That was it; bcmp and bcopy.

Shouldn't we be EOL'ing those in favor of memcmp and memcpy (and
memmove)?  It seems awfully silly to have an additional interface
in the user namespace which is completely redundant with an
interface in the implementation namespace.

(I wouldn't mind seeing memzero added, either.)

(No, this isn't really related to the problem, it just reminded me that
we have these functions and I don't understand why.)

Anyway, thanks for all the advice; sure enough, the updated
.S files compiled and built.

All is well.

Is it a bug or a feature that, if there exists a file named
libc.so.12.14, in the current directory, but none in /usr/lib,
and a program wants a libc.so.12.14, that it fails completely to
load with "no such file or directory" rather than loading .11?

-s