Re: UVM wait states

[Taylor R Campbell <campbell+netbsd-tech-kern%mumble.net@localhost>]

   Date: Wed, 24 Feb 2016 17:37:13 +0800 (PHT)
   From: Paul Goyette <paul%whooppee.com@localhost>

   On Wed, 24 Feb 2016, Christos Zoulas wrote:

   > In article <20160224031758.7A34B60321%jupiter.mumble.net@localhost>,
   > Taylor R Campbell  <campbell+netbsd-tech-kern%mumble.net@localhost> wrote:
   >> The rule is that wchan messages should be at most 8 characters (see,
   >> e.g., condvar(9)).  These wchan messages satisfy that, as far as I can
   >> tell.  Maybe top(1) should be adjusted to show all 8 characters
   >> instead?
   >
   > It is 6 now because of /<n>...

   Hmmm.  I still have a 24-core machine (dual AMD 12-core processors) 
   sitting in my too-expensive-to-run-and-cool pile.  If I were to fire 
   this beast up, would top(1) be limited to displaying only 5 chars of 
   waitchan?

Maybe top(1) should have an option to include or exclude the CPU, just
like for the per-CPU vs global summary?

(Not hard to imagine getting to three digits in the near future --
theoretically, we already do support up to 256 CPUs on x86, if I
recall correctly.)