A value of signo is in [1, SIGMAX].
Currenly signals are enumirated from 1 to SIGMAX, but SIGMAX
is not included. This patch fixes this mestake.
v2: * save backward compatibility
* set a correct value of SIGMAX = 64. It can not be in a
separate patch, because a format is changed again.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When resetting nr_in_progress for next stage no need
in waking up anyone. Nobody waits for it yet :)
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Switching to a new stage is 4-step procedure which
deserves its own helper. Besides, now the information
about how many tasks participage in each stage is
collected in one place.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
When finishing a stage we have to report this (decrement the
number of tasks in stage) and wait while stage switch. Write
a helper that does both.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This reverts commit ef3771d566dacb8ee9fe71b744d56f08674fe3db.
With new SO_BINDTODEVICE getting API it's not required.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
A parent process can change a few pages after forking a child and
all this pages should not be avaliable from the child.
Each vma has a bitmap of existent pages. Parent's and child's bitmaps
can be compared and all pages which are not present in a child bitmap
are dropped.
v2: don't check page_bitmap on NULL
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
All memory content are restored before entering in restorer.c.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's required for restoring copy-on-write regions.
The similar code will be removed from restorer.c.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
With this patch vma->shmid contains file id before mapping a region,
then it contains of a temporary address.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
All private vmas are placed in a premmapped region and
they are sorted by start addresses, so they should be shifted apart.
Here is one more problem with overlapped temporary and target regions,
mremap could not remap such cases directly, so for such cases a vma is
remapped away and then remapped on a target place.
v2: fix accoding with Pavel's comments
v3: add a huge comment with pictures
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Private vma-s are mapped before forking children, then they are
remapped to corrected places in restorer.c.
In restorer all unneeded vma-s are unmaped. VMA-s from premmapped
regions should not be unmaped.
v2: replace guard pages on arithmetic in restorer
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In this case private vma-s will be inherited by children,
it allows to restore copy-on-write reqions.
This code compares child and parent vma lists. If it found
two vma-s with the same start and end addresses, it decides
that the child inherites this vmas from the parent.
This code calculates a size of all private vma-s, then allocate
a memory region for all vma-s and maps them one by one. If a vma is
inherited it will be remaped to an allocated place.
As a result all vma-s will be placed in a continious memory region
and sorted by start addresses. This logic will be used for remap
vma-s to correct address.
v2: fix accoding with Pavel's comments ( clean up and simplify )
v3: simplify code and check that VMA-s are sorted
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
A private vma's should be inherited by children for
restoring copy-on-write regions.
v2: free parent's vma-s in this patch.
v3: split patch on two parts
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Those will be inherited from parent. Before this patch this list was
always empty, but it will change soon.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
read_vmas will be called bedore forking children to restore
copy-on-write memory.
v2: don't open an image one more time
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This makes code more readable, saves one ptr on stack and
lets us jump into restorer code using tags.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
To unify the code for both thread leader and regular threads
we move blocked signals for thread leader into threads argument
area and use restore_thread_common() helper.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Otherwise we might get nil dereference in sigreturn restore.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Three parts.
Proc: open of map_files' link doesn't work on sockets. We fstatat
it and check that it's a socket (it will be packet), then save
the socket inode on vma_area.
Dump: we resolve socket inode to socket id and save it on vma.
We use id, not inode, since on restore we'll have to mmap some
opened file, not just abstract socket with inode.
Restore: when reading vma-s we just need to find out on what fd
the respective packet socket is opened (i.e. -- no map-and-close
sockets supported by now) and dup() it to let restorer mmap it
back.
All this make it possible to c/r the tcpdump tool!
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The length of bootstrap in the print is old and wrong, we need to fix
it and unify the length variable.
Signed-off-by: Huang Qiang <h.huangqiang@huawei.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Writing to last_pid sysctl is CAP_SYS_ADMIN potected. Thus restoring
creds before it won't work in all the cases.
Fix this by making all threads restore creds themselves, and the
thread group leader -- after all of them.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Acked-by: Andrey Vagin <avagin@parallels.com>
The size of vector depends on the kernel config
so use the real size of a vector dumped. Otherwise
we might fail on restore.
Reported-by: Stanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The former is actually the parameters of thread group leader, so
it's natural to have them on-task.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
With some historical changes, the second page-aligned for
restore_thread_vma_len is reduplicate. So remove it.
Signed-off-by: Huang Qiang <h.huangqiang@huawei.com>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Many image files opened by open_image_ro weren't closed before return, fix
them all in this patch.
Signed-off-by: Huang Qiang <h.huangqiang@huawei.com>
Acked-by: Andrew Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It will be required to support socket bound to devices.
When restoring w/o net namespaces -- collect existing devices.
When restoring with them -- collect what is received from image.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We will need it for slave ttys migration. They serve for one purpose --
to clone self stdio descriptor and use it with tty layer, which will
be addressed in further patches.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
No magic here, just fetch info using getpriority and sched_getxxx calls.
Good news is that the mentioned syscalls take pid as argument and do work
with it, i.e. -- no need in parasite help here.
Restore is splitted into prep -- copy sched bits from image on restorer
args -- and the restore itself. It's done to avoid restoring tasks info
with IDLE priority ;) To make restorer not-fail sched bits are validated
for sanity on prep stage.
Minimal sanity test is also there.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In case if here no task found which would restore
controlling terminal -- exit with error instead of
continue with just error message.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Dumping them is performed via parasite, since calling the getgroups
is the only way of getting the complete list. Currently the nr of
groups to dump is limited explicitly with the size of shared memory
between crtools and parasite. This is MUCH more that we have seen
on real apps so far.
Restoring is done early, before restorer blob not to carry the undefined
array of grpous in there. This is OK, since groups do not affect us at
that point and are not affected by subsequent creds restore.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Otherwise there is a race between files with same names:
link(name -> ghost) link(name->ghost)
open(name)
unlink(name)
open(name) -> ENOENT
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Restore must not fail after unlocking connections.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Early we moved prepare_shared() to a root task,
because several preparation actions should be executed
in a target namespace set (e.g.: ghost files).
TCP sockets are a subset of init sockets,
they should be unlocked before resume. It's convient to do
from crtools.
An image can't be read more than one time, because we want to
send it via network.
For this two reasons prepare_shared is spitted in two parts,
one for crtools and one for a root task.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
TCP_REPAIR should be droppet when a network is unlocked.
A network should be unlocked at the last moment, because
after this moment restore must not failed, otherwise a state of
a tcp connection can be changed and a state of one side in our image
will be invalid.
v2: use xremalloc instead of mmap and remmap
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We need to do two non-trivial things with ttys -- interconnect
slaves to masters (or to each other) and setup ctl-tty restoring
task.
Now this is done in subsequently depending on each other steps:
1. collect ttys
2. interconnect slaves and mark ctl-tty tasks
3. collect fake fds for tty-ctl tasks
4. setup orphaned slaves
We can relax this logic in two ways:
1. don't split marking ctl-tty tasks and then creating fds for them
do it in one step at the end
2. don't interconnect slaves with masters and orphaned slaves in
two steps -- do it in one place after fds are collected
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
