Patch 1 fixes a small performance regression introduced when moving
our atomics to __atomic primitives. The regression can be measured
on RMO architectures (I used aarch64); the effect is very small
but consistently measurable: for instance, rcutorture performance
degraded by about 0.3%.
Patch 2 originates from a recent discussion[1] that led me to
look into what guarantees can be assumed from our atomic_read/set
implementations.
Thanks,
Emilio
[1] https://lists.gnu.org/archive/html/qemu-devel/2016-05/msg03088.html
