[openib-general] RFC: detecting duplicate MAD requests

[Michael S. Tsirkin]

Quoting r. Sean Hefty <sean.hefty at intel.com>:
> One of the ideas then, is for the kernel umad module to learn which MADs
> generate responses.

I thought about this a bit, this seems to add even more state for MAD processing
engine, which sounds like a wrong approach:

Would keeping around MADs in the done list consume significant extra memory
resources? What limits this memory? Would a small client that would normally
just send RMPP, get a response and exit will be slowed down significantly while
the agent learns?  Would a buggy application confuse the umad module, corrupting
MAD processing for all other applications?

The original approach by Jack of detecting, and dropping, duplicate responses
instead of duplicate requests seemed much easier to me.  The only disadvantage
it has that I'm aware of is a slight performance hit for duplicate processing of
each request. But all the done_list scans proposed seem even more CPU intensive.

Can we discuss that approach once again please? The patch is here:

https://openib.org/svn/trunk/contrib/mellanox/patches/mad_rmpp_requester_retry.patch


-- 
MST