[openib-general] [PATCH ] RFC IB/cm do not track remote QPN in timewait state

[Sean Hefty]

Michael S. Tsirkin wrote:
> It exposed a race in SDP. The patch itself does not lead to crashes -
> I re-attach it here for reference.
> As we discussed, this needs to be extended to handle DREQ retries
> properly.

I've committed this patch to SVN 9193.

The CM should already handle DREQ retries properly; however...

The CM timeout for a response can end up being close, or the same as the 
timewait time.  (All of my test apps will result in these values being the 
same.)  If a DREP is lost, the side that received the DREQ may enter and exit 
timewait before the DREQ is retried.  In this situation, the DREQ gets dropped 
repeatedly.

We will want to queue this patch for 2.6.19, if you can point Roland to your git 
tree.

Acked-by: Sean Hefty <sean.hefty at intel.com>