On 7/25/2011 9:43 AM, Tetsuo Handa wrote:

Casey Schaufler wrote:

quoted

quoted

OK. Then, the question is how to reduce performance loss by redundant
security_socket_sendmsg() calls.

Not to be splitting hairs, but if the packets are headed to
different destinations the calls to security_socket_sendmsg()
are not redundant, they are necessary and appropriate. What
you have with sendmmsg() is an optimization that sacrifices
correctness for performance.

Excuse me, but this thread is not trying to remove necessary and appropriate
security_socket_sendmsg() calls. Linux 3.0 was released without necessary and
appropriate security_socket_sendmsg() calls, and I'm trying to correct it (via
msg11504.html or msg11510.html) for Linux 3.0.x stable release.

I understand. Sorry if I did a poor job of jumping into
the thread.

quoted

I fear that you are going to find that the work you have
to do to reduce the number of calls is going to outweigh
the benefits of your optimization, as has been pointed out
earlier.

I fear it too. Unless many dozens (maybe some hundreds) of packets are sent by
sendmmsg(), msg11504.html might show better performance than msg11510.html .
But I don't have a machine to benchmark.

Is there some chance that the original authors could step up
to help with the benchmarking effort on this repair? Having been
on the end where I introduced problems more than once, I have a
good understanding of the principle "you broke it, you bought it".