[Mesa-dev] [PATCH] intel: Fix a case when mapping large texture fails

nobled nobled at dreamwidth.org
Thu Mar 15 17:04:54 PDT 2012


On Thu, Mar 15, 2012 at 2:26 PM, Anuj Phogat <anuj.phogat at gmail.com> wrote:
>
>
> On Wed, Mar 7, 2012 at 4:28 PM, Anuj Phogat <anuj.phogat at gmail.com> wrote:
>>
>> On Mon 27 Feb 2012 11:45:46 AM PST, Anuj Phogat wrote:
>> > This patch handles a case when mapping a large texture fails
>> > in drm_intel_gem_bo_map_gtt(). These changes avoid assertion
>> > failure later in the driver as reported in following bugs:
>> >
>> > https://bugs.freedesktop.org/show_bug.cgi?id=44970
>> > https://bugs.freedesktop.org/show_bug.cgi?id=46303
>> >
>> > Signed-off-by: Anuj Phogat <anuj.phogat at gmail.com>

I've been testing this since yesterday applied on top of the 8.0
branch, and compiz is still crashing for me (bug 46303).

>> > ---
>> >  src/mesa/drivers/dri/intel/intel_mipmap_tree.c |   26
>> > ++++++++++++++++-------
>> >  src/mesa/drivers/dri/intel/intel_regions.c     |    7 ++++-
>> >  2 files changed, 23 insertions(+), 10 deletions(-)
>> >
>> > diff --git a/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
>> > b/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
>> > index 5290da4..507702a 100644
>> > --- a/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
>> > +++ b/src/mesa/drivers/dri/intel/intel_mipmap_tree.c
>> > @@ -702,15 +702,20 @@ intel_miptree_map_gtt(struct intel_context *intel,
>> >     y /= bh;
>> >
>> >     base = intel_region_map(intel, mt->region, map->mode);
>> > -   /* Note that in the case of cube maps, the caller must have passed
>> > the slice
>> > -    * number referencing the face.
>> > -    */
>> > -   intel_miptree_get_image_offset(mt, level, 0, slice, &image_x,
>> > &image_y);
>> > -   x += image_x;
>> > -   y += image_y;
>> >
>> > -   map->stride = mt->region->pitch * mt->cpp;
>> > -   map->ptr = base + y * map->stride + x * mt->cpp;
>> > +   if (base == NULL)
>> > +      map->ptr = NULL;
>> > +   else {
>> > +      /* Note that in the case of cube maps, the caller must have
>> > passed the
>> > +       * slice number referencing the face.
>> > +      */
>> > +      intel_miptree_get_image_offset(mt, level, 0, slice, &image_x,
>> > &image_y);
>> > +      x += image_x;
>> > +      y += image_y;
>> > +
>> > +      map->stride = mt->region->pitch * mt->cpp;
>> > +      map->ptr = base + y * map->stride + x * mt->cpp;
>> > +   }
>> >
>> >     DBG("%s: %d,%d %dx%d from mt %p (%s) %d,%d = %p/%d\n", __FUNCTION__,
>> >         map->x, map->y, map->w, map->h,
>> > @@ -1063,6 +1068,11 @@ intel_miptree_map(struct intel_context *intel,
>> >
>> >     *out_ptr = map->ptr;
>> >     *out_stride = map->stride;
>> > +
>> > +   if (map->ptr == NULL) {
>> > +      mt->level[level].slice[slice].map = NULL;
>> > +      free(map);
>> > +   }
>> >  }
>> >
>> >  void
>> > diff --git a/src/mesa/drivers/dri/intel/intel_regions.c
>> > b/src/mesa/drivers/dri/intel/intel_regions.c
>> > index bc83649..982d6cb 100644
>> > --- a/src/mesa/drivers/dri/intel/intel_regions.c
>> > +++ b/src/mesa/drivers/dri/intel/intel_regions.c
>> > @@ -124,7 +124,7 @@ intel_region_map(struct intel_context *intel, struct
>> > intel_region *region,
>> >      */
>> >
>> >     _DBG("%s %p\n", __FUNCTION__, region);
>> > -   if (!region->map_refcount++) {
>> > +   if (!region->map_refcount) {
>> >        intel_flush(&intel->ctx);
>> >
>> >        if (region->tiling != I915_TILING_NONE)
>> > @@ -133,7 +133,10 @@ intel_region_map(struct intel_context *intel,
>> > struct intel_region *region,
>> >        drm_intel_bo_map(region->bo, true);
>> >
>> >        region->map = region->bo->virtual;
>> > -      ++intel->num_mapped_regions;
>> > +      if (region->map) {
>> > +         ++intel->num_mapped_regions;
>> > +      region->map_refcount++;
>> > +      }
>> >     }
>> >
>> >     return region->map;
>>
>> Any comments on this patch?
>
>
> I am planning to push this today.
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>


More information about the mesa-dev mailing list