[ebosido444-ai]

What

A title can relocate a registered Shader header — and the register, special and semantic arrays it points at — into memory of its own, then resolve the ABI with that copy, which it never registered (ResolveShaderAbi, ResolveGraphicsStagesAbi, ResolveGraphicsAbi). SameHeader accepted such a copy only while its pointers stayed put: RelocatedRegisters covers a moved register array alone, and the fixed-field comparison covers the case where only user_data moved. Once another referenced array moved with the header, RegisteredHeader matched nothing and the resolution threw graphics ABI refers to a replaced shader header. Astro's Playroom (PPSA01325) resolves its vertex ABI with such a copy (#2914, discussion #3232).

A copy is now accepted when every array it names still holds the registered bytes. The registered header then decodes exactly what the copy describes, so nothing is copied, no snapshot is published and ShaderSnapshot stays immutable.

What a copy's pointers cannot change is content, and that is what the check now compares:

• the fixed fields, with user_data not compared and the register, context-register, specials, input-semantic and output-semantic pointers masked;
• every array those pointers name, through its guest bytes, against the registered header's bytes at the registered pointer (unreadable arrays refuse).

A copy whose registers, context registers, specials or semantics differ, or whose fixed fields differ, still throws ... refers to a replaced shader header — exact-or-throw is unchanged.

Not the earlier attempts

#2969, #4043 and #5830 published or mutated a snapshot for the relocated header. This keeps the registry, the snapshot and its decoded program untouched: the acceptance test reads guest memory only, so a relocated call costs a bounded comparison and no allocation. There is no new snapshot, no PublishRegisteredShader call and no mutable member.

Tested

Windows 11 26200.8117, winlibs GCC 15.2.0, Ninja, Release. GPU: NVIDIA GeForce GTX 1650 Ti, driver 617.14 (Vulkan).

tests/execution/PreparedShaders.cpp gains RelocatedHeaders(), called from the default run beside SharedCodeHeaders():

• registers a vertex header, then resolves the ABI with a full copy of that header whose every embedded pointer points into the copy, and submits that draw;
• resolves with a copy that moves only its user_data pointer;
• requires three copies — one register patched, one special patched, one context register patched — to still throw replaced shader header.

On unmodified main (e9a3343) the same test fails at the first copy:

AGC driver: graphics ABI refers to a replaced shader header

With this change:

cmake --build build --parallel 8 --target agc_driver_prepared_shaders_tests
ctest --test-dir build -R ^agc_driver_prepared_shaders --output-on-failure
100% tests passed, 0 tests failed out of 11

The 11 modes are the default, indirect, uncached, debug, debug_failure, deferred, deferred_undecodable, registered_workgroup, registered_workgroup_sync, lazy and lazy_sync runs of the same binary; SharedCodeHeaders and Registration cases are inside them.

Checklist

• Based on current main; no other open PR implements the same functions (fix(agc): update snapshot header and bytes on relocated header ABI resolution #2969, fix(agc): publish relocated shader snapshot on ABI resolution #4043, fix(agc): refresh relocated shader snapshot headers #5830 are closed; the open PRs that share these files touch other behaviours)
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt (the entry in docs/dev/debt/libSceAgcDriver.md is extended)
• Unimplemented paths throw (a copy whose arrays differ or are unreadable still throws)
• No notes, investigation .md files or images added to the repository
• New third-party code is a submodule built from source (n/a)
• Generic behaviour, not specific to one title (PPSA21567's user_data-only copy and PPSA01325's relocated copy both take this path)
• Depends on: none
• Unblocks: PPSA01325 Astro's Playroom (PPSA01325) #3232
• GPUs tested: NVIDIA GeForce GTX 1650 Ti, driver 617.14, Windows 11 26200.8117 (Vulkan; agc_driver_prepared_shaders, all 11 modes)
• AI-assisted: yes

[github-actions]

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.

Pull request	Shared files not compared
PR 4448	docs/dev/debt/libSceAgcDriver.md
PR 5058	docs/dev/debt/libSceAgcDriver.md
PR 5115	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp core/libs/prx/libSceAgcDriver/tests/execution/PreparedShaders.cpp
PR 5431	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp
PR 5499	docs/dev/debt/libSceAgcDriver.md
PR 5925	docs/dev/debt/libSceAgcDriver.md
PR 6146	docs/dev/debt/libSceAgcDriver.md
PR 6188	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp
PR 6194	docs/dev/debt/libSceAgcDriver.md
PR 6223	docs/dev/debt/libSceAgcDriver.md
PR 6252	docs/dev/debt/libSceAgcDriver.md
PR 6280	docs/dev/debt/libSceAgcDriver.md
PR 6300	docs/dev/debt/libSceAgcDriver.md
PR 6457	docs/dev/debt/libSceAgcDriver.md
PR 6548	docs/dev/debt/libSceAgcDriver.md
PR 6549	docs/dev/debt/libSceAgcDriver.md
PR 6588	docs/dev/debt/libSceAgcDriver.md
PR 6612	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp docs/dev/debt/libSceAgcDriver.md
PR 6636	docs/dev/debt/libSceAgcDriver.md
PR 6679	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp
PR 6683	docs/dev/debt/libSceAgcDriver.md
PR 6727	docs/dev/debt/libSceAgcDriver.md
PR 6808	docs/dev/debt/libSceAgcDriver.md
PR 6867	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp core/libs/prx/libSceAgcDriver/tests/execution/PreparedShaders.cpp docs/dev/debt/libSceAgcDriver.md
PR 6868	docs/dev/debt/libSceAgcDriver.md
PR 6887	docs/dev/debt/libSceAgcDriver.md
PR 6941	docs/dev/debt/libSceAgcDriver.md
PR 6946	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp docs/dev/debt/libSceAgcDriver.md
PR 6947	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp docs/dev/debt/libSceAgcDriver.md
PR 6965	docs/dev/debt/libSceAgcDriver.md
PR 6973	docs/dev/debt/libSceAgcDriver.md
PR 6977	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp docs/dev/debt/libSceAgcDriver.md
PR 6987	docs/dev/debt/libSceAgcDriver.md
PR 6989	docs/dev/debt/libSceAgcDriver.md
PR 6995	docs/dev/debt/libSceAgcDriver.md
PR 7002	docs/dev/debt/libSceAgcDriver.md
PR 7085	docs/dev/debt/libSceAgcDriver.md
PR 7256	docs/dev/debt/libSceAgcDriver.md
PR 7332	docs/dev/debt/libSceAgcDriver.md
PR 7446	docs/dev/debt/libSceAgcDriver.md
PR 7449	docs/dev/debt/libSceAgcDriver.md
PR 7505	docs/dev/debt/libSceAgcDriver.md
PR 7604	docs/dev/debt/libSceAgcDriver.md
PR 7642	core/libs/prx/libSceAgcDriver/Execution/src/Driver/Shaders/ShaderRegistry.cpp docs/dev/debt/libSceAgcDriver.md
PR 7659	docs/dev/debt/libSceAgcDriver.md
PR 7683	docs/dev/debt/libSceAgcDriver.md
PR 7800	docs/dev/debt/libSceAgcDriver.md
PR 7844	docs/dev/debt/libSceAgcDriver.md
PR 7912	docs/dev/debt/libSceAgcDriver.md

[keravyn-ops]

AMD / Windows run of this PR at fb07b68: Radeon 780M (RDNA 3, 0x15bf), AMD driver 25.10.20.01 (32.0.21020.1007), Windows 11 Home 10.0.26100, WinLibs GCC 15.2.0 r7, Release.

Every ctest entry built from the test targets this PR changes passes (11):

• agc_driver_prepared_shaders (1.01 s): 591: prepared shader and transactional registration tests passed
• agc_driver_prepared_shaders_indirect (0.80 s): 784: prepared shader and transactional registration tests passed
• agc_driver_prepared_shaders_uncached (0.80 s): 785: prepared shader and transactional registration tests passed
• agc_driver_prepared_shaders_debug (0.80 s): 786: prepared shader and transactional registration tests passed
• agc_driver_prepared_shaders_debug_failure (0.81 s): 787: injected failure before registration
• agc_driver_prepared_shaders_deferred (0.20 s): 788: deferred shader preparation tests passed
• agc_driver_prepared_shaders_deferred_undecodable (0.19 s): 789: deferred undecodable shader tests passed
• agc_driver_prepared_shaders_registered_workgroup (0.19 s): 790: registered NGG preparation deferral passed; subgroup=64
• agc_driver_prepared_shaders_registered_workgroup_sync (0.18 s): 791: registered NGG preparation deferral passed; subgroup=64
• agc_driver_prepared_shaders_lazy (0.35 s): 792: lazy shader preparation tests passed
• agc_driver_prepared_shaders_lazy_sync (0.21 s): 793: lazy shader preparation tests passed

AI-assisted: yes (runs made on the device above).

[adriadam10]

Checked the diff (array bounds in ReferredBytesMatch, unreadable arrays refuse, exact-or-throw unchanged), that the new test pins the relocated-copy case and the three patched copies still throw, and the Debt entry. Rebased onto current main (only a test-function adjacency conflict in PreparedShaders.cpp); the agc prepared-shader tests pass on the rebase, and keravyn-ops' AMD/Windows run covers the same targets. merged.