Comment 10 for bug 2060740

Hi Sudip, thanks for the work here.

> [ Where problems could occur ]
> There is no change in the code, it only changes a build dependency
> which will make fsattr functional.

Mostly this looks ok, but with some diffoscope analysis I note that
some binaries are indifferent to this change (fsmount_kernel and
fsmove change only NT_GNU_BUILD_ID) but others do notice this builddep
change (fsattr as expected, but also fsremap looks different.)
Do you mind investigating a little on fsremap? I think we want a bit
of a regression test there to try to ensure that making fsattr better
doesn't make fsremap worse. WDYT?