SF: Fix repaint everything logic
This solves the same problem with dirty regions not being updated as Ia7a3dc15c96626a213c4459dac4f695d91540fb5, but it is better in a couple of ways: 1) It no longer updates dirty regions from a binder thread, which was likely causing the memory corruption in b/77919748. 2) It removes the distinction between repaintEverything and repaintEverythingLocked, which is not necessary since both operations in repaintEverything (updating an atomic and signaling a transaction) are threadsafe. Bug: 77335744 Bug: 77546473 Bug: 77919748 Test: Manual - on an ASAN build, played TouchLatency bouncy ball while toggling between GL and HWC composition; doesn't crash whereas it would crash in a matter of a couple minutes before Change-Id: Iecadfecb40e87e400a301de2dcad1664e154982d
Loading
Please register or sign in to comment