Chris Wilson enabled the compilation of SNA by default in a Git commit this afternoon. While that's happened, it's not being used by default. The SNA support and the different generational SNA back-ends are being compiled and built into the driver, but for now at least, it requires setting the AccelMethod within the xorg.conf to SNA rather than UXA or GLAMOR.
The SNA support is now being explictly built by default when using X.Org Server 1.10 or newer. It's too bad that it's not the default yet, but at least this will make it easy for users to switch to this means of faster and more robust 2D acceleration. This will be a change to be found in the xf86-video-intel 2.20 driver release.
SNA benchmarks for Ivy Bridge, benchmarks on Sandy Bridge, and even older generations of Intel hardware are quite positive towards SNA when compared to the default UXA or the new GLAMOR acceleration alternative. While the name might be confusing, SNA isn't limited to just Sandy/Ivy Bridge hardware but there's 2D back-ends going back to earlier Intel IGP generations.
There's no word yet when SNA may succeed UXA as the default choice, but today's Git activity is a step in that direction. Last I heard a few months back is that Intel was still resistant to changing the default 2D acceleration mode over fear of fallout and regressions for some users, especially after the problems in past years when transitioning from EXA to UXA, DRI2/KMS, GEM, and the other big underlying changes.