Added perfbd.py and block device performance sampling in bench-runner

Based loosely on Linux's perf tool, perfbd.py uses trace output with
backtraces to aggregate and show the block device usage of all functions
in a program, propagating block devices operation cost up the backtrace
for each operation.

This combined with --trace-period and --trace-freq for
sampling/filtering trace events allow the bench-runner to very
efficiently record the general cost of block device operations with very
little overhead.

Adopted this as the default side-effect of make bench, replacing
cycle-based performance measurements which are less important for
littlefs.

scripts/prettyasserts.py

@@ -42,14 +42,14 @@ LEXEMES = {
 }
 
 
-def openio(path, mode='r'):
+def openio(path, mode='r', buffering=-1):
     if path == '-':
         if mode == 'r':
-            return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
+            return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
         else:
-            return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
+            return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
     else:
-        return open(path, mode)
+        return open(path, mode, buffering)
 
 def write_header(f, limit=LIMIT):
     f.writeln("// Generated by %s:" % sys.argv[0])