There seems to be a regression in the upcoming Linux 3.16-rc2 release candidate that I bisected down to this first bad commit: [dbb42748ac4929987c1449ecb296b39ef8956b62] drm/i915: Move the C3 LP write bit setup to gen3_init_clock_gating() for KMS.
Can you attach the dmesg with rc2 and dbb4274 reverted?
Hmm, for some reason I started having a hard time reproducing the issue on commit dbb4274.
I still have it on the 3.16-rc2 though and I'm doing another bisect.
The issue comes and goes during the bisect so let's see where it leads... perhaps it turns out to be some configuration setting.