Commit graph

343 commits

Author SHA1 Message Date
kd-11 1d004f2788 rsx: Lower single-image-blit to a scaled coordinate fetch 2023-07-04 09:31:51 +03:00
kd-11 cef4ea786a rsx: Trivially detect single-image-blit transfer operations 2023-07-04 09:31:51 +03:00
kd-11 69b02f7e46 rsx: Implement atlas trivialization pass
- Decompose sequences of copies into the minimally required set.
  This does 2 things:
  a. Reduces GPU workload by doing less transfers
  b. Allows the clipping optimization to kick in and skip a ton of GPU work
2023-07-04 09:31:51 +03:00
kd-11 b788e05396 rsx: Implement GPU copy ellision through coordinate transform
- TODO: Some corner cases still exist where format may not be a match after a cache merge.
2023-07-04 09:31:51 +03:00
kd-11 d77a78cdf1 rsx: Rework texture coordinate handling to support clamping and a more sane scale-bias setup 2023-07-04 09:31:51 +03:00
kd-11 ca472b6750 Silence log spam about copy ellision 2023-06-27 16:41:30 +03:00
kd-11 b00b8ae71b rsx: Allow swizzle flag to propagate as an override to surface properties 2023-05-03 01:26:14 +03:00
kd-11 45f416c4f1 rsx: Apply swizzle flag on data source as well
- The fact that the attribute is tacked on to the dest is a semantic error.
  It's the context that is swz\ln, not the source or dest. They both are one or the other.
2023-05-03 01:26:14 +03:00
kd-11 c58149f5b6 vk: Propagate swizzle flag in hw blit engine 2023-05-03 01:26:14 +03:00
Megamouse 6555de6ef5 fix warnings 2023-04-21 16:41:06 +02:00
kd-11 04fb86556a rsx: Fix surface metadata life-cycle
- Beware of clone operations. Blindly inheriting the parent's metadata is wrong.
- It is possible, especially when reusing a pre-existing slice, that the parent and child info has diverged
2022-12-17 20:16:58 +03:00
kd-11 533f960854 rsx: Handle some more corner cases 2022-10-03 12:57:16 +03:00
kd-11 765208a181 rsx: Avoid clobbering CELL memory when splitting fbos 2022-10-03 12:57:16 +03:00
Nekotekina 4b787b22c8 Implement FN (lambda shortener)
Useful for some higher order functions.
Allows to make short lambdas even shorter.
2022-07-08 14:47:41 +03:00
RipleyTom 8316469cfc Update libusb to v1.0.26 2022-04-29 02:04:52 +02:00
kd-11 2b42895bc7 rsx: Reduce log spam a bit 2022-03-28 19:55:34 +03:00
kd-11 d98d152d23 rsx: Fix leaking surface cache refs from texture cache
- Lock surfaces in use by texture cache to prevent complete deletion
- Remove discarded surfaces from the reprotect cache to avoid uaf
2022-03-28 19:55:34 +03:00
kd-11 bc7ed8eaab rsx/vk: Rework MSAA implementation 2022-03-17 22:02:20 +03:00
kd-11 762b594927 rsx: Fully process texture if surface cache configuration changed 2022-03-08 22:06:26 +03:00
kd-11 8d3d290e33 rsx: Fix build 2022-03-08 22:06:26 +03:00
kd-11 0df903090d rsx: Optimize metrics a bit
- For some reason this has a massive impact on performance above some arbitrary threshold of calls
  Shows up under surface_cache::get_merged_memory_region when doing gathers.
2022-03-08 22:06:26 +03:00
kd-11 6812fa4764 rsx: Fix surface write coherency when MSAA is active 2022-03-08 22:06:26 +03:00
kd-11 00a1864a95 Revert "rsx: Downgrade depth-1 3D images to 2D (#11593)"
This reverts commit 6c096b72b5.
2022-03-01 21:51:55 +03:00
kd-11 6c096b72b5
rsx: Downgrade depth-1 3D images to 2D (#11593)
- Fixes problems with implicit view types derived from dimensions.
2022-03-01 10:45:50 +03:00
kd-11 ec3e8de780 rsx: End the current frame before performing cache cleanup to release in-flight data 2022-02-10 22:20:56 +03:00
kd-11 2d9f21a2ea rsx: Lower performance warnings to 'warn' level instead of 'error' level to avoid causing panic for users 2022-02-07 09:25:01 +03:00
kd-11 dca3d477c9 vk: Use image hot-cache for faster allocation times
- Creating new images is expensive.
- We can keep around a set of images that have been recently discarded and use them instead of creating new ones from scratch each time.
2022-02-06 15:49:50 +03:00
kd-11 86919ec0e1 rsx: Validate requested images before attempting to upload them
- Do not allow dimensions of 0 to reach the backend APIs
2022-01-30 14:58:51 +03:00
kd-11 3fa45ff994 Fix missing typeless info update 2022-01-26 12:08:36 +03:00
kd-11 2f7d38bb81 rsx: Improve coverage checking logic to handle 3D and cubemap resources 2022-01-23 00:03:03 +03:00
kd-11 4f8b5849b7 rsx: Take depth into account when calculating coverage 2022-01-23 00:03:03 +03:00
kd-11 6d737e61fd rsx: Use 32 bit integers for pitch
- RSX max pitch = 65536 which requires 17 bits
2022-01-10 12:27:30 +03:00
kd-11 83026fd263 rsx: use coverage ratio to determine when too much data is overlapping 2022-01-07 22:55:27 +03:00
kd-11 92824b6729 rsx: Rework invalidation tagging 2022-01-07 22:55:27 +03:00
kd-11 7563655221 rsx: Bump surface removal threshold values
- It is much slower to attempt surface removal than to render duplicates on the host GPU
2022-01-07 22:55:27 +03:00
kd-11 39ef39aa4e rsx: Exercise caution when testing for overlaps in invalidated sections 2021-12-24 15:13:33 +03:00
kd-11 94a3b1cfe8 rsx: Roll back some optimizations
- Just use RGB565 for all blit targets. Avoids really dumb transforms done by GPU hw.
- When X16 is used, all the channels get written to R channel alone. CmdBlit does perform format conversion!
- gl: Force image copy when blit is requested with compatible targets. Avoids format conversion issues.
2021-11-24 16:02:24 +03:00
kd-11 a21c6c4628 rsx: Fix handling of scaling requests for packed formats
- One does not simply interpolate RGB565 components as U16 data!
2021-11-24 16:02:24 +03:00
kd-11 97bd8f7bc1 rsx: Update sampler format class when inheriting mipmap slices/sections 2021-11-24 16:02:24 +03:00
Megamouse 4d0330bf82 rsx: fix possible segfault 2021-11-16 09:31:16 +01:00
kd-11 5b0ef401f7 rsx: Fix sampling in X when 0 pitch is given
- A pitch of 0 still allows 1-dimensional addressing.
2021-10-31 14:32:42 +03:00
kd-11 78bcb0fd53 rsx: Do not reuse/destroy sections that have references held
- Avoids a situation where blit-dst and blit-src have overlapping ranges. Uploading blit-dst destroys blit-src and vice-versa.
  This is not the end of the world, but blit-src should be kept around until the operation is completed to avoid stale references!
2021-10-27 12:30:43 +03:00
kd-11 f745971cc8
rsx: Fix coordinate scaling for shadow access (#10668)
- For shadow2DProj the 3rd coordinate is actually the depth value, do not scale
2021-08-06 22:49:50 +01:00
kd-11 99b6963fab rsx: Improve unnormalized coordinate sampling
- Improve rounding when sampling nearest neighbour. This is mostly a problem with NVIDIA
- Implement unnormalized 3D sampling
2021-08-03 00:36:04 +03:00
kd-11 6a9d1edee1 vk: Fix use-after-free hazard by checking if we're faulting from within the texture cache
- If we're using the texture cache, DO NOT delete resources.
2021-07-25 20:55:09 +03:00
kd-11 2524c35638 vk: Improve handling of texture cache temporary resources
- Temp resources from the texture cache are used to hold composite objects being sent to the GPU and can waste a lot of memory.
- Remove them if we run out of memory as they can linger around for a long time.
2021-07-15 18:05:35 +03:00
kd-11 a2f93b0696 rsx: Implement a simple cache eviction routine
- Can remove all non-essential textures from the cache except those passed as an exclusion list
2021-07-15 18:05:35 +03:00
kd-11 966aec7ad7 rsx: Resync excluded memory regions to avoid memory tests failing after flush events
- This is a mostly correct fix, but a corner case exists that can leak old data to the surface cache
2021-06-15 15:42:16 +03:00
kd-11 568af756cc rsx: Fix expired sampler descriptors
- Rebuilding when strict mode is enabled was incomplete.
  The copy has to be redone if the source has been updated.
2021-06-06 15:37:47 +03:00
Ani a49446c9e9
Replace gsl::span for std::span (c++20) (#7531)
* Replace gsl::span for std::span (c++20)
* Replace gsl::byte with std::byte

Co-authored-by: Bevan Weiss <bevan.weiss@gmail.com>
2021-05-30 17:10:46 +03:00