[sanxchep]

What

Size the guest-memory guard for a multisampled RenderTarget using 64 KiB blocks whose pixel dimensions account for fragment count. The old max(single-sample CB bytes × samples, raw pixel bytes × samples) bound can cover a neighboring allocation. For example, a 1920×1080 RGBA8 4× target occupies 0x1fe0000 bytes with fragment-aware blocks, while the previous guard spans 0x21c0000 bytes.

Keep the conservative bound for other tile modes. Add guard regressions for tiled 2× and 8× targets, a layout check for the 1920×1080 case, and document that console layout equivalence has not been measured.

This is independent of custom sample-location support in #2169. PR #2132 computes the same block dimensions for color-memory fills; if it merges first, this guard can reuse its footprint helper during integration.

Tested

Linux, Release, GCC 13.3:

• agc_driver_graphics — passed (3.40 s)
• agc_multisample_guest_access — passed (1.20 s)
• python3 tools/check_conventions.py --base origin/main — passed
• git diff --check — passed

A title was run with the patch and reported working by the PR author. Console layout equivalence was not separately measured.

Checklist

• Based on current main; no other open PR changes the guest-read guard
• One topic per PR; follow-ups go in a new PR
• No comments in code except technical debt
• Unimplemented paths throw; no silent stubs added
• No notes, investigation .md files or images added to the repository
• New third-party code is a submodule built from source; none added
• Generic behaviour, not specific to one title
• Depends on: none
• AI-assisted: yes

[MarcZX10]

The conventions run (check 114622799227) fails because the PR template’s AI-assisted field is unanswered. Please mark AI-assisted: yes or no in the checklist/body so the required template check can pass.

[sanxchep]

The conventions run (check 114622799227) fails because the PR template’s AI-assisted field is unanswered. Please mark AI-assisted: yes or no in the checklist/body so the required template check can pass.

done

[blawesom]

Rebase handoff — this branch was ~347 commits behind main (only a trivial TechnicalDebt.md union conflict). Rebased onto current main, authorship unchanged; both doc entries kept.

Verified on Linux (CachyOS, kernel 7.2.9, GCC 16.2.1, Ninja, Release, BUILD_TESTING=ON, AMD RX 9070 XT / RADV):

• builds; tools/check_conventions.py --base origin/main passes
• full suite: 949/949 passed

To pick it up:

git fetch https://github.com/blawesom/AnyPS5 rebase/4199
git checkout <your branch>
git merge --ff-only FETCH_HEAD   # or reset to it

Head is 5ab79cd51. (This PR also passed an earlier independent pre-review at its previous base — full suite was green there too.)

[sanxchep]

Rebase handoff — this branch was ~347 commits behind main (only a trivial TechnicalDebt.md union conflict). Rebased onto current main, authorship unchanged; both doc entries kept.

Verified on Linux (CachyOS, kernel 7.2.9, GCC 16.2.1, Ninja, Release, BUILD_TESTING=ON, AMD RX 9070 XT / RADV):

• builds; tools/check_conventions.py --base origin/main passes
• full suite: 949/949 passed

To pick it up:

git fetch https://github.com/blawesom/AnyPS5 rebase/4199
git checkout <your branch>
git merge --ff-only FETCH_HEAD   # or reset to it

Head is 5ab79cd51. (This PR also passed an earlier independent pre-review at its previous base — full suite was green there too.)

done

[blawesom]

Independent verification of your rebase — the new head builds and passes the full local suite on Linux (CachyOS, kernel 7.2.9, GCC 16.2.1, Ninja, Release, BUILD_TESTING=ON, AMD RX 9070 XT / RADV): 957/957, conventions green. From my side this is merge-ready.

[sanxchep]

@MarcZX100 should be ready to merge

[kr4y4]

Windows run, to go with the Linux runs above.

Setup: Windows 11 (26200), WinLibs GCC 15.2.0 r7, Release, CI configure flags, an RTX 4070 Laptop GPU and an Intel Raptor Lake-S iGPU. The head cff1f68 merged into main 2a5eca1, with no conflicts.

• agc_multisample_guest_access and agc_driver_graphics (with the new ColorTargetLayout check) pass on both GPUs with --repeat until-fail:3.
• With main's MultisampleTarget.cpp and the new tests, agc_multisample_guest_access fails on both GPUs: a CPU read of 0x…90000+0x40 overlaps multisampled color target 0x…80000+0x40000 (4 samples), whose samples are only on the host: guest access to it is not implemented.
• Full suite (ctest --timeout 120 --parallel 4): 973 of 973 on the RTX 4070. On the iGPU 972 of 973; the one failure, agc_driver_f32_dot2_rounding, fails the same way on main (no shaderFloat64, test(agc): skip the f32 dot2 rounding test on devices without shaderFloat64 #4528).

This touches guard() in MultisampleTarget.cpp like #4474, so whichever lands second will need a rebase there.

Written with AI assistance.

[github-actions]

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.

Pull request	Conflicting files (lines on main)	Same exports	Same new files
#4474	core/libs/prx/libSceAgcDriver/Graphics/src/MultisampleTarget.cpp L195-196 core/libs/prx/libSceAgcDriver/tests/Recorder.cpp L4661-4662