Thanks for hte review, Tyler. I've pushed a new package freeing the fuse data and ensuring that "cgroup" and the controller are separated by '/'.
The pid passed to pid_to_ns_wrapper() comes from the fuse context - it is the pid of 'current' which is doing the read or write operation. If that task were to exit, the operation would presumably be aborted. I have not yet completely verified (in the kernel and fuse sources) or tested that killing the writing task is not possible (or that the fuse operation would be canceled). If it were not that would appear to be a serious kernel bug which should be fixed there. This should be test- able with a small fuse library which sleeps during write, and a pair of sibling tasks where one writes while the other kills the first. I can test that on monday, but probably not before. Note that in this case I would consider that a serious kernel bug - not a bug in lxcfs. The uid/gid/pid of the acting task is (AIUI) meant to be trusted. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1413405 Title: [MIR] lxcfs To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/lxcfs/+bug/1413405/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
