Bug: angleproject:2747
Bug: angleproject:2763

Change-Id: I19f9e5ba760d3b5baf3997284523ee1db6bb8166
Reviewed-on: https://chromium-review.googlesource.com/1172206Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jie A Chen <jie.a.chen@intel.com>

D3D10+ has stricter rules on linking shader inputs with outputs than
the previous comments in the code indicated. Add comments and checks
related to this and structure the code in a way that makes it a bit
less prone to errors.

This page is intended to document the rules though it is somewhat
vague:
https://docs.microsoft.com/en-us/windows/desktop/direct3dhlsl/dx-graphics-hlsl-signatures

BUG=angleproject:2767
TEST=angle_end2end_tests

Change-Id: I236ec4cd5cbf3889fd2c97947ee81a6c5ae6a787
Reviewed-on: https://chromium-review.googlesource.com/1169818Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

This adds a new linking state to Program. If a Program is in linking
state, on the one hand the foreground thread may continue issuing more
GL calls, and on the other hand the background linking threads may be
accessing Program internally too. Without a proper constraint there
must be conflicts between them. For this purpose, we block any further
GL calls to Program until it's actually linked. In addition, we
prohibit parallel linking an active program, so that ProgramD3D does
not have to worry about such similar conflicts.

Also changes the WorkerThread to support limiting the number of
concurrently running worker threads.

BUG=chromium:849576

Change-Id: I52618647539323f8bf27201320bdf7301c4982e6
Reviewed-on: https://chromium-review.googlesource.com/1127495
Commit-Queue: Jie A Chen <jie.a.chen@intel.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Bug: angleproject:1944
Change-Id: I93f363dc5fb6096baec01584bbd2fb525a84909c
Reviewed-on: https://chromium-review.googlesource.com/1171502Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This completes the refactor for the D3D9/D3D11 back-ends.

Bug: angleproject:2752
Change-Id: Ie35a925f75c902d8f9fdab6bc33b3bb9c937f85b
Reviewed-on: https://chromium-review.googlesource.com/1167209
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Some Radeon cards in D3D11 would not report AMD or ATI in the
Renderer string. Fix this by also checking for 'Radeon'.

Bug: angleproject:2572
Change-Id: I866c3042e35448deff627efdb8e94fcbc68abf01
Reviewed-on: https://chromium-review.googlesource.com/1171363Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Implemented command buffer performance test based on
LunarG samples. Includes a lot of utility functions and
should build/run on Linux, Android

Bug: angleproject:2758
Change-Id: I8606802fb3e8e0621856860c45ef50ba836c26f8
Reviewed-on: https://chromium-review.googlesource.com/1162119
Commit-Queue: Omar El Sheikh <theoking@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

These tests need to use a valid program for some operations. Also make
them run in a D3D11 config since the OpenGL extension doesn't seem to
be available on the GPU config we run on the ANGLE CQ.

Also add tests that cover various subtle state change validation cache
bugs.

Bug: angleproject:2747
Change-Id: Idbd1d66854f2e3ca67b59836a370928e1fe8c3b3
Reviewed-on: https://chromium-review.googlesource.com/1167980Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

It's possible we were triggering a null pointer dereference in some
cases of lost device recovery. I was unable to reproduce this in a test
so this fix is speculative.

Bug: chromium:871178
Change-Id: I55b7790eabb81c25489bdfd105b5f2cbeedcc8db
Reviewed-on: https://chromium-review.googlesource.com/1169955Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Moving from State to Context allows the Context to update the State
Cache class directly. It also calls through to the State class to
update the Texture cache. This consolidates notification events into
the Context class. This is also in line with how we handle state event
updates in other gl classes.

Bug: angleproject:2747
Bug: angleproject:2763
Change-Id: Iff7dc7e46ee8768819235ebd151707cd2a03dfc9
Reviewed-on: https://chromium-review.googlesource.com/1166143
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

We don't support ES 3.0 or ES 3.1 in Vulkan yet. Remove Vulkan configs
from tests that don't support the higher features levels. We can
re-add them once we improve Vulkan support.

Bug: angleproject:2604
Change-Id: Icca57b5ec696aaeb2a571b79cd1a7869f806c176
Reviewed-on: https://chromium-review.googlesource.com/1169306Reviewed-by: Omar El Sheikh <theoking@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This also inlines a few accessors and changes the type of the texture
cache.

Bug: angleproject:2763
Change-Id: I82f3c508613f6284d835e8cb64808cfe26a56161
Reviewed-on: https://chromium-review.googlesource.com/1166142
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This mask cleans up some of the iteration logic in
State::syncProgramTextures. It will make it easier to optimize this
function in the future. This will also make it easier to recompute
the sampler type validation.

Leads to a 5% improvement in State::syncProgramTextures.

Bug: angleproject:2747
Bug: angleproject:2763
Change-Id: Ic9a555df843ad23b4c562e6e4a2d425bee58a856
Reviewed-on: https://chromium-review.googlesource.com/1164306
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

We needed to set the CWD for the mock ICD as well.

Bug: angleproject:2604
Change-Id: I1d285c478030056360fd2dceca78e231ac321219
Reviewed-on: https://chromium-review.googlesource.com/1169426Reviewed-by: Omar El Sheikh <theoking@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Instead of updating the cache directly we use event notifier functions.
For example instead of calling updateActiveAttribsMask in
Context::linkProgram we call onProgramExecutableChange. This makes the
code more self-documenting and easier to maintain.

Bug: angleproject:2747
Change-Id: Id27b58f646f6924db4c16e28609d6baf6b51c78e
Reviewed-on: https://chromium-review.googlesource.com/1164066
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This conflict should be validated even if the draw call would result in
no work being done on the GPU.

Bug: angleproject:2747
Change-Id: I6b32a07ff8a8598dd865270a3a935060f050a68f
Reviewed-on: https://chromium-review.googlesource.com/1162507
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

Split the parameter-dependent and parameter-independent validation.
This will more easily let us cache independent validation in Context.

Bug: angleproject:2747
Change-Id: I78a12798cd03a398392ca213bf51e2079295b97e
Reviewed-on: https://chromium-review.googlesource.com/1158610
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This perf test highlights the performance hotspots with
State::syncProgramTextures.

Also includes a fix to the perf test runner script.

Bug: angleproject:2763
Change-Id: I69ffa0cc0d5e023944495b7a1c844770a54f7ddc
Reviewed-on: https://chromium-review.googlesource.com/1166041
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Certain overflows wouldn't be detected when the attribute size was less
than the attribute size and we were drawing a small number of vertices.

Fix this and also set the sentinel value for an integer overflow to be
negative maxint instead of -1 to more effectily distinguish the cases.

Bug: angleproject:1391
Change-Id: I970d5e2a630c0a84c2c02ac0ac41ab1a395819fe
Reviewed-on: https://chromium-review.googlesource.com/1162264
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This enables the error refactor for D3D.

Bug: angleproject:2752
Change-Id: I7bb293d6c271d0bffb4d00b55029426b6cd0a42e
Reviewed-on: https://chromium-review.googlesource.com/1167208Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Bug: angleproject:2752
Change-Id: Ica77969c6e4b8d61d491396a4fd56b0864e4803c
Reviewed-on: https://chromium-review.googlesource.com/1163320
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

This is for reproducible build of angle related libraries.

Bug: 828669
Change-Id: I807c8fb01b976f6e7a71477ee934b9e0b0cdc684
Reviewed-on: https://chromium-review.googlesource.com/1164814Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This will be useful for validation state caching.

Bug: angleproject:2747
Change-Id: I0737adca7406f79b9e15429f30ae22e1299cd7e4
Reviewed-on: https://chromium-review.googlesource.com/1158611
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Refactor only.

Bug: angleproject:2747
Change-Id: I0ef7186daec54cd867ec2816f86f7205d836fd91
Reviewed-on: https://chromium-review.googlesource.com/1158609
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Also moves some code to the cpp.

Bug: angleproject:2747
Change-Id: I38468e6276d52d11d5751df6917c66f15ae61246
Reviewed-on: https://chromium-review.googlesource.com/1163822
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

BUG=864528

Change-Id: Iac418054ae56a2d9dd277d0474019997aeaa834b
Reviewed-on: https://chromium-review.googlesource.com/1163633Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

The improved cache locality improves performance by 5-10% on
command_buffer_perftests.

BUG=angleproject:2188

Change-Id: I6cce95481e2b1ce2605bce19b0a29d56718613a1
Reviewed-on: https://chromium-review.googlesource.com/1156853Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

This improves the performance of Framebuffer::hasTextureAttachment 3X.

BUG=angleproject:2188

Change-Id: I6acc7f4440fe0232438182274d7bba6075530f38
Reviewed-on: https://chromium-review.googlesource.com/1159171Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This can be updated in several places. Also adds a test which covers
some of the paths.

Bug: angleproject:2746
Change-Id: Id119e527fd0064998d7ad5011a9d8376e7b9dab0
Reviewed-on: https://chromium-review.googlesource.com/1153569
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

glProgramBinary may fail if the driver is incompatible with the binary
that is being loaded. After this ANGLE falls back to recompiling the
GLSL program from source, but this requires that the ProgramGL object
used is still valid after a failed glProgramBinary call.

Don't delete the GL program after a failed glProgramBinary call so
that the fallback works as intended.

BUG=angleproject:2751

Change-Id: I55c19d71414163b1cd9f898f304e4aa7052f6b16
Reviewed-on: https://chromium-review.googlesource.com/1160540Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Change-Id: Iee947a5abf4a833f155f7cb100577c92323a34e5
Reviewed-on: https://chromium-review.googlesource.com/1160863Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

Prepares for caching hasMappedBuffer. Also inclines several checks for
faster speed.

Bug: angleproject:2746
Change-Id: I74f9408d7b41e245c3f58d367dd2cc8fbace4a7a
Reviewed-on: https://chromium-review.googlesource.com/1150762
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This moves the client attribs check into a shared place. This cleans up
some of the surrounding code. Also get rid of the vertex count
parameter.

Bug: angleproject:1391
Change-Id: I9c688895c2cc5650d7b395497d69033093916f21
Reviewed-on: https://chromium-review.googlesource.com/1150517
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

Cache the minimum value for non instanced and instanced active
attributes. The cache is updated in the following places:

1. Context: bindVertexArray.
2. Context: any executable change (linkProgram/useProgram/programBinary).
3. Vertex Array: any state change call.
4. Buffer: a dependent buffer resize.

This greatly reduces the time we're spending in ValidateDrawAttribs.

Bug: angleproject:1391
Change-Id: I84bb222a1b9736e6165fe40e972cd4299ca1178d
Reviewed-on: https://chromium-review.googlesource.com/1150516
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:2668
BUG=angleproject:2714

Change-Id: I9f375caa605e13aae04e8051a779515f8ddfb93a
Reviewed-on: https://chromium-review.googlesource.com/1155105
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

ImageSiblings no longer inherit from RefCountObeject because they may be EGL
or GL objects and should handle their own deletion.

This can cause GL resources to outlive their contexts.  When the GL resource
is deleted, simply orphan the image.

BUG=angleproject:2668

Change-Id: I4a5c12bbe6e725f946209f9b48345a4097c9c91c
Reviewed-on: https://chromium-review.googlesource.com/1153601
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This avoids potential null-dereferences on the source if it is orphaned.

BUG=angleproject:2668

Change-Id: I5d591a941d114bb231044572a31a8c43cf3a9c4f
Reviewed-on: https://chromium-review.googlesource.com/1155104
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Container objects in this case are the Vertex Array and Read and Draw
Framebuffers. Instead of using a mutable dirty bit state with a const
function to notify of dirty bits we can use the Observer/Subject
pattern. This more cleanly allows us to do cache updates.

Bug: angleproject:1391
Change-Id: I88f863894ec3efa00322038f323a84850166107d
Reviewed-on: https://chromium-review.googlesource.com/1153399Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Bug: angleproject:2738
Change-Id: I1a699eda4c9795347fb43147275ac43d62523238
Reviewed-on: https://chromium-review.googlesource.com/1153634
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

Make the onDestroy method generic to use different context and error types.

This will allow us to have RefCountObjects for EGL that return EGL errors
and destroy with egl::Displays.

BUG=angleproject:2668

Change-Id: I77549cdc7119979d6eec96c7e6246242aed74830
Reviewed-on: https://chromium-review.googlesource.com/1153452
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>