feat(margin): Save every LTC2977 in one command, and stop lying about failures

margin_save discarded the result of every write, so a NACK printed
"Change Saved" exactly like a successful store. For the one permanent
operation in this tool that is the worst place to stay quiet. Each store
is now checked, named on failure, and reflected in the exit code.

STORE_USER_ALL also went out as 0x15 followed by a dummy 0x00 - an SMBus
write-byte where the command is a send-byte. The LTC2977 tolerated it,
but tolerance is not correctness. Now i2cset ... 0x15 c on the native
path and a no-data cb_pmbus_write on the FPGA path.

Two new subcommands:

  margin_save all      - all nine boards in settings/, then re-init
                         whichever board was loaded beforehand so a
                         following margin_status still reports the board
                         you were looking at. Does not stop on the first
                         failure; a half-saved set is harder to reason
                         about than a fully-attempted one.

  margin_save blanton  - this platform exposes the CB FPGA F3 I2C
                         channels as native Linux i2c buses (Ch6->13,
                         Ch7->14, Ch8->15, Ch9->16; CB stays on bus 4),
                         so all 18 LTC2977 can be stored with plain
                         i2cset: no margin_init, no FPGA channel setup,
                         no pcimem. 18 commands instead of 42.

The 18 raw commands were verified on the DUT 2026-08-27; the wrapper was
tested against stubbed hardware only. Release notes say so.

Board list is MARGIN_BLANTON_STORE at the top of margin.sh - comment out
what this DUT does not have rather than letting it talk to absent boards.
Missing bus and unresponsive chip are reported as distinct failures.

MARGIN_STORE_SETTLE (0.5s) sits between stores because nothing here polls
the part's busy bit, and blanton fires 18 in a row.

Script A -> V1.0.11.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014RetWKFZFG1ZHcitQAyhwM

CLAUDE.md

@@ -63,6 +63,14 @@ temp_all                           # CB + ICB 溫感
 pwr_data                           # SWB VRM 全軌
 margin_init Blanton_CB_CONN13 && margin_status
 
+# ── 電壓 margin 寫 NVM（⚠️ 永久，測試中絕不執行）─────
+margin_save                        # 只寫當前 margin_init 載入的那顆（2 個 LTC2977）
+margin_save all                    # settings/ 全部 9 顆；跑完自動切回原本載入的那顆
+margin_save blanton                # Blanton SONiC 捷徑：18 顆全走原生 i2c，不用 margin_init
+#   0x15 是 send-byte，不帶 data byte：i2cset -f -y <bus> <addr> 0x15
+#   bus = FPGA F3 通道 + 7 → Ch6=13, Ch7=14, Ch8=15, Ch9=16（CB 本來就是 bus 4）
+#   ⚠️ 沒裝滿的機器：改 margin.sh 頂端的 MARGIN_BLANTON_STORE，把沒有的板子註解掉
+
 # ── SWB loopback 流量（bcmcmd，unit 0/1）─────────────
 # 不帶 -u 就是預設 TL_UNITS="0 1"，兩塊 switch board 都做（TTL 三支腳本即如此）
 blanton_traffic_linespeed ps             # 對照線的 link/speed
@@ -178,6 +186,11 @@ Tera Term 連 COM port (115200-8-N-1)
 - ⚠️ **I2C 讀一律 Repeated START**（不是 STOP + START），否則 PMBus/SMBus 裝置（VRM、EFUSE）不理。
 - ⚠️ **`margin_save` / `STORE_USER_ALL` 會永久寫 NVM**，斷電保留 → **測試流程中絕不執行**，
   以免覆蓋客戶出廠 config。`margin.sh` 的 auto-enable（`ON_OFF_CONFIG=0x1a`）只改 RAM，斷電復原。
+  v2.7.0 新增的 **`margin_save all` / `margin_save blanton` 一次寫 18 個 LTC2977**，
+  同一個警告放大 9 倍；另外 `0x15` 現在送 send-byte，不再帶多餘的 `0x00` data byte。
+- ⚠️ **SWB 的 LTC2977 有兩條路可以到**：`cb_pmbus_*`（CB FPGA F3 Ch6/7/8/9 → pcimem）與
+  原生 i2c bus（`i2cset -f -y 13/14/15/16`，Ch+7 = bus）。`margin_save blanton` 走後者，
+  其餘功能走前者。同一顆晶片，**不要同時用兩條路**。
 - ⚠️ **改電壓會弄壞 DUT**：套 profile 前確認該軌 OV/UV limit，且 servo DAC 注入電阻有 populate
   （本板部分 margin 電阻標 PROTO，沒 populate 的軌電壓不會動，不是腳本壞掉）。
 - ⚠️ **`ARB_LOST` ≠ 有裝置**：`vi2c_scan` 只看 `RX_ACK`，仲裁失敗會被誤報成裝置存在。