On Tue, 2010-03-30 at 16:30 -0700, Roland Dreier wrote:
> > OpenFabrics added a IB_CQ_REPORT_MISSED_EVENTS flag to detect this
> 
> OpenFabrics?  Funny, I thought I added that flag.

You did. I can change the comment if you like.

> (This is not a serious objection ... I'm just being annoying.  Being
> serious, I'm happy to see this used outside of IPoIB and this looks like
> a good optimization)
> 
>  > This patch by itself should improve RDS performance over QLogic HCAs
> 
> Out of curiousity have you measured an improvement?

It is a great improvement over a bunch of printks and then
hitting BUG_ON() in __rds_ib_ring_used().

An actual performance difference will have to wait until
the underlying race is better understood. I thought this
patch might spark a discussion since the code in
rds_ib_send_cq_comp_handler() calls printk_ratelimit()
so the 0xdead message must have been seen before enough
to warrant adding the call and the comment:
/* In the error case, wc.opcode sometimes contains garbage */


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to