Commit Graph
381 Commits
Author SHA1 Message Date
swung0x48 112c1d8dfc [MG_Test] (Disaggregated): P5e vi - the two source decisions, the window rule and the client-array lane
- SanityTest's DirectGLESVertexInputDraw pair drives the PRODUCTION decisions
  (ResolveDrawVertexBuffersFromRecord / ResolveDrawIndexBufferFromRecord) with the frontend
  VAO deliberately pointing at a buffer the record does not name - the state a run-ahead
  client makes ordinary and one process cannot otherwise produce. Both functions take the
  applier state and nothing else, so the only revert is a frontend read inside them and
  the cases then go red naming the field (ID-102's rule about what a red-once may test).
  G2/G14 skip twins in the pull arm.
- VertexInputEmit pins ID-95 / ruling 19 on the producer side: set_vertex_buffers' window
  must cover every attribute create_vertex_elements declares ENABLED, with a sparse
  configuration so "covers the enabled set" is not "is as long as the array". The sink half
  is tx2's CheckUnitWindows.
- ClientVertexArrayScenario is BRIEF-P5E's open item (a), answered: NO existing scenario
  exercises a client-memory vertex array under split (DoublePrecisionScenario uses the
  binding-model calls against real buffers), so the lane had no entry that could reach the
  refusal at all. Registered in the split lane with a private log path, and it is also the
  positive control for keeping the upload arm alive under lockstep.
2026-09-18 08:10:26 -04:00
swung0x48 1f8de61be1 [MG_Remote, MG_Pipe, MG_Impl, MG_State, MG_Backend, MG_Test, MG_IntegrationTest] (Disaggregated): P5d round 3 - the profile-directed cuts on both roles
- The round-2 report guessed the apply thread's ~48% mutex time was the backend
  pendingMutex; the symbolized call graph at 56a77348 says it is the idle poll in
  ServerLoop::ApplyThreadMain taking m_controlMutex twice per iteration
  (ControlIsPending in the ready lambda, PumpControlRequest), with DrainRing at
  29% of the thread. Every package below is cut from that graph, not from self%.
- T (transport): the control mailbox gets an atomic shadow bit m_controlPosted so
  the idle poll takes no lock (handshake, C2 critical section and the
  NOT_INITIALIZED answer unchanged); Doorbell::Wait spins on a once-calibrated
  iteration budget and reads no clock in steady state (the park phase alone
  computes deadlines); client/server wait and park counts ride the stats line
  as wait[srv= srvpark= cli= clipark=] and the wire ledger as cliwait=/clipark=.
- D (role guards): OnApplyThread() is inline - one relaxed atomic key against the
  caller's thread pointer (aarch64 __builtin_thread_pointer, else
  std::this_thread::get_id) - and m_applyThreadId, the function-static guard and
  the pthread_equal call are gone; the three scope depth counters stop being
  thread_local (their only reader is the apply-thread guard, so they count only
  when OnApplyThread() is true; a GL-thread scope no longer exempts the apply
  thread, pinned by a new case); the guard chains are ordered cheapest-first.
- B (persistent maps): the shadow allocator rounds the SIZE to a page multiple, so
  TrackWriteMap aligns OUTWARD when the shadow is page-granular - a sub-page map
  is one protected page, edges and their per-draw XXH3 disappear; the inward arm
  stays as the fallback. Review found a hole the edge hashes had hidden since
  round 1: a map that faulted but was not consumed by the walk that advanced the
  epoch was never pushed. The walk now drains every marked member before the
  binding walk, registration bumps the epoch so a fresh map ships once on the
  draw path, and a non-4 KB kernel page refuses the mprotect arm by name.
- C (client per verb): an image-unit high-water mark (push build only, fed by
  glBindImageTexture) replaces the 192-unit sweep per draw; the residual fill's
  7-term supplied conjunction is a memo keyed on its four inputs; the server's
  63-field stamp per verb is a constexpr per-class mask table.
- generate_mipmap no longer skips its wait under MOBILEGL_IPC_BATCH_WAITS: its sink
  is the backend GenerateMipmap, which reads MGB_CTX->GetActiveTextureUnit(), and
  a batched publish let the next validate move the serial under it -
  Fatal{UnmigratedPipeInput, "GetActiveTextureUnit@ReadPixels"} on the gate tree.
  The rule is written at the site: a record may skip its wait only if its apply
  reads nothing the residual fill writes. The texture-death scenario gets the
  framebuffer case's fence before it reads the object_death tally.
- Redmi 2f7cbe2e, FCL + Minecraft 26.3-rc-3, VD12, CPU pinned: inproc 64.3 fps
  (client 15.0 ms/frame, apply 12.4) -> 103-114 fps (client 8.7-9.2 ms/frame on a
  mid core, apply 7.0-7.4); monolith 115 (120 Hz vsync-capped runs) / 206 (one
  uncapped run, 5.0 ms/frame on the big core). Gates: unit 2203/2203,
  integration-split 111/111 (the two formerly flaky cases 6/6 on repeat),
  generators, include closure, dirty surface and doc citations green.
2026-09-18 04:28:25 -04:00
swung0x48 cb06538c06 [MG_Remote, MG_Pipe, MG_Impl, MG_Test, MG_IntegrationTest, Docs] (Disaggregated): inproc performance package - dirty-driven persistent-map push, batched barrier waits, spin-clock batching
VD12 Minecraft on the Redmi (FCL, DirectGLES): the split's in-world frame rate
13 -> 60-66 fps against monolith's 110-114, and persistent-map wire bytes
25 MB/frame -> ~0.23 MB/frame. The residual split-specific costs are named at
the end, not hidden: the R-1 serialization (WaitForApplied ~12%, P11's
versioned-inputs design) and a split-only mutex complex still under
investigation.

Persistent-map push, three arms picked per buffer at map time:
- mprotect dirty tracker: the mapped range's page-aligned interior is
  protected PROT_READ; the app's next write faults, the SIGSEGV handler marks
  the page in the slot's bitmap and unprotects, and the push ships one 64 KB
  block per block containing a faulted page. Ownership is three tests -
  SEGV_ACCERR only (a userspace raise and an unmapped page chain verbatim),
  inside a live tracked range, and not a refault on a still-set bit (a write
  we unprotected succeeds, so that is a non-write access and it crashes
  honestly). The alignment is INWARD: an outward-aligned edge page holds
  foreign bytes, and a signal-blocking worker thread faulting on one is a
  process kill - that was the ctest-only SmallRing segfault. Edge spans
  (at most two per map) are hashed per push and ship their containing block
  only on change (unconditional edges measured 112 MB/frame). Re-arm is one
  mprotect per contiguous run. The first push ships everything once.
- hash-suppress arm (MOBILEGL_IPC_PERSISTENT_HASH_SUPPRESS, default 1):
  per-block XXH3 against the last push's hashes; untracked buffers only.
- PushDrawConsumers: only the buffers the upcoming draw/dispatch can read
  (VAO attribs/element, indexed binding points, indirect/parameter slots,
  TexBuffer backing - the last also retires the per-frame-cadence cloud
  twitch), plus the fault-epoch skip: an unmoved epoch proves no interior
  byte changed, tracked edges are hashed in place, and any hash-arm member
  vetoes. The membership map caches the tracker slot pointer (a 64-slot
  rescan per buffer per draw was measured and removed).

MOBILEGL_IPC_BATCH_WAITS (default 1; MOBILEGL_PIPE_VERIFY forces 0): a
value-class record (kCtxState/kCtxCso/kCtxObject with no reply slot) is
published without waiting for its own apply; the barrier is taken at the next
pull-reading verb, the only place BARRIER-PULLED fields are read, so the R-1
fence posts are untouched and only the round trips between them change. Two
consequences are pinned rather than argued: kCtxObject verbs the tests read
synchronously get explicit fences (GenerateMipmapFieldsCross,
FramebufferDeathCrosses), and a frontend object whose last SharedPtr drops on
the apply thread - impossible before batching - now enqueues its
(kind, lifetimeId) on the deferred destroy queue and the GL thread replays
the DestroyAndFree helper at the verb hooks (CopyImageLayeredScenario aborted
Fatal{RoleViolation, MGPipeSlots} without it).

Doorbell::Wait's spin reads the clock once per 64 yields: on the workload
device a steady_clock read is a real syscall (no vDSO) and the per-iteration
reads were ~20% of the whole process at ~1000 waits/frame.

Gates: gen_pipe --check/--self-test green; unit 2188/2188;
integration-split 110/111, the one red being GenerateMipmapDepthPixels whose
DirectGLES and DirectVulkan MONOLITH siblings flake identically with every
code path of this package provably inert (PushIsArmed false, no ClientSession,
deferral's monolith early-out).
2026-09-18 01:58:59 -04:00
swung0x48 b88e848761 [MG_Remote, MG_State, MG_Impl, MG_Backend, MG_IntegrationTest] (Disaggregated): P5c census triage - the three real regressions the merge owed
The broad inproc census against a same-machine 11ac3de6 baseline found 183 newly-failing;
172 are the design-red the contract rules (the legacy-arm refusal under transport: the
subsystem-off and legacy control lanes, and Magma's P7 surface - every one carries its
Fatal name in ~/p5c-fatal-map.tsv), and the other eleven were three REAL regressions,
fixed here:

1. The surface-changed event drained at the first verb's EmitAndWait, but a pre-verb
   glGetFramebufferAttachmentParameteriv on the default framebuffer was answered from the
   placeholder attachments, and buffers allocated from that answer are blit-incompatible
   with the real surface (DepthStencilReadbackAttachmentShapeScenario). The event ring
   gains its SECOND drain-point class: the blocking EGL lifecycle RPC returns
   (BackendObject_Remote's surface create/resize and make-current) - the same known-idle
   premise as EmitAndWait's post-barrier point. CONTRACT-P5C section 4.1 amended.
2. A whole-buffer writeback larger than the event ring could never fit one record
   (capacity/2), aborting the server with Fatal{EventRingOverflow} on the large-arena
   readbacks. The client now slices the request at a quarter of the ring (header and
   co-posted small events ride along), each slice round-tripping its own barrier + drain;
   the whole-buffer flag is cleared by hand after the last slice lands, which is exactly
   what WritebackFromBackend's clear says. CONTRACT-P5C section 4.4 amended.
3. ScatterCapturedRecords / ReadbackCapturedRanges read the capture buffers through the
   frontend MappedData() (a layer-1 surface since gt) and wrote back by direct call. The
   XFB capture path now reads the server's staged shadow (RequireStagedCoverage asserted,
   syncedChangeSerial bumped so no later draw overwrites the capture with stale shadow)
   and writes back through OnBufferWriteback, the Ops_H_Readback order.

CtWireScenario's two death cases skip on DirectVulkan by name: object_death's producer is
the Espryt-side death-notice ops and Magma installs none until P7.

Evidence: unit 2187/2187; integration-split 111/111; the three families' named cases green
again; census newly-failing reduced to the evidenced design-red set.
2026-09-17 13:13:41 -04:00
swung0x48 a7e55ee62a [MG_Pipe, MG_Remote, MG_Backend, MG_Impl, MG_IntegrationTest] (Disaggregated): P5c ct - the applier_reset and object_death control records
Two appended wire records replace the two cross-role control bypasses:
applier_reset (opcode 77, MGPApplierReset{Uint64 ContextSerial}, kScreen) is emitted by
the GL thread at the FreshlyPrimed make-current edge, before the client-side resets, so
the record's barrier orders the server's MGPipeApplierReset against every verb that
follows; a direct MGPipeApplierReset() from a non-apply thread with a live session is
Fatal{RoleViolation, "g_applier"} (MGPipeApply.cpp carries the layer-2 guard; the
apply-thread and no-session bring-up shapes keep the direct call verbatim).
object_death (opcode 78, reuses MGPipeHandleOnly, kCtxObject) is the framebuffer
family's first wire delete opcode: the client thread resolves the dying object's handle
in its OWN allocator, emits nothing for a handle the server never saw, and EmitAndWaits
otherwise - the wait preserving the death's ordering against in-flight records that
name the handle, the only property the blocking mailbox hop provided. The sink releases
per-kind twins through a handle-keyed ReleaseTwinByHandle (seven kinds, SamplerViewCso's
idempotent second path kept); a null arriving handle is
Fatal{ProtocolCorruption, "ObjectDeath.Handle"}. A documented NoSession fallback arm
delivers a death to a server-only fixture's loop; monolith keeps the mailbox hop and
the FreshlyPrimed direct call character for character.

Evidence: gen_pipe --check/--self-test and gen_pipe_field_ownership --check/--self-test
green; unit 2155/2155; integration-split 111/111 (107 plus four new CtWireScenario
entries, all genuinely run, each with its own log path and an EXPECT_EXIT death case);
red-once - shorting the emission to the direct call turns the four Ct cases red with
Fatal{RoleViolation, "g_applier"} by name, then restores green verbatim.
CONTRACT-P5C sections 1, 5.1-5.2, 7.
2026-09-17 06:08:54 -04:00
swung0x48 a79a0af69d [Docs] (Disaggregated): consolidate the five documents at the P5b close - drop superseded narrative, the -O0 device tables and process history, renumber MEASUREMENTS by phase, correct the catalogue and emit-table counts to the head, and record the Redmi pin profile 2026-09-16 23:08:39 -04:00
swung0x48 348d22a4b9 [Fix] (DirectGLES): verify generated mip storage from pushed descriptors 2026-09-16 13:56:55 -04:00
swung0x48 f018c0ef05 [Feat] (MG_Remote, P5b sync): migrate fence lifecycle and wait replies to the apply thread 2026-09-16 13:46:49 -04:00
swung0x48 56c6e49d1d [Feat] (MG_Remote): lower named framebuffer blits through scoped wire bindings 2026-09-16 13:44:35 -04:00
swung0x48 12905a7eb8 [Merge] (P5b): land r1 core fixes with multi-tail command-ring retry 2026-09-16 13:41:40 -04:00
swung0x48 db8b347952 [Fix] (Remote): preserve server ownership and wait for ring retirement 2026-09-16 13:39:50 -04:00
swung0x48 9606466ab6 [Merge] (MG_Remote, P5b): combine d1 draw migration with f1 i1 t2 and r2 2026-09-16 13:30:01 -04:00
swung0x48 5443bced9a [Merge] (P5b): integrate t2 and reconcile all migrated emitter families 2026-09-16 13:27:15 -04:00
swung0x48 2deca994b2 [Merge] (P5b): integrate i1 alongside f1 and r2, preserving the combined slot partition 2026-09-16 13:25:44 -04:00
swung0x48 a523d6195a Merge branch 'p5b/r2' into p5b/integrate-codex 2026-09-16 13:24:41 -04:00
swung0x48 d50183cb1a [Fix] (IntegrationTest): assert actual staging backpressure in the small ring lane 2026-09-16 12:34:33 -04:00
swung0x48 6fa6a925bd [Fix] (CI): make split controls mandatory and isolate runner evidence 2026-09-16 12:34:33 -04:00
swung0x48 fc5b52e456 [Merge] (P5b, t2): merge feat/disaggregated at 37fc4fdb (x2 via x2m, ID-65) into p5b/t2 2026-09-16 12:24:02 -04:00
swung0x48 c831ab0390 [Test] (MG_Test, MG_IntegrationTest, P5b/t2): pin the span order and the accounting fields, the six flipped slots by name, and run the tessellation-capture and buffer-reuse scenarios under inproc 2026-09-16 12:23:57 -04:00
swung0x48 fbe17d06d6 [Test] (MG_IntegrationTest, P5b d1): keep GuiBatchScenario out of the d1 split lists - past its draws it stops at Fatal{UnmigratedVerb, "MemoryBarrier"}, which is i1's slot to flip and register 2026-09-16 12:23:35 -04:00
swung0x48 e564939615 [Feat] (MG_Remote, P5b d1): flip the nineteen indexed / instanced / multi-draw / indirect draw slots to class B on draw_vbo - the emitters plan the record from the GL call and the bound element / indirect / parameter buffers, stage a client index array as the kDrawHasUserIndices span, refuse +CLIENT_INDICES / +CLIENT_COMMANDS / +INDEX_OFFSET by name, and the sink dispatches every shape to the backend slot the monolith calls (a base vertex of 0 is DrawElements, never DrawElementsBaseVertex with 0: ANGLE has no glDrawElementsBaseVertex) 2026-09-16 12:10:38 -04:00
swung0x48 e352e9a6c9 [Merge] (P5b, i1): merge feat/disaggregated at 37fc4fdb (x2 via x2m, ID-65) into p5b/i1 2026-09-16 12:10:25 -04:00
swung0x48 a72a96d85f [Test] (MG_Test, MG_IntegrationTest, P5b): pin i1 seven class-B slots and the served compute program, and arm 24 inproc lane entries whose results only exist if the records crossed 2026-09-16 12:10:09 -04:00
swung0x48 fe5ac006f8 [Merge] (Remote): bring f1 up to the integrated P5 wire ledger 2026-09-16 11:50:00 -04:00
swung0x48 7ffa4998ab [Feat] (Remote): migrate f1 clear copy and mipmap verbs under inproc 2026-09-16 11:49:41 -04:00
swung0x48 bd11ab189f [Merge] (MGPipe, P5): land x2 - the E2 draw-drop control, R-10 maxrec and the small-ring wrap counters, resolved against j0 (ID-65) 2026-09-16 11:30:12 -04:00
swung0x48 579118a1b3 [Test, CI] (scripts/ci, workflows): E2's draw-drop control requires the SSIM red AND the library's dropped-record count, E3(a) requires its private-log line as well as the pixels, and the smoke red-check finally perturbs the split control (ID-65) 2026-09-16 11:09:21 -04:00
swung0x48 a7b43cbbcd [Test] (MG_IntegrationTest): exit gate E3(e) drives the small ring past its own capacity and asserts the wrap, and both split arms of TriangleScenario publish R-10's maximum record bytes against the ring's cap (ID-65) 2026-09-16 11:09:15 -04:00
swung0x48 cc20de37da [Fix] (IntegrationTest): reject skipped split control selections 2026-09-16 09:31:39 -04:00
swung0x48 68d8ef9f34 [Fix] (IntegrationTest): gate private split logs and repair control red-checks 2026-09-16 09:23:27 -04:00
swung0x48 b2dcae713f [Fix] (IntegrationTest): give Split entries private diagnostic logs 2026-09-16 08:48:30 -04:00
swung0x48 6cabc6b070 [Fix, Test] (MG_IntegrationTest, MG_Test): make the persistent-map membership expectation a function of the arm the map landed in, and pin the arm-independent half on IsLivePersistentMap where both answers can be produced 2026-09-16 06:03:52 -04:00
swung0x48 8b425a9718 [Test] (MG_IntegrationTest): arm the split lanes from the running process - MG_Config::Transport, ClientSession::Active() and ImplementedVerbCount() - instead of from a grep over c0's stub messages, which a one-word rename disarmed into eight green monolith runs, and require the encoder's record ordinal to move before any split case may pass 2026-09-11 15:30:17 -04:00
swung0x48 7ad0ff3c33 [Test] (MG_IntegrationTest): assert which persistent-map arm the scenario landed in through a peek at IsBackendPersistentMapped, not by inferring it from the push byte counter - "the push was never wired" and "the store was adopted" are different defects with different owners and the counter reads the same for both 2026-09-11 14:40:30 -04:00
swung0x48 eafa47ea46 [Test] (MG_IntegrationTest): add the Triangle and PersistentCoherentMap scenarios and the DirectGLES.Split. lanes - the split family skips naming c0's remaining signature stubs rather than going green against the monolith path it would otherwise run 2026-09-11 14:00:14 -04:00
swung0x48 03082b8367 [Fix, Test] (clientsp, MG_Pipe, Espryt): produce kMGPipeBindSampler and kMGPipeBindShaderImage where D-A4 places them - nothing set either bit, so ImageBindableHint was always 0, the metadata respecify had no live trigger and the remint pull the hint prevents was neither prevented nor counted (final review M-A); the sampler-view resolution notes SAMPLER, glBindImageTexture's state setter notes SHADER_IMAGE at the bind (so the hint precedes the first sync) and the image walk notes it too, both through a contract door since neither may include TextureEmit.h, and Espryt counts every re-mint of storage it already held as tex-remint-pulls (trp= on the stats line, ROADMAP open question 2's number) 2026-09-09 00:09:30 -04:00
swung0x48 a7d56a2fe5 [Fix, Test] (MG_Pipe, clientfb, clientsp): retire every emitter's entry at the object's death - the six death helpers freed the slot and told no emitter, so a dead-but-unrecycled texture handle still resolved to the freed ITextureObject* (the allocator's generation moves only at the next hand-out) and the drain list kept the level: glTexImage2D; glDeleteTextures; <any verb> called a virtual on freed memory from the next validate point (final review C-2); the helpers now forward to the texture, renderbuffer, framebuffer, sampler-view and shader-CSO emitters between the wire delete and the free, ResolveTexture refuses a dead slot loudly on IsLive, the sticky-mask producers stamp the generation they write under, and the delete-then-use sequence is pinned for every kind, for a recycled slot, and under MALLOC_PERTURB_ on both backends 2026-09-08 23:59:48 -04:00
swung0x48 4bb5e082d3 [Fix, Test] (clientfb, MG_Pipe): pass the level a texture respecify redefines - every per-level glTexImage*D and glGenerateMipmap grow took the applier's whole-resource arm and dropped the texture's every pending upload, including a level the applier had accepted at an earlier verb with the client's dirty flag already clear, so L0; draw(other); L1; draw(T) read a black level 0 on the handle arm (final review C-1); the storage entry points now state the scope (one level, a chain cut, the whole resource), the emitter builds wire's MGPRespecifiedLevel with the drain's packed target, a per-level call is never deduped on the descriptor, and the applier drops a named level whether or not the descriptor moved 2026-09-08 23:55:44 -04:00
swung0x48 a25b997380 [Fix] (Espryt, Integration): compile the in-place storage-redefinition generation bump under MOBILEGL_PIPE_PUSH - it is Espryt code the pull build shares and G1 keeps the pull library byte-identical to the P4a baseline (0/0/2/0 on 419f9941), so the pull build keeps the pre-P4a hole until the bump lands on dev on its own and the F-3 texture cases decline by name there 2026-09-08 21:11:11 -04:00
swung0x48 ceb190d1c9 [Fix] (Espryt): re-sync a framebuffer's cross-object masks after an attachment's storage is redefined in place - a mutable texture regenerated on its driver id or a renderbuffer re-storaged on its moved neither the FBO memo's frontend versions nor the backend-id generation, so the pre-handle arm kept the widening masks of the storage the object was attached with; the renderbuffer half of the F-3 scenario asserts on the handle arm only, where the resource record closes D-D2 (found by the F-3 scenarios on the 0x1ff, 0 and pull lanes) 2026-09-08 20:50:25 -04:00
swung0x48 e9133d9e4e [Fix, Test] (Espryt): drive a unit's sampler through the CSO's own twin on the handle arm - the record arm looked a content-addressed handle up in the identity-keyed registry and could never hit, so only the pre-handle program pass ever put a glBindSampler'd object on the driver (P4a seam F-4) 2026-09-08 20:29:24 -04:00
swung0x48 0b1a540e91 [Fix, Test] (Pipe): mix the program identity into the shader-image shutter's plain arm - the per-program counter is equal across programs and against no program, so a switch never moved the image window and a buffer image never reached set_shader_images (P4a seam F-2, esprytdraw SD-4) 2026-09-08 20:28:56 -04:00
swung0x48 0f82c9d757 [Fix, Test] (Pipe): mix the program in use and the params aggregate into the sampler-view shutter - set_sampler_views is resolved for the current program and a glUseProgram alone never re-emitted it, so a texture bound to an empty slot under one program was never synced for the next (P4a seam F-1 / F-1b) 2026-09-08 20:28:29 -04:00
swung0x48 54566e806f [Fix, Test] (MG_State, Integration): move the framebuffer aggregate from a texture's or renderbuffer's storage definition - set_framebuffer_state inlines an attachment's format and a respecify while attached left Espryt's handle arm answering its four cross-object masks from the stale copy (P4a seam F-3) 2026-09-08 20:15:55 -04:00
swung0x48 c397b99d0d [Fix] (Test): warm the whole content space before the SamplerCso leak baseline, so the case measures the death path and not the content-addressed cache filling up 2026-09-08 19:08:12 -04:00
swung0x48 82484f615d [Test] (Pipe): pin that the applier's no-consumer belt never fires behind the client's gate, on the backend that has a consumer and on the one that has none 2026-09-08 19:08:12 -04:00
swung0x48 1ea5d472e8 [Test] (Espryt): take G9's reading while the texture is still attachment-only - the applier's params record, Espryt's applied value and the absence of a sampler view, before the sample that would repair all three 2026-09-08 19:08:12 -04:00
swung0x48 afcee801cc [Fix] (Test): read the DIRECTION of the subsystem refusal, so the two dependency lanes stop accepting each other's sentence 2026-09-08 19:08:12 -04:00
swung0x48 3049dc1d53 [Fix, Test] (Pipe): read the shader composite band's own counters in the composite leak case, arm the ABA flip from anywhere under a backend, match Espryt's exact refusal line and pin the bit-10-requires-bit-11 arm 2026-09-08 19:08:12 -04:00
swung0x48 ab4384a007 [Test] (Pipe): switch P4a's four object subsystems off against the shipping mask and record the texture upload shape the two counters make visible 2026-09-08 19:08:12 -04:00