1]: *** [crtools] Error 1
| make: *** [all] Error 2
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Kinsbursky Stanislav <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Ad addition to commit 63000f5bb20e12ced2bc9662482775d163c97c5a,
this place was missed to fix.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
- add test for flags recovery
- drop sockets, not needed
- simplify test for owners and signumber
v2: remove polling (by avagin@)
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
SIGABRT may be blocked. We do that in the restorer.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
v5:
1) make change directory messages supressed.
v4:
1) Fixed apply conflicts.
v3:
1) "crtools" goal dependences fixed.
v2:
1) wrong "protobuf/tmp.pb-c.h" inclusion removed from cr-dump.c
Without this patch "touch parasite.c" will lead to rebuild of the whole
project.
This patch also introduces "pie" goal as a subsystem. This will be usefull,
when other subsystems will apper, because it allows to build them in parallel.
"all" becomes default goal. This allows to skip deps inclusion for empty,
"all", "zdtm" and "test-legacy" goals.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
v2: Synchronize the argument type of __alloc_pstree_item and the
values you put into it. I.e. int-int or bool-bool, not bool-int.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Taking into account the way the dump saves pstrees in the image.
If pstree.img isn't edited, a slow path should not be executed at all.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
because they describes a process TREE.
It's usefull, when we dump tasks from another pid namespace,
because a real pid is got from parasite. In previous version
we need to update pid in two places one is in a pstree_item and
one is in a children array.
A process tree will be necessery to restore sid and pgid,
because we should add fake tasks in a tree. For example if
a sesion leader is absent.
v2: fix rollback actions
v3: fix comments from Pavel Emelyanov
* add macros for_each_pstree_item
* and a few bugs
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Code to test unix dgarm sockets (extracted from
previous sockets00 test case).
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Code to test socketpair (extracted from
previous sockets00 test case).
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Don't mix various types of sockets in a single
test case, in case of problems it become harder
to fins which kind of socket failed.
The removed test cases will be addressed in
further patches.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The reason for this patch is that only phony goal can depend on phony targets.
But with protobuf we will have to execute phony targets prior to main target
build.
So, this patch set add few new goals and dependences:
1) "all" become default goal.
2) new goals "deps" and "preq" introduced.
3) "all" depends on "deps" and "crtools".
4) All global goals like "zdtm" and "test-legacy" now depends on "all".
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Instead of using wildcard expressions like "*.c" lets implicitely specify
desired dependence files.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Also, .PHONY have to be specified prior to desired goals,
And it's better to have one line for all goals instead on one line per goal.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Replace dances with "|" by simple dependence rule both for DEPS and OBJS.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use fdtype_ops facility to c/r inet sockets.
v2:
- Use BUG_ON if socket is attempted to be dumped
several times
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Use fdtype_ops facility to c/r unix sockets.
v2
- BUG_ON added in dump_one_unix_fd if socket
is already dumped since we never should dump
same socket several times
- The order of restore remains as it was before,
the lookup is done via socket inode numbers
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The task is not complete - this is just a part of what have to be done. I.e.
looks like a lot of excessive deps can be fixed.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Need increase buffer in msgque.c to avoid overflow via memcpy,
because now mtext[20] less than ANOTHER_TEST_STRING "Yet another test sysv5 msg".
Signed-off-by: Victor Konyashkin <vkonyashkin@parallels.com>
Acked-by: Andrey Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
