On 11/27/2012 01:04 PM, Ian Romanick wrote:
> On 11/21/2012 05:43 PM, Chad Versace wrote:
>> Before this patch, intelInitScreen2 set DRIScreen::api_mask with the hacky
>> heuristic below:
>>
>> if (gen >= 3)
>> api_mask = GL | GLES1 | GLES2;
>> else
>> api_mask = 0;
>>
>> I don't understand why this hack works with gen2 (i830), or even if it
>> works properly at all. I don't care enough to investigate. On first
>> glance, it appears that this will cause every EGLConfig on i830 to have
>> EGL_RENDERABLE_TYPE=0, and thus prevent eglCreateContext from ever
>> succeeding. Anyway, moving on to living drivers...
>
> I think this doesn't work. My guess is that EGL just won't work on i8xx.
>
>> With the arrival of EGL_OPENGL_ES3_BIT_KHR, this heuristic is now
>> insufficient. We must enable the GLES3 bit if and only if the driver is
>> capable of creating a GLES3 context. This requires us to determine the
>> maximum supported context version supported by the hardware/driver for
>> each api *during initialization of intel_screen*.
>>
>> Therefore, this patch adds four new fields to intel_screen which indicate
>> the maximum supported context version for each api:
>> max_gl_core_version
>> max_gl_compat_version
>> max_gl_es1_version
>> max_gl_es2_version
>>
>> The api mask is now correctly set as:
>>
>> if (max_gl_core_version > 0 || max_gl_compat_version > 0)
>> api_mask |= GL;
>> if (max_gl_es1_version > 0)
>> api_mask |= GLES1;
>> if (max_gl_es2_version > 0)
>> api_mask |= GLES2;
>>
>> Tested against gen6 with piglit egl-create-context-verify-gl-flavor.
>> Verified that this patch does not change the set of exposed EGL context
>> flavors.
>>
>> Signed-off-by: Chad Versace <chad.vers...@linux.intel.com>
>> +static void
>> +set_max_gl_versions(struct intel_screen *screen)
>> +{
>> +#ifdef TEXTURE_FLOAT_ENABLED
>> + bool has_texture_float = true;
>> +#else
>> + bool has_texture_float = false;
>> +#endif
>> +
>> + if (screen->gen >= 7) {
>> + screen->max_gl_es1_version = 11;
>> +
>> + if (has_texture_float && screen->kernel_has_gen7_sol_reset) {
>> + screen->max_gl_core_version = 31;
>> + screen->max_gl_compat_version = 30;
>> + screen->max_gl_es2_version = 30;
>> + } else {
>> + screen->max_gl_core_version = 0;
>> + screen->max_gl_compat_version = 21;
>> + screen->max_gl_es2_version = 20;
>> + }
>> +
>> + } else if (screen->gen >= 6) {
>> + screen->max_gl_es1_version = 11;
>> +
>> + if (has_texture_float) {
>> + screen->max_gl_core_version = 31;
>> + screen->max_gl_compat_version = 30;
>> + screen->max_gl_es2_version = 30;
>> + } else {
>> + screen->max_gl_core_version = 0;
>> + screen->max_gl_compat_version = 21;
>> + screen->max_gl_es2_version = 20;
>> + }
>> +
>> + } else if (screen->gen >= 4) {
>> + screen->max_gl_core_version = 0;
>> + screen->max_gl_compat_version = 21;
>> + screen->max_gl_es2_version = 20;
>> + screen->max_gl_es1_version = 11;
>> +
>> + } else if (screen->gen >= 3) {
>> + bool has_fragment_shader = driQueryOptionb(&screen->optionCache,
>> "fragment_shader");
>> + bool has_occlusion_query = driQueryOptionb(&screen->optionCache,
>> "stub_occlusion_query");
>> +
>> + screen->max_gl_core_version = 0;
>> + screen->max_gl_es1_version = 11;
>> +
>> + if (has_fragment_shader && has_occlusion_query) {
>> + screen->max_gl_compat_version = 21;
>> + } else {
>> + screen->max_gl_compat_version = 15;
>> + }
>> +
>> + if (has_fragment_shader) {
>> + screen->max_gl_es2_version = 20;
>> + } else {
>> + screen->max_gl_es2_version = 0;
>> + }
>> +
>> + } else if (screen->gen >= 2) {
>> + screen->max_gl_core_version = 0;
>> + screen->max_gl_compat_version = 13;
>> + screen->max_gl_es2_version = 0;
>> + screen->max_gl_es1_version = 11;
>> +
>> + } else {
>> + assert(false);
>> + }
>
> I'd make this just "else" with 'assert(screen->gen == 2)'. I also think this
> would be better as a switch-statement.
Agreed that a switch statement is better. Expect a v2 of this patch.
>> @@ -1115,18 +1193,17 @@ __DRIconfig **intelInitScreen2(__DRIscreen *psp)
>>
>> intel_override_separate_stencil(intelScreen);
>>
>> - api_mask = (1 << __DRI_API_OPENGL);
>> -#if FEATURE_ES1
>> - api_mask |= (1 << __DRI_API_GLES);
>> -#endif
>> -#if FEATURE_ES2
>> - api_mask |= (1 << __DRI_API_GLES2);
>> -#endif
>> + intelScreen->hw_has_swizzling = intel_detect_swizzling(intelScreen);
>>
>> - if (IS_9XX(intelScreen->deviceID) || IS_965(intelScreen->deviceID))
>> - psp->api_mask = api_mask;
>> + set_max_gl_versions(intelScreen);
>>
>> - intelScreen->hw_has_swizzling = intel_detect_swizzling(intelScreen);
>> + psp->api_mask = 0;
>> + if (intelScreen->max_gl_compat_version > 0)
>> + psp->api_mask |= (1 << __DRI_API_OPENGL);
>
> It seems like this should be unconditional. I know that's different behavior
> than the existing code, but desktop OpenGL is supported by every piece of
> hardware this driver can drive.
I'll make the change. Unless we start supporting building the driver *without*
desktop GL, there's no reason why it shouldn't be unconditional.
>
>> + if (intelScreen->max_gl_es1_version > 0)
>> + psp->api_mask |= (1 << __DRI_API_GLES);
>> + if (intelScreen->max_gl_es2_version > 0)
>> + psp->api_mask |= (1 << __DRI_API_GLES2);
>>
>> psp->extensions = intelScreenExtensions;
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
