8d00abe10f
'Unsetting' the title specific user settings. Fixes UnrealEngine-based games try to decompress 3000 null bytes and end up barfing all over the stack instead.
2016-06-20 15:45:12 -05:00
afda577484
Adjusted user profile settings so that we can know about settings, but have their data unset.
2016-06-20 15:44:11 -05:00
8965fda402
RSqrt for VEC128_TYPE.
2016-06-20 13:31:11 -05:00
53e37c3167
Constant propogation for DOT_PRODUCT_3.
2016-06-20 13:30:19 -05:00
dd18112905
Assert the CPU index on set.
2016-06-20 12:53:07 -05:00
27610989a2
Fixed fake CPU handling.
2016-06-20 12:52:36 -05:00
6a082c42d3
Fixed top-down memory allocations when alignment > page size.
2016-06-20 12:02:48 -05:00
919d84ad73
Fixed ToXdkProtectFlags returning bad protection flags when
...
nocache/writecombine were present.
2016-06-20 11:27:53 -05:00
6729ec182b
These clears are redundant.
2016-06-19 21:32:31 -05:00
f194d261b1
TextureCache: Speed up descriptor writes (no more vector usage)
...
WIP TextureCube uploading skeleton
2016-06-19 21:27:30 -05:00
450ca87120
Added comparison operator for --break_on_instruction.
2016-06-19 21:01:18 -05:00
b5bef8c54d
Undoing the previous commit in favor of finding a better solution (xmm1/xmm2 can't safely be used here).
2016-06-19 02:33:52 -05:00
d04a6cc983
Added support for when both src1 and src2 are constants in EmitCommutativeBinaryXmmOp and EmitAssociativeBinaryXmmOp.
2016-06-19 02:10:40 -05:00
807c030f86
Add title request functions to Emulator
2016-06-18 21:17:37 -05:00
6bd19518eb
GL4: Fix incorrectly calculating the base offset for IssueCopy
2016-06-18 21:16:10 -05:00
0c69e05eb1
Handle load requests made by the guest
2016-06-18 21:00:01 -05:00
23b2e25fb8
XThread: Don't release ourselves before calling thread_->Terminate!
2016-06-18 20:58:58 -05:00
b82f6a990a
Core: Add a file open dialog and refactor logic around loading new games
2016-06-18 20:42:28 -05:00
dcd71c1613
Fixing format_core for %hc/%hC/%wc/%wC/%hs/%hS/%ls/%lS/%ws/%wS type specifiers.
2016-06-18 19:24:49 -05:00
85e38b7657
Invalidate physical memory passed to NtReadFile.
2016-06-18 18:26:23 -05:00
23b72e4888
Complain loudly when ReadFile() returns ERROR_NOACCESS.
2016-06-18 18:26:23 -05:00
d5dd96d626
Invalid iterator in control_flow_simplification_pass
2016-06-18 16:34:46 -04:00
096febb3bb
Oops. Unbreaking shim_utils.h.
2016-06-18 14:02:38 -05:00
4fc3bdf213
Fixed AppendParam for X_EX_TITLE_TERMINATE_REGISTRATION,
...
X_EXCEPTION_RECORD.
Added AppendParam for X_OBJECT_ATTRIBUTES.
2016-06-18 13:44:52 -05:00
2d60d6f672
Split out trampoline generation to make it reusable.
2016-06-18 13:41:49 -05:00
8ea9aea84c
Actually there's yet one more argument to VdInitializeScalerCommandBuffer.
2016-06-18 08:16:49 -05:00
890a0513ef
Tagged NtSetInformationFile as high frequency.
2016-06-18 08:03:19 -05:00
f39eb9f047
Tagged NtReadFile as high frequency.
2016-06-18 08:02:18 -05:00
9477cef006
Corrected offset in comment for X_DISPLAY_INFO.
2016-06-18 08:00:41 -05:00
b12579c64c
Fixed arguments to VdInitializeScalerCommandBuffer.
2016-06-18 08:00:09 -05:00
d526d2febf
GPU: In auto-indexed draws, VGT_INDX_OFFSET appears to be reused as a vertex offset.
2016-06-17 20:19:38 -05:00
2e34a98cef
Update the vulkan loader and headers.
2016-06-17 19:32:21 -05:00
27c16b1936
GPU: Fix incorrectly calculating the primary ringbuffer's size
2016-06-16 21:30:06 -05:00
af870999fa
SPIR-V: Clear shader interface IDs between invocations.
2016-06-13 18:15:59 -05:00
5937e603d7
SPIR-V: Fix vec4 Dst
2016-06-13 15:29:07 -05:00
d5865740e2
Add a few more vector tests
2016-06-12 21:45:51 -05:00
0ef16b10e2
Constant Propagation for OPCODE_VECTOR_ADD
2016-06-12 21:41:02 -05:00
7ef7cc39f5
SPIR-V: Declare all inputs/outputs in OpEntryPoint
2016-06-12 11:18:00 -05:00
841f6fe239
Merge branch 'vulkan'
2016-06-11 19:23:23 -05:00
323b993bbc
Revert the default graphics backend to GL4 (in preparation of merge to master)
2016-06-11 19:13:45 -05:00
32011a841e
Formatting.
2016-06-05 17:48:06 -05:00
b01903ccc7
Set basePipelineIndex to -1 instead of 0 to indicate invalid.
2016-06-05 12:55:19 -05:00
6f5f3534c4
Rebuild shaders with the latest compiler
2016-06-04 15:17:31 -05:00
254acf2a67
RenderCache: Hardcode surface height to 2560
...
Fix a couple of other things
2016-06-03 20:01:49 -05:00
2a924d2b05
Pass vertex buffer endianness into the BufferCache
...
IssueCopy: Actually issue the pipeline barrier to transition the image
2016-06-03 20:00:28 -05:00
ff5529714c
clang-format to Google Standards
...
Added formatting to code
2016-05-29 13:22:14 +09:30
fd37112db8
VK Immediate Drawer: Properly transition texture layouts
...
Support wrapping of textures not created here
2016-05-28 10:26:40 -05:00
f8d9472872
TextureCache: Fix trying to erase a resolve texture that isn't in the resolve textures list.
2016-05-26 14:46:18 -05:00
c85756981b
TextureCache: Fix a few null pointer bugs
...
Ordering of in-flight descriptor sets
Change staging buffer size
Free all samplers on exit
2016-05-25 19:55:50 -05:00
8611417219
Initialize rasterDiscardEnable to VK_FALSE
2016-05-25 17:45:38 -05:00
008167fa66
VK: Enable full rotation / Set render_state to nullptr on failure / Fix format info in texture resolves
2016-05-25 13:49:36 -05:00
a187a4931a
Whoops - fix setting the wrong host base for some graphics allocations.
2016-05-25 13:14:03 -05:00
9d1e66ab16
Don't write to color targets in depth-only mode.
2016-05-24 21:58:02 -05:00
ad83a1994d
VK: Remove alphaToOne feature requirement
2016-05-24 12:53:25 -05:00
5f764730ae
Vulkan CP: Override frame traces if renderdoc is attached
2016-05-23 13:16:13 -05:00
729152a58b
VK: Enable independentBlend feature
2016-05-23 09:59:37 -05:00
6e21d88250
Fixup circular buffers for full rotation
2016-05-23 09:58:46 -05:00
c6e905db2f
Fix a memory leak in fenced pools.
2016-05-22 22:14:45 -05:00
7004f83665
CP: Don't check for shader validity here
...
Fix a lousy typo in PrepareTextureSet
2016-05-22 20:08:35 -05:00
bd27835a3d
Pipeline Cache: Translate shaders when program cntl register is available
2016-05-22 20:03:13 -05:00
d1b4d61b52
SPIR-V: Use the register count from the program control register
...
Workaround for broken OpBitFieldUExtract on NVIDIA drivers
kRcpc/kRcpf/kRsqc/kRsqf
Fix broken ps_ usage
2016-05-22 20:01:42 -05:00
d94ff6eb25
Shaders: Track the register count from the program control register (if available)
2016-05-22 19:58:50 -05:00
1faf5a813a
Fix ALU scalar swizzles (Possibly)
2016-05-22 19:57:05 -05:00
b025790207
Disable Vulkan native MSAA by default for now.
2016-05-17 05:58:52 -05:00
f2ca689ea0
Expanded vec128 and VectorSub
...
Added double precision (f64) values to the vec128 structure
Added cases for all formats of data for vector subtraction in the
Value::VectorSub function
NOTE: Unsure what the saturate function is for however maintained
original functionality
2016-05-17 05:34:09 +09:30
b9a40d1a00
Use Vulkan as the default graphics backend.
2016-05-15 12:08:29 -05:00
2bb52ef86b
SPIR-V: WIP shader compiler / optimizations / alpha test implementation
2016-05-15 12:01:38 -05:00
c06a7cdf81
BaseFencedPool::has_open_batch (and other uncommitted changes)
2016-05-07 19:17:56 -05:00
79f1193130
Vulkan CP: Fix calculating an invalid copy destination base address when sizeof(texel) != 4
2016-05-06 00:12:47 -05:00
7c5042add7
Vulkan CP: Add in separate swap-chain images
...
Some other changes I can't remember
2016-05-03 14:12:05 -05:00
f2af28c322
TextureCache: Fix up some synchronization flaws (deleting in-use textures/etc)
...
Fix texture binding IDs not matching fetch instruction IDs.
Fix some bad texture format matching.
Add access watches
2016-05-03 14:10:15 -05:00
8e8df2e778
PipelineCache: Support shader disasm dumps for nvidia cards.
...
Fix MSAA 2X multiplier.
2016-05-03 14:07:20 -05:00
d18c99aab6
RenderCache: Account for MSAA when calculating tile sizes.
...
Add a new flag to enable native MSAA (this does not work properly at the moment)
2016-05-03 14:05:34 -05:00
aa038fbf23
Skip the wrapping packet end after parsing IB end (to avoid false draws appearing)
2016-05-01 15:48:31 -05:00
720f8b0dc2
GL4: Track the internal format of render targets and match them based on that.
2016-05-01 11:23:54 -05:00
cbccc785cc
TraceViewer: Build a tree of all command buffers and display that instead of a flat list.
2016-05-01 10:15:33 -05:00
6101b70641
Fix the Vulkan immediate drawer not drawing lines.
2016-04-29 13:09:39 -05:00
9b2e2a7275
SPIR-V: Hack in OpSelectionMerge as hints to NVidia's shader compiler (TODO: Make a Shader Compiler)
2016-04-13 23:17:03 -05:00
2bd603bf18
CircularBuffer: use std::list for allocations instead of a vector.
2016-04-09 21:40:18 -05:00
4811ebc2ce
BufferCache: Use a CircularBuffer as the transient buffer.
2016-04-09 21:27:32 -05:00
b7f2c93d73
SPIR-V: Batch predicated instructions together into a single block.
...
Add Post-Translation validation.
Fix a couple of type-related typos.
2016-04-09 21:03:44 -05:00
a1c9540063
SPIR-V Validator util class
2016-04-09 18:35:00 -05:00
3726064af5
Can't use CmdCopyBufferToImage or vice versa for depth and stencil.
2016-04-01 22:03:29 -05:00
f9a634ad25
CircularBuffer remove Discard functionality and allow rotation
2016-04-01 21:53:46 -05:00
50f72b4e42
Enable native MSAA
...
Copy back EDRAM buffers in order by base offset.
2016-04-01 21:52:39 -05:00
2eca3ce9e6
Texture uploads/basic formats
...
Fixed swizzle one/zero mismatch
Sampler setup
Remove samplers from the descriptor set layout
2016-04-01 21:51:17 -05:00
1ea72c5e06
FencedPool::CancelBatch
2016-04-01 21:49:58 -05:00
44cffab389
SPIR-V Max4
2016-03-25 18:23:45 -05:00
fc1bd0f379
Fix texture uploads
2016-03-25 17:29:39 -05:00
692d666d57
Wipe the buffer cache in ClearCache for now.
2016-03-25 16:50:06 -05:00
0e44cda961
Update the rectangle list shader
2016-03-25 16:49:41 -05:00
d7599c817f
Formatting.
2016-03-25 16:44:25 -05:00
a5a31cf123
VulkanShader::Prepare - return false if vkCreateShaderModule failed.
2016-03-25 16:37:24 -05:00
2bb40c122d
Vulkan util Fence class
2016-03-25 16:36:21 -05:00
f75e5fec24
CP: Use a single command buffer for every frame, reuse render passes/pipelines if not dirty
...
Hook up resolves and swaps
2016-03-25 16:35:34 -05:00
1e1da1eb6c
PipelineCache::ConfigurePipeline - Inform the caller if the pipeline is dirty or they can reuse the previously bound pipeline.
...
Make SetDynamicState public.
2016-03-25 16:34:14 -05:00
b2457d7e72
Basic texture uploads/address lookups/etc
...
Freeing of descriptor sets when the GPU is finished with them.
2016-03-25 16:32:29 -05:00
0e41774e36
RenderCache::dirty() - used to tell if we need to begin a new pass
...
Round all pixel pitch/heights up before dividing.
2016-03-25 16:31:12 -05:00
181b2af5a4
Vulkan Circular Buffer
2016-03-25 13:49:07 -05:00
e72e283e79
Primitive type makes rasterization state dirty too!
2016-03-23 16:20:06 -05:00
7b962e59a4
SPIR-V Dst
...
Fix a few bugs in the translator
2016-03-20 14:21:55 -05:00
38b94dd9e2
Add in Xenos events
2016-03-17 21:58:23 -05:00
2512a6360e
Pass the physical frontbuffer address into the CP
2016-03-17 21:55:47 -05:00
0e3c113375
Physical write watches -> access watches (read and/or write watching)
2016-03-17 21:55:16 -05:00
1831e7a936
Pipeline stencil state
2016-03-16 14:45:40 -05:00
54f89825d9
SPIR-V Dp2Add/Dp3
2016-03-15 00:30:39 -05:00
245102e9e5
RenderCache::BlitToImage
2016-03-12 22:04:32 -06:00
822d61c3d9
Fix a few stale data usage bugs in the pipeline cache.
...
Hook up part of depth/stencil tests/writes
2016-03-12 22:03:11 -06:00
5ba04b9e55
RenderCache::ClearEDRAMColor/ClearEDRAMDepthStencil
2016-03-12 11:49:59 -06:00
635d095b87
RenderCache: Track color target / depth target usage, refactor RawCopyToImage
2016-03-10 20:39:46 -06:00
f903a559b3
Blit Vulkan CP output to the main window's swap chain
2016-03-10 13:01:39 -06:00
f839a1293f
Add a specialized copy command buffer to the vulkan swap chain
2016-03-10 12:59:48 -06:00
90b54bbaa6
Fix usage of mix in fragment shader
...
The boolean value must be a vector too, where each bool component
selects whether the value of the output vector component is from the
first or the second source vector.
2016-03-10 01:14:42 -05:00
86cb40f0c6
Change how the render cache renders into EDRAM.
...
Rendering directly into the EDRAM buffer is bad because we don't know how the GPU lays out memory when it draws. Instead, we create temporary render targets and copy EDRAM contents to/from those temporary RTs before and after each draw.
2016-03-08 18:02:03 -06:00
4e27539709
Fix Vulkan texture drawing.
2016-03-08 17:57:04 -06:00
af7fc20c38
Beginnings of texture conversion/uploads
2016-03-05 22:09:18 -06:00
38094ac819
Updated local clang-format.
2016-03-02 21:17:34 -06:00
740c70f270
Scalar logc, fix log
2016-03-01 12:52:34 -06:00
48cf270724
Use spv::NoPrecision instead of DecorationInvariant
...
Set samplers/images as uniform constants
2016-02-27 16:21:37 -06:00
c648e54539
Short-circuit draw calls if the render target's pitch is 0
2016-02-27 11:30:50 -06:00
6109e0b03a
Fix incorrect images/samplers definitions
...
ps_param_gen and fix interpolators being copied incorrectly
2016-02-25 17:41:41 -06:00
08a287d2ad
Whoops
2016-02-21 22:30:31 -06:00
06d5a28615
Fix texture fetches
2016-02-21 22:27:23 -06:00
5e31c51a3d
Textures are now in descriptor set 1.
2016-02-21 20:03:51 -08:00
3b20c757ff
Merge branch 'vulkan' of github.com:benvanik/xenia into vulkan
2016-02-21 19:56:30 -08:00
62931f8c8e
Adding texture setup and a dummy grid texture.
2016-02-21 19:56:23 -08:00
8a29330f8c
First-pass image sampling
2016-02-21 20:42:37 -06:00
2785a94fea
SetpXXPush
2016-02-21 19:20:53 -06:00
9030c87386
SetpPop/SetpRstr
2016-02-21 19:20:53 -06:00
5c2b5123ac
Floors/Frcs/Truncs
2016-02-21 19:20:53 -06:00
0680e451bc
Exp2
2016-02-21 19:20:53 -06:00
cbac9b2f4e
Use vec4 zero rather than float zero where needed
2016-02-21 19:20:53 -06:00
13049912ee
Vec Log
2016-02-21 19:20:53 -06:00
1d4190af02
MaxA
2016-02-21 19:20:53 -06:00
568845e81d
Dp4
2016-02-21 19:20:53 -06:00
3877afe90a
MaxAsf
2016-02-21 19:20:53 -06:00
d217f7b3c3
MaxAs/SetpClr
2016-02-21 19:20:53 -06:00
fa3ca4a5d7
Short-circuit if the store has no writes.
2016-02-21 19:20:53 -06:00
d2e3b5533d
Whoops
2016-02-21 19:20:53 -06:00
9b805b929e
Scalar kill ops
2016-02-21 19:20:53 -06:00
e78537571f
Vector kill ops
2016-02-21 19:20:53 -06:00
633746b5e4
Actually preserve pv/ps if predicate fails
2016-02-21 19:20:53 -06:00
cda08ff826
Setup descriptor set for textures.
2016-02-21 17:17:57 -08:00
1def5bece9
Fix viewport inversion.
2016-02-21 15:59:35 -08:00
ace895d4f3
Fixing 1/w0 correction.
2016-02-21 15:44:09 -08:00
24743bf676
Actually configure constant descriptor set.
2016-02-21 15:20:35 -08:00
d57f974e2e
Fix buffer alloc alignment and framebuffer comparison.
2016-02-21 14:43:59 -08:00
06ba273492
Warn and be ok with symbol services failing.
2016-02-21 14:26:16 -08:00
8bf5eba098
Capture exe with renderdoc and pass --vulkan_renderdoc_capture_all.
2016-02-21 12:44:16 -08:00
b83b9c28a1
Fix push constants declaration
2016-02-21 13:19:08 -06:00
a7880645be
Merge branch 'vulkan' of github.com:benvanik/xenia into spv_translator
2016-02-21 12:47:02 -06:00
bd0fb81b22
Apply window scaling
2016-02-21 12:46:54 -06:00
32c4f3ce24
Fixing pool shutdown.
2016-02-21 10:46:47 -08:00
b3faba50a3
Pushing constants.
2016-02-21 10:38:58 -08:00
71440a4137
Add code to correct vertex format
...
Move translated shader code to its own function
Add push constants
2016-02-21 12:06:59 -06:00
b1e56ee4e5
Vector Cndxx
2016-02-20 23:45:01 -06:00
ff373a83a4
Private->Function variables, fragment output, geometry shader tweaks.
2016-02-20 20:44:11 -08:00
2629ae4a14
Fix using incorrect result types for compares
2016-02-20 21:55:32 -06:00
ca01bb2311
ALU predicated discard
2016-02-20 21:53:22 -06:00
08a173e5ec
Scalar Sxxx
2016-02-20 21:28:25 -06:00
6d65439efe
Merge branch 'vulkan' of github.com:benvanik/xenia into vulkan
2016-02-20 19:19:41 -08:00
cd02cdfc70
Making memory API less error prone; fixes buffer/constant uploads.
2016-02-20 19:19:29 -08:00
3130d4b38e
Add the vulkan backend as a choice in xenia-app
2016-02-20 21:02:28 -06:00
8777d22b72
Add name information to constants
...
Fix Rcp/Rsq OpSelect incorrect usage
2016-02-20 19:31:28 -06:00
1d82d7bd92
For now, all 512 constants live in the same block.
2016-02-20 17:05:53 -08:00
4aff1c19a7
(WIP) SPIR-V Shader Translator
2016-02-20 18:44:37 -06:00
b5a0c4715b
Porting GL4 geometry shaders. Likely not working.
2016-02-20 16:35:21 -08:00
769c58a9b2
Adding `xb genspirv` to do glsl->.h.
2016-02-20 16:24:42 -08:00
5759f82276
Adding pipeline caching.
2016-02-20 15:00:11 -08:00
731ff52773
Implementing a lot of the pipeline state setting.
2016-02-20 12:59:18 -08:00
97174dbe4d
Skeleton leaky hacky hardcoded pipeline setup.
2016-02-19 23:23:58 -08:00
990b600f53
Buffer uploads (without reclaiming, yet).
2016-02-19 21:08:50 -08:00
de1e4661ff
Not crashing (but also likely not working) EDRAM emulation.
2016-02-19 20:36:10 -08:00
8820c73532
A night's worth of work: documented EDRAM. Seems mostly right.
2016-02-19 10:38:11 -08:00
1dcc84a472
Starting render passes.
2016-02-18 18:18:31 -08:00
52a6f795f7
Shader modules and plumbing.
2016-02-18 18:18:31 -08:00
f9ca0b86a8
Query to see if the physical device supports a swapchain surface to silence a validation layer error.
2016-02-18 20:10:14 -06:00
3bf1c91eaa
Those are really pointers - need those bits.
2016-02-18 16:52:52 -08:00
4c4a641096
WIP rough sketch of vulkan backend structure.
2016-02-18 16:43:41 -08:00
35e08d9428
Switching from fork to main glslang spirv builder.
2016-02-18 16:43:41 -08:00
00594da417
Adding mutliple queue/shared queue support.
2016-02-18 16:43:41 -08:00
ecd257b34c
Fix headers include guards.
2016-02-18 16:43:41 -08:00
618ca80011
Count packed bytes used by a constant map.
2016-02-18 16:43:41 -08:00
5ab0af9e6d
Implementing shader constant register map construction.
2016-02-18 16:43:41 -08:00
05107d2d3e
Skeleton xenia::gpu::vulkan implementation, enough to start trace viewer.
2016-02-18 16:43:41 -08:00
ca5902c111
Recreate swap chain on window resize.
2016-02-18 16:43:41 -08:00
a97fa36512
Fixing swap chain scissoring; surface is still not resized right.
2016-02-18 16:43:41 -08:00
9f52f2e819
Initial vulkan context and immediate drawer.
...
Extremely rough, just checking in so DrChat can snoop.
2016-02-18 16:43:41 -08:00
27bace0e5c
Actually disable writewatch kill in ConvertTexture
2016-02-17 18:04:10 -06:00
b14c4ad681
Disable write watch kill in ConvertTexture
2016-02-17 17:57:56 -06:00
8bf1bf57fd
Fix GL4 alpha test, use blend disable flag
...
If color clear is enabled, grab the color target no matter what.
Update texture readback code on copy (but not enabled)
Track PA_SC_VIZ_QUERY in UpdateRasterizerState
2016-02-17 17:56:28 -06:00
f5eef128f8
Merge pull request #539 from DrChat/vizquery_wip
...
Viz Query WIP Notes
2016-02-16 17:21:39 -08:00
864b1b5fac
Merge pull request #538 from DrChat/trace_viewer_misc
...
Trace Viewer Zoomed View / Fix Alpha Test
2016-02-16 17:19:26 -08:00
a2dd1c48ae
Add some WIP viz query notes to the command processor.
...
Also implement PM4_SET_BIN_MASK/SELECT.
2016-02-16 18:36:43 -06:00
253e164753
Move EnableDepthTest functionality into a bool alpha_blend var on ImmediateDraw.
2016-02-16 18:24:37 -06:00
c667dfc596
Don't put kernel stacks in GPU memory.
2016-02-16 15:05:11 -08:00
e77af94c7c
Ignore color target / depth target / texture alpha in the trace viewer.
2016-02-16 16:45:49 -06:00
a8dfd6a21a
Add a flag to ImTextureID that disables alpha test
2016-02-16 16:43:12 -06:00
65d1ea0250
Add a zoomed image view for color targets and depth targets.
2016-02-16 16:41:20 -06:00
c5de61860d
Merge pull request #531 from DrChat/gl4_shader_cache
...
GL4 Shader Cache
2016-02-16 07:20:33 -08:00
f16a288f30
Merge pull request #532 from DrChat/profiler_flip_decouple
...
Decouple MicroProfilerFlip from Profiler::Present
2016-02-16 07:19:20 -08:00
cc00b2d7a5
Merge pull request #533 from sephiroth99/noamd
...
More shader changes
2016-02-16 07:18:55 -08:00
a442bb50b4
Prevent undefined behaviour with thread_safe_gl enabled.
...
CreateOffscreenContext already calls ClearCurrent. When it is done twice,
the global GL mutex is unlocked a second time, and undefined behavour
happens. In this particular case, a deadlock occured; unlocking the already
unlocked mutex locked it.
2016-02-11 23:50:11 -05:00
205f64eb8b
Fix compilation with VS2015 Update 2 CTP
2016-02-11 23:47:11 -05:00
00674736fe
Fix minor syntax error in shaders.
...
Discovered with glslangValidator.
2016-02-08 20:42:09 -05:00
1520869a82
Revert "AMD people: have at it. I ain't gonna guarantee it'll work, though."
...
This reverts commit 8c0614b202
2016-02-08 20:40:52 -05:00
4335e41c14
Decouple MicroProfileFlip from Profiler::Present (and start a new frame in XE_SWAP)
2016-02-07 19:25:30 -06:00
a64e53fba6
Fix include order.
2016-02-05 16:31:01 -06:00
6cb7f0c23d
GL4CommandProcessor: Use the new GL4ShaderCache
2016-02-05 16:14:35 -06:00
406ec8c6da
GL4ShaderCache util class - caching to filesystem disabled by default
2016-02-05 16:11:00 -06:00
e9dd0ce9de
GL4Shader load from binary support
2016-02-05 16:02:14 -06:00
1066362ada
ShaderTranslator::GatherAllBindingInformation
2016-02-05 16:00:50 -06:00
ccb6a9146c
Expose microcode data hash in Shader
...
Return paths to files that were dumped from Dump
2016-02-05 15:59:07 -06:00
f75a4cab0e
Apply a few optimizations to generated GLSL shaders.
2016-01-30 19:39:47 -06:00
9a09a61874
Merge pull request #526 from DrChat/command_processor_rb
...
Command Processor RingBuffer Replacement
2016-01-30 11:07:27 -08:00
bf2c3d34da
Fix XMAInitializeContext is_stereo flag.
2016-01-30 13:04:52 -06:00
6ce28ffd8c
Command Processor: Replace RingbufferReader with RingBuffer
...
Report any packet processing errors
PM4_MEM_WRITE
2016-01-24 22:54:26 -06:00
6a67632af1
Update RingBuffer - Add immediate read/write and some error checking asserts.
2016-01-24 22:39:05 -06:00
9b1e6d6cb3
Resolve a module name to its full path before comparing names.
2016-01-24 09:50:48 -06:00
8a89a93eab
Fix improperly reading 2_10_10_10 bitfields (or so I hope)
...
Check at runtime if a vertex fetch index doesn't match gl_VertexID
Fix trying to instantiate a vec1
2016-01-22 18:49:17 -06:00
b2d8837e52
Fix depth clear when not sourcing from a depth target.
2016-01-22 18:47:27 -06:00
6c6f7c6eca
Merge pull request #522 from sephiroth99/lf3
...
Linux compilation/link and compatibility fixes
2016-01-20 19:41:40 -08:00
a74c88851b
Rework RtlTime with portable time conversion functions
2016-01-20 12:26:49 -05:00
bf910c2d9a
Add missing definitions for XAPC constants
...
Fixes linking errors for these symbols.
2016-01-20 11:26:35 -05:00
f17b80c61e
Use stdlib 'hardware_concurrency' to get logical processor count
...
This has the benifit of being multiplatform.
2016-01-20 11:04:28 -05:00
5d9a4b6830
Handle all other cases in switch statements
2016-01-20 11:00:37 -05:00
1c14964739
Add missing override on virtual function
2016-01-20 10:57:50 -05:00
4c0f9e6342
Add missing includes
2016-01-20 10:36:01 -05:00
d78317847d
Fix link error when profiling is disabled
...
These four functions are needed by microprofile even when profiling
is disabled.
2016-01-20 09:35:27 -05:00
8710b105bf
Draw smaller floating point formats in the trace viewer.
2016-01-19 21:18:48 -06:00
9ed0934123
Swap to modern OpenGL for TFB usage.
2016-01-19 18:31:36 -06:00
47f5896089
Add TFB support to the trace viewer
2016-01-18 22:22:15 -06:00
a34e82c77f
Disable GL cull on rectangle lists
2016-01-18 21:00:58 -06:00
3859b0a091
Add an option for the draw batcher to collect vertex shader output. Disabled by default.
2016-01-18 20:57:36 -06:00
55a8964428
Fix shaders reading 10_11_11 bitfields backwards.
2016-01-18 15:46:41 -06:00
2e5c64fb13
Prevent some divide by zeros in shaders. Thanks @vlj!
...
Fixes #518 .
2016-01-18 11:51:18 -08:00
6777ce6668
Reconcile debugger and save state stuff into a single implementation.
...
Fixes #497 and fixes #496 .
Still rough edges, but at least less duplication.
2016-01-18 11:48:21 -08:00
34e03294ef
Fix vfetch_mini instructions using stale data when following a vfetch_full that doesn't fetch anything
2016-01-16 22:39:00 -06:00
e8719dd949
Fix an overflow crash in DiscImageDevice::Verify
2016-01-14 15:18:41 -06:00
9473f20c7f
Fixing travis lint diff.
2016-01-13 22:55:28 -08:00
2e826f9321
Supporting constants in Haswell codepaths for mul_add/sub.
2016-01-13 22:29:39 -08:00
0771eea0b3
Change comments to comply with local regulations.
2016-01-13 19:41:15 -06:00
bd6bf16bd1
New flag: disable_global_lock - Disables global lock usage in guest code.
2016-01-13 19:40:00 -06:00
a5b37dce1d
Shaders: Emit an unimplemented warning for indexed vertex fetch instructions
2016-01-13 01:50:20 -06:00
14c5e48120
Enable speculative PACK_TYPE_FLOAT16_3
2016-01-13 01:48:18 -06:00
36372d1eae
Enforce xex memory protections
2016-01-13 01:47:06 -06:00
f0802d75fa
Use atomic compare exchange in stdcx/stwcx (still under the global lock)
2016-01-13 01:45:31 -06:00
51dde62f09
x64 backend vector optimizations
2016-01-13 01:43:23 -06:00
7b6f1388d6
OPCODE_ATOMIC_COMPARE_EXCHANGE
2016-01-13 01:41:58 -06:00
3ccb2a978d
Propagate vector constants.
2016-01-13 01:40:44 -06:00
gibbed