errseq: eliminate special limitation for macro MAX_ERRNO

Current errseq implementation depends on a very special precondition that
macro MAX_ERRNO must be (2^n - 1).

Eliminate the limitation by

- redefining macro ERRSEQ_SHIFT
- defining a new macro ERRNO_MASK instead of MAX_ERRNO for errno mask.

There is no plan to change the value of MAX_ERRNO, but this makes the
implementation more generic and eliminates the BUILD_BUG_ON().

Link: https://lkml.kernel.org/r/20250407-improve_errseq-v1-1-7b27cbeb8298@quicinc.com


Signed-off-by: Zijun Hu <quic_zijuhu@quicinc.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>