Re: [-mm] Add an owner to the mm_struct (v8)

["Paul Menage" <menage@google.com>]

On Sat, Apr 5, 2008 at 10:48 AM, Balbir Singh <balbir@linux.vnet.ibm.com> wrote:
> Paul Menage wrote:
>  > On Sat, Apr 5, 2008 at 7:47 AM, Balbir Singh <balbir@linux.vnet.ibm.com> wrote:
>  >>  Repeating my question earlier
>  >>
>  >>  Can we delay setting task->cgroups = &init_css_set for the group_leader, until
>  >>  all threads have exited?
>  >
>  > Potentially, yes. It also might make more sense to move the
>  > exit_cgroup() for all threads to a later point rather than special
>  > case delayed group leaders.
>  >
>
>  Yes, that makes sense. I think that patch should be independent of this one
>  though? What do you think?

Yes, it would probably need to be a separate patch. The current
positioning of cgroup_exit() is more or less inherited from cpusets.
I'd need to figure out if a change like that would break anything.

>  >
>  > Yes, I agree it could potentially happen. But it seems like a strange
>  > thing to do if you're planning to be not have the same groupings for
>  > cpu and va.
>
>  It's easier to set it up that way. Usually the end user gets the same SLA for
>  memory, CPU and other resources, so it makes sense to bind the controllers together.
>

True - but in that case why wouldn't they have the same SLA for
virtual address space too?

>
>  >>  I measured the overhead of removing the delay_group_leader optimization and
>  >>  found a 4% impact on throughput (with volanomark, that is one of the
>  >>  multi-threaded benchmarks I know of).
>  >
>  > Interesting, I thought (although I've never actually looked at the
>  > code) that volanomark was more of a scheduling benchmark than a
>  > process start/exit benchmark. How frequently does it have processes
>  > (not threads) exiting?
>  >
>
>  I could not find any other interesting benchmark for benchmarking fork/exits. I
>  know that volanomark is heavily threaded, so I used it. The threads quickly exit
>  after processing the messages, I thought that would be a good test to see the
>  overhead.

But surely the performance of thread exits wouldn't be affected by the
delay_group_leader(p) change, since none of the exiting threads would
be a group leader. That optimization only matters when the entire
process exits.

Does oprofile show any interesting differences?

Paul