On Thu, 28 Mar 2013 10:41:56 +0100 Daniel Vetter <daniel.vetter at ffwll.ch> wrote: > > + /* DP has a bunch of special case unfortunately, so mark the pipe > + * accordingly. */ > + bool has_dp_encoder; Looks pretty good, but I don't think this field is used anywhere? Maybe it belongs in a later patch? Also, this makes me wonder if we should be clearing the m_n regs somewhere and asserting for them in a few places. Definitely looks better than the current code though; CPU vs PCH writes sprinkled all about, and FDI thrown into the mix. It's probably a bit unfair to say the HSW mode set is confused though; it's just using existing code as best it can. These new bits are definitely clearer though. -- Jesse Barnes, Intel Open Source Technology Center
