On Fri, 8 Oct 2010 17:46:49 -0700, Bryan Freed <[email protected]> wrote: > Has this been reported before?
No, smells like a hardware issue. Though since we will do a full modeset every time, there are quite a few paths that may fail. How many cycles does it take to trigger the error? > I dumped a bunch of i915 registers, but I do not see a significant > difference between "working" and "not working". Just some (undocumented, as > far as I can tell) sequential counter at 0x61100 and a scanline counter at > 0x71000. 0x61100 is the Analog Display Port Register (ADPA, or an FDI reg on Ironlake+). Some reserved bits but nothing that indicates a sequential counter. 0x71000 is indeed the pipe B Display Scan Line (DSL). -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
