- 02 Mar, 2016 1 commit
-
-
Andrew Vagin authored
Signed-off-by:
Andrew Vagin <avagin@virtuozzo.com> Signed-off-by:
Pavel Emelyanov <xemul@virtuozzo.com>
-
- 01 Mar, 2016 8 commits
-
-
Tycho Andersen authored
Travis uses cpusets in such a way [1] that we can't actually write to cpuset.cpu_exclusive ever, so none of these tests will work. They'll still work in jenkins, though, so disabling them is probably ok. Closes #118 [1]: https://github.com/travis-ci/worker/blob/master/backend/docker.go#L66Signed-off-by:
Tycho Andersen <tycho.andersen@canonical.com> CC: Andrew Vagin <avagin@virtuozzo.com> Signed-off-by:
-
Tycho Andersen authored
Some libcs buffer writes to FILE*, which means that we error on fclose instead of write, which makes it hard to figure out what property actually failed writing. Also shorten the error path a bit. Hopefully this patch will help with debugging #118 Signed-off-by:
Andrew Vagin <avagin@openvz.org> Signed-off-by:
-
Andrew Vagin authored
*** CID 158458: Memory - corruptions (NEGATIVE_RETURNS) /criu/pie/parasite.c: 321 in get_proc_fd() 315 316 ret = sys_readlinkat(AT_FDCWD, "/proc/self", buf, sizeof(buf)); 317 if (ret < 0 && ret != -ENOENT) { 318 pr_err("Can't readlink /proc/self (%d)\n", ret); 319 return ret; 320 } >>> CID 158458: Memory - corruptions (NEGATIVE_RETURNS) >>> Using variable "ret" as an index to array "buf". 321 buf[ret] = 0; 322 323 /* Fast path -- if /proc belongs to this pidns */ 324 if (pie_atoi(buf) == sys_getpid()) 325 return sys_open("/proc", O_RDONLY, 0); 326 Signed-off-by: -
Andrew Vagin authored
*** CID 158459: Uninitialized variables (UNINIT) /criu/proc_parse.c: 2218 in parse_task_cgroup() 2212 2213 int parse_task_cgroup(int pid, struct parasite_dump_cgroup_args *args, struct list_head *retl, unsigned int *n) 2214 { 2215 FILE *f; 2216 int ret; 2217 LIST_HEAD(internal); >>> CID 158459: Uninitialized variables (UNINIT) >>> Declaring variable "n_internal" without initializer. 2218 unsigned int n_internal; 2219 struct cg_ctl *intern, *ext; 2220 2221 f = fopen_proc(pid, "cgroup"); 2222 if (!f) { 2223 pr_perror("couldn't open task cgroup file"); ** CID 158458: Memory - corruptions (NEGATIVE_RETURNS) /criu/pie/parasite.c: 321 in get_proc_fd() Signed-off-by: -
Tycho Andersen authored
The restore process uses these modules as well, so let's modprobe them. This prevents: (00.217856) 1: Running ip rule delete (00.218970) 1: Running ip rule delete (00.220059) 1: Running ip rule delete (00.221695) 1: Running ip rule restore (00.223068) 1: Running iptables-restore for iptables-restore (00.439385) 1: Running ip6tables-restore for ip6tables-restore modprobe: ERROR: could not insert 'ip6_tables': Operation not permitted ip6tables-restore v1.6.0: ip6tables-restore: unable to initialize table 'filter' Error occurred at line: 2 Try `ip6tables-restore -h' or 'ip6tables-restore --help' for more information. Signed-off-by:
-
Tycho Andersen authored
Signed-off-by:
-
Andrew Vagin authored
Signed-off-by:
-
Cyrill Gorcunov authored
When we run "make install" the python's setup script prepares all directories for modules but if we need to run crit from the source tree without its install then we fall in trouble because python doesn't know where the fetch pycriu from. Thus simply provide the symlink to the modules emulating that instalation complete. Note this is for developers conveniency only because for end users "make install" always must has place. Reported-by:
Cyrill Gorcunov <gorcunov@openvz.org> Signed-off-by:
-
- 29 Feb, 2016 15 commits
-
-
Tycho Andersen authored
At one point in the cgns patchsets I had removed this, but somehow it got lost in the shuffle. Since we support this now, let's remove this restriction. Signed-off-by:
-
Tycho Andersen authored
Signed-off-by:
-
Andrey Ryabinin authored
We don't have a way to dump proccess blocked in vfork(), hence mark this test as expected to fail. Signed-off-by:
Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Andrew Vagin <avagin@virtuozzo.com> Signed-off-by:
-
Andrey Ryabinin authored
execlp() fails when we run vfork00 test inside namespace because we don't have '/bin/true' there. Instead of execlp() in vfork-child we can just _exit(). Signed-off-by:
-
Andrey Vagin authored
$ echo test//home/avagin/git/criu test//home/avagin/git/criu v2: use double quotes to run pwd Signed-off-by:
-
Tycho Andersen authored
Consider the case where --freeze-cgroup=/lxc/foo, but (e.g. with systemd in lxc), all of the tasks actually live in a set of sub cgroups, e.g. /lxc/foo/init.scope and others. In this case, we will have a multi-headed controller, since there is nothing in the common parent. We should just save the freezer value in all of these heads instead of failing. Note that this doesn't address the larger problem that only the top level freezer.state file is c/r'd, or waht happens when the container itself has frozen tasks but not at the top level. After some discussion, there is no nice way to atomically test-and-set the cgroup freezer, so we'll need some other kernel help. But I'll ignore this for now :) Signed-off-by:
-
Pavel Emelyanov authored
Signed-off-by: -
Andrew Vagin authored
Signed-off-by:
-
Andrew Vagin authored
Signed-off-by:
-
Andrew Vagin authored
Signed-off-by:
-
Andrew Vagin authored
Signed-off-by:
-
Tycho Andersen authored
As with the socket diag modules, since we might be using the ip*filter_tables modules, we should preload those as well, in case the host system hasn't already loaded them. Really, I should implement netlink buffer dumping so we can get rid of this hack :) v2: remember to close /dev/null fd after using it Signed-off-by:
-
Cyrill Gorcunov authored
We better should switch to nmk usage. But lets c/p for now. Signed-off-by:
-
Dmitry Safonov authored
Fixes commit e7d89a60 ("add openat() to syscall list"). Nip: move sys_seccomp for numerical order Signed-off-by:
Dmitry Safonov <dsafonov@virtuozzo.com> Acked-by:
Cyrill Gorcunov <gorcunov@gmail.com> Signed-off-by:
-
Andrew Vagin authored
We can do this, but we need to be sure that all structures are consistent in any moment and we need to block alarm when they are inconsistent. I don't think that we really want to do this now. I suggest to interrupt a current syscall if an alarm signal is triggered. v2: print an error message before exiting Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Signed-off-by:
-
- 26 Feb, 2016 11 commits
-
-
Cyrill Gorcunov authored
They are running inside dumpee space so should not be injected with Gcov instructions. Signed-off-by:
-
Cyrill Gorcunov authored
This cause problem on ppc64 | gcc -c -O2 -g -Wall -Werror -DCONFIG_PPC64 -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE | -fno-strict-aliasing -iquote /home/cyrill/criu/criu/include -iquote /home/cyrill/criu/images | -iquote /home/cyrill/criu/criu/pie -iquote /home/cyrill/criu/criu/arch/ppc64 | -iquote /home/cyrill/criu/criu/arch/ppc64/include -iquote /home/cyrill/criu/ -I/usr/include/libnl3 | -iquote ppc64 -DCONFIG_PPC64 -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE parasite-syscall.c -o parasite-syscall.o | parasite-syscall.c: In function ‘parasite_dump_cgroup’: | parasite-syscall.c:1283:2: error: size of unnamed array is negative | ca = parasite_args(ctl, struct parasite_dump_cgroup_args); | ^ | /home/cyrill/criu/scripts/nmk/scripts/rules.mk:53: recipe for target 'parasite-syscall.o' failed Just use size parasite known to support. Signed-off-by:
Laurent Dufour <ldufour@linux.vnet.ibm.com> Acked-by:
-
Pavel Emelyanov authored
In commit 736a2940 there was added a collection of non-root task's cgsets. But criu's cgset should anyway be excluded from it. Signed-off-by:
-
Cyrill Gorcunov authored
Signed-off-by:
-
Laurent Dufour authored
The tags target should only create tag file using ctags. For those who like to run emacs, the new target etags is generating the TAGS file. Cc: Dmitry Safonov <dsafonov@virtuozzo.com> Cc: Cyrill Gorcunov <gorcunov@openvz.org> Cc: Pavel Emelyanov <xemul@virtuozzo.com> Signed-off-by:
-
Sergey Bronnikov authored
After refactoring of build system we broke target for code coverage report generation in the main Makefile. This patch fix it. Signed-off-by:
Sergey Bronnikov <sergeyb@openvz.org> Signed-off-by:
-
Tycho Andersen authored
Signed-off-by:
-
Kir Kolyshkin authored
The problem with --no-print-directory set by default is when we run :make from vim and compilation performed by a sub-make fails, vim can't figure out the correct path to a file with error, so instead of jumping to the position of the first error, it opens a new empty file. When make prints "Entering directory ..." and "Leaving directory ..." messages, vim :make works as intended. [Alternatively, we could add -w to $(MAKE) invocations where needed, but the problem is --no-print-directory overrides -w even if -w comes later in the options list, so -w is not working. I was also thinking about detecting whether we are running from vim, but when we'll need to add detection for emacs etc., so let's keep things simple.] Cc: Andrey Vagin <avagin@openvz.org> Cc: Cyrill Gorcunov <gorcunov@openvz.org> Signed-off-by:
Kir Kolyshkin <kir@openvz.org> Reviewed-by:
-
Andrey Vagin authored
Signed-off-by:
-
Tycho Andersen authored
I have no idea if this is really the right long term fix, but at least it allows me to build on make 4.1 :) Closes #117 Signed-off-by:
-
Dmitry Safonov authored
In glibc 2.23 PTRACE_{GET,SET}SIGMASK are defined at <ptrace.h>: http://fossies.org/diffs/glibc/2.22_vs_2.23/sysdeps/unix/sysv/linux/powerpc/sys/ptrace.h-diff.html Move them under ifdef to fix the following errors: In file included from arch/x86/crtools.c:14:0: ~/tools/criu/criu/include/ptrace.h:47:0: error: "PTRACE_GETSIGMASK" redefined [-Werror] #define PTRACE_GETSIGMASK 0x420a ^ In file included from include/ptrace.h:5:0, from arch/x86/crtools.c:14: /usr/include/sys/ptrace.h:156:0: note: this is the location of the previous definition #define PTRACE_GETSIGMASK PTRACE_GETSIGMASK ^ In file included from arch/x86/crtools.c:14:0: ~/tools/criu/criu/include/ptrace.h:48:0: error: "PTRACE_SETSIGMASK" redefined [-Werror] #define PTRACE_SETSIGMASK 0x420b ^ In file included from include/ptrace.h:5:0, from arch/x86/crtools.c:14: /usr/include/sys/ptrace.h:159:0: note: this is the location of the previous definition #define PTRACE_SETSIGMASK PTRACE_SETSIGMASK ^ cc1: all warnings being treated as errors Fixes #119 Signed-off-by:
-
- 25 Feb, 2016 1 commit
-
-
Pavel Emelyanov authored
Signed-off-by:
-
- 24 Feb, 2016 4 commits
-
-
Cyrill Gorcunov authored
Will beautify later -- just explicitly link with syscalls and arch-lib, cause we're switched to phonytargets for deps sake. Signed-off-by:
-
Laurent Dufour authored
Cc: Cyrill Gorcunov <gorcunov@gmail.com> Signed-off-by:
-
Kir Kolyshkin authored
With GNU make, using .PHONY leads to the same result as using the "FORCE" hack, but unlike FORCE is not a hack. Since we do not plan to use any other make implementation than GNU make, let's not provide FORCE in nmk. For more details about why PHONY is better than FORCE, see http://www.gnu.org/software/make/manual/make.html#index-FORCESigned-off-by:
-
Kir Kolyshkin authored
With GNU make, using .PHONY leads to the same result as using the "FORCE" hack, but unlike FORCE is not a hack. Since we do not plan to use any other make than GNU make, let's just mark phony targets as such. For more details about why PHONY is better than FORCE, see http://www.gnu.org/software/make/manual/make.html#index-FORCE [v2: rebased and splitted] Signed-off-by:
-
