The follows( ) function uses pa->os_lign when determining
whether an output section pa immediately follows another
section pb.
However, emits( ) was not using this alignment information
when laying out and padding the output sections. This seems
to be a bug.
I suspect that mach/arm/cv/cv.c might need a similar fix.
Copy rhead() and rsect() from aslod to aelflod, so aelflod can work
for machine with 64-bit long.
In aelflod, fix ELF header so file(1) no longer reports "corrupted
section header size".
rhead() and rsect() had assumed sizeof(long) == 4, but OpenBSD/amd64
has sizeof(long) == 8. The problem revealed itself when sect->os_lign
became zero, and align() divided by zero.
