This is a big change, yes. Dump unix sockets in the same manner
as all the other files are done now. A few notes however.
1. We explicitly drop names for connected stream sockets. This is
done to avoid conflicts with names -- accepted sockets share their
names with the listening parent. This can be done later by binding
a socket to a name, them renaming it to some temporary uniq one
and at the very very end renaming some back to original.
2. Interconnected sockets are restored via socketpair() call. This is
correct, but names are dropped. Need to bind() sockets after this
(yes, this can be done), but for this we need to implement the trick
with renames described before.
3. FD for socket queues is constantly re-opened not to resolve fd
conflicts. Need to use service fds engine for this later.
4. Some code cleanup is still required, yes (will follow shortly).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
A pipe buffer has 16 slots. A slot is page, offset and size.
When we use splice and data is not aligned, splice connects
a page from file cache and set offset. For this reason we loose
a part of buffer.
If a data size is more than 15 pages, data will be aligned in a image.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Information about pipe's file structs saved in one global file and
fdinfo_entry is saved for each descriptor
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This was required when pages were stored in elf files for
exec. Now we can stop reading it on eof.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Now every inetsk fd dump results in a new entry in the fdinfo.img file. Sockets itself are
dumped into inetsk.img global image file. On restore the generic fdinfo redistribution algo
is used and inet sockets are opened only when required.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Each fdset item now has the callback which will show a contents of a magic-described
image file. Per-task and global show code is reworked to walk the respective fdsets
and calling ->show on each file.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
After we removed the pid from pstree image file the -t or -p option for show
command no longer makes sense. Make 'show' mode rely on -D option to find out
where to find the root (i.e. pstree.img) file.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This contains reg-files and sk-queues images, as they contain data
which is potentially generated by every task, so keep it open all
the time dump goes.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Current fdsets are ugly, limited (bitmask will exhaust in several months) and
suffer from unknown problems with fdsets reuse :(
With new approach (this set) the images management is simple. The basic function
is open_image, which gives you an fd for an image. If you want to pre-open several
images at once instead of calling open_image every single time, you can use the
new fdsets.
Images CR_FD_ descriptors should be grouped like
_CR_FD_FOO_FROM,
CR_FD_FOO_ITEM1,
CR_FD_FOO_ITEM2,
..
CR_FD_FOO_ITEMN,
_CR_FD_FOO_TO,
After this you can call cr_fd_open() specifying ranges -- _FROM and _TO macros,
it will give you an cr_fdset object. Then the fdset_fd(set, type) will give you
the descriptor of the open "set" group corresponding to the "type" type.
3 groups are introduced in this set -- tasks, ns and global.
That's it.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Write two helpers for opening an fdset for task and one for ns.
This probably can be done with some "generic" macro(s), but this
time it's simpler not to produce more code of that type.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's not required any longer. Now fdsets are allocated one-by-one only
when required and there's no need in adding new fds to existing sets.
Thus just remove the last arg from cr_fdset_open.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This fd is global, so make it such. It will stop being just a global
variable soon.
Plus, remove the pid arg from format.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Pid number is redundant - this file is one for the whole tree.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Open the exec link at fd restore stage as yet another service fd,
then pass it to restover via args and just call prctl on it.
This is good for several reasons -- the amount of code required for
this is less and opening files should better happen before we switch
to restorer (opening will be complex and it's MUCH easier to open all
we need in one place).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The core image now contains only core per-task stuff.
The new file resurrects Tula magic number removed earlier.
Acked-by: Andrey Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's a rudiment from old times, when restore worked via ececve.
Now we modify the core file in place to fixup vma-s.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
vma_entry contains shmid and all shared memory are dumped in own files.
The most interesting thing is restore.
A maping is restored by process with the smallest pid. The mamping
is created before executing restorer.
We map a full mapping and restore it's conten, then we open a file from
/proc/pid/map_files and store a descriptor in vma_info. The mapping is
unmaped. Now we can map any region of this mapping in the restorer.
We use this trick, because a target process may have this mapping in
some places and the restorer has not function to open proc files.
v2: fix error hangling
xemul: Fixed static-s and args for cr_dump_shmem
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It will be used to restore shared mappings
v2: clean up
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Now a name of an image file is hard coded ("smth-%d.img", pid),
but the images of namespaces, shared memery, etc belong to
not one task, so they may have other formats of names, which
will describe objects.
For example a image of shared memory content may have name like
this ("pages-shmem-%ld.img", shmid)
v2: fix comment
v3: rebase
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Using absolute paths for this is dangerous - while doing c/r we should
be extremely carefully and not change tasks' roots and mount namespaces
too early. Sometimes it will not work -- when restoring containers we'll
be unable to switch to new CT and still have the ability to open images.
Rework the images opening via openat and keep the image dir fd open all
the time as the service fd (introduced earlier).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
These are the fds that help us to do c/r. We want them not to intersect
with any "real-life" ones and thus store them close the the file rlimit
boundary. For now only the logfd one is such.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Reserve more mem for bootrstrap code and put all self vmas at its tail.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Remove CR_TASK_XXX states, use the TASK_XXX ones (for image). This is
required to unseize tasks properly in the next patches.
Plus, make sure that pstree_list and the seized set coincide (i.e.
handle error in collect_task).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
When we've seized all the tasks and threads found in /proc check for
the /proc contents be the same. Do it one-by-one as we descend the tree.
This is OK, since tasks cannot create kids for anyone but themselves or
their parents (reparent will be handled later).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
While we try to seize task it can die and give its pid to
somebody else. This can break pstree consistency. Check for
parent being valid after task is seized.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Based on xemul@ patches.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
This patch was designed to be generic and thus usable for all kinds of
sockets. Not sure, thah this goal has been reached, but at least I tried.
Key ideas:
1) On-stack structure for collecting sockets queues and then passing them to
parasite code.
2) Singly linked list is used for collecting structures, representing sockets
of any kind (!) with queues.
Based on xemul@ patches.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
v2: New "MSG_STEAL" functionality is used
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
v2: wrappers names become less obfuscating
This patch:
1) Updates function cr_fdset_open() to be suitable for handling fdset creation
for dump and show stages.
2) Replaces cr_fdset_open() by new wrapper function cr_fdset_dump().
3) Replaces prep_cr_fdset_for_restore() by new wrapper function cr_fdset_show().
Signed-off-by: Stanislav Kinsbursky <skinsbursky@parallels.com>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
This patch removes collect stage and dumps tunables object right after
collect.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@parallels.com>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
This commit brings the former "Rewrite task/threads stopping engine"
commit back. Handling it separately is too complex so better try
to handle it in-place.
Note some tests might fault, it's expected.
---
Stopping tasks with STOP and proceeding with SEIZE is actually excessive --
the SEIZE if enough. Moreover, just killing a task with STOP is also racy,
since task should be given some time to come to sleep before its proc
can be parsed.
Rewrite all this code to SEIZE task and all its threads from the very beginning.
With this we can distinguish stopped task state and migrate it properly (not
supported now, need to implement).
This thing however has one BIG problem -- after we SEIZE-d a task we should
seize
it's threads, but we should do it in a loop -- reading /proc/pid/task and
seizing
them again and again, until the contents of this dir stops changing (not done
now).
Besides, after we seized a task and all its threads we cannot scan it's children
list once -- task can get reparented to init and any task's child can call clone
with CLONE_PARENT flag thus repopulating the children list of the already seized
task (not done also)
This patch is ugly, yes, but splitting it doesn't help to review it much, sorry
:(
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Stopping tasks with STOP and proceeding with SEIZE is actually excessive --
the SEIZE if enough. Moreover, just killing a task with STOP is also racy,
since task should be given some time to come to sleep before its proc
can be parsed.
Rewrite all this code to SEIZE task and all its threads from the very beginning.
With this we can distinguish stopped task state and migrate it properly (not
supported now, need to implement).
This thing however has one BIG problem -- after we SEIZE-d a task we should seize
it's threads, but we should do it in a loop -- reading /proc/pid/task and seizing
them again and again, until the contents of this dir stops changing (not done now).
Besides, after we seized a task and all its threads we cannot scan it's children
list once -- task can get reparented to init and any task's child can call clone
with CLONE_PARENT flag thus repopulating the children list of the already seized
task (not done also)
This patch is ugly, yes, but splitting it doesn't help to review it much, sorry :(
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
v2: strlen() check removed from parse_ns_string()
Now '-n' option must be followed by namespaces tags, separated by commas.
Currently, only "uts" namespace is supported.
Signed-off-by: Stanislav Kinsbursky <skinsbursky@parallels.com>
Acked-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Dumping is simple. All but secbits can be read from proc, secbits
are got from parasite.
Restoring is a bit tricky -- when you change anything on kernel
cred's struct it performs sophisticated checks and can change
some more stuff than requested, so the creds restoration procedure
is carefully commented step-by-step.
Another thing to mention is that creds are restored after everything
else, i.e. right before performing final threads sync and sigreturns.
This is done to avoid potential problems with insufficient caps for
restoring other stuff (e.g. CAP_DAC_OVERRIDE or zero euid is most
likely required for opening any image file and the notorious control
/proc/sys/kernel/ns_last_pid, which in turn is performed till the
very last moment).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
