Fix block bitmap utilisation

[vinc]

I found another issue after fixing #636. We were using only 1/8 of every block bitmap to mark allocated blocks resulting in an exception when trying to use more than one eight of the disk:

~
> dhcp
[23.345445] NET IP 10.0.2.15/24
[23.348445] NET GW 10.0.2.2
[23.352444] NET DNS 10.0.2.3

~
> dsk u -b
size: 8.0M
used: 108K
free: 7.9M

~
> time http 10.0.2.2:8000/1000kb.txt => 1
Executed 'http 10.0.2.2:8000/1000kb.txt' in 14.685895s

~
> dsk u -b
size: 8.0M
used: 1.1M
free: 6.9M

~
> time http 10.0.2.2:8000/1000kb.txt => 2
DEBUG: panicked at src/sys/fs/bitmap_block.rs:25:13:
index out of bounds: the len is 512 but the index is 512
qemu-system-x86_64: terminating on signal 2

[vinc]

And now there's another issue around half of the disk:

~
> time http 10.0.2.2:8000/1000kb.txt => 3
Executed 'http 10.0.2.2:8000/1000kb.txt' in 14.993001s

~
> dsk u -b
size: 8.0M
used: 3.1M
free: 4.9M

~
> time http 10.0.2.2:8000/1000kb.txt => 4
DEBUG: panicked at src/sys/fs/block_device.rs:69:43:
index out of bounds: the len is 16360 but the index is 4294967295