- 19 Feb, 2014 4 commits
-
-
Jamie Madill authored
We had accidentally left out the y and l files from our generation scripts, causing us to miss several instances of updated enum names. BUG=angle:550 Change-Id: I8790742fbaab5435e4c0db4f61c3e8194a231550 Reviewed-on: https://chromium-review.googlesource.com/186972Reviewed-by:
Nicolas Capens <nicolascapens@chromium.org> Tested-by:
Jamie Madill <jmadill@chromium.org>
-
Jamie Madill authored
We had accidentally blocked mipmap generation for all unsized formats in ES3, because they were marked as non-renderable. The spec specifically allows for mipmaps with unsized formats in GenerateMipmap. BUG=angle:550 Change-Id: Iee7eddd011f4fe86b690422723620593f8049ba6 Reviewed-on: https://chromium-review.googlesource.com/186971Reviewed-by:
Shannon Woods <shannonwoods@chromium.org> Tested-by:
-
Jamie Madill authored
We had accidentally marked 'smoooth', 'centroid' and 'case' as reserved in ES2, when they are available as identifiers. Additionally 'sampler2DShadow' was marked as free, instead of reserved. BUG=angle:550 Change-Id: I33ec1b790ad5b4d9a6f4d2ea2e91cec1ca8e7320 Reviewed-on: https://chromium-review.googlesource.com/186970Reviewed-by:
-
Jamie Madill authored
This patch turns on the dynamic conversion logic for integer formats that we must convert to floating point for ESSL 1.0 compatibility. The converted formats are 1, 2 and 4 channel unnormalized integer. BUG=angle:560 Change-Id: I3bcc45f1103290613029bc54fdd6012dddb0a03d Reviewed-on: https://chromium-review.googlesource.com/186700Reviewed-by:
Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
- 18 Feb, 2014 1 commit
-
-
Geoff Lang authored
This moves the start-up and tear-down code for D3DCompiler.dll into an object that Renderer9 and Renderer11 use. This will help future efforts to remove references to HLSL at the GL/Renderer interface level. BUG=angle:558 Change-Id: I18fcf9b237265d69c1d7d2ea345696c8fd31df29 Reviewed-on: https://chromium-review.googlesource.com/185568Tested-by:
-
- 14 Feb, 2014 7 commits
-
-
Jamie Madill authored
We can use the format tables and input layout to determine when we need to generate conversion code in the vertex shader. BUG=angle:560 Change-Id: Ib64fb16823bf78ed6432fba283b0d24bcb673a76 Reviewed-on: https://chromium-review.googlesource.com/185195Reviewed-by:
-
Jamie Madill authored
The patch refactors the vertex input HLSL to allow for dynamic conversion in the future. Using a placehold keyword, the code replaces that stub with the vertex input conversion logic. BUG=angle:560 Change-Id: I100c50cb4decd45b0f0c10d7c2c22583368b773e Reviewed-on: https://chromium-review.googlesource.com/185194Reviewed-by:
-
Jamie Madill authored
Our caching of compiled vertex shaders according to their input signature needs a consistent input signature to match up shaders. Since the linking step currently shuffles attributes according to API specified locations we need a separate list. BUG=angle:560 Change-Id: Icc4d5df1d37ae895d5c882b86e60f05b5c268461 Reviewed-on: https://chromium-review.googlesource.com/185193Reviewed-by:
-
Jamie Madill authored
With dynamic vertex conversion the GPU, we will have different input layouts resulting in different executables. This patch adds a way of mapping the input layouts to vertex executables. BUG=angle:560 Change-Id: Ie36f2f8ac2dfcb96f562af577d31f57d6d89b447 Reviewed-on: https://chromium-review.googlesource.com/185192Reviewed-by:
-
Jamie Madill authored
DynamicHLSL encapsulates all HLSL generation we need outside the the shader translator, such as for linking between input and output stages and point sprite geometry shaders. BUG=angle:560 Change-Id: Ib5079aa102000a7c37b166bcbe26b09cc82f8932 Reviewed-on: https://chromium-review.googlesource.com/185191Reviewed-by:
-
Geoff Lang authored
Instead of doing a masked clear when a texture has mismatched channel counts, set the clear color to the default for the missing channels and do a regular clear. This catches es3proto up to master in terms of clear performance. BUG=angle:557 Change-Id: I4f16976154af7735f5bb2b0b17adb3a264872ec0 Reviewed-on: https://chromium-review.googlesource.com/186611Reviewed-by:
-
Jamie Madill authored
Port changes from master branch with commit hash embedding to the development branch. BUG=angle:529 Change-Id: I307bf44103913aaaeb875a2e3ade1fab77838c48 Reviewed-on: https://chromium-review.googlesource.com/186390Reviewed-by:
-
- 13 Feb, 2014 1 commit
-
-
Jamie Madill authored
Certain vertex formats using fewer than 4 bytes could confuse our buffer code, giving us improperly aligned offsets. Aligning vertex data across the board should fix the problem, and minimal cost. Change-Id: I6884b4ddcfb91ef309a970593aaf7932cb65712e Reviewed-on: https://chromium-review.googlesource.com/185196Reviewed-by:
-
- 12 Feb, 2014 4 commits
-
-
Zhenyao Mo authored
ANGLEBUG=554 TEST=webgl conformance test on mac: shaders-with-varyings.html r=kbr@chromium.org,nicolascapens@chromium.org cc=alokp@chromium.org,shannonwoods@chromium.org Change-Id: I2e692d43fb15f1cf3ade3e398020d1fedb2b32f0 Reviewed-on: https://chromium-review.googlesource.com/185922Reviewed-by:
Zhenyao Mo <zmo@chromium.org> Tested-by:
Kenneth Russell <kbr@chromium.org> Reviewed-by:
-
Geoff Lang authored
Change-Id: I0ead14916ec677c97afb53f4c31a903e71fbc70f Reviewed-on: https://chromium-review.googlesource.com/185862Reviewed-by:
-
Jamie Madill authored
We were checking for extra conditions that apply to ES3 within all ES2 contexts, causing the webgl depth texture conformance test to fail. BUG=angle:550 Change-Id: Iae9a4880b9efc6be8c6857bce27a47d42fecc23a Reviewed-on: https://chromium-review.googlesource.com/186122Reviewed-by:
-
Geoff Lang authored
Change-Id: I56e0f5ff11cebd8b9415081b930a804dcce9530c Reviewed-on: https://chromium-review.googlesource.com/185566Tested-by:
-
- 11 Feb, 2014 8 commits
-
-
Geoff Lang authored
Change-Id: Ied10d866627485c6cd3eb09e23c14937efc465f4 Reviewed-on: https://chromium-review.googlesource.com/185864Tested-by:
-
Geoff Lang authored
Change-Id: If09d621cc249754caa57b057f76ff84726c18265 Reviewed-on: https://chromium-review.googlesource.com/185861Tested-by:
-
Geoff Lang authored
Change-Id: I020bbf8179c08574f5fc0a207db03d7456eeeba1 Reviewed-on: https://chromium-review.googlesource.com/185860Reviewed-by:
-
Geoff Lang authored
Change-Id: I1f836575824ff9df17fe3832997f06acc7464746 Reviewed-on: https://chromium-review.googlesource.com/185032Reviewed-by:
-
Geoff Lang authored
Change-Id: I83b68286b6f3dd369081cb2188af27c86b3de6eb Reviewed-on: https://chromium-review.googlesource.com/185865Reviewed-by:
-
Geoff Lang authored
Change-Id: I67add42cc69ec372840c67656bd2c77c623b23d0 Reviewed-on: https://chromium-review.googlesource.com/185567Reviewed-by:
-
Geoff Lang authored
Change-Id: I0e4c79cb7eaacf5403203b7b65094a54405ad47a Reviewed-on: https://chromium-review.googlesource.com/185863Reviewed-by:
-
Geoff Lang authored
Change-Id: I0aab43b146f87259898db57acb9f74fdbfdb1b96 Reviewed-on: https://chromium-review.googlesource.com/184580Reviewed-by:
-
- 10 Feb, 2014 6 commits
-
-
Geoff Lang authored
Cleaned up the agument parsing and added support for excluding files/folders. BUG=angle:552 Change-Id: I6dcf965f9b92ce4aa1c0b64bd4f9a0420883bcf6 Reviewed-on: https://chromium-review.googlesource.com/185204Reviewed-by:
-
Geoff Lang authored
BUG=angle:552 Change-Id: I9f357c68c258c7731cf0d5c2ceed3a2875b21228 Reviewed-on: https://chromium-review.googlesource.com/185142Reviewed-by:
-
Geoff Lang authored
BUG=angle:498 Change-Id: Ib60c39e206003ae67c93769e35f7f9ef790ce9f4 Reviewed-on: https://chromium-review.googlesource.com/184396Reviewed-by:
-
Geoff Lang authored
Change-Id: I66f36623eb5fb60fbb92d16af4f9854688b9484c Reviewed-on: https://chromium-review.googlesource.com/185190Tested-by:
-
Jamie Madill authored
Our restriction to 4-byte alignments caused us to always convert formats like R8 or R16, or R8G8 instead of allowing D3D to handle these formats naturally. From experimentation the proper rule seems to be "4-bytes, unless the vertex format is 1- or 2-byte size." I couldn't find supporting documentation on MSDN, but it seems common sense. Change-Id: I9e38fe86790e34024d1ae912a809a3d3cdc02ccf Reviewed-on: https://chromium-review.googlesource.com/184523Tested-by:
-
Jamie Madill authored
Since we now have Renderer methods to query formats for vertex inputs, we can get rid of the remaining calls to VertexBuffer which we used to query properties of vertex formats. Change-Id: I9e57d1d9bf80296aef271877fdd97a16d29114af Reviewed-on: https://chromium-review.googlesource.com/184522Tested-by:
-
- 07 Feb, 2014 1 commit
-
-
Nicolas Capens authored
BUG=339860 Change-Id: I64600db4bef62236c6acd8728902471df06a8f0a Reviewed-on: https://chromium-review.googlesource.com/185380Tested-by:
-
- 06 Feb, 2014 2 commits
-
-
Jamie Madill authored
With the new support for vertex formats living in the format utils, we can use the Renderer class to query vertex properties instead of the VertexBuffer class. Change-Id: Id76bd4dbf63855016faccef703521887f50da3b9 Reviewed-on: https://chromium-review.googlesource.com/184521Reviewed-by:
-
Jamie Madill authored
Vertex format information is useful for querying format properties with the new vertex translation logic with dynamic shaders. Change-Id: I1935c75f1e3920dc1d4a364a2d6968c82692e476 Reviewed-on: https://chromium-review.googlesource.com/184520Reviewed-by:
-
- 05 Feb, 2014 6 commits
-
-
Jamie Madill authored
Moving the vertex format information to a globally accessible helper file makes them available for the dynamic shaders logic. Change-Id: I67d6280c0603d9c601ca504751c8905677c2426f Reviewed-on: https://chromium-review.googlesource.com/184399Reviewed-by:
-
Jamie Madill authored
A VertexFormat is similar to a simplified VertexAttribute, without extra unncessary data. VertexFormats will be useful for GPU vertex conversions with dynamic shader compilation. Change-Id: I4e60e0a12d5f081fb52d95c8977c43d481b33bff Reviewed-on: https://chromium-review.googlesource.com/184398Reviewed-by:
-
Jamie Madill authored
With dynamic shaders we may have multiple shader executables per program binary. We must store the uniforms outside the executable, in the program binary, to be consistent between variations. Change-Id: I1b29a5d78c72dede8562d4878569b609536ba074 Reviewed-on: https://chromium-review.googlesource.com/183586Reviewed-by:
-
Jamie Madill authored
With dynamic shaders we may have multiple shader executables per program binary. We must store the uniforms outside the executable, in the program binary, to be consistent between variations. Change-Id: I1bec83dfb83ee9bb422448b157338a46e8e7ba56 Reviewed-on: https://chromium-review.googlesource.com/183585Tested-by:
-
Geoff Lang authored
In preparation for transform feedback, BufferStorage11 needs to be able to handle a non-staging buffer being updated. Each D3D11 buffer now has an incrementing data revision associated with it so the most up-to-date buffer is always known. Staging buffers are now represented like any other DirectBuffer. Change-Id: I6e881867cb2bd02d600213d08cce3ebba316c525 Reviewed-on: https://chromium-review.googlesource.com/184395Tested-by:
-
Jamie Madill authored
With certain selection statements with a vertex input in the condition and side-effects in the else-block, we'd run in to a D3D9 compiler bug which would cause incorrect results. We can work around this bug in D3D9 by selectively rewriting these statements to use an 'else if' clause instead of 'else'. BUG=322794 Change-Id: I93c96fb201ff4959c00d9a36321faac7e0343278 Reviewed-on: https://chromium-review.googlesource.com/184681Reviewed-by:
-
