K 10
svn:author
V 6
marcel
K 8
svn:date
V 27
2002-05-19T20:19:07.000000Z
K 7
svn:log
V 433
Simplify IA64_CMPXCHG to avoid having braced-groups in expressions.
As a minor positive side-effect, code at -O0 is more optimal. As a
minor negative side-effect, certain boundary cases yield no better
code than non-boundary cases. For example, atomic_set_acq_32(p, 0)
does a useless logical OR with value 0. This was previously elimina-
ted as part of if/while optimizations. Non-boundary cases yield
identical code at -O1 and -O2.

END