Eric your patch seems closest to correct.
You'll probably want to keep the resize2fs() call after the zeroing?
Also there is the caveat that if size is not a multiple of 4MiB the
skip value may be too small and overwrite FS structures.
So you may need another dd invocation to deal with the slop at the start.
Eric your patch seems closest to correct.
You'll probably want to keep the resize2fs() call after the zeroing?
Also there is the caveat that if size is not a multiple of 4MiB the
skip value may be too small and overwrite FS structures.
So you may need another dd invocation to deal with the slop at the start.