sysview: bench-run the WCH post-mortem dump recipe on ch32v307v_r1_1v0 The recipe was committed syntax-checked only. Run as written (cdc_msc, since neither WCH family builds the FreeRTOS examples): halt-only attach, 4096-byte ring dumped and WrOff-split, decoded to a 0.16 s window with overflow 0, ISR 85 n=68 p50 5.0 us, tud_task n=69 p50 8.0 / p99 17.0 us. ch582m is out of this recipe's scope - its ring does not survive OpenOCD's attach-reset.
diff --git a/.claude/skills/sysview/SKILL.md b/.claude/skills/sysview/SKILL.md index a3581b6..832429a 100644 --- a/.claude/skills/sysview/SKILL.md +++ b/.claude/skills/sysview/SKILL.md
@@ -185,13 +185,16 @@ `boards.md` has `JLINK_DEVICE` `—`, so on `ch32v20x`/`ch32v30x` the dump is hand-driven over OpenOCD: one attach reads the same channel-1 ring descriptor and dumps the same bytes, and the WrOff split below is the script's -linearization verbatim. `ch32v10x` and `ch583` reset on every WCH-Link attach, -which wipes the ring — no dump route there either (`boards.md`). Build with the -same `cmake` lines as above (there is no `-jlink` flash target here); `<uid>` is -the board's `flasher.uid` in `test/hil/tinyusb.json`. +linearization verbatim. `ch32v10x` resets on every WCH-Link attach and ch582m's +ring does not survive OpenOCD's attach-reset (`boards.md`), so this recipe is +for `ch32v20x`/`ch32v30x` only. Neither family builds the FreeRTOS examples, so +the post-mortem image is `cdc_msc`; `<uid>` is the board's `flasher.uid` in +`test/hil/tinyusb.json`. ```bash -ELF=build-pm/cdc_msc_freertos.elf +cmake -B build-pm -DBOARD=<board> -G Ninja -DCMAKE_BUILD_TYPE=MinSizeRel \ + -DSYSVIEW=4 -DSYSVIEW_POST_MORTEM=1 examples/device/cdc_msc && cmake --build build-pm +ELF=build-pm/cdc_msc.elf OUT=/tmp/sysview-pm; mkdir -p $OUT OOCD=(openocd -c "tcl_port disabled" -c "gdb_port disabled" -c "telnet_port disabled" -c "adapter serial <uid>" -c "adapter usb vid_pid 0x1a86 0x8010" @@ -227,10 +230,10 @@ ``` The core is left halted; recover the board by reflashing pristine firmware — -`reset run` under SDI never comes back. **Not bench-run yet**: the OpenOCD/Tcl -and the split are syntax-checked against openocd 0.12.0+dev and against -`sysview_dump.py`'s linearization, but no WCH post-mortem has been captured on -the rig. +`reset run` under SDI never comes back. Bench-run as written on +`ch32v307v_r1_1v0` (2026-09-04, openocd 0.12.0+dev-02620): a 4096-byte ring +halted mid-CDC-burst decoded to a 0.16 s window, overflow 0, ISR 85 n=68 +p50 5.0 µs, `tud_task` n=69 p50 8.0 / p99 17.0 µs. ## Build options
diff --git a/.claude/skills/sysview/boards.md b/.claude/skills/sysview/boards.md index 690e880..9bc1dfc 100644 --- a/.claude/skills/sysview/boards.md +++ b/.claude/skills/sysview/boards.md
@@ -52,7 +52,7 @@ | raspberry_pi_pico2 | rp2040 | rp2350_m33_0 | DWT | validated — see caveat | OpenOCD | | — | — | — | — | | same54_xplained | samd5x_e5x | ATSAME54P20 | DWT | validated end-to-end | J-Link | | — | — | — | — (prose below) | | mimxrt1064_evk | imxrt | MIMXRT1064xxx6A | DWT | validated (dual) | J-Link | 65536 | — | — | — | — | -| ch32v307v_r1_1v0 | ch32v30x | — | QingKe SysTick | tick 996.1 Hz (−0.39%) | dump | | — | — | — | — | +| ch32v307v_r1_1v0 | ch32v30x | — | QingKe SysTick | tick 996.1 Hz (−0.39%) | dump | 4096 | 0.16 s | 0 | 85, 5.0 | 8.0 / 17.0 | | ch582m_evt | ch583 | — | QingKe SysTick | tick 997.2 Hz (−0.28%) | dump | 2048 | — | — | — | — | | nanoch32v203 | ch32v20x | — | QingKe SysTick | tick 838.6 Hz (−16.1%) | dump | 2048 | — | — | — | — | @@ -64,7 +64,9 @@ `ci.lan:~/sysview-v2/out/campaign-final/`); the rate column keeps each board's original measurement, some of which predate the route switch. `—` capture cells: the board was not in the campaign (J-Link-only validations, host-role -pico2, dump-route WCH). Reading the overflow column: zero on 65536 with a fast +pico2, dump-route WCH — ch32v307's cells are its 2026-09-04 post-mortem dump, +a 4096-byte ring halted mid-burst, hence the 0.16 s window). Reading the +overflow column: zero on 65536 with a fast probe path; small-buffer boards (2048/4096) drop by design; f723/fruit_jam/ feather lose on drain-path throughput (drain-rate caveat below). metro's short window is the attach_only mid-stream join (2.4–5.6 s across runs).