Dave Hansen [hansendc@us.ibm.com] wrote:
| What's different about this patch from earlier ones?  What should I go
| look at, first?  Is the performance any better than before?
| 

Sorry, I missed this mail. 

Following are the major changes since my last post to Containers@.
They are listed in the Changelog section of [PATCH 0/17]. Would it
be better to list them elsewhere ?

        2.6.22-rc4-mm2-pidns2:

        - CLONE_NEWPID value has changed to 0x20000000 and so the pidns_exec.c
          example from previous version should be updated (or attached
          pidns_exec.c must be used).

        - Implement signal support with child pid namespaces.

        - Modify is_global_init() implementation to use a global variable

        - Bugfix: leaking 'struct pids'

        - Limit to and optimize for MAX_NESTED_PID_NS levels of nested pid
          namespaces which is currently set to 4. Support for deeper nesting
          requires trivial changes to the macro and to a global variable,
          pid_cahce_names[], and a kernel rebuild. clone() fails with -EINVAL
          when attempting to exceed MAX_NESTED_PID_NS levels of nesting.
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers