[Mesa-dev] [PATCH] i965: Fix segfault in WebGL Conformance on Ivybridge

[kalyan kondapally]

The patch seems to fix the crash I used to see while running this
WebGL conformance test on my machine (Intel(R) Core(TM) i7-4600U CPU @
2.10GHz). and now I can see all the tests pass. I believe 78770 is a
different issue (Added my comments to 78770) and not actually dealing
with fixing the crash we saw on 64 bit machines.

Br,
Kalyan

On Mon, Nov 10, 2014 at 8:02 AM, Chad Versace <chad at kiwitree.net> wrote:
> Fixes regression of WebGL Conformance test texture-size-limit [1] on
> Ivybridge Mobile GT2 0x0166 with Google Chrome R38.
>
> Regression introduced by
>
>     commit 6c044231535b93c5d16404528946cad618d96bd9
>     Author: Kenneth Graunke <kenneth at whitecape.org>
>     Date:   Sun Feb 2 02:58:42 2014 -0800
>
>         i965: Bump GL_MAX_CUBE_MAP_TEXTURE_SIZE to 8192.
>
> The test regressed because the pointer offset arithmetic in
> intel_miptree_map_gtt() overflows for large textures. The pointer
> arithmetic is not 64-bit safe.
>
> This patch fixes the bugzilla ticket below on Ivybridge. This patch
> doesn't close the ticket, though, because the bug report is against
> Sandybridge, and QA cannot confirm the fix on that hardware.
>
> [1] https://github.com/KhronosGroup/WebGL/blob/52f0dc240f04dce31b1b8e2b8107fe2b8332dc90/sdk/tests/conformance/textures/texture-size-limit.html
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=78770
> Fixes: Intel CHRMOS-1377
> Reported-by: Lu Hua <huax.lu at intel.com>
> Signed-off-by: Chad Versace <chad at kiwitree.net>
> ---
>  src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 11 ++++++++++-
>  1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/intel_mipmap_tree.c b/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> index 8fda25d..24e217c 100644
> --- a/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> +++ b/src/mesa/drivers/dri/i965/intel_mipmap_tree.c
> @@ -1769,7 +1769,16 @@ intel_miptree_map_gtt(struct brw_context *brw,
>        y += image_y;
>
>        map->stride = mt->pitch;
> -      map->ptr = base + y * map->stride + x * mt->cpp;
> +
> +      /* The variables in below pointer arithmetic are 32-bit. The arithmetic
> +       * overflows for large textures.  Therefore the cast to intptr_t is
> +       * needed.
> +       *
> +       * TODO(chadv): Fix this everywhere in i965 by fixing the signature of
> +       * intel_miptree_get_image_offset() to use intptr_t.
> +       */
> +      map->ptr = base + (intptr_t) y * (intptr_t) map->stride
> +                      + (intptr_t) x * (intptr_t) mt->cpp;
>     }
>
>     DBG("%s: %d,%d %dx%d from mt %p (%s) %d,%d = %p/%d\n", __FUNCTION__,
> --
> 2.1.0-rc0
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev