RFR: -Xlog:gc+ergo and heuristics UX fixes (pack of three)

[Aleksey Shipilev]

http://cr.openjdk.java.net/~shade/shenandoah/gc-ergo-ux/webrev.01/

This does three UX changes that turns gc+ergo and heuristics less noisy and more correct:

The pack includes:

 *) adaptive-cset: UX: Cleanup (adaptive) CSet selection message
      This unclutters the message, and aligns it better format-wise.

 *) shorter-ergo:  UX: Shorter gc+ergo messages from CSet selection
      Ditches some noisy lines that we don't really care about, and cleans up
      related code. Computes total garbage on the fly, without involving another
      walk over the regions, also skips CSet selection if we know we do not needed
      due to immediate garbage. This gives a marginal performance boost.

 *) pacer-tax:     UX: Pacer reports incorrect free size
      Fixes Pacer to report correct "free". Currently it reports free after subtracting
      non-taxable base, which is confusing.

Testing: hotspot_gc_shenandoah

--- Before:

 Adaptive CSet selection: free target = 35840M, actual free = 27968M; min cset = 7872M, max cset =
20976M
 Total Garbage: 55264M
 Immediate Garbage: 0M, 0 regions (0% of total)
 Garbage to be collected: 53777M (97% of total), 1927 regions
 Live objects to be evacuated: 7862M
 Live/garbage ratio in collected regions: 14%
 Free: 27968M, 874 regions (27% of total)
 Pacer for Evacuation. CSet: 7862M, Free: 25149M, Non-Taxable: 2794M, Alloc Tax Rate: 1.1x

--- After (with collectable garbage):

 Adaptive CSet Selection. Target Free: 35840M, Actual Free: 27904M, Target CSet: [7936M, 20928M]
 Collectable Garbage: 53824M (97% of total), 7927M CSet, 1930 CSet regions
 Immediate Garbage: 0M (0% of total), 0 regions
 Free: 27904M, 872 regions
 Pacer for Evacuation. CSet: 7927M, Free: 27904M, Non-Taxable: 2790M, Alloc Tax Rate: 1.1x

--- After (with only immediate garbage):

 Immediate Garbage: 11430M (99% of total), 2858 regions
 Free: 4908M, 1227 regions
 Pacer for Evacuation. CSet: 0M, Free: 4912M, Non-Taxable: 491M, Alloc Tax Rate: 1.1x


Thanks,
-Aleksey