modified chacha20 to do inline enc/dec with slice results.

chacha_mem.out

original: 
go test -count 1 -bench=Original -benchmem -memprofilerate=1 -memprofile=memoriginal.pprof
BenchmarkOriginal-8   	 1000000	     25551 ns/op	41038.30 MB/s	     592 B/op	      13 allocs/op

(pprof) top10
565.58MB of 565.64MB total (  100%)
Dropped 85 nodes (cum <= 2.83MB)
Showing top 10 nodes out of 11 (cum >= 565.58MB)
      flat  flat%   sum%        cum   cum%
  274.66MB 48.56% 48.56%   274.66MB 48.56%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20.NewWithRounds
  122.07MB 21.58% 70.14%   396.73MB 70.14%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20poly1305.(*chacha20Key).initialize
   91.55MB 16.19% 86.32%    91.55MB 16.19%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20poly1305.tag
   76.22MB 13.47% 99.80%   320.36MB 56.64%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20poly1305.(*chacha20Key).Seal
       1MB  0.18%   100%   565.58MB   100%  github.com/wirepair/chachazero.BenchmarkOriginal
    0.08MB 0.014%   100%   244.22MB 43.18%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20poly1305.(*chacha20Key).Open
         0     0%   100%   564.58MB 99.81%  github.com/wirepair/chachazero.original
         0     0%   100%   274.66MB 48.56%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20.New
         0     0%   100%   565.62MB   100%  runtime.goexit
         0     0%   100%   565.58MB   100%  testing.(*B).run1.func1




inline:
go test -count 1 -bench=Inline -benchmem -memprofilerate=1 -memprofile=meminline.pprof
BenchmarkInline-8   	 1000000	     20347 ns/op	51534.39 MB/s	     544 B/op	       9 allocs/op

(pprof) top10
519.80MB of 519.86MB total (  100%)
Dropped 94 nodes (cum <= 2.60MB)
Showing top 10 nodes out of 11 (cum >= 519.80MB)
      flat  flat%   sum%        cum   cum%
  274.66MB 52.83% 52.83%   274.66MB 52.83%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20.NewWithRounds
  122.07MB 23.48% 76.31%   396.73MB 76.31%  github.com/wirepair/chachazero.(*chacha20Key).initialize
   91.55MB 17.61% 93.93%    91.55MB 17.61%  github.com/wirepair/chachazero.tag
   30.52MB  5.87% 99.80%   274.66MB 52.83%  github.com/wirepair/chachazero.(*chacha20Key).Seal
       1MB  0.19%   100%   519.80MB   100%  github.com/wirepair/chachazero.BenchmarkInline
         0     0%   100%   244.14MB 46.96%  github.com/wirepair/chachazero.(*chacha20Key).Open
         0     0%   100%   518.80MB 99.80%  github.com/wirepair/chachazero.inline
         0     0%   100%   274.66MB 52.83%  github.com/wirepair/chachazero/vendor/github.com/codahale/chacha20.New
         0     0%   100%   519.84MB   100%  runtime.goexit
         0     0%   100%   519.80MB   100%  testing.(*B).run1.func1