This patch can be committed instead of:
[PATCH 1/6] cr-dump: move parasite_drain_fds_seized out of dump_task_files
[PATCH 2/6] cr-dump: fix dumping file locks in a mount namespace
readlink is not required here and a file can be unavailable,
if a process is in another mnt namespace
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's no longer required to use this option -- two currently
supported cases (tasks on host and tasks in containers) can
be detected automatically. Keep this option for future.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's required to know whether the root task lives in
namespaces very very early (e.g. -- to lock the network
properly). Thus we have to collect task IDs right at
the time we collect the tasks themselves.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The recent kernels allow to get namespaces IDs by reading proc-ns links.
Use this to generate IDs for tasks' namespaces (I do generate them, since
IDs provided by kernel look ugly :( ).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
These functions are designated to convert a native pointer
to uint64_t used to store a virtual address in protobuf messages
and vice versa in a machine-independent way.
Signed-off-by: Alexander Kartashov <alekskartashov@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Dump file locks' necessary entries to the image, we only support flock and
posix file lock right now.
Changelog since the initial version:
We got file lock info from global list, so the dump_task_file_locks
can be much simpler.
Originally-signed-off-by: Zheng Gu <cengku.gu@huawei.com>
Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We collect all file locks to a golbal list, so we can use them easily
in dump_one_task. For optimizaton, we only collect file locks hold by
tasks in the pstree.
Thanks to the ptrace-seize machanism, we can aviod the blocked file lock
issue, makes the work simpler.
Right now, the check handles only one situation:
-- Dumping tasks with file locks hold without the -l option.
This covers for the most part. But we still need some more work to make
it perfect robust in the future.
Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
For fanotify dumping we need to find mount points the path
lay on, moreover we save a mount point device number in the
image, thus collect mount point information on dump stage.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
We will be handling both inotify and fanotify
objects here thus to make less confusion rename
the files to fsnotify.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This is the merge and a slight rework (no TI_SP macro) of Alexander's patches
about the subj.
Signed-off-by: Alexander Kartashov <alekskartashov@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Currenly crtools supports a case when a child shared a fd table
with parent.
Here is only two interesting things.
* Service descriptors should be cloned for each process
who shared one fd table.
* One task should restore files and other tasks should sleep in this
* time.
v2: * allocate fdt_lock from shared memory
* don't wait a child, if it doesn't share fdtable
v3: * don't move ids on the pstree image
v4: * save ids in a separate image
* save fdinfo per id instead of pid
v5: fix alignment of service_fd_id
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It will be used for determing which resources are shared
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It is read together with pstree items for checking what kind of
resources should be shared. Core is too big for reading it in
this place.
v2: fix check_core
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Currently fdinfo dumps for each task, so CR_FD_FDINFO is in cr_fdset.
A few tasks can share one fd table and the set of descriptors will be
dumped once and a image name will contain files_id instead of pid.
In this case CR_FD_FDINFO will go away from cr_fdset.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Dump the with "new" prlimit syscall that works on arbitrary pid.
Restore is done in restorer _after_ mappings mixup and _before_
caps drop to make it set any max value.
The RLIM_INFINITY is handled explicitly to help future 64<->32
bits migration.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The number of arguments used to carry data via them is too
big already. Just fill the required core fields inside.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The size of an auxv is the machine pointer but a 64-bit integer is reserved
in a MmEntry protobuf message to store an auxv. Moreover the number of auxv's
varies from one architecture to another. So the following is proposed
to alleviate the issue.
* Introduced the type auxv_t representing a machine-pointer sized integer.
* The size of auxv array is extracted from a MmEntry message instead of using
the value of the macro AT_VECTOR_SIZE.
Signed-off-by: Alexander Kartashov <alekskartashov@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
* The following files goes into the directory arch/x86/include/asm unmodified:
- include/atomic.h,
- include/linkage.h,
- include/memcpy_64.h,
- include/types.h,
- include/bitops.h,
- pie/parasite-head-x86-64.S,
- include/processor-flags.h,
- include/syscall-x86-64.def.
* Changed include directives in the source files that include the headers
listed above.
* Modified build scripts to reflect the source moves.
Signed-off-by: Alexander Kartashov <alekskartashov@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The dumping of FPU state is done with help of ptrace
facility. There are two cases which we need to handle
depending on which features are available on host machine
1) The dump via ptrace(PTRACE_GETFPREGS ...)
In this case the kernel will use fxsave approach
inside the kenrel and provides us back the data
encoded in i387_fxsave_struct format.
2) The dump via ptrace(PTRACE_GETREGSET ...)
In this case the kernel will use xsave approach
inside the kernel and provides us back the data
encoded in xsave_struct format.
In any case we decode data and save it in protobuf format.
This is why core.proto file has been extended to keep new
entries.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
And don't forget to undef them once they are not needed.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Actually it was never used, just drop it.
Because of backward compatibility problem we
can't just zap it in protofile.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The cpu we're running on must at least support fxsave feature.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
I've moved dump_thread helper a bit lower in file
since I've to call for find_thread_state helper.
After all this groups all thread related functions
in one slab.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This patch starts using parasite_init_threads_seized and
parasite_fini_threads_seized helpers to save per-thread
data in parasite and remove it on cure procedure.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The per-thread information requires own space in parasite data.
In particular we will keep the blocked signals bound to thread
pids.
For this sake the caller need to provide the parasite how many
threads will be used to calculate space.
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 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 dumping of tty peers is somewhat tricky. And it became more
complex once we allowed to migrate/inherit sessions.
It's being found (in screen c/r) that we've a problem in looking
up of session leaders while dumping tty.
Let me explain with more details. Here is an example of screen
session
PID GID SID
20567 20567 20567 SCREEN
20568 20568 20568 pts/3 \_ /bin/bash
The screen opens master peer (ptmx) and then provides
bash the slave peer (pts/3) where bash sets up a session
leader on it.
Thus we get interesting scenario -- our pstree construction
is done in lazy fashion, we run parasite code to fetch sid/pgid
of a process tree item only when we're really dumping the task.
Thus when we start dumping ptmx peer (which belongs to SCREEN)
we've not yet constructed the process tree item for children
(ie /bin/bash) and the lookup function in tty code (which walks
over all process items in a tree) simply fails to find sid of
child, because we've not yet dumped it.
Thus, to resolve such situation we verify tty sids at late stage
of dumping.
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>
