[CRIU] [PATCH 0/5] parasite: don't daemonize non-leader threads

[Alexander Kartashov]

On 05/27/2013 04:38 PM, Andrey Vagin wrote:
> For dumping thread we need to execute only one command, so its
> demonization is overload.
>
> This patch blocks all signals and dumps all thread related parameters in
> PARASITE_CMD_INIT_THREAD, then registers are dumped for this thread and
> the last step is unblocking thread signals.
>
> This patch simplifies a parasite code, because it removes execution of
> commands for threads.
>
> Thanks Pavel for this idea.
>
> Andrey Vagin (5):
>    parasite: return ability to execute code in a thread with specified
>      pid
>    parasite: don't daemonize threads
>    parasite: remove code which used for daemonized threads
>    parasite: remove struct parasite_thread_ctl
>    parasite: restore signal blocking mask on error paths
Thank you Andrew for this patchset, it fixed a multithreaded C/R on ARM :)
Nevertheless, the previous version of the daemonized parasite discovered
a stack corruption in the ARM kernel in the sys_futex() syscall.

>
>   arch/arm/crtools.c         |   4 +-
>   arch/x86/crtools.c         |   4 +-
>   cr-exec.c                  |   2 +-
>   include/parasite-syscall.h |  35 ++---
>   include/parasite.h         |  13 +-
>   mem.c                      |   4 +-
>   parasite-syscall.c         | 342 +++++++++++++++++++--------------------------
>   pie/parasite.c             | 218 +++++++----------------------
>   8 files changed, 221 insertions(+), 401 deletions(-)
>
> --
> 1.8.2
>
> _______________________________________________
> CRIU mailing list
> CRIU at openvz.org
> https://lists.openvz.org/mailman/listinfo/criu


-- 
Sincerely yours,
Alexander Kartashov

Intern
Core team

www.parallels.com

Skype: aleksandr.kartashov
Email: alekskartashov at parallels.com