- Jul 26, 2019
-
-
Lina Versace authored
FINISHME: anv: Fix VkPhysicalDeviceExternalImageFormatInfo for VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT
-
Lina Versace authored
-
Lina Versace authored
-
Lina Versace authored
-
Lina Versace authored
-
Lina Versace authored
-
Lina Versace authored
Rename it to 'base_info', to distinguish it from 'drm_info' introduced in upcoming commit.
-
Lina Versace authored
-
Lina Versace authored
anv_get_image_format_features() already reports that we do not support linear images with depth/stencil formats. For consistency, teach anv_get_format_plane() to do the same.
-
Lina Versace authored
Replace field 'isl_tiling_flags' with 'drm_format_mod'. This cleanup makes the override features of anv_image_create_info behave more similarly to VkImageDrmFormatModifierExplicitCreateInfoEXT. The field 'isl_tiling_flags' offered more flexibility than needed. The driver set at most one flag, and when set the driver translated the sole flag into a DRM format modifier.
-
Lina Versace authored
Refactor only. No intended behavioral change. Much of the existing code assumes, with good justification, that tiling is either VK_IMAGE_TILING_LINEAR or VK_IMAGE_TILING_OPTIMAL. But VK_EXT_image_drm_format_modifier will void that assumption. Some examples where currently valid assumptions become invalid: - Code written as if (tiling != VK_IMAGE_TILING_OPTIMAL) but whose true intent is if (tiling == VK_IMAGE_TILING_LINEAR) will break when tiling is VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT. - Some code written as if (tilng == VK_IMAGE_TILING_LINEAR) will need updating to if (tiling == VK_IMAGE_TILING_LINEAR || (tiling == VK_IMAGE_TILING_DRM_FORMAT_MODIFIER_EXT && drm_format_mod == DRM_FORMAT_MOD_LINEAR)) - Many query functions with a VkImageTiling parameter will require an additional parameter for the DRM format modifier. The invalidity of many of the above cases become clearer if we generally abandon using VkImageTiling to indicate tiling, and instead use a new type that includes an optional DRM format modifier: struct anv_tiling { VkImageTiling vk; uint64_t drm_format_mod; }; This patch does not attempt to fix code with soon-to-be invalid assumptions. Instead, it restricts itself to the following changes: - Replace type of: - anv_get_format_features()::tiling - anv_get_format_plane()::tiling - anv_get_isl_format()::tiling - anv_image::tiling - Fold anv_image::drm_format_mod into anv_image::tiling. - Replace variables 'vk_tiling' with 'tiling.vk'.
-
Lina Versace authored
To allow use of DRM_FORMAT_MOD_INVALID in future inline functions for struct anv_tiling.
-
- Jul 24, 2019
-
-
Faith Ekstrand authored
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Instead of building a constant mask (which depends on knowing the subgroup size), we build an expression. Because the pass uses the nir_shader_lower_instructions helper, subgroup lowering will be run on any newly emitted instructions as well as the previously existing instructions. In particular, if the subgroup size is known, the newly emitted subgroup_size intrinsic will get turned into a constant and a later constant folding pass will clean it up. Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
The rules for gl_SubgroupSize in Vulkan require that it be a constant that can be queried through the API. However, all GL requires is that it's a uniform. Instead of always claiming that the subgroup size in the shader is 32 in GL like we have to do for Vulkan, claim 8 for geometry stages, the maximum for fragment shaders, and the actual size for compute. Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Instead of lowering the subgroup size so early, wait until we have more information. In particular, we're going to want different subgroup sizes from different stages depending on the API. We also defer lowering of subgroup masks because the ge/gt masks require the subgroup size to generate a subgroup mask. Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Faith Ekstrand authored
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
-
Sagar Ghuge authored
Tested on Gen > 9. v2: 1) Fix lowering 2) Keep a consistent i/u order (Matt Turner) Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com> Reviewed-by: Matt Turner <mattst88@gmail.com>
-
Add freeing of SubroutineIndexes to the _mesa_free_shader_state. Fixes: 4566aaaa ("mesa/subroutines: start adding per-context subroutine index support (v1.1)") Signed-off-by: Yevhenii Kolesnikov <yevhenii.kolesnikov@globallogic.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com>
-
warning: struct 'fs_reg' was previously declared as a class Fixes: e64be391 ("intel/compiler: generalize the combine constants pass") Reviewed-by: Eric Engestrom <eric.engestrom@intel.com> Reviewed-by: Matt Turner <mattst88@gmail.com> Signed-off-by: Andrii Simiklit <andrii.simiklit@globallogic.com>
-
Qiang Yu authored
temp read/write use negtive offset, and handle alignment==1 case. Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com> Signed-off-by: Qiang Yu <yuq825@gmail.com>
-
Eric Engestrom authored
Fixes: ed23335a ("gallium: use enums in p_shader_tokens.h (v2)") Signed-off-by: Eric Engestrom <eric.engestrom@intel.com> Reviewed-by: Marek Olšák <marek.olsak@amd.com>
-
Eric Engestrom authored
Fixes: b8e077da ("util: no-op __builtin_types_compatible_p() for non-GCC compilers") Signed-off-by: Eric Engestrom <eric.engestrom@intel.com> Reviewed-by: Matt Turner <mattst88@gmail.com>
-
Samuel Pitoiset authored
When a pipeline uses transform feedback, the driver fallbacks to the legacy path because NGG support for streamout is a non-trivial amount of work. AMDVLK also uses the legacy path for streamout, while RadeonSI uses the new NGG path. Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
-
Samuel Pitoiset authored
NGG GS for streamout requires a bunch of work, so enable it with the legacy path only for now. Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
-
Samuel Pitoiset authored
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
-
Samuel Pitoiset authored
Required for legacy streamout. Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
-
Samuel Pitoiset authored
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
-
Samuel Pitoiset authored
For some reasons, InstanceID is VGPR3 although StepRate0 is set to 1. Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com> Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
-
- Jul 23, 2019
-
-
Dave Airlie authored
The parameters were getting messy and I have to add a few more for compute shaders, so clean it up before proceeding. Reviewed-by: Roland Scheidegger <sroland@vmware.com>
-
Faith Ekstrand authored
I can't find a single place where nir_lower_io is called after going out of SSA which is the only real reason why you wouldn't do this. Returning SSA defs is more idiomatic and is required for the next commit. Reviewed-by: Matt Turner <mattst88@gmail.com>
-
Dylan Baker authored
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111016 Fixes: a47c525f ("meson: build glx") Acked-by: Emil Velikov <emil.velikov@collabora.com> Reviewed-by: Eric Anholt <eric@anholt.net>
-
Jan Zielinski authored
Ensure constant attributes stay constant with barycentric interpolation. Reviewed-by: Alok Hota <alok.hota@intel.com>
-
Jan Zielinski authored
Reviewed-by: Alok Hota <alok.hota@intel.com>
-
Jan Zielinski authored
Reviewed-by: Alok Hota <alok.hota@intel.com>
-
Jan Zielinski authored
Clipping is not correctly handled by the rasterizer - fixing this. Reviewed-by: Alok Hota <alok.hota@intel.com>
-