While trying to resurrect the test cases in test/other/rpc one test case
was checking for returned errors during restore failures. The process
was dumping itself and while the process kept on running it tried to
restore from that checkpoint. That should fail as the PID is in use of
the process doing the test. CRIU, however, killed the test process
during restore failure.
This is the fix as proposed by Andrei Vagin on the CRIU mailing list.
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
For easier understanding what is failed.
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
With new cpu-cap='op=noxsaves' mode on x86 we should use
compel's instance of rt info since only it carries
features masked.
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Currently even if kernel supports compact xsave frame a user
can disable it by passing noxsaves argument as a boot option.
Thus cpuid instruction will report its presence but in real
it gonna be masked from kernel pov. Lets do the same and
allow a user to mask it via --cpu-cap=noxsaves option
(valid for x86 only).
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Will need them to mask some of the features from
command line options.
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
We don't have yet support of compacted xsave frames so report
error on cpu-check, checkpoint, restore actions. Basically
it is done in cpu_init routine which is called in the sites
we're interested in.
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Tracking cpuid features is easier when sync'ed with kernel
source code. Note though that while in kernel feature bits
are not part of ABI, we're saving bits into an image so
as result make sure they are posted in proper place together
with keeping in mind the backward compatibility issue.
Here we also start using v2 of cpuinfo image with more
feature bits.
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
To be close to the kernel code.
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
zdtm/static/maps06 failed on systems with different page-size than 4096.
This changes maps06 to use sysconf(_SC_PAGESIZE) instead.
Signed-off-by: Adrian Reber <areber@redhat.com>
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
We operate by long variables in out bit arithmetics, so our constants
should be marked as long too.
Cc: Adrian Reber <areber@redhat.com>
Reported-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Tested-by: Adrian Reber <areber@redhat.com>
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
We alwasy exclude vsyscall vma-s for x32 tests, but
this part was broken by python3 patches.
1: Old maps lost: set([])
1: New maps appeared: set([u"ffffffffff600000-ffffffffff601000 [u'r-xp']"])
Reported-by: Mr Jenkins
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Improve code readability and take into account that even if
uffdio_{copy,zeropage} has failed, some pages had been updated.
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
The criu_status_in is not always used and it may be -1 when the signal
handler closes it. With lazy-pages we hit a corner case which clobbers the
errno value. This happens when we resume the process inside glibc syscall
wrapper and get the signal before the page containing errno is copied. In
this case, signal handler is invoked before the syscall return value is
written to errno and the actual value of errno seen by the process becomes
-EBADF because of close(-1) in the signal handler.
Let's ensure that close() in signal handler does not fail to make Jenkins
happier while the proper solution for the lazy-pages issue is found.
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
is_any_service_fd didn't check for service file descriptor
being marked in sfd_map in result we may end up leaving
some garbage file obtained from the program executing
criu.
[root@sqvm0910 criu-stable.git]# ll /proc/354877/fd
lr-x------ 1 root root 64 May 16 19:03 114 -> /
lr-x------ 1 root root 64 May 16 19:03 115 -> /
l--------- 1 root root 64 May 16 19:03 117 -> /18311 (deleted)
lr-x------ 1 root root 64 May 16 19:03 119 -> /vzt/del/vzctl-rm-me.X1JH4vI (deleted)
We are to test if the file descriptor was indeed installed
explicitly.
https://jira.sw.ru/browse/PSBM-83892
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Drop the constants for default cache host/port and page size because
they are not used anywhere.
Signed-off-by: Radostin Stoyanov <rstoyanov1@gmail.com>
Reviewed-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
I am regularly running zdtm.py after updating the CRIU rpm to test if
the new rpm still works. Until know I usually did:
-criu_bin = "../criu/criu"
-crit_bin = "../crit/crit"
+criu_bin = "/usr/sbin/criu"
+crit_bin = "/usr/bin/crit"
This commit adds two arguments to zdtm.py:
--criu-bin CRIU_BIN Path to criu binary
--crit-bin CRIT_BIN Path to crit binary
It still defaults to the old values from above, but can now easily be
changed to the CRIU binary provided by the packaging system.
This change was more complicated than expected which is probably related
to the fact that zdtm restarts itself in namespaces.
v2:
- rebase
v3:
- fix errors reported by 'make lint'
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
check_tun_netns_cr() should return 0 on CONFIG_TUN=n kernels.
That is, when we fail to open tun device.
Fixes: #496
Signed-off-by: Dmitry Safonov <dima@arista.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
test/dump/zdtm/static/msgque/43/1/ipcns-msg-12.img decode fails: object of type 'NoneType' has no len()
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Running the test suite on aarch64 showed two tests failing (static/fd
and transition/maps0007). Debugging showed that the
atomic_{sub,add}_return() functions did not return the value but the
lower 32bit of the address of that value.
Simply dereferencing the input solves the two mentioned test cases.
Signed-off-by: Dmitry Safonov <dima@arista.com>
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Syscalls do return negative value in case of an error.
But errno contains the error code itself.
Add uint32_t to sys_futex() definition.
sizeof(unsigned int) might be not 4 bytes.
Signed-off-by: Dmitry Safonov <dima@arista.com>
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Send pre-dump notify to 36
Traceback (most recent call last):
File "zdtm.py", line 2161, in <module>
do_run_test(tinfo[0], tinfo[1], tinfo[2], tinfo[3])
File "zdtm.py", line 1549, in do_run_test
cr(cr_api, t, opts)
File "zdtm.py", line 1264, in cr
test.pre_dump_notify()
File "zdtm.py", line 490, in pre_dump_notify
fdin.write(struct.pack("i", 0))
TypeError: write() argument 1 must be unicode, not str
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Here is an example of ldd output:
/musl # ldd /usr/lib/libxtables.so.12
ldd (0x7fae8e06c000)
libc.musl-x86_64.so.1 => ldd (0x7fae8e06c000)
We need to skip the "ldd (0x7fae8e06c000)" line.
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
We don't tune tracers, so we don't expect to get anything.
In docker containers, debugfs isn't mounted and this command failed.
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
fprintf in Alpine Linux calls writev, which is prohibited in the
SECCOMP_MODE_STRICT mode.
===================== Run zdtm/static/seccomp_strict in h ======================
Start test
Test is SUID
./seccomp_strict --pidfile=seccomp_strict.pid --outfile=seccomp_strict.out
make: *** [Makefile:385: seccomp_strict.pid] Error 1
Test zdtm/static/seccomp_strict FAIL at start: [Errno 2] No such file or directory: 'zdtm/static/seccomp_strict.pid'
Test output: ================================
22:45:24.661: 38: ERR: seccomp_strict.c:97: read (errno = 2 (No such file or directory))
22:45:24.662: 37: ERR: test.c:315: Test exited unexpectedly with code 1
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
Otherwise one library will be imported twice by two different names: py
and pycriu, because pycriu is used in the library.
https://github.com/checkpoint-restore/criu/issues/495
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
In case if dump_namespaces failed we may have
ret shadowed and erroneously report checkpoint
success.
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
On fedora rawhide seccomp_metadata for some
reason is not defined (while in kernel it introduced
together with PTRACE_SECCOMP_GET_METADATA). So
lets do a trick for a while -- define own alias.
Once system headers get settled down we might find
more suitable solution. Because it's a part of kernel
API we're on the safe side.
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
After CR_STATE_RESTORE_SIGCHLD stage triggered we are
not allowed to exit, just yield the BUG instead.
Reported-by: Andrei Vagin <avagin@virtuozzo.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
For python2/python3 compatibility install additional (python2-future)
and different (python2-ipaddress) packages during test.
Also switch Fedora tests to explicitly install versioned python packages
(python2-* instead of python-*).
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
As python3 imports are working a bit differently the structure of the
imported files needs to change. Instead of having symlinks to criu.py
and rpc_pb2.py which does not match the import structure just create
symlink to ../lib/py (like crit) to make zdtm usable again with the
python3 compatibility.
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
With this last commit of the crit with python3 series it is possible to
either use python2 or python3 with CRIU.
Now the basic build system functionality (make and make install) are
python2/python3 aware. zdtm.py and criu-coredump are still python2, but
as they are not part of 'make install' those parts have not yet been
ported from python2 to python3.
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
