On Thu, Dec 08, 2016 at 04:26:41PM +0000, Chris Wilson wrote: > On Thu, Dec 08, 2016 at 06:17:28PM +0200, Ville Syrjälä wrote: > > On Thu, Dec 08, 2016 at 08:45:31AM +0000, Chris Wilson wrote: > > > On Wed, Dec 07, 2016 at 07:28:10PM +0200, ville.syrjala@xxxxxxxxxxxxxxx wrote: > > > > From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> > > > > > > > > Use two-dimensional arrays and named initializers to make the > > > > overlay filter coefficient tables easier to parse for humans. > > > > > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx> > > > > > > Spot checking didn't reveal any typos. I presume that since they are u16 > > > array of arrays, gcc is not adding any padding between rows? > > > > Didn't see any. I suppose we could > > > > BUILD_BUG_ON(sizeof(y_static_hcoeffs) != N_PHASES * N_HORIZ_Y_TAPS * 2); > > BUILD_BUG_ON(sizeof(uv_static_hcoeffs) != N_PHASES * N_HORIZ_UV_TAPS * 2); > > > > for extra paranoia? > > I don't think it matters. Maybe a __packed and just leave it to gcc's > whims? Apparently __packed isn't a thing for arrays. So I went ahead and pushed the patch as is. I pushed all the other ones that had been reviewed as well. That leaves us with the mm.interruptible thing which probably needs actual thought, and the "nuke 830 MI_OVERLAY_OFF w/a" patch. I tries to slap on some cc:stables to the oops fixes since the offenders have by now reached v4.9. Thanks for the reviews. -- Ville Syrjälä Intel OTC _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
