swung0x48/MobileGL
1
0
Fork 0
mirror of https://github.com/MobileGL-Dev/MobileGL synced 2026-11-07 18:13:23 +09:00
Code Issues Packages Projects Releases Wiki Activity
Files
575be3532d24d4d64ec8bb764f192ba52bb77e06
MobileGL/docs/Disaggregated/notes/p5e/TASK-ra.md
T
swung0x48 aba329cc53 [Docs] (Disaggregated): P6's plan and contract draft; the working notes move from a local directory into the repository 
Two things land together: the plan and contract draft P6 needs before it starts, and the
working notes that until now existed on exactly one machine.

1. P6's plan and contract draft (neither started; not one line of code)

     docs/Disaggregated/P6-SPAWN-PLAN.md      the package plan
     docs/Disaggregated/P6-CONTRACT-DRAFT.md  the contract draft (moves to
                                              MG_Remote/CONTRACT-P6.md when c6 lands)

   The three findings that mattered, all written into the documents:

   a. P6 is SMALLER than its roadmap cell suggests. The transport primitives are already in
      the tree and already tested: SocketDoorbell including hangup detection and the two
      lost-wakeup fences (Doorbell.cpp:211-360, four cases in FdPassingTest), SCM_RIGHTS fd
      passing, ShmSegment::Adopt, and the SurfaceOp / WindowKind schema already defined in
      protocol.fbs. inproc uses a real second shm mapping rather than an alias, and
      SessionRings.h:22-28 says that was done so today is not the first time it runs. P6 is
      assembly plus a control plane plus a process, not "write a socket transport".

   b. The evidence for "P6 is only a transport swap" has EXPIRED. It was bought by P5c, whose
      audit found 59 direct memory accesses in a tree that already looked clean. P5e changed
      the tree again. So the first package, a6, is a read-only audit that writes no code, and
      every contract row that depends on it is marked PENDING a6 rather than decided in
      advance.

   c. P6 does not get a game running on the phone. ANativeWindow* is meaningless in the child,
      and protocol.fbs:186-196's own comment says Android transfers the window out of band -
      that band is P12. P6 lands pbuffer / surfaceless / offscreen, which is what HeadlessGL,
      the split-equivalent lane and trace replay need, and which is exactly what P6's own exit
      gate measures. A real window reaching the server is a named refusal. This gets its own
      section because "P6 is done" would otherwise read as "the game runs now".

   The contract adds rule G: a role may not name a process-local handle. It extends P5c's
   rule E (a role may not name the other role's memory) from memory to identity - EGLDisplay /
   EGLSurface / EGLContext / ANativeWindow* / fd number / pid / mapped address either cross as
   a token this contract defines or they do not cross, and a descriptor crosses only through
   SCM_RIGHTS, where the kernel renumbers it.

   What P5e left for P6 gets its own section. The latch semantics are already designed
   (ARCHITECTURE.md:504) and are copied unchanged. What is new is that after run-ahead the
   client no longer waits while the peer process can DIE: the client holds records that were
   published and will never be applied, and Android's freezer suspends the child without
   closing anything. So slow and dead must be distinguishable WITHOUT a timeout threshold.
   That is P6's only new design; everything else is carriage. It shares a predicate with E1's
   outstanding re-specification (ID-122), and the two should be designed together.

   Two gates carried over from P5e's lessons: G2 requires integration-spawn's test-name set to
   match integration-split's name for name, because a differing set means the lane is quietly
   running less; and the arm-proof gate requires every entry to record the child pid and
   transport=spawn in its OWN private log - ID-124 established that this kind of false green
   really happens, and a spawn lane that actually ran monolith would pass every other gate.
   S2's shape must be settled before it is written, and the document says why: E1 is the
   cautionary case.

2. The working notes move into the repository (docs/Disaggregated/notes/, 415 .md files)

   ROADMAP / CURRENT_STAGE_PROGRESS / MEASUREMENTS cite ruling numbers, audit line numbers and
   the provenance of gate figures from ~/w7/notes/, a directory that exists on one machine and
   is under no version control. A record cited hundreds of times that lives only locally is not
   a record, so it moves in whole, with its directory structure intact.

   Only .md moves. Scripts, run logs, census and gate result JSON, patches, APKs and nm dumps
   stay where they are: they are evidence DATA rather than documents - large, regenerated per
   run, and mostly carrying absolute paths and device serials that mean something only on the
   machine that produced them. The mapping and this split are written in
   docs/Disaggregated/notes/README.md.

   Two exceptions are recorded there too: P6's BRIEF has been promoted to a first-class
   document and keeps no copy under notes/, so the two cannot drift apart; and recovered/ holds
   material salvaged from failed workflow runs, which may duplicate the like-named phase
   directories and is kept because a few of those were the only complete version at the time.

   CI's citation lint takes docs/Disaggregated/*.md only (non-recursive), and notes/ is
   deliberately outside it: those notes' file:line citations were written against their own
   code heads, and resolving them against today's HEAD would fail wholesale - which is not an
   error, it is their correct state as a historical record. To check one, use the revision it
   declares (--rev).

3. Inconsistencies fixed in passing

   - README.md's status line still said P5e was in progress while the stage table already said
     closed.
   - CURRENT_STAGE_PROGRESS's next-step item 1 still listed the exit gate as pending, though
     (a) and (b) were run in the previous commit; rewritten to closed plus the one item that
     actually remains (E1).
   - ROADMAP's P6 row still had the pre-phase plan in three of its cells; it now carries the
     package order, the scope boundary and the two new gates.

Checked with check_doc_citations.py: 125 citations across docs/Disaggregated/*.md, 0 problems.
2026-09-18 23:31:39 -04:00

5.0 KiB
Raw Blame History

Package ra — the wait rule, present credit, server-owned gPipeInputs (P5e; parallel, lands last)

Read in this order, all of it: PACKAGE-PREAMBLE-P5E.md, INTEGRATOR-DECISIONS-P5E.md (ID-83, ID-84, ID-91, ID-92, ID-93 are yours), BRIEF-P5E.md (§0, §1, §2 "ra", §3 the integration gate, §4 the device exit, §6 rulings 3, 4, 13, 14, 15, 17, 18), CONTRACT-P5E.md §2 and §3 (all of both), your scout report scout-S6.md (all of it), and lists (b) and (c) of kimi-audit.md. Notes: //wsl.localhost/Arch/home/swung/w7/notes/p5e/.

Your slug is ra; your tree is ~/w7/p5e-ra (branch p5e/ra, based on the landed c0e head — NOT on the family packages; you are parallel to them and rebase at landing). c0e landed the WaitClass column, MGPipeWaitClassFor(op), the RunAhead / PresentCredit knobs, kCapRunAheadApply (bit 10) and kMGPipeP5eRunAheadReady = false; id landed the CurrentRecordBarriered line in ApplyOne.

What "done" means

The mechanism exists and is INERT until the integration commit: with kMGPipeP5eRunAheadReady false the server never publishes bit 10, the client never arms run-ahead, and every gate is exactly as green as today. Your last commit flips that constant — but only after the family packages have landed and the integrator says so, so structure your branch that way (the flip is its own final commit, trivially revertable).

  1. The predicate. MGPipeBarriered(op, payload, applierState) (c0e declared it) computed identically by both roles: the static WaitClass column, plus XFB-active from the context values and the kDrawClientArrays flag (ruling 3 / ID-83). The server's ApplyOne already sets CurrentRecordBarriered from it (id's line).
  2. The client wait rule. EmitAndWaitTails after PublishAndNotify: reply rows wait; barriered rows wait; unbarriered rows return. Present pacing through WaitForPresentAck with credit (default 1, range 1..8; ID-92) — the plumbing exists end to end and has no production caller today; MGPPresent::FrameSerial minted client-side; OnPresent returns the credit. Drains and forced waits per contract §2.5 (MapBuffer / GetBufferSubData / glFinish / fence client-wait / the EGL forwarders / the glGetError window — name the relaxation in the contract and in code).
  3. resource_subdata's buffer half stops waiting (ruling 15 / ID-93): the reply Bool is discarded at the call site today, so every 64 KB persistent-map block push is a hidden round trip. The row keeps its kReplySlot on the wire; the client neither waits nor reads. The texture half keeps its reply (trailing).
  4. gPipeInputs becomes server-role memory for unbarriered records (contract §3): the client's residual fill and its stamp bookkeeping do not run for them (the tracker walk and the emitters still run — they produce records); the stamp becomes server-private; MGPipeInputUnfreshRead becomes an UNCONDITIONAL detector for unbarriered applies (a named Fatal, not a counter), which is what turns the strict lane into a hard gate. Barriered records keep P5C's semantics exactly (ID-84 / ruling 4).
  5. Event-ring flow control (contract §2.6): the ring's overflow is a Fatal today on the premise that the client drains at every verb — under run-ahead that premise is gone. Make the producer block/flow-control instead, and keep a named refusal for the impossible case.
  6. Lifetime under run-ahead (contract §2.7): slot reuse vs records in flight, object_death ordering, the deferred-destroy queue (ID-91 / ruling 13: it stays; an unbarriered enqueue is a finding — pin it), Forget() of a persistent map while its pushes are in SEG_STAGE.
  7. The strict lane (.github/workflows/test.yml): rewritten as a hard green gate with the contract §7 allowlist (<field>@<verb> for the barriered rows that keep pulling).
  8. Docs: docs/Disaggregated/ARCHITECTURE.md §11, §14, §17 — the wait rule, the credit, the event ring, gPipeInputs ownership.

Red-once (execute, revert, quote)

  1. Encode present without WaitForPresentAck → the credit-1 client behind a stalled apply thread publishes a third present; Waits() unchanged → red by count.
  2. Leave one CopyField in an unbarriered fill → Fatal{RoleViolation, "gPipeInputs"} by name.
  3. Count instead of Fatal under an unbarriered record → the strict lane turns red on the first pulled row (that is the gate's own red-once).
  4. Keep the event-ring Fatal → a unit case publishing ~300 KiB of events behind an unwaited sequence aborts; with flow control it completes.
  5. Publish bit 10 for Magma → MagmaPipeIdentityTest red.
  6. MOBILEGL_IPC_RUN_AHEAD=0 renders identically; MOBILEGL_IPC_VERB_BARRIER=0 under run-ahead is the documented red.

Gate (before the flip)

build, unit, isplit, gens, strict — all exactly as green as the base, because every branch is inert with the bit unpublished. Your report must state, per numbered item above, what is wired and what it does when the bit flips; the integrator runs the device exit.

Reference in New Issue View Git Blame Copy Permalink