Commit Graph
4490 Commits
Author SHA1 Message Date
swung0x48 a6144426d0 [MG_Pipe] (Disaggregated): P5f fm - correct texture staging record type 2026-09-19 22:45:29 -04:00
swung0x48 a204b357e7 [MG_Pipe] (Disaggregated): P5f fm - correct texture staging record type 2026-09-19 22:45:02 -04:00
swung0x48 c747fa1fc2 [MG_Backend] (Disaggregated): P5f fm - preserve mipmap view windows and array extents 2026-09-19 22:44:39 -04:00
swung0x48 3e1f1b71f3 Merge commit '01e00bc4' into codex/p5f-wave2 2026-09-19 22:43:51 -04:00
swung0x48 2e3088e6c4 [MG_Pipe] (Disaggregated): P5f fm - retain named null-data mip definitions 2026-09-19 22:43:44 -04:00
swung0x48 2c9b9d2a9f [MG_Backend] (Disaggregated): P5f fm - retire draw attachments after GPU completion 2026-09-19 22:43:35 -04:00
swung0x48 01e00bc495 [MG_Test] (Disaggregated): P5f fm - verify hookless texture staging lifecycle 2026-09-19 22:43:07 -04:00
swung0x48 777f706ca2 [MG_Pipe] (Disaggregated): P5f fm - retain texture staging without buffer backend ops 2026-09-19 22:42:07 -04:00
swung0x48 29ede72346 Merge commit '6a138a52' into codex/p5f-wave2 2026-09-19 22:41:57 -04:00
swung0x48 2d507d5bd4 [MG_IntegrationTest] (Disaggregated): P5f fm - cover bufferless draw and layered copies under role split 2026-09-19 22:41:47 -04:00
swung0x48 6a138a5221 [MG_Backend] (Disaggregated): P5f fm - fence remaining buffer debt and preserve native transfer layouts 2026-09-19 22:40:04 -04:00
swung0x48 d4ae45d6b8 Merge branch 'p5f/fm' of /mnt/c/Users/geekerwan/AndroidStudioProjects/FoldCraftLauncher/.git/modules/MobileGL into codex/p5f-wave2
# Conflicts:
#	MobileGL/MG_Pipe/PipeApply.h
#	MobileGL/MG_Remote/Server/PipeApplier.cpp
2026-09-19 22:36:56 -04:00
swung0x48 eb882ebf18 Merge commit '203f59cc' into codex/p5f-wave2 2026-09-19 22:34:51 -04:00
swung0x48 fec49a54d8 [MG_Backend, MG_IntegrationTest] (Disaggregated): P5f fm - apply bufferless draws and compute from server records 2026-09-19 22:34:28 -04:00
swung0x48 a6223c6509 [MG_Backend] (Disaggregated): P5f fm - order wire image writes and preserve implicit alpha 2026-09-19 22:34:04 -04:00
swung0x48 d9fc45bf4c [MG_Backend] (Disaggregated): P5f fm - use shared image access constant 2026-09-19 22:32:57 -04:00
swung0x48 f37380d0dc [MG_Backend] (Disaggregated): P5f fm - resolve uniform descriptors from server records 2026-09-19 22:32:10 -04:00
swung0x48 2abc561721 p5f fm: cover GPU texture writes and view subresources 2026-09-19 22:29:36 -04:00
swung0x48 277d1a07f0 [MG_Test] (Disaggregated): P5f fm - pin server program overlays and identities 2026-09-19 22:29:33 -04:00
swung0x48 867c5c5efd [MG_Backend] (Disaggregated): P5f fm - expose archive stage and capture-mode queries 2026-09-19 22:27:51 -04:00
swung0x48 31795675f1 p5f fm: build Vulkan samplers from server-owned values 2026-09-19 22:27:43 -04:00
swung0x48 62bf26bf8c [MG_Pipe] (Disaggregated): P5f fm - admit registered server object consumers independently of buffer ops 2026-09-19 22:27:03 -04:00
swung0x48 6ae5944831 [MG_Backend] (Disaggregated): P5f fm - build Magma programs from server archive views 2026-09-19 22:26:23 -04:00
swung0x48 a6283f40f2 [MG_Backend, MG_Impl] (Disaggregated): P5f fm - map texture view attachments through wire identities 2026-09-19 22:23:05 -04:00
swung0x48 203f59ccad [Docs, MG_Backend] (Disaggregated): P5f fe - report carrier retirement and verified gates 2026-09-19 22:20:57 -04:00
swung0x48 4b3afe7840 [MG_Backend] (Disaggregated): P5f fm - retain blit scissor clipping across wire arm 2026-09-19 22:20:53 -04:00
swung0x48 911cc74a57 [MG_Backend, MG_Impl, MG_Remote, CI] (Disaggregated): P5f fm - connect family records and preserve GPU texture writes 2026-09-19 22:18:24 -04:00
swung0x48 31e95b0794 [MG_Backend] (Disaggregated): P5f fe - clarify readback pack-buffer boundary 2026-09-19 22:16:02 -04:00
swung0x48 0ddfd18aac Merge commit '63c2a331' into codex/p5f-wave2 2026-09-19 22:15:14 -04:00
swung0x48 63c2a331a9 [MG_Remote] (Disaggregated): P5f fe - document handle consumers and capture ordering 2026-09-19 22:14:38 -04:00
swung0x48 40f083cfc6 Merge commit '19a4400a' into codex/p5f-wave2 2026-09-19 22:13:37 -04:00
swung0x48 19a4400ae8 [MG_Pipe] (Disaggregated): P5f fe - retain live capture snapshots across make-current 2026-09-19 22:13:14 -04:00
swung0x48 81d20554aa [MG_Test] (Disaggregated): P5f fe - expose capture snapshot loss across make-current 2026-09-19 22:12:36 -04:00
swung0x48 20bbf9647e [MG_Backend, MG_IntegrationTest] (Disaggregated): P5f fm - carry framebuffer copies and exercise Magma wire verbs 2026-09-19 22:11:38 -04:00
swung0x48 8a6e124f12 Merge commit 'b5c60d5a015c68348725ae93297b13a8fff27723' into codex/p5f-wave2 2026-09-19 22:10:26 -04:00
swung0x48 db9f95576e [MG_Test, MG_IntegrationTest] (Disaggregated): P5f fe - close scalar freshness and strict ratchets 2026-09-19 22:09:32 -04:00
swung0x48 b5c60d5a01 [MG_Remote, MG_Test, Docs] (Disaggregated): fix P5f fc wire replies and control-frame guards 2026-09-19 22:03:55 -04:00
swung0x48 a0004db9cb [MG_Pipe, MG_Test] (Disaggregated): P5f fe - derive scalar forwards and ratchet Espryt fatals 2026-09-19 22:02:23 -04:00
swung0x48 4593719a3c [MG_Backend] (Disaggregated): P5f fm - share native depth readback and retain masked clear behavior 2026-09-19 22:01:44 -04:00
swung0x48 036c3a3d5e [MG_Backend, MG_Pipe, MG_Remote] (Disaggregated): P5f fm - route framebuffer verbs through server records 2026-09-19 22:00:14 -04:00
swung0x48 881e050e2c [MG_Pipe, MG_Test] (Disaggregated): P5f fe - verify XFB snapshot fields and preserve spans 2026-09-19 21:58:27 -04:00
swung0x48 07bf8d7ce0 [MG_Pipe, MG_Remote, MG_Backend] (Disaggregated): P5f fe - carry XFB capture snapshots across Begin 2026-09-19 21:57:45 -04:00
swung0x48 fa1e1cc427 Merge commit 'd5ed139cc50f75a6e7f20eb0cff4a36c7e8b46be' into codex/p5f-wave2 2026-09-19 21:52:55 -04:00
swung0x48 85f3104731 [MG_Backend, MG_Remote] (Disaggregated): P5f fe - retire existing-carrier Espryt pulls 2026-09-19 21:52:35 -04:00
swung0x48 7d8c8b04b1 [Docs] (Disaggregated): P5f handoff - where the interrupted session stopped, and how to resume 2026-09-19 21:44:30 -04:00
swung0x48 d5ed139cc5 [MG_Remote, MG_Test] (Disaggregated): P5f fc - the EGL control plane crosses as value frames, not function pointers
The twelve Server* forwarders' one-slot mailbox carried a raw function
pointer plus a void* to a stack-local Args struct (P5c audit row G4);
neither means anything across a process boundary. The slot now carries
one SurfaceControlFrame by value: kind is the op, scalars are the
arguments, and ok/eglMajor/eglMinor come back in the same value. The
blocking handshake, the shadow, the doorbell order and the C2 exit
block are unchanged.

- protocol.fbs: append-only - SurfaceOpKind gains SetSwapInterval=8,
  ReleaseResources=9, SetWindowHandle=10; WindowKind gains MetalLayer=6;
  SurfaceOp gains readSurface and context (the MakeCurrent four-tuple).
  protocol_generated.h regenerated with the pinned flatc.
- Server/SurfaceControlFrame.{h,cpp}: the frame (trivially copyable by
  static_assert) and the op enum - ten wire kinds plus four inproc-only
  kinds (InitCapabilities is answered by CapsSnapshot on the wire,
  SwapBuffers/InitWindowSurface have no production caller - f0-egl F8,
  kept for the inproc test lane).
- Protocol/SurfaceOpCodec.{h,cpp}: the frame <-> SurfaceOp/SurfaceReply
  codec, the explicit WindowBackend <-> WindowKind table, and the wire
  entry ServerApplyWireSurfaceOp whose refusals are named:
  Fatal{UnmigratedSurface, "AndroidNativeWindow@P12"} and
  Fatal{ProtocolCorruption, "SurfaceOp"}.
- ServerLoop: RunOnApplyThread(fn, void*) -> RunSurfaceControlFrame(frame)
  plus the dispatch ApplySurfaceControlFrame; the old per-forwarder Args
  bodies moved into the dispatch unchanged (C7/ID-54, N-3, R-12 included).
  RunProbeOnApplyThreadForTesting is the test seam through the same
  channel (the hook lives in a member, never in the slot).
- Tests: RemoteClientTest/ServerLoopTest migrated to the probe seam; new
  SurfaceControlFrameTest suite (mapping tables, wire round-trips, the
  two named refusals) plus ServerLoopTest's AVoidForwarderCrossesAsOne-
  DispatchedFrame (the frame channel's red-once handle) and
  ServerLoopEglTest's TheInitializeDisplayReplyArrivesThroughTheFrame.
2026-09-19 20:30:21 -04:00
swung0x48 418c765c58 [Docs, MG_IntegrationTest] (Disaggregated): P5f f1 report - the census ratchet's first version (6 pairs), CONTRACT-P5E section 3.2 marked landed, and the gate numbers 2026-09-19 10:39:40 -04:00
swung0x48 cfcc12f73a [MG_Backend, MG_Impl, MG_Remote, MG_Test, MG_IntegrationTest, CI] (Disaggregated): P5f f1 - the dual-block rehearsal: per-role PipeInputs, the MOBILEGL_IPC_ROLE_SPLIT_STATE knob, and the named-Fatal arm that makes the red a census 2026-09-19 09:59:33 -04:00
swung0x48 e75e00cb7a [Docs] (Disaggregated): P5f f0 - the eight-way read-only census reports
Statics, reverse channel, Magma cross-role sites, EGL control plane,
registry rekey, build/gate mechanics, the 15 BARRIER_PULLED fields'
carrier rulings, and the dual-block mechanism's design recon.
2026-09-19 09:08:56 -04:00
swung0x48 8b68b92c76 [Docs] (Disaggregated): P5f section 1 - the pointer fact, and one draw's full lifecycle through gPipeInputs
The section explained what gPipeInputs IS and what breaks under spawn, but not the one fact
that decides the shape of the fix, and not how a call actually reaches the driver through it.
Both are now in the document rather than only in the conversation that found them.

- THE POINTER FACT, now stated first because everything else follows from it: several
  BARRIER_PULLED fields do not hold values, they hold pointers into client-owned objects.
  PipeFill.cpp:274 fills m_textureUnitBase with `&ctx.GetTextureUnitObject(0)` - a raw base
  into the frontend's array - and :245 fills m_programForDraw with a SharedPtr<ProgramObject>.
  PipeInputs.h:669 dereferences it as `m_textureUnitBase[unit]`. PipeFill.cpp:2591-2592 says
  the same of the shader-buffer field: "four raw bases into the frontend's binding-point
  table".
  The consequence gets its own paragraph in 1.4, because it is the step most likely to be got
  wrong: this CANNOT be fixed by moving the block into shared memory. Mapping the same pages
  into both processes does not make a client heap address meaningful in the server, and a
  SharedPtr's refcount does not cross at all. The 15 fields each need a different SOURCE, not
  a shared location.

- NEW 1.2, one draw end to end: glBindTexture marks a tracker bit and sends nothing, because
  push only happens at a verb boundary; glDrawArrays enters MGPipeValidateForVerb
  (PipeFill.cpp:3347) whose four steps are barrier decision, dirty walk, emission (records,
  carrying handles), and the residual fill - which is where BARRIER_PULLED happens; the record
  goes on SEG_CMD; the apply thread dispatches to the backend; PrepareForDraw reads either a
  record-carried handle against the server's own tables, or gPipeInputs; then the real driver
  call. Plus the field lifetime: stamped with verb N's serial, read during verb N's apply,
  Fatal{UnmigratedPipeInput} if read at verb N+1 without a refill - one verb, never longer.

- The same example carries the Espryt/Magma split concretely: the texture binding reaches
  Espryt as a record with a handle (P5e retired GetTextureUnitObject for it) and reaches Magma
  through gPipeInputs into the client's array. That is FieldOwnership.def's own
  "P5e (Espryt unbarriered), P7 (Magma)" made legible, and it is why section 2.1's table has
  eight rows in that state.

Sections renumbered 1.1-1.4. Doc citation lint: 129 citations, 0 problems.
2026-09-19 06:02:28 -04:00