This brings hardness into syscall trasition to asm code,
pass this constants in callers.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
On some systems PAGE_SIZE is declared as sysconf(_SC_PAGESIZE) in <sys/user.h>
this is non-constant expression, so it cannot be used in type declarations.
This breaks compilation with a very non-obvious error message:
CC parasite-syscall.o
In file included from parasite-syscall.c:30:0:
./include/parasite.h:90:8: error: variably modified ‘fds’ at file scope
crtools doesn't uses anything from <sys/user.h>, so we can drop its usage.
Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Command below was executed several times:
sed 's/\(pr_.*[^%,x,X]\)\(\%[0-9,l,L]*x\)/\10x\2/g' -i *.c
Signed-off-by: Stanislav Kinsbursky <skinsbursky@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Util we have kernel support.
[ xemul: MySQL uses runaway pgid and sid and we cannot restore it
gracefully with exiting API :( Byt MySQL seem not to care about
pgid and sid change after restore, so ignore this for a while ]
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Completely unlinked file is the one with n_link count being zero.
Such files only allow to read their contents and carry with us.
In order to dump this thing I introduce the "path remap" technology.
For reg file a remapping entry is dumped which describes, that at
restore stage before opening a regfile->path this path should be
linked to some other name and then (after open) unlinked.
For completely unlinked files the remap path would be a path to
a "ghost" file, i.e. a file which is created only at the time of
restore and which is removed completely at the end of it.
Partially unlinked files (i.e. those having n_link != 0, but a
path by which we see them in someone's fd is not accessible) should
be handled in another way.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This is preriquisity for terminals handling and just a good
practice to save and restore everything we can :)
Not all combinations are supported. All the problems we still
have come from the inability to attach to group/session with
ID no tasks own as its PID.
This can be workarounded by fork()-ing this pid temporarily,
but we'd rather think in the direction of modifying the kernel
to give us direct syscall for this (oh my...)
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
I store them on _entry since sids can only be inherited or
set to current's pid. Thus the best we can do it restore sids
at fork time, thus save them in the image we use to fork.
Maybe when we submit patches that will give us ability to set
arbitrary pgid and sid we'll change this, but this is in the
future.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
New stage CR_STATE_FORKING. This is required to restore pgids
properly -- we need to make sure a task with pid whose pgid we
are about to enter is alive. And this task is not necesserily
our parent, thus wait for everyone to appear.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
The mm_xxx bits are per-mm_struct, not per-task_struct in kernel.
Thus, when we support CLONE_VM we'd better have these bits in a
separate image file.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Why? Because one day we'll support various CLONE_ flags and
for fdtable and fs info we'd like to have separate images (since
these objects are separate in kernel).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's an O(n) algorithm.
Now we iterate both lists simultaneously to find a hole.
[xemul: Discussion making the patch more understandable:
Cyrill:
If s_vma is the last one on self_vma_list you could break immediately, no?
And the snippet I somehow miss is -- how the situation handled when
hole
a b
source |----| |-----|
target |----| |-----|
c d
the hole fits the requested size but the hole is shifted
in target, so that you've
prev_vma_end = a
and then you find that a - d > vma_len and return a
as start address for new mapping while finally it
might intersect with address c.
Or I miss something obvious?
Andrey:
Look at "continue" one more time.
prev_vma_end is returned only if both condition are true
if (prev_vma_end + vma_len > s_vma->vma.start) {
....
if (prev_vma_end + vma_len > t_vma->vma.start) {
...
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Looks-good-to: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
[ xemul: The fix effectively is -- stop scanning the 2nd vma list
once we see, that the hint's end hits the next vma ]
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
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>
pipe_entry is encapsulated in pipe_info.
All pipe_info-s connects in the list pipes.
All pipe_info-s with the same piep_id connects to pipe_list,
it a circular list without a defined head.
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
[ xemul: I don't know how to make this with incremental changes either,
and just go with it :( ]
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Collect pstree_item-s on restore in big list. This lets
us not lseek this file on restore and simplifies the code
a little.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
In commit 71cc2733a79efba65d3466f784b19d17805cf50d
I occasionally dropped the ability to abort on waiting
(because we used signed -1 value to inform waiters that
something is wrong and waiting should be aborted, but
the type was changed to unsigned one and as result
this condition never triggers).
So to resolve it futex_abort_and_wake() is added and
should be used explicitly where appropriate instead
if signess hack.
Reported-by: Andrew Vagin <avagin@parallels.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Andrew 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>
Same as prev 2 patches now for the unix sockets. They are still in per-pid image files, but
this is going to change soon (I hope).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Instread of re-reading this image again and again on every fd restore, pull the
reg-files.img in early and store the entries in a hash. This will simplify the
further fd restoring fixes and will allow for dump/restore via a stream (socket).
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
It's wrong to treat restore_thread_exec_start as arguments
area (I managed to overlook this problem in commit
014841825acb14a1d695569b9fe3575f5de6442b) it's rather
a function start address.
The thread arguments area allocated dynamically after the
restorer blob itself.
We didn't hit any problems earlier simply because there
were a few bytes owerwritten in function prologue.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
CC: Andrey Vagin <avagin@openvz.org>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
This is not good to update images while restoring.
Thus, read vma_entry-es once into a list, put opened (when required) fds
in there and make restorer walk the entries in mem, not those read from
the image file.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
These chunks of memry, which transit into restorer code gets unmapped one-by-one and thus each of them
should be page-aligned.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
To be consistent. Mutexes are futex based but have
own semantics so better to be able to distinguish
the types.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Andrey Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Instead of open-coded u32 variables poking lets use
futex_t type and appropriate helpers where needed.
This should increase readability.
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Acked-by: Andrey Vagin <avagin@parallels.com>
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>
Since the collect_shmems updates start address for vmas, for
two shared mappings in one task we'll try to dup() the 1st
restoration attempt, since the si's start will be set to the
2nd one, which is not yet restored.
Thus we should map-and-open the first one being restored, not
the one with matched address and dup() all the rest.
[avagin@: There's no such thing, since the collect stage checks
for pid being less _or_ _equal_ and this only the first vma's start
will be saved. But anyway, this makes it more obvious.]
Acked-by: Andrey Vagin <avagin@parallels.com>
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Just make the fixup_vma_fds read and write vma images and
those called by it provide and fd for this.
Acked-by: Andrey Vagin <avagin@parallels.com>
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>
