Over in bug 1086964, I'm getting rid of the no_pgo hack on Layers.cpp and LayerTreeInvalidation.cpp that is no longer needed. In review, I was asked to try moving them from SOURCES back to UNIFIED_SOURCES since no_pgo was the reason for being there in the first place, but my attempt to do so on Try didn't go swimmingly. https://treeherder.mozilla.org/logviewer.html#?job_id=167999450&repo=try --> Fixed with a couple gfx:: sprinkled on SingleTiledContentClient.cpp https://treeherder.mozilla.org/logviewer.html#?job_id=167999443&repo=try https://treeherder.mozilla.org/logviewer.html#?job_id=167999462&repo=try https://treeherder.mozilla.org/logviewer.html#?job_id=167999461&repo=try --> Fixed with a couple mozilla::ipc:: in RefCountedShmem.cpp and a TextureForwarder.h include in TextureDIB.cpp

This has gone through a number of iterations on Try already, but doing one more final run for kicks. https://treeherder.mozilla.org/#/jobs?repo=try&revision=4d40be67682f1cf470b53b837c8e2b1dbfe5559c&group_state=expanded

Pushed by ryanvm@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/bfe4b0a2e835 Fix some non-unified build bustage in gfx/layers. r=lsalzman