On Mon, Mar 23, 2015 at 10:35:16AM +0100, Daniel Vetter wrote: > On Sun, Mar 22, 2015 at 03:41:04PM +0000, Chris Wilson wrote: > > Delay the expensive read on the FPGA_DBG register from once per mmio to > > once per forcewake section when we are doing the general wellbeing > > check rather than the targetted error detection. This almost reduces > > the overhead of the debug facility (for example when submitting execlists) > > to zero whilst keeping the debug checks around. > > > > Signed-off-by: Chris Wilson <ch...@chris-wilson.co.uk> > > Cc: Mika Kuoppala <mika.kuopp...@intel.com> > > Cc: Paulo Zanoni <paulo.r.zan...@intel.com> > > Unfortunately the unclaimed reg stuff is most useful for debugging display > power well issues (it catches those while nothing else really does), and > this removes that facility. Can't we do tricks with using raw reads/writes > for forcewaked registers or something like that?
How about beefing up intel_uncore_check_errors() to enable one-shot mmio_debug as well? Then if it is an i915 error we would catch it on the next cycle. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
