- 01 Apr, 2015 2 commits
-
-
Cyrill Gorcunov authored
We have them in vma_entry_is_private() already. Signed-off-by:
Cyrill Gorcunov <gorcunov@openvz.org> Signed-off-by:
Pavel Emelyanov <xemul@parallels.com>
-
Cyrill Gorcunov authored
We have two helpers for VMA type testing: privately_dump_vma() and vma_priv(). They work with different types but basically do the same: check if we should dump VMA into the image and restore it back then. Lets unify they both into common vma_entry_is_private() helper and vma_area_is_private() for working with vma_area type. Signed-off-by:
-
- 31 Mar, 2015 4 commits
-
-
Tycho Andersen authored
For some undiagnosed reason, this test hangs for me when using SIGINT. Since we always want to kill this process anyway, use SIGKILL instead. Signed-off-by:
Tycho Andersen <tycho.andersen@canonical.com> Signed-off-by:
-
Tycho Andersen authored
*can* reads that it might fail but won't. Instead, this comment means that any code added below it should not fail, because the network is unlocked and so the restored process is exposed to the world. Signed-off-by:
-
Tycho Andersen authored
Signed-off-by:
-
Andrey Vagin authored
libcontainer saves PID in a state file. Signed-off-by:
Andrew Vagin <avagin@openvz.org> Signed-off-by:
-
- 30 Mar, 2015 6 commits
-
-
Oleg Nesterov authored
This is pure theoretical, especially in this particular case when we actually want to (likely) free the unused memory. Still the code which ignores potential error doesn't look good. Signed-off-by:
Oleg Nesterov <oleg@redhat.com> Reviewed-by:
-
Oleg Nesterov authored
1. parse_mountinfo_ent() mixes "return -1" and "goto err" on failure, this looks confusing and inconsistent. 2. And buggy. It forgets to free(opt) if parse_mnt_flags() fails. Signed-off-by:
-
Oleg Nesterov authored
The caller will do this on failure too. So this is unnecessary and wrong because we do not nullify ->mountpoint. Signed-off-by:
-
Oleg Nesterov authored
1. parse_mountinfo() forgets to free(fst) if parse_mountinfo_ent() succeeds. 2. The usage of fst/r_fstype is ovecomplicated for no reason. Just change the parse_mountinfo() paths to populate/use/free this fsname unconditionally, and move the ownership to the caller. There is no reason to check FSTYPE__UNSUPPORTED and/or fallback to ->name. Better yet, we could even turn fsname into the local "char []" and avoid %ms and free(), but then we would need to pass the length of this buffer to parse_mountinfo_ent(). Signed-off-by:
-
Oleg Nesterov authored
Preparation to simplify the review. parse_mountinfo() assumes that: 1. The "err:" block does all the necessary cleanups on failure. This is wrong, see the next patch. 2. We can never skip the mountpoint. This is true, but we are going to change this. s/goto err/goto end/ in the main loop, add the "end:" label which inserts the new mount_info into the list and then checks ret != 0 to figure out whether we need to abort. Signed-off-by:
-
Andrey Vagin authored
This is required to use criu swrk in libcontainer. v2: remove useless function declaration allow to set inherit_fd only for swrk v3: check swrk out of loop Cc: Saied Kazemi <saied@google.com> Signed-off-by:
-
- 27 Mar, 2015 8 commits
-
-
Andrey Vagin authored
Use the format specifier PRIx64 instead of %lx to print uint64. integer. Reported-by: Mr Travis CI Signed-off-by:
-
Cyrill Gorcunov authored
This is logging routine, lets it not affect the global @errno. Signed-off-by:
Andrew Vagin <avagin@parallels.com> Signed-off-by:
-
Cyrill Gorcunov authored
The __print_on_level routine may modify global @errno variable which is inacceptable: this is logging routine which must be transparent to the rest of the program code. Thus save @errno in local @__errno variable and restore it on return. Signed-off-by:
-
Andrey Vagin authored
Here is a real example how it's used. Cc: Tycho Andersen <tycho.andersen@canonical.com> Signed-off-by:
-
Oleg Nesterov authored
validate_mounts() prints ->mnt_id in hex when it reports the failure. This complicates the understanding because this ->mnt_id is printed as decimal elsewhere, including /proc/$pid/mountinfo. parse_mountinfo() adds "0x" at least and this is just pr_info(), but lets change it too. Signed-off-by:
-
Andrey Vagin authored
mkdir: cannot create directory ‘zdtm.mpAWGL/holder’: File exists Reported-by:
-
Andrey Vagin authored
parse_smaps() is too big for easy reading. In addition, we are creating a new interface to get information about processes, which is called taskdiag, so parse_smaps() will do only what it should do accoding with the name. All other should be moved in a separate functions which will be reused to work with task_diag. Signed-off-by:
-
Andrey Vagin authored
parse_smaps() is too big for easy reading. In addition, we are creating a new interface to get information about processes, which is called taskdiag, so parse_smaps() will do only what it should do accoding with the name. All other should be moved in a separate functions which will be reused to work with task_diag. Signed-off-by:
-
- 26 Mar, 2015 4 commits
-
-
Cyrill Gorcunov authored
For some reason it returns ENOENT sometime when run in namespace. Need to figure out why. Disable it for a while to continue testing the rest. Signed-off-by:
-
Andrey Vagin authored
inotify_irmap creates files in /etc so it should be able to do this from userns. Signed-off-by:
-
Andrey Vagin authored
Currenlty if criu fails before prepare_cgroup_properties(), cgyard isn't umounted. I think it's déjà vu, but it isn't :) commit 28b0e16d Author: Andrew Vagin <avagin@openvz.org> Date: Mon Aug 25 14:29:00 2014 +0400 cgroup: call fin_cgroup() on error paths Cc: Tycho Andersen <tycho.andersen@canonical.com> Cc: Saied Kazemi <saied@google.com> Signed-off-by:
-
Oleg Nesterov authored
1. The -ERESTART_RESTARTBLOCK case in get_task_regs() depends on kernel internals too much, and for no reason. We shouldn't rely on fact that a) we are going to do sigreturn() and b) restore_sigcontext() always sets restart_block->fn = do_no_restart_syscall which returns -EINTR. Just change this code to enforce -EINTR after restore, this is what we actually want until we teach criu to handle ERESTART_RESTARTBLOCK. 2. Add pr_warn() to make the potential bug-reports more understandable, a sane application should handle -EINTR correctly but this is not always the case. Signed-off-by:
-
- 25 Mar, 2015 2 commits
-
-
Andrey Vagin authored
/proc/pid/pagemap is protected by the global CAP_SYS_ADMIN, so it can't be opened in a non-root userns. 00:34:48.260 Execute ns/user/static/cow01 ... 00:34:48.260 12:37:03.951: 4: ERR: cow01.c:87: Unable to open child pagemap file /proc/5/pagemap (errno = 1 (Operation not permitted)) Reported-by: Mr Jenkins Signed-off-by:
-
Cyrill Gorcunov authored
The idea is simply make sure that timeour has not been screwed during c/r cycle. v2: - Drop unneeded early poll() call. - Make poll loop in cycle until caller passes c/r cycle. v3: - Use test_go helper - Do poll in cycle Signed-off-by:
-
- 24 Mar, 2015 6 commits
-
-
Tycho Andersen authored
== is a bashism, we should use = for /bin/sh; this prevents output like: /home/ubuntu/criu/test/post-dump.sh: 3: [: post-dump: unexpected operator /home/ubuntu/criu/test/post-dump.sh: 3: [: network-unlock: unexpected operator Signed-off-by:
-
Tycho Andersen authored
We need to finalize the cg yard both on successful cgroup restore and on a failed restore. Further, we should restore the cgroup properties before allowing the task to continue in all modes (previously properties were only restored correctly in --restore-detached mode). CC: Saied Kazemi <saied@google.com> Signed-off-by:
-
Andrey Vagin authored
# Build criu in a docker container $ make direct-build # Execute tests in a docker container $ make direct-test Signed-off-by:
-
Andrey Vagin authored
* redirect both stdout and stderr bash: >& and &> dash: command > file 2>&1 or command 2>&1 | othercommand Signed-off-by:
-
Andrey Vagin authored
mawk doesn't handle '\<...\>' Ubuntu: $ awk -W version mawk 1.3.3 Nov 1996, Copyright (C) Michael D. Brennan $ ldd /bin/ps linux-vdso.so.1 => (0x00007fffb290a000) $ ldd /bin/ps | awk '/\<linux-vdso\.so\>/ { print $1 }' $ ldd /bin/ps | awk '/linux-vdso\.so/ { print $1 }' linux-vdso.so.1 Fedora: $ ldd /bin/ps linux-vdso.so.1 => (0x00007fff61fcb000) $ ldd /bin/ps | awk '/\<linux-vdso\.so\>/ { print $1 }' linux-vdso.so.1 $ awk --version GNU Awk 4.1.1, API: 1.1 Signed-off-by: -
Andrey Vagin authored
Vladimir said that a page can be unmapped from one process, but can be mmaped in another one. In this case we can't understand whether both processes link to the same page or they don't. So we can try read both pages to take them from swap. Cc: Vladimir Davydov <vdavydov@parallels.com> Reported-by: Parallels' QA Signed-off-by:
-
- 23 Mar, 2015 1 commit
-
-
Andrey Vagin authored
cow01 uses /proc/pid/pagemap which has been protected by CAP_SYS_ADMIN. """ commit ab676b7d6fbf4b294bf198fb27ade5b0e865c7ce pagemap: do not leak physical addresses to non-privileged userspace As pointed by recent post[1] on exploiting DRAM physical imperfection, /proc/PID/pagemap exposes sensitive information which can be used to do attacks. """ Signed-off-by:
-
- 20 Mar, 2015 2 commits
-
-
Oleg Nesterov authored
Before the recent "x86_64,signal: Fix SS handling for signals delivered to 64-bit programs" kernel patch, sigreturn paths forgot to restore ->ss after return from the signal handler. Now that the kernel was fixed, restore_gpregs() has to initialize ->ss too, it is no longer ignored. Note: this is the minimal fix. In the long term we probably should not dump/restore the segment registers at all. We can use sigcontext filled by the target kernel and modify the general-purpose regs. Reported-and-tested-by:
Andrey Wagin <avagin@gmail.com> Signed-off-by:
-
Ruslan Kuprieiev authored
This is common for log files, when a number of processes are writing something into one file. Currently, if someone wrote something to log since dump happened, on restore criu will complain about "File has bad size" and refuse to restore. If file is opened with O_APPEND|O_WRONLY flags it is going to jump to the EOF anyway. v2, use O_ACCMODE and separate helper Signed-off-by:
Ruslan Kuprieiev <rkuprieiev@cloudlinux.com> Signed-off-by:
-
- 19 Mar, 2015 1 commit
-
-
Pavel Emelyanov authored
It's obsoleted and is not required. Signed-off-by:
-
- 18 Mar, 2015 2 commits
-
-
Kirill Tkhai authored
Print all cap values in case of test is failed. Also check for capset() and capsget() return values. Signed-off-by:
Kirill Tkhai <ktkhai@parallels.com> Signed-off-by:
-
Saied Kazemi authored
If the --restore-detached command line option is not specified during restore, CRIU should unmount and remove the temporary cgyard directory tree before waiting for the restored process to exit. Otherwise, all the temporary cgyard mount points will remain mounted and visible. Signed-off-by:
Saied Kazemi <saied@google.com> Signed-off-by:
-
- 16 Mar, 2015 2 commits
-
-
Andrey Vagin authored
Lazy images are opened on a first attempt of using. 00:01:18.534 Test: zdtm/live/static/pipe00, Result: FAIL 00:01:18.537 ==================================== ERROR ==================================== 00:01:18.538 Test: zdtm/live/static/pipe00, Namespace: 1 00:01:18.538 Dump log : /var/lib/jenkins/jobs/CRIU/workspace/test/dump/ns/static/pipe00/13536/1/dump.log 00:01:18.540 --------------------------------- grep Error --------------------------------- 00:01:18.543 (00.026666) Error (include/image.h:153): BUG at include/image.h:153 00:01:18.543 (00.050663) Error (namespaces.c:801): Namespaces dumping finished with error 134 00:01:18.543 (00.050918) Error (cr-dump.c:1979): Dumping FAILED. 00:01:18.545 ------------------------------------- END ------------------------------------- 00:01:18.548 ================================= ERROR OVER ================================= Reported-by: Mr Jenkins Signed-off-by:
-
Pavel Emelyanov authored
Currently on dump we generate too many image files, effectively all the stuff from the GLOB set is created. The thing is that sometimes some of created images can be empty (just contain the magic number at the head). Thos images are useless and just waste the space. When applied after the "empty images" set, this introduces the lazy images -- when we call open_image() the actual file is only created (and the magic number is written into it) when the very first object goes into it. For example for the simplest test we have, then static/env00 one, the created image files are core-7290.img creds-7290.img fdinfo-2.img fs-7290.img ids-7290.img inventory.img mm-7290.img pagemap-7290.img pages-1.img pstree.img reg-files.img sigacts-7290.img Signed-off-by:
-
