On Tue, Jan 09, 2018 at 10:05:42PM -0800, John Stultz wrote:
> In order to get the hikey960, which uses the mali bifrost driver
> working with drm_hwcomposer, its needed to tweak some extension
> and funciton names used in the shaders.
>
> Specifically:
> * GL_OES_EGL_image_external_essl3 instead of
> GL_OES_EGL_image_external
> * texture() instead of texture2D()
>
> Which is configured using a build time definition.
Build time is kinda uncool, at least in the spirit of multiarch kernels
:-) Can't we try a few alternatives until the glsl compiler takes it? For
extensions you could/should even query them upfront and then pick the
right one.
-Daniel
> Credit to Matt Szczesiak for suggesting these changes to get
> hikey960 working!
>
> I'm a bit new to all this, and I expect there may be a better
> way to do this, so I'd love any feedback or comments!
>
> Change-Id: I2c8f08341ad086479b66241b903c79b00f2a0feb
> Cc: Marissa Wall <marissaw@xxxxxxxxxx>
> Cc: Sean Paul <seanpaul@xxxxxxxxxx>
> Cc: Dmitry Shmidt <dimitrysh@xxxxxxxxxx>
> Cc: Robert Foss <robert.foss@xxxxxxxxxxxxx>
> Cc: Matt Szczesiak <matt.szczesiak@xxxxxxx>
> Cc: Liviu Dudau <Liviu.Dudau@xxxxxxx>
> Cc: David Hanna <david.hanna11@xxxxxxxxx>
> Cc: Rob Herring <rob.herring@xxxxxxxxxx>
> Signed-off-by: John Stutlz <john.stultz@xxxxxxxxxx>
> ---
> glworker.cpp | 15 +++++++++++++--
> 1 file changed, 13 insertions(+), 2 deletions(-)
>
> diff --git a/glworker.cpp b/glworker.cpp
> index ca726bf..c35d1b6 100644
> --- a/glworker.cpp
> +++ b/glworker.cpp
> @@ -41,6 +41,17 @@
>
> #define MAX_OVERLAPPING_LAYERS 64
>
> +#ifdef USE_TEXTURE_FN
> + #define TEXTURE_STR "texture"
> +#else
> + #define TEXTURE_STR "texture2D"
> +#endif
> +
> +#ifdef USE_IMAGE_EXTERNAL_ESSL3
> + #define IMAGE_EXTERNAL_STR "GL_OES_EGL_image_external_essl3"
> +#else
> + #define IMAGE_EXTERNAL_STR "GL_OES_EGL_image_external"
> +#endif
> namespace android {
>
> // clang-format off
> @@ -237,7 +248,7 @@ static std::string GenerateFragmentShader(int layer_count) {
> std::ostringstream fragment_shader_stream;
> fragment_shader_stream << "#version 300 es\n"
> << "#define LAYER_COUNT " << layer_count << "\n"
> - << "#extension GL_OES_EGL_image_external : require\n"
> + << "#extension " << IMAGE_EXTERNAL_STR << " : require\n"
> << "precision mediump float;\n";
> for (int i = 0; i < layer_count; ++i) {
> fragment_shader_stream << "uniform samplerExternalOES uLayerTexture" << i
> @@ -257,7 +268,7 @@ static std::string GenerateFragmentShader(int layer_count) {
> fragment_shader_stream << " if (alphaCover > 0.5/255.0) {\n";
> // clang-format off
> fragment_shader_stream
> - << " texSample = texture2D(uLayerTexture" << i << ",\n"
> + << " texSample = " << TEXTURE_STR << "(uLayerTexture" << i << ",\n"
> << " fTexCoords[" << i << "]);\n"
> << " multRgb = texSample.rgb *\n"
> << " max(texSample.a, uLayerPremult[" << i << "]);\n"
> --
> 2.7.4
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@xxxxxxxxxxxxxxxxxxxxx
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel
