console: add flash benchmark command

[jwise]

On Asterix, here's two different runs at main, both in airplane mode:

running flash read benchmark
benchmarking 4 bytes...
  -> 4 bytes: 16384 iters in 501 ticks = 29 us/iter
benchmarking 5 bytes...
  -> 5 bytes: 8192 iters in 421 ticks = 50 us/iter
benchmarking 16 bytes...
  -> 16 bytes: 8192 iters in 341 ticks = 40 us/iter
benchmarking 64 bytes...
  -> 64 bytes: 8192 iters in 427 ticks = 50 us/iter
benchmarking 256 bytes...
  -> 256 bytes: 4096 iters in 477 ticks = 113 us/iter
benchmarking 1024 bytes...
  -> 1024 bytes: 4096 iters in 1272 ticks = 303 us/iter

running flash read benchmark
benchmarking 4 bytes...
  -> 4 bytes: 16384 iters in 500 ticks = 29 us/iter
benchmarking 5 bytes...
  -> 5 bytes: 8192 iters in 421 ticks = 50 us/iter
benchmarking 16 bytes...
  -> 16 bytes: 8192 iters in 341 ticks = 40 us/iter
benchmarking 64 bytes...
  -> 64 bytes: 8192 iters in 429 ticks = 51 us/iter
benchmarking 256 bytes...
  -> 256 bytes: 4096 iters in 476 ticks = 113 us/iter
benchmarking 1024 bytes...
  -> 1024 bytes: 4096 iters in 1271 ticks = 303 us/iter

And on asterix-stop-mode:

running flash read benchmark
benchmarking 4 bytes...
  -> 4 bytes: 4096 iters in 691 ticks = 164 us/iter
benchmarking 5 bytes...
  -> 5 bytes: 4096 iters in 775 ticks = 184 us/iter
benchmarking 16 bytes...
  -> 16 bytes: 4096 iters in 739 ticks = 176 us/iter
benchmarking 64 bytes...
  -> 64 bytes: 4096 iters in 725 ticks = 172 us/iter
benchmarking 256 bytes...
  -> 256 bytes: 4096 iters in 900 ticks = 214 us/iter
benchmarking 1024 bytes...
  -> 1024 bytes: 4096 iters in 1706 ticks = 406 us/iter

So this really explains a lot. I'll try to smooth out some of that newtimer stuff while I'm on the plane.