Fastfetch reports much smaller `used spaces` for disk module

[CarterLi]

We changed the logic of calculating byteUsed from f_blocks - f_bavail to f_blocks - f_bfree in #508, but it seems problematic for (at least) macOS

Currently the result of fastfetch matches neo(wo)fetch's, but it doesn't match the hardware.

Really strange.

[CarterLi]

CC @xzl01