f36d8a3b23
RISC-V support
...
With some addition to macros the project is able to build on riscv64
machine.
plthook has added RISC-V support upstream [1]. However upstream has
changed their code quite a bit, so I just modified the current vendored
version.
[1]: https://github.com/kubo/plthook/commit/a564738a4707fc5eb4d388b9e97eeab51b8c9d56
2023-09-29 17:26:11 +01:00
2886f4a282
Fx mesh viewer being opened repeatedly causing crashes on D3D12
2023-09-29 17:24:19 +01:00
c46a9d0c73
When removing library flag remove any dependent flags as well
2023-09-29 17:23:50 +01:00
9a5e4aa5af
Bump version to v1.30
2023-09-29 17:23:36 +01:00
0202a066a9
Update year in about dialog
2023-09-28 16:37:17 +01:00
7f92b0241e
fix copyrights
2023-09-28 13:52:23 +01:00
5c1971272c
Make sure to account for missing draws even with MaxCount==1
...
* If we don't account for an execute with MaxCount==1 and actual count 0 then
our EID tracking gets mismatched.
2023-09-27 15:13:41 +01:00
adfcedff81
Sync any pending data uploads before any queue executes on D3D12
2023-09-27 14:47:54 +01:00
5addf044d7
Properly release all D3D12 objects
2023-09-27 12:40:31 +01:00
2576a7b7f6
Don't hook into AMD driver
...
AMD driver was creating and presenting a D3D12 swapchain when creating a GL context
2023-09-25 17:17:21 +01:00
c859da8f1a
Update compiler for spirv-plugins yet again
2023-09-25 16:20:09 +01:00
23bcc88a1a
Record barriers under original command list data for state lookups
2023-09-22 14:49:02 +01:00
fe6b702ff4
Allow out of bounds GPU addresses when serialising
...
* Worst case this is just as invalid as an application, if it uses a totally
bogus VA. However in D3D12 it is apparently valid to refer to VAs out of
bounds of any resource as long as it's within bounds of an underlying heap. To
handle this without serialising VAs as Heap+offset we instead just allow the
address lookup to run out of bounds and pick the next lowest buffer. If the
offset is greater than the buffer size then we're probably no worse than the
application.
2023-09-22 14:47:37 +01:00
e8a7560b51
Add checks for GL maps failing
2023-09-22 12:47:09 +01:00
d78565df33
Coalesce unbound sparse bindings on D3D12
2023-09-21 15:47:25 +01:00
81cbd7f404
Add delay to empty capture tests to ensure the capture propagates
2023-09-20 12:06:21 +01:00
b262958d53
Update message for linux on nvidia to clarify to copy symlink
2023-09-20 11:58:11 +01:00
5f9bd7e51b
Flush the context after garbage collecting D3D11 state objects
2023-09-19 18:52:13 +01:00
5432a0576b
Functional tests cannot be built in CMake >= 3.12
...
In development/CMakeLists.txt:477, the Python interpreter and libs are discovered. It uses the 'old' package discovery mechanism and the new depending on CMake's version.
However the old mechanism sets different variables to the new, so the new also sets the old vars for compatibility, however it is missing two variables.
The first missing var causes the Python minor version to be hardcoded to zero (renderdoccmd/CMakeLists.txt:9) and the second causes the later conditional (CMakeLists.txt:495) to be malformed.
2023-09-19 14:46:59 +01:00
d911ec8e3d
Update docker prepare script to remove expired root CA certificate
2023-09-19 10:49:46 +01:00
30a013907d
Cannot run functional tests using Python >= 3.10
...
In replay/entry_points.cpp RENDERDOC_RunFunctionalTests(..), it uses a simple char-replacement algorithm for finding the Python modules depending on the Python minor version.
Unfortunately that mechanism relies on there being only a single character to replace which doesn't work once we reach double figures.
2023-09-18 17:39:56 +01:00
d88eff15f3
Update execute indirect test to verify more complex EI behaviour
2023-09-18 12:19:36 +01:00
df591c09ec
Fix some Vulkan draw indirect bugs that were revealed in D3D12 refactor
2023-09-18 11:21:50 +01:00
55dae7ebc6
Refactor D3D12 ExecuteIndirect handling to patch arguments on GPU
2023-09-18 11:21:50 +01:00
cb7048e1a4
Fix validation error on dummy pipeline that was missing an RP
2023-09-18 11:21:50 +01:00
815c88b932
Update Ubuntu dependencies to prioritise newer (18.04 and up) versions
2023-09-13 16:26:29 +01:00
3a2b7db54f
Properly initialise dummy pipeline
...
* This prevents a possible crash if the uninitialised memory is used for
destroying these objects
2023-09-13 16:26:04 +01:00
8cd37f019e
Avoid warning on newer cmakes
2023-09-13 16:25:24 +01:00
92f4aeacde
Fix compile error on some clangs with non-const string literals
2023-09-13 16:25:09 +01:00
02272f98ac
Garbage collect D3D state objects before creation
...
* If we only garbage collect after a successful create, applications which
deliberately leak state objects and only free them when they see a failed
create will not have a change to garbage collect.
2023-09-12 21:53:23 +01:00
1bfe6d1492
Add dedicated test for VK_Counters so GPU duration isn't degenerate
...
* Some cards run it so fast that the triangle takes "0" microseconds
2023-09-12 21:53:23 +01:00
832bb4fb74
Add an early-out for DXBC chunks with the wrong size
2023-09-12 21:53:22 +01:00
c72a79988d
Only do the overlay mip subresource tests once
2023-09-10 09:39:19 +01:00
93506245c4
Add format name to overlay test python logging
2023-09-10 09:39:19 +01:00
78a64f8640
Extend Overlay tests to handle more depth stencil formats : D3D12
...
D32_S8, D24_S8, D32, D24, D16
2023-09-08 22:38:01 +01:00
cfdb349cdb
Extend Overlay tests to handle more depth stencil formats : D3D11
...
D32_S8, D24_S8, D32, D16
2023-09-08 22:38:01 +01:00
8459229441
Extend Overlay tests to handle more depth stencil formats : Vk
...
D32_S8, D24_S8, D32, D24, D16
2023-09-08 22:38:01 +01:00
40090fb779
Extend Overlay tests to handle more depth stencil formats : GL
...
D32_S8, D24_S8, D32, D24, D16
2023-09-08 22:38:01 +01:00
163f2ff747
Fix validation error in Vk_Overlay_Test
2023-09-07 09:48:57 +01:00
ae8e923a7c
Fix calculation of resource use EIDs on D3D12
2023-09-06 10:18:47 +01:00
a7a4521b63
Fix 'alignas' use
...
I think the correct syntax is to have this at start, otherwise it
upsets more strict compilers (like clang).
2023-09-05 15:21:10 +01:00
401c6617dc
Fix typo in front/back face stencil mask check
2023-09-05 15:21:10 +01:00
18d75ff493
Don't source clang_format script, run via bash
2023-09-05 11:10:06 +01:00
9d39b8e1a8
Reformat code for clang-format 15
2023-09-05 11:02:08 +01:00
4d94209ec1
Change fixed clang-format version to 15
...
* We commit statically linked x64 binaries for clang-format-15 for windows and
linux to make the transition easier for anyone running a linux distro that
doesn't yet support llvm 15, or in future for distros that drop support for
it.
2023-09-05 11:02:08 +01:00
5e7b25efca
GL: Fallback for uniform lookup for arrays of dimension 1 Closes #3047
...
ie.
struct A {
vec3 m1;
vec3 m2;
vec3 m2;
};
uniform A a[1];
with a[0] being passed to a function.
Previously it would look for "a.m1", if that is not found then it looks for "a[0].m1" etc.
2023-09-05 09:36:11 +01:00
bfe89e4c17
Extended test to include a test for GL uniform array of size 1
...
ie.
uniform float2_struct M[1];
2023-09-05 09:36:11 +01:00
df3b17a265
Fix handling of specialization Op::LogicalNot Closes #3046
2023-09-04 21:37:25 +01:00
0b669b61e9
Improve GIT versiion retrieval code to work with worktrees
2023-09-04 21:37:25 +01:00
08a9c1dac2
minimal change to fix Mac CI
2023-09-04 20:50:50 +01:00
Eric Long