swung0x48/MobileGL
1
0
Fork 0
mirror of https://github.com/MobileGL-Dev/MobileGL synced 2026-11-07 01:53:24 +09:00
Code Issues Packages Projects Releases Wiki Activity
Files
d2efd1d79d9fac7fb12307250abd39b2f85d79da
MobileGL/docs/Disaggregated/notes/p12
T
History
BZLZHH d2efd1d79d [Perf] (MG_Remote/Client): the create window lands OFF - the deferral cannot be collected, 8 reply slots against 59,673 replies in one frame - measured on the device at 4,536 blocking creates either way 
The window was the handoff's remaining item: let up to N resource_create answers be
outstanding, read a provisional accept, collect the answers in order, and un-latch an
object whose late answer is DECLINED. It is a sound mechanism and the gates below pin
it - and on the device it does nothing at all, which is why the default is 1.

MEASURED, same device, same frame, same 101,421,396 B staged, same binary:

    MOBILEGL_IPC_CREATE_WINDOW=4   by_op 2 = 4,536   wall 28.64 s
    MOBILEGL_IPC_CREATE_WINDOW=1   by_op 2 = 4,536   wall 28.64 s
    (the handoff's pre-window baseline: 31.3 s)

by_op 2 counts the creates that took the BLOCKING path (LinkMetricsBeginReply is
called with ownsReplySlot = rowCarriesReplySlot && wantReply), so it not moving means
no create deferred. A temporary counter in the row said why: pushed=24, blocked=4,476,
took=20, readfail=8,951, pending pinned at 4 - the drain READ the answers back and
failed every time.

WHY, IN THE PROTOCOL'S OWN ARITHMETIC. An answer is readable only until its pool slot
is reused, and the pool is kDefaultReplySlotCount = 8. The server answers EVERY
reply-owning record, not only the rows that wait, because it cannot know which those
are (ClientSession.cpp:1897-1900: "The server posts either way"). This frame posts
59,673 of them: create 4,536 + respecify 20,633 + params 18,056 + sub-data 12,671 +
the rest. 8 slots against 59,673 replies churns 7,459 times per frame, i.e. 13.2
replies between two creates, so an answer is overwritten after 8 replies - 0.61 of a
create. The deferral is not slow, it is UNREADABLE, and no window depth fixes that.

AND IT IS NOT NEUTRAL: with the window on, the provisional accept latches an object
the applier may have refused, and the un-latch that corrects it lives in that same
unreadable drain. A refusal would become an object the client believes is published -
the silent divergence the suspect list exists to prevent. So the machinery is kept
(one experiment away from usable: a deeper pool, or a "do not answer me" bit, both
wire changes) and the default is off.

Worktree note: the window implementation itself was uncommitted in the tree at
82ed1df9 (MOBILEGL-CS-HANDOFF-2.md section 2, "can compile, gates mostly green, never
run on a device"). This commit lands it in the state the measurement supports. The
row also gains the heal the suspect list was missing - a BLOCKING create that lands
takes its object back out of it, or kCreateSuspectMax recoveries would retire the
window for the rest of the process - and the clamp that ties the environment's value
to the array, with a static_assert tying the array to half the reply pool so "someone
raises kCreateWindowMax to 8 and steps on the pool" is a compile error.

Gates (RemoteClientControls.inc), all green, 171/171 RemoteClientTest:
  ACreateWindowDefersTheAnswersAndTheWindowOffDoesNot - window on = 12 records 0 waits;
    the SAME CHILD turns the knob to 1 and it is 12 records 12 waits (the negative
    control); pending <= effective after every create; 99 clamps to the array; the
    array is <= half the pool
  ARefusedCreateUnlatchesItsObjectAndItsNextCreateBlocks - a real refusal (the
    server's consumed mask narrowed) clears the latch, +1 suspect, the object's next
    create BLOCKS (2 records 2 waits), and restoring the consumer HEALS it (latch
    back, suspect -1)

Docs: CONTRACT-P5E.md section 2.5 rewritten (it still said all three rows keep
kWaitReply), plus
docs/Disaggregated/notes/p12/CREATE-WINDOW-MEASURED.md (the numbers, the arithmetic,
the picture evidence: both arms SSIM 0.977) and
docs/Disaggregated/notes/p12/UNPUBLISH-GUARD-SIGNALS.md (handoff section 3: neither
caps generation nor FreshlyPrimed is a signal for "the applier dropped its records",
and the residual gap is one line - the dedup at TextureEmit.h:838 does not read the
latch, so an UNCHANGED descriptor after a drop produces no record and therefore no
refusal to self-heal from).
2026-09-26 00:47:19 +08:00
..
android
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
device
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
gate
[Docs] (P12 review): ID-P12-14 records the review round's host gate - the new cases' red-without-fix demonstrations, and every remaining red traced to the container's CMake/CTest, not to the fixes
2026-09-24 14:03:48 +00:00
BRIEF-onscreen.md
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
CREATE-WINDOW-MEASURED.md
[Perf] (MG_Remote/Client): the create window lands OFF - the deferral cannot be collected, 8 reply slots against 59,673 replies in one frame - measured on the device at 4,536 blocking creates either way
2026-09-26 00:47:19 +08:00
DEVICELOST-AND-UPLOAD-WAITS.md
[Perf] (MG_Pipe, MG_Remote/Client): a texture respecify for a CONFIRMED object stops waiting - the row pays its round trip once, at first use - and the frame drops from 130.9 s to 31.3 s
2026-09-25 23:29:08 +08:00
fix-icd.sh
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
g1-base-names.sh
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
gate-container.sh
[Docs] (P12 review): the review round's host gate log and the container variant of the gate script
2026-09-24 14:03:48 +00:00
gate.sh
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
INTEGRATOR-DECISIONS-P12.md
[Docs] (P12 review): ID-P12-15 - G1 holds for the review round: the pull build at cb0404f2 and at this head, same machine and compiler, has .text 0x9b1bf2 on both sides and 30614 symbols with 0 added / 0 removed
2026-09-24 14:03:48 +00:00
map-apk-process.md
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
map-client-surface.md
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
map-docs-contract.md
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
map-server-surface.md
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
p203.sh
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
PLAN-P12.md
[Docs] (P12): record post-review verification
2026-09-25 01:34:29 -04:00
README.md
[Docs] (P12): the device-lost reply tolerance and the texture-upload wait granularity get their own record - what landed, the gates that can go red, and what is still owed
2026-09-25 19:44:51 +08:00
setup-tree.sh
[Docs] (P12): the on-screen server window's plan, audit, harness and device evidence move into the tree, and the status page says what this line is and what it still owes - CURRENT_STAGE_STATUS records that the P12 subset is implemented but NOT handed off (both exit gates unrun: no cross-machine TCP, no FCL + kill-server device-lost), the ROADMAP P12 row stops reading "待排", and notes/p12 carries the brief, the four read-only audits, the gate and device harness, and the curated device evidence (check transcripts, replay verdicts, on-screen captures; logcat left out per notes/README.md's evidence-data rule)
2026-09-24 07:44:40 -04:00
UNPUBLISH-GUARD-SIGNALS.md
[Perf] (MG_Remote/Client): the create window lands OFF - the deferral cannot be collected, 8 reply slots against 59,673 replies in one frame - measured on the device at 4,536 blocking creates either way
2026-09-26 00:47:19 +08:00

README.md

P12 — Android 生产窗口路径（「server 自有屏幕窗口」子集已实现，未收官）

本页是该阶段的阶段汇总：阶段表行、实测与落地形状的完整文本（2026-09-24 从上层索引移入，原文照录）都在这里；上层只留摘要——ROADMAP.md、MEASUREMENTS.md、ARCHITECTURE.md。文中 file:line 与"今天""当前"按各段写成时的头理解；文中的 MEASUREMENTS.md §N / ARCHITECTURE.md §17.x 编号保留，对照表见 ../README.md。

摘要

  • 当前阶段。状态摘要在 ../../CURRENT_STAGE_PROGRESS.md；计划与交接 PLAN-P12.md；任务书 BRIEF-onscreen.md；裁定 INTEGRATOR-DECISIONS-P12.md（ID-P12-1..15）；只读审计 map-*.md 四份。
  • 形状：server APK 自建 ANativeWindow（MobileGLDisplayActivity 的 SurfaceView，进程内 TCP server，会话串行）把渲染流上屏；离屏路径保留、同一时刻一条活跃；client 以 WindowKind::ServerOwned（控制修订 3）完全无头接入，几何由 server 回传；失窗 → Fatal{ServerWindowLost} → 干净 device-lost。
  • 进展：15 个 (P12) 提交 + 审查轮修复（10 个问题，ID-P12-5..11）已并入 feat/disaggregated；主机门绿（ID-P12-14）、G1 成立（ID-P12-15）；审查后在 1fb18d9e 上完成 P12 定向 CTest 46/46 和 Redmi 真机七项复测。当前分支 0fe01588 后续提交只改了文档、文档检查范围和基准脚本说明文字，未改 P12 实现。
  • 2026-09-25 追加（会话交接 §4/§5.1）：device-lost 的 DECLINED 回复容忍与纹理上传的等待粒度已落地，各有自己的 red-once 门（前者就是进世界后 Fatal{ProtocolCorruption, "Fence.reply"} 那条崩溃；后者让单条纯上传的回包等待 1 → 0）。真机端到端（进世界不崩、那一帧墙钟）与 G1 本机复核仍待做，见 DEVICELOST-AND-UPLOAD-WAITS.md。
  • 未完成：出口门 (a) FCL 同机 spawn + 杀 server 的 device-lost、(b) 跨机 TCP 入世界 + P6.5 必测数；CONTRACT-P12.md；阶段表行里的其余条目（unix: 监听、DirectGLES 去全局、freezer、多 context、FCL 开关接线、D8 白名单）。

阶段表行（原 ROADMAP.md）

  • 阶段：P12 Android 生产窗口路径
  • 状态：子集已实现、未收官（2026-09-25）：上屏 server 窗口、WindowKind::ServerOwned、无头 client、失窗 device-lost 已落地；审查轮 10 个问题已修（两个大项：批内失窗、会话间 Espryt 单元影子悬空；裁定 INTEGRATOR-DECISIONS-P12.md ID-P12-5–11）。审查后 P12 定向 CTest 46/46、真机七项复测通过；出口门 (a) FCL + 杀 server 与 (b) 跨机 TCP 仍未完成（见 PLAN-P12.md §3）。
  • 落地什么 / 范围：按重审 §6 改小：Service 收 Java Surface 删掉 Surface 传递那一半——server app 拥有自己的 SurfaceView，client 根本没有窗口，于是 minSdk 26 没有扁平化 ANativeWindow 的 NDK API 这条约束不再成立；WindowKind::ServerOwned；两个后端各一条 CreateEGLWindowSurface 臂；m_windowHandle 四处写点 + sameHandle 去重键（a6 §6）；DirectGLES g_Display / g_Surface / g_Context 去全局；kEventSurfaceChanged 的 Width / Height 缺口（Espryt 只发布格式，server 拥有显示时 client 默认 FBO 会停在 512×512 占位值）；client 侧不再无条件发 SetWindowHandle（BackendObject_Remote.cpp:215——今天 FCL 里开 spawn，第一次 eglCreateWindowSurface 就在 server 进程 abort）；server app 监听 unix: 与 tcp://（后者 Ph 之后）；server 生命周期绑 Activity；cached-app freezer（SPAWN-PLAN §8.2）；多 context（契约 §11）；FCL env 与 plugin APK 开关表接线
  • 验收门 / 证据：两个门：(a) Minecraft 经 FCL 同机 spawn 在 Adreno 830 双后端入世界，杀 server 产生干净 device-lost latch；(b) 另一台机器上的 client 经 TCP 在 Redmi 上进世界（数据面 stream），并记录 P6.5 的必测数

2026-09-25 复测补记

复测针对 P12 实现提交 1fb18d9e（控制协议修订 3）；当前 feat/disaggregated@0fe01588 相对它只改了文档、文档检查范围和基准脚本说明文字。测试 APK 为 26.09.1fb18d9-trace（独立包名 top.mobilegl.plugin.p12verify.trace），设备为 Redmi 2f7cbe2e（Adreno 830）。本记录只覆盖 P12 定向用例，不表示两个收官出口门已通过。

主机 P12 定向 CTest：46/46 通过。 包含 ServerDisplayTest、ServerLoopTest、InProcessServer、SupervisorChildren 与 ServerOwnedSurface；不是整套 P12 主机门重跑。

真机七项检查：

# 检查 审查后结果
1 Espryt 上屏（OpenRA） 通过，SSIM 1.000000，server-owned window 640×480
2 Magma 上屏（OpenRA）及同进程第二会话（Minecraft startup） 通过，SSIM 1.000000 / 0.999999511
3 串行会话与 Busy 第二次 OpenRA 通过，SSIM 1.000000；并发客户端收到 Refuse{Busy}（rc=134），长 Minecraft 会话继续并以 SSIM 0.999999511 完成
4 推流时按 HOME 触发 surfaceDestroyed ServerWindowLost 后客户端干净 device-lost；Activity 进程存活，surface 在 15 ms 内释放；重附窗口后的 OpenRA 会话 SSIM 1.000000
5 离屏 service（pbuffer） OpenRA 通过，SSIM 1.000000
6 离屏 server 收到 ServerOwned 请求 服务端和客户端都返回 NoServerDisplay；后续 pbuffer 会话通过，SSIM 1.000000。首次尝试与上一会话回收竞态命中 Busy；临时测试副本等待 2 秒后重跑通过，仓库脚本未修改
7 Activity 与离屏 service 互斥 启动 Activity 会停止 service 及其活动会话；启动 service 会停止 :mglwin，之后 service 正常监听

设备检查通过 adb forward 由 trace_replay 驱动；失窗检查使用 HOME，不是 FCL 杀 server。出口门 (a) 的 FCL 同机 spawn / 杀 server，以及出口门 (b) 的跨机 TCP 入世界与 P6.5 必测数仍未验证。

运行证据保存在测试机器的 /home/swung/w7/logs/p12-verify-1fb18d9e/（build-host/Testing/Temporary/LastTest.log 与 device/）；按证据数据规则，大日志、APK 和截图未复制进仓库。

状态记录（原 CURRENT_STAGE_STATUS.md，2026-09-24 历史快照）

本段保留当日记录；其中“审查后未复测”的状态已由上方 2026-09-25 复测补记更新。

更新：2026-09-24（P12 子集） · 计划与交接 notes/p12/PLAN-P12.md。P12「server 自有屏幕窗口」子集已实现、未收官，已并入 feat/disaggregated（15 提交自 9f669e52 起，加审查轮 10 个问题的修复——两个大项：批内失窗、会话间 Espryt 单元影子悬空；裁定 notes/p12/INTEGRATOR-DECISIONS-P12.md ID-P12-5–15，主机门与 G1 见 ID-P12-14/15，审查轮之后未上真机复测，ID-P12-13）：Android 上的 server 自建 ANativeWindow（自己的 SurfaceView）把渲染流上屏，离屏路径保留，同一时刻只一条活跃；client 以新的 WindowKind::ServerOwned 完全无头接入（控制修订 3）。主机门除一条既有环境敏感项外全绿，G1（pull 构建 .text 0xa52203、符号 0/0）已独立复核；真机 Redmi 2f7cbe2e 上七项编号检查过了双后端上屏、串行/Busy、失窗 device-lost、离屏不回归、无显示具名拒绝、一台设备一个 server。两个出口门都还没打：跨机 TCP（门 b）完全未做，FCL + 杀 server 的 device-lost（门 a）未做——本轮的真机是 trace_replay 经 adb forward 环路驱动的。上一阶段 P7 已收官（2026-09-23，ID-P7-63）：§8 九项全绿——终局窗口 3（p7w8-78e71f2b）门 3 PASS 36/36、门 5 PASS；门 4 以 # P13 读法关闭。

  • P12 子集：APK 26.09.29b7b28-trace（控制协议修订 3，与 p7w8 的 server 不兼容），证据 notes/p12/device/（九格拼图 evidence-strip.png）。七个编号检查在三个 head 上跑过，trial-*/ 是按 head 归档的旧轮次。

当前状态详情（2026-09-24，原 CURRENT_STAGE_PROGRESS.md 正文）

顶层 CURRENT_STAGE_PROGRESS.md 只留摘要；带裁定编号的落地表、门表、余项清单与证据位置在这里。

1. 一句话（2026-09-24 历史快照；现状见上方复测补记）

Android 上的 server 自建窗口（自己的 SurfaceView）把 IPC 渲染流上屏，client 以 WindowKind::ServerOwned 完全无头接入。该快照记录当时审查轮修复后尚欠真机复测；复测现已在上方补记，两个出口门仍未打。

2. 已落地

面 内容 出处
wire WindowKind::ServerOwned、控制修订 2 → 3、SurfaceReply 带回窗口真实尺寸或具名拒绝 PLAN-P12 §1
server 窗口 ServerDisplay（钩子 + 租约 + 有界 3 s Detach）；会话表面模式首次创建即闩定（SurfaceModeMismatch）；无显示具名拒绝（NoServerDisplay）；失窗 → Fatal{ServerWindowLost} → client 读干净 device-lost ID-P12-2、ID-P12-3
client MOBILEGL_IPC_SURFACE=server → 无头窗口 surface，不发 SetWindowHandle；resize 变成对 server 窗口的几何请求，采用 server 回的真实尺寸 ID-P12-10
进程内 server MobileGLDisplayActivity（:mglwin）里一个线程跑 TCP server：会话串行、会话间重置闩、后端按进程钉住；与离屏 MobileGLServerService 互斥，一台设备同一时刻一个 server ID-P12-4
审查轮 10 个问题已修；两个大项：批内失窗检查（DrainRing 每次 pop 前）、会话间 Espryt 单元影子悬空；另有推流中停机、会话结束清理、陈旧尺寸、日志转发 ID-P12-5 … 11

3. 门

门 状态 证据
G1 ✅ pull 构建同机前后 .text 一致、符号 0 增 0 减 ID-P12-15
主机门 ✅ unit 2605/2605，spawn / tcp 与三条 Magma 双进程车道全绿；剩余红全部归容器的 CMake/CTest 版本 ID-P12-14
真机七项检查（Redmi 2f7cbe2e） ✅ 审查轮之前（29b7b284）及审查后（1fb18d9e）；审查后详细结果见上方复测补记 PLAN-P12 §2、ID-P12-13
出口门 (a) ❌ FCL 同机 spawn、双后端入世界、杀 server 产生干净 device-lost（本轮的驱动是 trace_replay，device-lost 由按 HOME 触发） PLAN-P12 §3
出口门 (b) ❌ 另一台机器的 client 经 TCP 在 Redmi 入世界，并记录 P6.5 必测数（本轮 client 是 adb forward 环路） PLAN-P12 §3

4. 在跑

无。

5. 下一步

  1. 出口门：先 (b) 跨机 TCP 入世界 + P6.5 必测数，再 (a) FCL + 杀 server。
  2. 收尾：写 MG_Remote/CONTRACT-P12.md；修 notes/p12/gate.sh 的 G1 符号比较（与基准不同源，会印出误导的 added=N removed=N）；按惯例做一次收官审查。
  3. 阶段表行里的余项（交接时在树上核过，均未做）：in-process server 的 unix: 监听、DirectGLES g_Display / g_Surface / g_Context 去全局、cached-app freezer、多 context、FCL env 与 plugin 开关表接线、D8 窗口种类白名单。
  4. 其后按 ROADMAP.md：monolith 跑道 P8；IPC 跑道 P9 → P10 → P11；P6.5 残余（39 例 device 矩阵 + 必测数）与 P3b/P4b 余项并行；CONTRACT-P7.md §12 的记录债按阶段认领。

6. 阻塞 / 需要人

  • 没有需要用户决策的阻塞。真机复测需要一台装有 NDK、连着 Redmi 的机器（审查轮所在的云容器没有）。
  • 已知环境敏感项：IterationRPProgram203Scenario（integration-magma-full-split）偶发红，按 ID-P7-22 / ID-P12-12 读作"既有、环境敏感"（钉 lavapipe ICD 即绿），不阻塞 P12。

7. 真机与证据

  • 当前 APK 26.09.29b7b28-trace（控制修订 3，与 P7 终局窗口的 p7w8 server 不兼容）；设备 Redmi 2f7cbe2e（Adreno 830），定频协议 devices/pin-verification-2026-09-07.md。
  • 真机证据 notes/p12/device/（evidence-strip.png 九格拼图，trial-*/ 是按 head 归档的旧轮次）；主机门日志 notes/p12/gate/；设备脚本 notes/p12/android/。

本目录

文件 内容
DEVICELOST-AND-UPLOAD-WAITS.md 2026-09-25 落地记录：DECLINED 回复容忍（A）与纹理上传等待粒度（B），含门、red-once 与未做项
BRIEF-onscreen.md P12 subset "on-screen server window" — implementation brief (2026-09-23)
fix-icd.sh 钉 lavapipe ICD
g1-base-names.sh G1 基准符号名生成
gate-container.sh 主机门的容器变体
gate.sh 主机门脚本
INTEGRATOR-DECISIONS-P12.md P12 集成者裁定日志（「server 自有屏幕窗口」子集）
map-apk-process.md MobileGL-disagg: Android APK and process model for an on-screen server path (commit 9f669e52)
map-client-surface.md Client-side surfaces in split mode: map, open questions and change list (commit 9f669e52)
map-docs-contract.md Design docs and contracts for a server-owned on-screen window (P12 subset)
map-server-surface.md Server-side surface path and both backends' window surfaces (commit 9f669e52)
p203.sh IterationRPProgram203Scenario 复核
PLAN-P12.md P12「server 自有屏幕窗口」子集 — 计划与交接（2026-09-24）
setup-tree.sh 包树搭建
android/ 27 个文件：真机与构建脚本（dev.sh、check*.sh、final-run.sh、red-once 工具等）
device/ 85 个文件：真机七项检查证据（evidence-strip.png 九格拼图；trial-*/ 为按 head 归档的旧轮次）
gate/ 6 个文件：三个 head 的构建日志与主机门日志（含审查轮、CMake 4 对照）
Reference in New Issue View Git Blame Copy Permalink