Tests found excessive cache thrashing on x86-64 architectures. The
problem was traced to the alignment of sections. This patch sets up
an architecture-specific alignment of 512 bytes for x86-64 and 16
bytes for all others. It uses preprocessor directives that, hopefully,
make it easy to tweak for other architectures.
Starting early April, a number of commits were pushed to the mob branch
that did not reflect the interest, wishes, or code quality of the tcc
community. This commit reverts those commits, while cherry-picking
grishka's commit for win64 linkage issues.