On Thu, 15 Jun 2023 08:48:29 +0200 Christoph Hellwig <hch@xxxxxx> wrote: > Hi Song, > > the md bitmap file support is very problematic in how it bypasses the > file system for file access. I looked into fixing it in preparation > for making buffer_head optionals but had to give up because it is so > convoluted. This series includes the cleanups I've started which seem > useful even for the internal bitmap support, then makes the bitmap file > support conditional and adds a deprecation warning. > Hi Christoph, I think that it is worthy to make mdadm aware of that. For example, by requiring "--force" to make the volume with bitmap file now. Thanks, Mariusz
