fix(vk): release pages ending at a GPU import boundary #469

Open

zeldaootr wants to merge 4 commits from zeldaootr/Longshot:import-exclusive-end into main

pull from: zeldaootr/Longshot:import-exclusive-end

merge into: MagicBeanDev:main

MagicBeanDev:main

Conversation 5 Commits 4 Files changed 12 +98 -1

zeldaootr commented 2026-11-14 16:04:02 +00:00

Contributor

• Copy link
• Copy Markdown

Type: fix · Risk: low · Tested on: Steam Deck OLED, SteamOS 3.9.5, RADV; Linux x86-64 (game-free) · Needs: Windows, any GPU

Why

A free that ends exactly where a live GPU import starts keeps the entire range backed: the renderer counts the touching import as an overlap. This happens during the Kokiri Forest to Hyrule Field transition on the Steam Deck (twice per transition in an instrumented run, with 8 KiB and 64 KiB frees).

How to reproduce

Seen on: Steam Deck OLED (RADV), an instrumented RC3 renderer at the Field transition; Linux x86-64, game-free
Settings and switches: defaults
Steps: on main aa95d35f with this PR's test added (d30476db), run python -I -B tests/test_guest_imported.py --cxx g++: it asks about [0x1000, 0x2000) with an import starting at 0x2000
Expected: not imported: the ranges only touch
Actual: Assertion '!memory_guest_imported(0x1000, 0x2000)' failed

What changed

The ownership check now treats a freed range as half-open, so an import that starts exactly where the free ends no longer keeps the freed pages backed.

• vk memory: memory_guest_imported() in gpu/vk/vk_memory.cpp starts its backwards search at imported.lower_bound(hi) for the half-open range [lo, hi), so an import starting at hi no longer counts as an overlap; an empty range answers 0 before the pool lock is taken.
• tests: tests/test_guest_imported.py (new) cuts the function out of the source and checks both touching boundaries and true overlaps.
• docs/tested: the earlier instrumented-backport runs, and the Deck runs of this branch against main (the four main records are the same files in #468 and #472: one shared main arm, which merges as one).

Measurements

Verdict: passes: no change in fps, frame times or memory on a Steam Deck at the Field transition (20.60 against 20.56 fps, inside the run-to-run spread; no fault in 4 runs each): the fix frees too little there to show; the game-free test fails on main and passes here.

Hardware: Steam Deck OLED, AMD Custom APU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), 16 GB shared, SteamOS 3.9.5 (kernel 7.2.7-valve2)
Scene: Hyrule Field (child), arriving from Kokiri Forest by teleport to the Lost Woods bridge entrance, day; 1280×720 output, render resolution option 4, no frame generation, precompile off, 1536 MiB texture budget, 256 MiB mirror budget, headless
Method: one scripts/dev_packs.py base at main 9c9d69ed (release optimisation, x86-64-v3); main is the base's runtime with no pack, this PR a renderer pack (longshot_vk.so only; the host binary is byte-identical). One warm-up pass of main made the caches, copied fresh into every run. Each run: child Kokiri save settled 20 s, teleport to HyruleField_Child, 20 s warm-up, 45 s stationary sample. Four rounds in one session, interleaved main, #468, #469, #472; medians of the 4 runs per arm. FPS and present intervals from HOS_PRESENT_TRACE in the 45 s window; GPU clock and busy from sysfs every 0.5 s; guest memory backing from the guest memfd's allocated blocks (st_blocks).
Rebased since: the branch is now on main aa95d35f, the fix at 075e9c24 (the same change as bed7d855, which 76382425 carries, git patch-id). Not measured again on aa95d35f. Since 9c9d69ed main changed gpu/vk/vk_memory.cpp only in how the mirror budget is set (#735), not in memory_guest_imported(); everything else it changed both arms would take alike, so these numbers still compare the change. The game-free test was run again on aa95d35f (below).
How to reproduce: build a base at main and a pack of this branch (python -I scripts/dev_packs.py base|pack|rt, docs/development.md "Testing a change against a base"), start each runtime with HOS_CONTROL and a child save at Kokiri spawn, then longshot_ctl.py teleport HyruleField_Child, wait 20 s and sample 45 s.

	Before: main @ 9c9d69ed	After: this PR @ 76382425
FPS, Field sample (median of 4)	20.60	20.56
Present interval median (ms)	48.56	48.63
Present interval p95 (ms)	51.79	51.76
GPU clock median (MHz) / busy	1040 / 99%	1040 / 97%
Guest memory backed before the teleport (MiB)	5431	5430
Guest memory backed at sample end (MiB)	5867	5860
Lowest available memory (MiB)	2185	1950
Device faults, run memory guard, crashes	none in 4 runs	none in 4 runs

Per run, FPS: main 20.56, 20.54, 20.69, 20.64; this PR 20.59, 20.66, 20.52, 20.38. Guest memory backed at the end: main 5861-5909 MiB, this PR 5832-5900 MiB. Both differences are inside the run-to-run spread: at this scene the fix frees too little to show in the backing or the frame rate. That matches the earlier instrumented runs (below), which saw the boundary hit twice per transition, with 8 KiB and 64 KiB frees: the fix is for correctness, not for a measurable saving here.

Earlier, an instrumented v3.0.0-rc.3 renderer (SHA-256 prefix ce5646666a7e) with a test switch choosing the old or new decision (RX_TEST_IMPORT_EXCLUSIVE_END, not in this PR), one run each on the same Deck: 20.09 FPS old, 20.33 FPS new, the boundary hit twice per transition (docs/tested/2026-11-14-field-renderer-zeldaootr-import-old.md, -import-fixed.md). On a Linux sparse shared memfd, freeing a 64 MiB range next to a simulated live import left 64 MiB backed with the old predicate and released it with the fix (three densities, three repeats, adjacent and one-page-overlap controls, 36 trials).

main-r2:   [hos host] longshot v3.0.0-rc.3-75-g9c9d69ed (source 6b76ded3)
           vk: 5160 presents, 20.6 fps; a frame: 1176 draws (0 skipped), 145 dispatches, 155 passes, ...
           vk: imports: 48 live (2016 MB), 0 retired still imported (0 MB), 13 retired given back (1056 MB)
pr469-r2:  renderer v3.0.0-rc.3-78-g76382425, host as main's
           vk: 5040 presents, 20.9 fps; a frame: 1185 draws (0 skipped), 145 dispatches, 154 passes, ...
           vk: imports: 52 live (1984 MB), 0 retired still imported (0 MB), 11 retired given back (816 MB)

$ python3 -I -B tests/test_guest_imported.py --cxx g++      (at d30476db, main aa95d35f plus the test; rounds 1 and 2)
check.exe: check.cpp:38: main()::<lambda()>: Assertion `!memory_guest_imported(0x1000, 0x2000)' failed.
$ python3 -I -B tests/test_guest_imported.py --cxx g++      (at afd5ea15; rounds 1 and 2)
(passes, exit 0)

Flags and settings

None

Testing

Tested: On the Steam Deck OLED above, 4 runs of this PR interleaved with 4 of main through the Kokiri Forest to Hyrule Field transition: Field reached every time (status map HyruleField_Child, 918-930 presents in each 45 s sample), no device fault, no VK_ERROR, no crash record, the run's low-memory guard never tripped, every run exited 0. In one run of this PR (round 3) the renderer gave back 16 idle imports at its system-heap limit (vk: system heap 2648 of 3072 MB: 16 idle imports given back (352 MB)); main's runs did not log that line, and the run's frame rate was not affected (20.52 FPS). Game-free, on this branch rebased onto main aa95d35f: tests/test_guest_imported.py fails at the test-only commit d30476db and passes at afd5ea15, twice each, alternating, and under tests/run_all.py with g++ 13.3 and with clang 23.1.5 at both commits; the 43 checks that read gpu/vk/vk_memory.cpp or the new test pass at the tip, 43 of 43. With #468 and #472, as rebased, the branch merges cleanly (git merge-tree).

Not tested: Unique pages reclaimed during gameplay (the backing measured includes everything the game holds); Windows and other GPUs and drivers in the game.

Must be tested:

• Windows 10/11, any GPU: build this PR (with llvm-mingw), load a child save in Kokiri Forest, go to Hyrule Field, stay 1-2 minutes and back, expecting no device lost, no VK_ERROR, no crash record in longshot_log.txt and fps like main at the same spot (by hand, test request T-S2R3: post the fps, any vk: system heap lines and the GPU, driver and OS)

Risks

What could break: a free that only touches a live import is now released where it was kept; the existing synchronization between the ownership check and the host discard remains. Review alongside the pending-import lifetime fix, #468.

Turning it off: revert the fix commit

Compatibility: Nothing changes

Issues

None filed. Overlaps #468 (the same empty-range guard; the two merge cleanly) and #431 (the 1 GiB stop two lines below).

Checklist

Everyone

• One topic.
• The title is a Conventional Commit (type(scope): description) of the status line's type.
• Why gives the reason plainly and links the issues that fit (the tracker searched).
• Measurements has a verdict, or one line saying why there is nothing to measure.
• Testing names what was tested, what wasn't, and what must still be tested, where and how.
• Rebased on current main. Onto main 94abdd57 (head 33ca425e) with no conflict, with the commit subjects made Conventional Commits, which CI's lint job now checks for every commit, on a base that has CI's config (#417); each commit's change is the same as before by git patch-id. The commits named in this description are from before this rebase: d30476db is now 351d28ba, 075e9c24 is now 5adc36c1, 7508a3f2 is now 00e03697, afd5ea15 is now 33ca425e. Re-run on 94abdd57: test_guest_imported fails at the test-only commit 351d28ba and passes at the head, with g++ 13.3 and clang 23.1.5; of the 46 checks that are this PR's or read its files (g++ 13.3), 46 pass.
• python -I -B tests/run_all.py passes (any failure that also fails on main is named). The full suite on main aa95d35f (Linux, g++ 13.3, the Vulkan headers set): 528 passed, 15 failed, 50 skipped; on this branch the new test and the 43 checks that read the changed files ran, with no failure that main does not have (see Testing). The 15 that fail on main: test_appimage_player_build, test_bake_portability, test_cpm_discovery, test_crash_dialog, test_env_on, test_lift_cache, test_lift_cleanup_error, test_make_game_cache, test_mods_size_table, test_options_budget_once, test_pipeline_tour, test_readback_late, test_vk_pgo, test_walk_key and nx2/test_texture_init_barrier.py.
• No game files, dumps, build output or generated files.
• third_party/ untouched (a new folder with a PROVENANCE.md instead).
• The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled.

When it applies

• AI assistance recorded with Assisted-by: / Co-Authored-By: trailers.
• Numbers from the same PC, scene and settings, at least 2 interleaved runs each, the median reported.
• A file in docs/tested/ for each new measured run (format in docs/tested-configurations.md).
• A game-free test added for the fix, failing on main and passing here.
• Host code builds on both Windows (llvm-mingw) and Linux, with UTF-8 paths. Before the rebase, built on Linux for both: RX_VK_ONLY=1 RX_RELEASE=1 RX_TARGET=windows with llvm-mingw 20261104 linked longshot_vk.dll, and the same without RX_TARGET linked longshot_vk.so; after the rebase, gpu/vk/vk_memory.cpp at afd5ea15 checked whole (-fsyntax-only -Wall -Wextra) with clang 23.1.5 for Linux and llvm-mingw 20261104 for Windows: the same diagnostics as main aa95d35f's (54 and 43), none new. The diff has no path handling.

**Type:** fix · **Risk:** low · **Tested on:** Steam Deck OLED, SteamOS 3.9.5, RADV; Linux x86-64 (game-free) · **Needs:** Windows, any GPU ## Why A free that ends exactly where a live GPU import starts keeps the entire range backed: the renderer counts the touching import as an overlap. This happens during the Kokiri Forest to Hyrule Field transition on the Steam Deck (twice per transition in an instrumented run, with 8 KiB and 64 KiB frees). ## How to reproduce **Seen on:** Steam Deck OLED (RADV), an instrumented RC3 renderer at the Field transition; Linux x86-64, game-free **Settings and switches:** defaults **Steps:** on `main` aa95d35f with this PR's test added (d30476db), run `python -I -B tests/test_guest_imported.py --cxx g++`: it asks about `[0x1000, 0x2000)` with an import starting at `0x2000` **Expected:** not imported: the ranges only touch **Actual:** `Assertion '!memory_guest_imported(0x1000, 0x2000)' failed` ## What changed The ownership check now treats a freed range as half-open, so an import that starts exactly where the free ends no longer keeps the freed pages backed. - vk memory: `memory_guest_imported()` in `gpu/vk/vk_memory.cpp` starts its backwards search at `imported.lower_bound(hi)` for the half-open range [lo, hi), so an import starting at `hi` no longer counts as an overlap; an empty range answers 0 before the pool lock is taken. - tests: `tests/test_guest_imported.py` (new) cuts the function out of the source and checks both touching boundaries and true overlaps. - docs/tested: the earlier instrumented-backport runs, and the Deck runs of this branch against main (the four main records are the same files in https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/468 and https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/472: one shared main arm, which merges as one). ## Measurements **Verdict:** passes: no change in fps, frame times or memory on a Steam Deck at the Field transition (20.60 against 20.56 fps, inside the run-to-run spread; no fault in 4 runs each): the fix frees too little there to show; the game-free test fails on `main` and passes here. Hardware: Steam Deck OLED, AMD Custom APU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), 16 GB shared, SteamOS 3.9.5 (kernel 7.2.7-valve2) Scene: Hyrule Field (child), arriving from Kokiri Forest by teleport to the Lost Woods bridge entrance, day; 1280×720 output, render resolution option 4, no frame generation, precompile off, 1536 MiB texture budget, 256 MiB mirror budget, headless Method: one `scripts/dev_packs.py` base at main 9c9d69ed (release optimisation, x86-64-v3); main is the base's runtime with no pack, this PR a renderer pack (`longshot_vk.so` only; the host binary is byte-identical). One warm-up pass of main made the caches, copied fresh into every run. Each run: child Kokiri save settled 20 s, teleport to HyruleField_Child, 20 s warm-up, 45 s stationary sample. Four rounds in one session, interleaved main, #468, #469, #472; medians of the 4 runs per arm. FPS and present intervals from `HOS_PRESENT_TRACE` in the 45 s window; GPU clock and busy from sysfs every 0.5 s; guest memory backing from the guest memfd's allocated blocks (`st_blocks`). Rebased since: the branch is now on main aa95d35f, the fix at 075e9c24 (the same change as bed7d855, which 76382425 carries, `git patch-id`). Not measured again on aa95d35f. Since 9c9d69ed `main` changed `gpu/vk/vk_memory.cpp` only in how the mirror budget is set (#735), not in `memory_guest_imported()`; everything else it changed both arms would take alike, so these numbers still compare the change. The game-free test was run again on aa95d35f (below). How to reproduce: build a base at main and a pack of this branch (`python -I scripts/dev_packs.py base|pack|rt`, docs/development.md "Testing a change against a base"), start each runtime with `HOS_CONTROL` and a child save at Kokiri spawn, then `longshot_ctl.py teleport HyruleField_Child`, wait 20 s and sample 45 s. | | Before: main @ 9c9d69ed | After: this PR @ 76382425 | |---|---|---| | FPS, Field sample (median of 4) | 20.60 | 20.56 | | Present interval median (ms) | 48.56 | 48.63 | | Present interval p95 (ms) | 51.79 | 51.76 | | GPU clock median (MHz) / busy | 1040 / 99% | 1040 / 97% | | Guest memory backed before the teleport (MiB) | 5431 | 5430 | | Guest memory backed at sample end (MiB) | 5867 | 5860 | | Lowest available memory (MiB) | 2185 | 1950 | | Device faults, run memory guard, crashes | none in 4 runs | none in 4 runs | Per run, FPS: main 20.56, 20.54, 20.69, 20.64; this PR 20.59, 20.66, 20.52, 20.38. Guest memory backed at the end: main 5861-5909 MiB, this PR 5832-5900 MiB. Both differences are inside the run-to-run spread: at this scene the fix frees too little to show in the backing or the frame rate. That matches the earlier instrumented runs (below), which saw the boundary hit twice per transition, with 8 KiB and 64 KiB frees: the fix is for correctness, not for a measurable saving here. Earlier, an instrumented v3.0.0-rc.3 renderer (SHA-256 prefix ce5646666a7e) with a test switch choosing the old or new decision (`RX_TEST_IMPORT_EXCLUSIVE_END`, not in this PR), one run each on the same Deck: 20.09 FPS old, 20.33 FPS new, the boundary hit twice per transition (`docs/tested/2026-11-14-field-renderer-zeldaootr-import-old.md`, `-import-fixed.md`). On a Linux sparse shared memfd, freeing a 64 MiB range next to a simulated live import left 64 MiB backed with the old predicate and released it with the fix (three densities, three repeats, adjacent and one-page-overlap controls, 36 trials). ``` main-r2: [hos host] longshot v3.0.0-rc.3-75-g9c9d69ed (source 6b76ded3) vk: 5160 presents, 20.6 fps; a frame: 1176 draws (0 skipped), 145 dispatches, 155 passes, ... vk: imports: 48 live (2016 MB), 0 retired still imported (0 MB), 13 retired given back (1056 MB) pr469-r2: renderer v3.0.0-rc.3-78-g76382425, host as main's vk: 5040 presents, 20.9 fps; a frame: 1185 draws (0 skipped), 145 dispatches, 154 passes, ... vk: imports: 52 live (1984 MB), 0 retired still imported (0 MB), 11 retired given back (816 MB) $ python3 -I -B tests/test_guest_imported.py --cxx g++ (at d30476db, main aa95d35f plus the test; rounds 1 and 2) check.exe: check.cpp:38: main()::<lambda()>: Assertion `!memory_guest_imported(0x1000, 0x2000)' failed. $ python3 -I -B tests/test_guest_imported.py --cxx g++ (at afd5ea15; rounds 1 and 2) (passes, exit 0) ``` ### Flags and settings None ## Testing **Tested:** On the Steam Deck OLED above, 4 runs of this PR interleaved with 4 of main through the Kokiri Forest to Hyrule Field transition: Field reached every time (status map `HyruleField_Child`, 918-930 presents in each 45 s sample), no device fault, no `VK_ERROR`, no crash record, the run's low-memory guard never tripped, every run exited 0. In one run of this PR (round 3) the renderer gave back 16 idle imports at its system-heap limit (`vk: system heap 2648 of 3072 MB: 16 idle imports given back (352 MB)`); main's runs did not log that line, and the run's frame rate was not affected (20.52 FPS). Game-free, on this branch rebased onto `main` aa95d35f: `tests/test_guest_imported.py` fails at the test-only commit d30476db and passes at afd5ea15, twice each, alternating, and under `tests/run_all.py` with g++ 13.3 and with clang 23.1.5 at both commits; the 43 checks that read `gpu/vk/vk_memory.cpp` or the new test pass at the tip, 43 of 43. With https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/468 and https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/472, as rebased, the branch merges cleanly (`git merge-tree`). **Not tested:** Unique pages reclaimed during gameplay (the backing measured includes everything the game holds); Windows and other GPUs and drivers in the game. **Must be tested:** - Windows 10/11, any GPU: build this PR (with llvm-mingw), load a child save in Kokiri Forest, go to Hyrule Field, stay 1-2 minutes and back, expecting no device lost, no `VK_ERROR`, no crash record in `longshot_log.txt` and fps like main at the same spot (by hand, test request T-S2R3: post the fps, any `vk: system heap` lines and the GPU, driver and OS) ## Risks **What could break:** a free that only touches a live import is now released where it was kept; the existing synchronization between the ownership check and the host discard remains. Review alongside the pending-import lifetime fix, https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/468. **Turning it off:** revert the fix commit **Compatibility:** Nothing changes ## Issues None filed. Overlaps https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/468 (the same empty-range guard; the two merge cleanly) and https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/431 (the 1 GiB stop two lines below). ## Checklist ### Everyone - [x] One topic. - [x] The title is a Conventional Commit (`type(scope): description`) of the status line's type. - [x] Why gives the reason plainly and links the issues that fit (the tracker searched). - [x] Measurements has a verdict, or one line saying why there is nothing to measure. - [x] Testing names what was tested, what wasn't, and what must still be tested, where and how. - [x] Rebased on current `main`. Onto main 94abdd57 (head 33ca425e) with no conflict, with the commit subjects made Conventional Commits, which CI's lint job now checks for every commit, on a base that has CI's config (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/417); each commit's change is the same as before by `git patch-id`. The commits named in this description are from before this rebase: d30476db is now 351d28ba, 075e9c24 is now 5adc36c1, 7508a3f2 is now 00e03697, afd5ea15 is now 33ca425e. Re-run on 94abdd57: `test_guest_imported` fails at the test-only commit 351d28ba and passes at the head, with g++ 13.3 and clang 23.1.5; of the 46 checks that are this PR's or read its files (g++ 13.3), 46 pass. - [x] `python -I -B tests/run_all.py` passes (any failure that also fails on `main` is named). The full suite on main aa95d35f (Linux, g++ 13.3, the Vulkan headers set): 528 passed, 15 failed, 50 skipped; on this branch the new test and the 43 checks that read the changed files ran, with no failure that `main` does not have (see Testing). The 15 that fail on main: `test_appimage_player_build`, `test_bake_portability`, `test_cpm_discovery`, `test_crash_dialog`, `test_env_on`, `test_lift_cache`, `test_lift_cleanup_error`, `test_make_game_cache`, `test_mods_size_table`, `test_options_budget_once`, `test_pipeline_tour`, `test_readback_late`, `test_vk_pgo`, `test_walk_key` and `nx2/test_texture_init_barrier.py`. - [x] No game files, dumps, build output or generated files. - [x] `third_party/` untouched (a new folder with a `PROVENANCE.md` instead). - [x] The license is OK: GPL-3.0-or-later, nothing incompatible and nothing decompiled. ### When it applies - [x] AI assistance recorded with `Assisted-by:` / `Co-Authored-By:` trailers. - [x] Numbers from the same PC, scene and settings, at least 2 interleaved runs each, the median reported. - [x] A file in `docs/tested/` for each new measured run (format in `docs/tested-configurations.md`). - [x] A game-free test added for the fix, failing on `main` and passing here. - [x] Host code builds on both Windows (llvm-mingw) and Linux, with UTF-8 paths. Before the rebase, built on Linux for both: `RX_VK_ONLY=1 RX_RELEASE=1 RX_TARGET=windows` with llvm-mingw 20261104 linked `longshot_vk.dll`, and the same without `RX_TARGET` linked `longshot_vk.so`; after the rebase, `gpu/vk/vk_memory.cpp` at afd5ea15 checked whole (`-fsyntax-only -Wall -Wextra`) with clang 23.1.5 for Linux and llvm-mingw 20261104 for Windows: the same diagnostics as `main` aa95d35f's (54 and 43), none new. The diff has no path handling.

zeldaootr added 3 commits 2026-11-14 16:04:04 +00:00

reproduce adjacent GPU import overlap ... 3db151f733

Assisted-by: OpenAI Codex

exclude touching ranges from GPU import checks ... 4d6d25abe2

Assisted-by: OpenAI Codex

measure adjacent frees during the Field transition ... a6337d73ed

Assisted-by: OpenAI Codex

zeldaootr referenced this pull request 2026-11-14 16:05:47 +00:00

fix(vk): retain guest pages during import handoff and retirement #468

zeldaootr referenced this pull request 2026-11-14 16:06:21 +00:00

docs(tested): measure partially covered guest frees on Deck #471

mooki added the

area/renderer

kind

bug

priority

high

labels 2026-11-14 16:28:08 +00:00

mooki added the

platform/linux

platform/windows

labels 2026-11-14 18:24:13 +00:00

zeldaootr force-pushed import-exclusive-end from a6337d73ed to 3046b83159 2026-11-14 18:57:53 +00:00 Compare

zeldaootr commented 2026-11-14 18:57:55 +00:00

Author

Contributor

• Copy link
• Copy Markdown

Rebased onto main 1be60eae, the run records renamed to the tester handle, and measured on a Steam Deck OLED against main: four interleaved rounds, medians 20.60 fps on main and 20.56 here, guest memory backing the same within spread, no fault. One run here logged 'system heap 2648 of 3072 MB: 16 idle imports given back', which no main run did; it is noted in the description. Still WIP: not built for Windows yet; the test request asks for that.

Rebased onto main 1be60eae, the run records renamed to the tester handle, and measured on a Steam Deck OLED against main: four interleaved rounds, medians 20.60 fps on main and 20.56 here, guest memory backing the same within spread, no fault. One run here logged 'system heap 2648 of 3072 MB: 16 idle imports given back', which no main run did; it is noted in the description. Still WIP: not built for Windows yet; the test request asks for that.

zeldaootr referenced this pull request 2026-11-14 18:57:58 +00:00

perf(vk): scan pending uniform updates in place #472

zeldaootr commented 2026-11-14 19:24:32 +00:00

Author

Contributor

• Copy link
• Copy Markdown

Built for Windows from Linux with llvm-mingw 20261104 (the renderer alone, RX_VK_ONLY, as the CI job does) and for Linux: both link, and the changed file has no new warning under -Wall -Wextra on either. The Windows box is ticked; the Windows test request stays open for a run on real hardware.

Built for Windows from Linux with llvm-mingw 20261104 (the renderer alone, RX_VK_ONLY, as the CI job does) and for Linux: both link, and the changed file has no new warning under -Wall -Wextra on either. The Windows box is ticked; the Windows test request stays open for a run on real hardware.

zeldaootr changed title from WIP: fix(vk): release pages ending at a GPU import boundary to fix(vk): release pages ending at a GPU import boundary 2026-11-14 19:24:33 +00:00

zeldaootr referenced this pull request 2026-11-15 00:19:25 +00:00

fix(vk): give memory back under Linux memory pressure and after the precompile #223

zeldaootr commented 2026-11-15 16:41:39 +00:00

Author

Contributor

• Copy link
• Copy Markdown

Rebased onto main aa95d35f with no conflict (the same four changes, git patch-id; commits now under the noreply identity) and the description moved to the current template.
The game-free test was run again on the new base (fails at the test-only commit, passes at the tip, two rounds, g++ and clang; the 43 checks that read the changed files pass); the Field numbers still hold because main changed vk_memory.cpp since 9c9d69ed only in how the mirror budget is set (#735), not in memory_guest_imported(). T-S2R3 (Windows) still waits for the testing wanted label.

Rebased onto `main` aa95d35f with no conflict (the same four changes, `git patch-id`; commits now under the noreply identity) and the description moved to the current template. The game-free test was run again on the new base (fails at the test-only commit, passes at the tip, two rounds, g++ and clang; the 43 checks that read the changed files pass); the Field numbers still hold because `main` changed `vk_memory.cpp` since 9c9d69ed only in how the mirror budget is set (#735), not in `memory_guest_imported()`. T-S2R3 (Windows) still waits for the `testing wanted` label.

zeldaootr force-pushed import-exclusive-end from 3046b83159 to afd5ea15aa 2026-11-15 16:43:11 +00:00 Compare

HumanBean referenced this pull request from a commit 2026-11-15 18:42:23 +00:00

docs(tested): measure partially covered guest frees on Deck (#471)

HumanBean referenced this pull request 2026-11-15 19:04:54 +00:00

fix(vk): retain guest pages during import handoff and retirement #468

HumanBean approved these changes 2026-11-15 19:05:01 +00:00

Dismissed

HumanBean left a comment

• Copy link
• Copy Markdown

Code review only (nothing run). The change is correct. imported is keyed by lo, so with upper_bound(hi) the backwards walk started at a pool whose lo == hi, and it->second.hi > lo then counted a pool that only touches the freed range. lower_bound(hi) starts at the last pool with lo < hi, which is the half-open answer, and the 1 GB early-break is unaffected. The empty-range early return is fine. The pools in retired_pools and evicted_pools already use strict < against hi, so only this loop had the bug.

Fitting with #468: the two are complementary, since 468 adds the pending-import sources (all half-open) and this fixes the imported loop. They touch the same lines, so the second one to merge needs a rebase. Both add the same if (hi <= lo) return 0;, and this PR rewrites the neighbouring upper_bound line. Each PR's test also builds the same fixture (test_guest_imported.py here, test_pending_guest_import.py there). They are compatible, since the fixtures define the same bg_* and dead_imports shapes, and once both land the touching-boundary cases are covered twice. That is harmless.

Approving. The description matches the diff and is honest that the fix shows no measurable change on the Deck. The test fails on main and passes here.

Code review only (nothing run). The change is correct. `imported` is keyed by `lo`, so with `upper_bound(hi)` the backwards walk started at a pool whose `lo == hi`, and `it->second.hi > lo` then counted a pool that only touches the freed range. `lower_bound(hi)` starts at the last pool with `lo < hi`, which is the half-open answer, and the 1 GB early-break is unaffected. The empty-range early return is fine. The pools in `retired_pools` and `evicted_pools` already use strict `<` against `hi`, so only this loop had the bug. Fitting with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/468: the two are complementary, since 468 adds the pending-import sources (all half-open) and this fixes the `imported` loop. They touch the same lines, so the second one to merge needs a rebase. Both add the same `if (hi <= lo) return 0;`, and this PR rewrites the neighbouring `upper_bound` line. Each PR's test also builds the same fixture (`test_guest_imported.py` here, `test_pending_guest_import.py` there). They are compatible, since the fixtures define the same `bg_*` and `dead_imports` shapes, and once both land the touching-boundary cases are covered twice. That is harmless. Approving. The description matches the diff and is honest that the fix shows no measurable change on the Deck. The test fails on main and passes here.

zeldaootr force-pushed import-exclusive-end from afd5ea15aa to 33ca425e46 2026-11-15 19:14:30 +00:00 Compare

zeldaootr commented 2026-11-15 19:14:30 +00:00

Author

Contributor

• Copy link
• Copy Markdown

Rebased onto main 94abdd57 (head 33ca425e), with no conflict; the commit subjects are now Conventional Commits, which CircleCI's lint job checks for every commit of a pull request (#417), and the base has the CI's config, which a fork's pull request needs at its own head. No other change: each commit's change is the same as before by git patch-id.
Re-run on 94abdd57: test_guest_imported fails at the test-only commit 351d28ba and passes at the head, with g++ 13.3 and clang 23.1.5; of the 46 checks that are this PR's or read its files (g++ 13.3), 46 pass. The description names the new commits.

Rebased onto `main` 94abdd57 (head 33ca425e), with no conflict; the commit subjects are now Conventional Commits, which CircleCI's lint job checks for every commit of a pull request (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/417), and the base has the CI's config, which a fork's pull request needs at its own head. No other change: each commit's change is the same as before by `git patch-id`. Re-run on 94abdd57: `test_guest_imported` fails at the test-only commit 351d28ba and passes at the head, with g++ 13.3 and clang 23.1.5; of the 46 checks that are this PR's or read its files (g++ 13.3), 46 pass. The description names the new commits.

zeldaootr dismissed HumanBean's review 2026-11-15 19:14:30 +00:00

Reason:

New commits pushed, approval review dismissed automatically according to repository settings

Some checks are pending

Hide checks

PR description / Description follows the template (pull_request_target) Waiting to run

Details

CI / Conventional Commit title (pull_request) Blocked by required conditions

Details

CI / Python source checks (ubuntu-24.04, part 1/2) (pull_request) Blocked by required conditions

Details

CI / Python source checks (ubuntu-24.04, part 2/2) (pull_request) Blocked by required conditions

Details

CI / SPIR-V fixtures (global narrow load/store, direct surface atomics, F2I immediate) (pull_request) Blocked by required conditions

Details

CI / Workflow lint (actionlint) (pull_request) Blocked by required conditions

Details

Guide screenshots / Guide images and local UI source review (pull_request) Blocked by required conditions

Details

Third-party parts / As their notes say (pull_request) Blocked by required conditions

Details

Third-party parts / The same as upstream (pull_request) Blocked by required conditions

Details

Website / Website on Cloudflare (pull_request) Blocked by required conditions

Details

This pull request doesn't have enough approvals yet. 0 of 1 approvals granted.

Some workflows are waiting to be reviewed.

This branch is out-of-date with the base branch

You are not authorized to merge this pull request.

View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.

git fetch -u 'import-exclusive-end:zeldaootr-import-exclusive-end'

git switch 'zeldaootr-import-exclusive-end'

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels   

area/accessibility

Barriers for people with disabilities

  

area/audio

Sound and music

  

area/input

Controllers, keyboard, mouse

  

area/install

Installing, updating, building from your game

  

area/menu

In-game menu and settings

  

area/renderer

Vulkan renderer, shaders, effects

  

area/saves

Save files

  

area/upscaling

DLSS, FSR, frame generation

  

beanstalk: approved @ 1697603

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 2c5a62e

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 4fd82b7

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 64d6d09

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 6f91fae

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 70a78f9

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 8015c9a

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 8d361dc

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ 9d8356e

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ abf2227

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ d82f202

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ e5f1e40

Set by Beanstalk for show; the farm's database is what counts

  

beanstalk: approved @ ee8ab5f

Set by Beanstalk for show; the farm's database is what counts

  

ci: approved

A maintainer approved CircleCI for this fork PR (sticky; removed after 14 quiet days)

  

good first issue

A good start for a new contributor

  

gpu/amd

AMD GPUs

  

gpu/igpu

Integrated GPUs

  

gpu/intel

Intel GPUs

  

gpu/nvidia

NVIDIA GPUs

  

help wanted

Extra hands or knowledge wanted

  

kind

bug

Something doesn't work as it should

  

kind

chore

Build, CI, dependencies, tidying

  

kind

docs

Documentation

  

kind

feature

A new feature or an improvement

  

kind

performance

Frame rate, frame time, loading, memory

  

kind

question

A question rather than a change

  

needs measurements

Waiting for before/after measurements

  

platform/linux

Affects Linux (a change for every system carries both Windows and Linux)

  

platform/macos

macOS (Apple silicon or Intel Macs)

  

platform/steam-deck

Steam Deck

  

platform/windows

Affects Windows (a change for every system carries both Windows and Linux)

  

preview

Maintainers: deploy a preview of this PR's website (read its changes first). Removed after the run.

  

priority

critical

Crashes, data loss, or blocks most players

  

priority

high

Important: next in line

  

priority

low

Nice to have

  

regression

Worked in an earlier build, broken in a newer one: highest priority, above other bugs

  

status

blocked

Waits on something else first

  

status

confirmed

Reproduced or agreed on

  

status

duplicate

Already reported or proposed

  

status

invalid

Not a problem in this project

  

status

needs-info

Waiting for more information from the reporter

  

status

wontfix

Will not be worked on

  

testing wanted

Wanted: a test on other hardware (posts a test request in #test-requests)

  

ci: approved

No labels

area/accessibility

area/audio

area/input

area/install

area/menu

area/renderer

area/saves

area/upscaling

beanstalk: approved @ 1697603

beanstalk: approved @ 2c5a62e

beanstalk: approved @ 4fd82b7

beanstalk: approved @ 64d6d09

beanstalk: approved @ 6f91fae

beanstalk: approved @ 70a78f9

beanstalk: approved @ 8015c9a

beanstalk: approved @ 8d361dc

beanstalk: approved @ 9d8356e

beanstalk: approved @ abf2227

beanstalk: approved @ d82f202

beanstalk: approved @ e5f1e40

beanstalk: approved @ ee8ab5f

ci: approved

good first issue

gpu/amd

gpu/igpu

gpu/intel

gpu/nvidia

help wanted

kind

bug

kind

chore

kind

docs

kind

feature

kind

performance

kind

question

needs measurements

platform/linux

platform/macos

platform/steam-deck

platform/windows

preview

priority

critical

priority

high

priority

low

regression

status

blocked

status

confirmed

status

duplicate

status

invalid

status

needs-info

status

wontfix

testing wanted

ci: approved

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

MagicBeanDev/Longshot!469

No description provided.