The 5.15-rc4 kernel prepatch is out for
testing.
One thing standing out in the diffs might be the m68k ‘set_fs()’
removal – not really a regression fix, but it has been pending for
a while, and it turned out that the problems attributed to it were
due to an entirely unrelated m68k signal handling issue. So with
that fixed, we could get rid of set_fs from another architecture.
See this article for information on
set_fs() and its removal.