On Fri, May 3, 2013 at 12:02 AM, Paulo Zanoni <przanoni at gmail.com> wrote: >> static void valleyview_crtc_enable(struct drm_crtc *crtc) >> @@ -5800,6 +5798,9 @@ static void haswell_modeset_global_resources(struct drm_device *dev) >> /* XXX: Should check for edp transcoder here, but thanks to init >> * sequence that's not yet available. Just in case desktop eDP >> * on PORT D is possible on haswell, too. */ >> + /* Even the eDP panel fitter is outside the always-on well. */ >> + if (I915_READ(PF_WIN_SZ(crtc->pipe))) > > So this patch made it to drm-intel-next-queued and the line above > causes "unclaimed register" errors. This happens even if we don't > disable the power well, and we were 100% clear of these error messages > on that case, so I guess I can call this a "regression". Can you > please provide a fix? Well, that hunk switches from checking dev_priv->pch_pf_size (which is now in crtc->config, so we'd need to check the right pipe/cpu_transcoder first) to the register value, which is just plain bogus. While at it we could fix the XXX in haswell_crtc_disable, which is now possible with the pch pfit tracking. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch
