Now they seem to look OK.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

The first is similar to existing snapshot test -- it calls
several pre-dumps on memory-changing zdtm test, then dumps
and check restore goes OK.

The second test is aimed at measuring frozen-time -- it
takes zdtm test with big mappings, dumps one, then start the
test again and calls one pre-dump plus dump. The sum of
frozen times in the 2nd dump (well, pre-dump + dump) should
be much smaller then the one of the 1st "just dump".
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

With this action criu will seize tasks, grab all its memory into
page-pipes, rest dirty tracker and will then release them. Writing
the memory from page-pipes would occur after tasks are unfreezed
and thus the frozen time should become reasonably small.

When pre-dump is in action, the dirty tracking is forcedly turned
off as well as tasks are resumed afterwards, not killed, by default.

This is a prerequisite for iterative migration.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Dumping memory is draining it from parasite, for pre-dump
this time would be reasonably small. _Writing_ the memory
would occur _after_ tasks unseize and resume.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

For pre-dump we'll need to cure the target task(s), but keep the
parasite_ctl (for its local args mapping) for memory image write.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

There will be more tests on memory snapshot.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

For pre-dump we won't need IDs, but will need items only,
thus make these two calls separate.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

They are not documented, thus OK for now. Two options --
* one to specify where the parent images are
* one to reset dirty memory tracking
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This is required for (soon to be) pre-dump command -- this command
will have to keep parasite args with pagemap (iovecs) some time after
parasite is executed. Since we call mprotect cmd _after_ pages dump
we can thus spoil these iovecs. To address this vmas to mprotect and
iovecs to dump are located in one parasite args area one after another
without intersections.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

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

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

Options are collected in global variable, no need to have them on stack.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Basic timings for dump (freeze, frozen, mem dump).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

It's not perfect, only the depth is taken into account,
not the field number, but this is OK for now.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

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

We'll have one more "image" file generated by dump and (surprisingly)
restore commands -- the stats one. It will contain in a single pb
object all the statistics collected by dump/restore.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

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

Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

...after commit 1e1bf240Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

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

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

and warn if it is used.

v2: fix help text as well
Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

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

This thing is global, we can address one explicitly.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

We can "guess" one w/o addirional (obfuscating) argument to
the pb showing engine.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

For consistency sake.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

The infected code must be executed in the ARM mode but the dumpee
may run in the Thumb mode while being dumped so the bit CPSR.T
may be set when the control is transfered to the infected blobs
so this code is executed in the Thumb mode too.

We have to clear the bit CPSR.T before transfering control
to the infected code to prevent the behavior described above.
Signed-off-by: Alexander Kartashov <alekskartashov@parallels.com>
Reported-by: Chanho Park <chanho61.park@samsusng.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

In case if we are migrating a session leader task with a solely
slave peer we should not try to restore a group on the terminal
but rather simply inherit it, because a session of the group of
the inherited terminal is not the same as a session of our leader
thus kernel refuse to set any different group here.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Since we're migrating session leader here, test for session has not
been changed instead.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

This is easier for debugging, the former SID/PGID remains untouched
and printed out in logs.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

If enabled CONFIG_VFP, we should use PTRACE_GETVFPREGS instead of GETFPREGS.
Most of arm hardwares(since then armv6) use VFP. If we need runtime checking of
vfp, we should parse /proc/cpu_info or use any other techniques.
Signed-off-by: Chanho Park <chanho61.park@samsung.com>
Signed-off-by: Myungjoo Ham <myungjoo.ham@samsung.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Avoid memleak in err case
Signed-off-by: Libo Chen <libo.chen@huawei.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>

Note to package maintainers:

1 Yes we do support DESTDIR.

2 From packaging scripts, use something like this:

	make install DESTDIR=$RPM_BUILD_ROOT PREFIX=/usr

(v2: add PREFIX, install to /usr/local not /usr by default)
(v3: fix criu man section -- it is 8 not 1)
Signed-off-by: Kir Kolyshkin <kir@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>