On Mon, 5 Oct 2015 21:26:39 +0200 Jesper Dangaard Brouer [off-list ref] wrote:

On Fri, 2 Oct 2015 14:50:44 -0700 Andrew Morton [off-list ref] wrote:

[...] 

quoted

Deleting the BUG altogether sounds the best solution.  As long as the
kernel crashes in some manner, we'll be able to work out what happened.
And it's cant-happen anyway, isn't it?

To me WARN_ON() seems like a good "documentation" if it does not hurt
performance.  I don't think removing the WARN_ON() will improve
performance, but I'm willing to actually test if it matters.

I tested removing BUG/WARN_ON altogether, and it gives slightly worse
performance. The icache-misses only increase approx 14% (not 112% as
before).  This, I'm willing to attribute to some code alignment issue.

Thus, let us just keep the WARN_ON() and move along.

-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  Author of http://www.iptv-analyzer.org
  LinkedIn: http://www.linkedin.com/in/brouer