punch.py

#!/usr/bin/python

import ctypes
import ctypes.util
import os
import re

null_re = re.compile('^\0*$')

c_off_t = ctypes.c_int64

def make_fallocate():
    libc_name = ctypes.util.find_library('c')
    libc = ctypes.CDLL(libc_name)

    _fallocate = libc.fallocate
    _fallocate.restype = ctypes.c_int
    _fallocate.argtypes = [ctypes.c_int, ctypes.c_int, c_off_t, c_off_t]

    del libc
    del libc_name

    def fallocate(fd, mode, offset, len_):
        res = _fallocate(fd.fileno(), mode, offset, len_)
        if res != 0:
            raise IOError(res, 'fallocate')

    return fallocate

fallocate = make_fallocate()
del make_fallocate

FALLOC_FL_KEEP_SIZE = 0x01
FALLOC_FL_PUNCH_HOLE = 0x02


def punch(filename, verbose, blocksize=None):
    if verbose:
        print "processing", filename
    with open(filename, 'r+') as f:
        if not blocksize:
            blocksize = os.fstatvfs(f.fileno()).f_bsize
        offset = 0
        length = 0
        while True:
            buf = f.read(blocksize)
            if not buf:
                break
            if null_re.match(buf):
                if verbose:
                    print "punching hole at offset", offset, "length", len(buf)
                fallocate(f, FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE,
                          offset, len(buf))
            offset = offset + blocksize

if __name__ == '__main__':
    import sys
    import argparse
    parser = argparse.ArgumentParser(
        description = "Punch out the empty areas in a file, making it sparse")
    parser.add_argument('file', metavar='FILE',
                        help='file(s) to modify in-place', nargs='+')
    parser.add_argument('-v', '--verbose', action="store_true", default=False,
                        help='be verbose')
    parser.add_argument('-b', '--blocksize', metavar='BS', default=None,
                        help='override detected filesystem blocksize')
    args = parser.parse_args()
    for filename in args.file:
        punch(filename, args.verbose, args.blocksize)
        

fixing a bug... setting -b results in an error

$ python punch.py.orig -b 4096 "$d"
Traceback (most recent call last):
  File "punch.py.orig", line 69, in <module>
    punch(filename, args.verbose, args.blocksize)
  File "punch.py.orig", line 46, in punch
    buf = f.read(blocksize)
TypeError: an integer is required

--- punch.py.orig       2014-07-18 10:01:01.189842667 +0200
+++ punch.py    2014-07-17 17:39:53.012197857 +0200
@@ -40,6 +40,8 @@
     with open(filename, 'r+') as f:
         if not blocksize:
             blocksize = os.fstatvfs(f.fileno()).f_bsize
+        else:
+            blocksize = int(blocksize)
         offset = 0
         length = 0
         while True:
@@ -67,4 +69,4 @@
     args = parser.parse_args()
     for filename in args.file:
         punch(filename, args.verbose, args.blocksize)