llvmpipe: enable coherent FB fetch extension
llvmpipe's fragment shaders are always run sequentially and in API order for a single tile, so it's impossible to have out of order render target writes requiring fetch barriers.
Bugs fixed in this series were actually breaking most tests for coherent variant of GL_EXT_shader_framebuffer_fetch
. All deqp and piglit tests (barring 8/16xMSAA related skips) for it are passing now.