Repository navigation
Conversation
|
Other open pull requests touch the same code. Merging one will break the other, or both implement the same thing. If this one builds on them, list them in Depends on.
Some comparisons could not be completed because a pull request conflicts with the base branch. Review these shared files manually before concluding that there is no overlap.
|
|
Thanks, the mip addressing and the per-texel The PR now decides "3D target" by Repro: replace the A linear depth-1 volume without an existing host import reaches the same refusal. It used to render, and linear 3D at depth 1 has the same layout as 2D. Could you:
|
ce9264c to
a96dad4
Compare
|
Thanks for catching the depth-one regression. Addressed in a96dad4ef, after rebasing onto mainf663 and removing the now-merged #7294 dependency. All three requested changes are covered: volumes bypass the2D byte comparison; single-level depth-one linear volumes keep their valid existing copy path; AI-assisted: Codex; author responsible. |
a96dad4 to
bf98421
Compare
github-actions
Bot
added
unblocks title
|
Apple GPU / MoltenVK run of the new Machine: Apple M2 Max, macOS 27.2, x86-64 build under Rosetta. Head bf98421. Method: main 81cb863 + @mugurc's
Not done: the full suite with the PR. GPUs tested: Apple M2 Max (MoltenVK 1.4.2; MoltenVK main + SPIRV-Cross #2704) Written with AI assistance (Claude Code). I ran the builds and tests above on my machine. |
This branch has not been deployed
3 participants
What
Render into one mip of an uncompressed3D color surface, preserving the base dimensions and volume identity, including depth1. Use its3D mip layout, bound its slice view by the selected mip's depth, validate the full allocation, and create the resident attachment view at that mip. Unverified mipmapped3D DCC/CMASK metadata paths still throw.
Depends on: none. #7294 has merged. Head
bf984211f5rebases this topic's three commits onto mainbb2624a5331ad814616e24958a434eb9c1945200. The first conflict combines the3D mip layout with main's new CMASK mip-tail accounting; both depth-one review commits replay unchanged. Later main changes to write-watch stamping/SRT plans are included, with the same three topic patches. The prior published a96 head is preserved; local canonical branches now match the published heads.Sonic Unleashed Recompiled PPSA99902, #7260 / #4443, requests a two-level128x128x8 RGBA8_UINT
SW_4KB_S_Xtarget after the opening movie. The ordinary bring-up composition stops at the guard still present on main:The diagnostic captured
CB_COLOR_INFO=0x00070428,ATTRIB2=0x101fc07f,ATTRIB3=0x0ea54007,VIEW=0x0000e000(mip0, slices0..7). #7274 merged the diagnostic; this capability change replaces the general refusal. Its64x4x4/mip1 descriptor test verifies the selected32x2 one-slice view. No title-specific branch or game modification is added.Review correction: non-mipmapped depth-one volumes now bypass the2D resident-byte comparison as all other volumes do; their complete3D allocation was already validated against
DescribeSurface. A single-level linear volume of depth1 retains the existing linear copy path, whose2D/3D layout is identical. Other nonresident volume cases continue to throw. Four new--mipscases cover96x32x1/one level in Linear, Standard4KB, RenderTarget and S4KBX modes, with exact per-texel readback.Overlap review: #7199 is merged, and its2D CMASK accounting/refusals are retained. #6959 (mesh layer windows), #7254/#7256 (array attachment windows) are separate capabilities. Existing mesh/vertex cases and selected-mip bounds remain.
Tested
Require(maxMip == 0, "mipmapped 3D color targets are unsupported"). The volume group fails3/3 on AMD at that exact refusal. Restore the implementation:3/3 pass on both GPUs. The earlier depth-one review control at3294c685f also failed3/3 on the linear refusal; correcting only that exposed the Standard4KB resident-layout mismatch, and the volume predicate correction made all14 cases pass.agc_driver_layered_volume_vertex_no_gplfails3/3: slice1 red remains0x40, expected255. Its GPL-on variant passes3/3. Same result on this PR; existing #5271, maintained fresh validation. No workaround or new cause claim. This is not an all-green suite.10de:2c02,610.62/32.0.16.1062, and AMD Radeon(TM) Graphics1002:13c0,32.0.21036.18; Windows11 Enterprise26200. Linux: Mesa25.2.8 llvmpipe LLVM20.1.2. Khronos1.4.363 loaded, Vulkan required, project shader cache/pipeline journal disabled, driver caches uncontrolled; serial30s bounds. Zero logged VUID/validation errors, but maintenance9 forward-compatibility advisories remain:150 in the NVIDIA/Linux selections,124 in AMD's shorter failing selection.WARNING-VkImageSubresourceRange-layerCount-compatibilityasks for VK_REMAINING_ARRAY_LAYERS if maintenance9 is enabled later. No warning-clean or performance claim.ctest --test-dir build -V -j1 --repeat until-fail:3 --timeout 30 -R '^agc_(write_tracking|shader_memory|driver_(graphics|dcc_keys_first_write|image_xor_swizzle_volume|layered_volume_(draw|submit|vertex|vertex_no_gpl)|volume_mip_targets))$'Title evidence: the separate45dafde integration, containing this PR's a96 implementation and other bring-up fixes, reached a level scene with Sonic/Chip and the gameplay HUD (timer00:00:00) in its final capture. Its3600s bound ended the run; movement/interactive gameplay and FPS remain unverified.117 captures/post-run state were archived,142 prior runtime/state/config files hash-restored. Zero logged VUID/errors,8 maintenance9 advisories and12 guest fence failures remain. Current source pin, input and limits. This is composed title evidence; no PR-only title A/B, full-suite, console-equivalence or performance claim.
Checklist
main; no other open PR implements the same functionsNotImplemented_nid_no_patch); silent stubs are listed under Silent stubs in their library's technical debt file.mdfiles or images added to the repository (attach them to this PR)