On Tue, Jun 13, 2017 at 04:31:26PM -0700, Nanley Chery wrote:
> On Tue, Jun 13, 2017 at 05:49:59PM +0300, Topi Pohjolainen wrote:
> > Signed-off-by: Topi Pohjolainen <topi.pohjolai...@intel.com>
> > ---
> > src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 29
> > +++++++++++++++++++++++++--
> > 1 file changed, 27 insertions(+), 2 deletions(-)
> >
> > diff --git a/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> > b/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> > index 253d833b13..78a223a7f3 100644
> > --- a/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> > +++ b/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> > @@ -285,6 +285,26 @@ intel_depth_format_for_depthstencil_format(mesa_format
> > format) {
> > }
> > }
> >
> > +static bool
> > +create_mapping_table(GLenum target, unsigned first_level, unsigned
> > last_level,
> > + unsigned depth0, struct intel_mipmap_level *table)
> > +{
> > + for (unsigned level = first_level; level <= last_level; level++) {
> > + const unsigned d = target == GL_TEXTURE_3D ? depth0 >> level :
> > depth0;
>
> There's a bug here. If the target is GL_TEXTURE_3D we should
> minify(depth0, level) to avoid setting a depth of 0.
Oops, definitely.
>
> This seems to be more than a refactor. Prior to this patch,
> brw_miptree_layout_gen6_hiz_stencil wouldn't shrink the number of slices
> per mipmap level as the level increases, but does so now.
Right. I actually missed that. How do want to handle that? I could write a
patch against brw_miptree_layout_gen6_hiz_stencil() doing the same thing there
(modifying the argument given to intel_miptree_set_level_info() but keeping
actual allocation size as it was in order to have space for level 0 qpitch).
>
> -Nanley
>
> > +
> > + table[level].slice = calloc(d, sizeof(*table[0].slice));
> > + if (!table[level].slice)
> > + goto unwind;
> > + }
> > +
> > + return true;
> > +
> > +unwind:
> > + for (unsigned level = first_level; level <= last_level; level++)
> > + free(table[level].slice);
> > +
> > + return false;
> > +}
> >
> > /**
> > * @param for_bo Indicates that the caller is
> > @@ -424,6 +444,12 @@ intel_miptree_create_layout(struct brw_context *brw,
> > }
> > }
> >
> > + if (!create_mapping_table(target, first_level, last_level, depth0,
> > + mt->level)) {
> > + free(mt);
> > + return NULL;
> > + }
> > +
> > /* Set array_layout to ALL_SLICES_AT_EACH_LOD when array_spacing_lod0
> > can
> > * be used. array_spacing_lod0 is only used for non-IMS MSAA surfaces on
> > * Gen 7 and 8. On Gen 8 and 9 this layout is not available but it is
> > still
> > @@ -1103,9 +1129,8 @@ intel_miptree_set_level_info(struct intel_mipmap_tree
> > *mt,
> > DBG("%s level %d, depth %d, offset %d,%d\n", __func__,
> > level, d, x, y);
> >
> > - assert(mt->level[level].slice == NULL);
> > + assert(mt->level[level].slice);
> >
> > - mt->level[level].slice = calloc(d, sizeof(*mt->level[0].slice));
> > mt->level[level].slice[0].x_offset = mt->level[level].level_x;
> > mt->level[level].slice[0].y_offset = mt->level[level].level_y;
> > }
> > --
> > 2.11.0
> >
> > _______________________________________________
> > mesa-dev mailing list
> > mesa-dev@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/mesa-dev
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
