Timo
a0f6e875be
otbr-agent ran for hours, then died on a burst of Matter traffic: [W] P-RadioSpinel-: radio tx timeout [C] P-RadioSpinel-: Failed to communicate with RCP otbr-agent exited with code 6 The RCP itself was fine -- its frame timestamps showed 18 hours of uninterrupted uptime across the crash, so nothing on the device reset. The host simply lost bytes and desynchronised. Nordic's nrf54l15dk overlay runs uart20 at 1 Mbaud, and this board's overlay copied that, but the DK pairs it with hw-flow-control. The XIAO cannot: uart20's pinctrl defines only TX and RX, and the CMSIS-DAP bridge carries no RTS/CTS. Without back-pressure a single late RX interrupt drops a byte, and a dropped byte corrupts the HDLC frame around it. 460800 keeps roughly 4x headroom over peak 802.15.4 throughput while widening the per-byte service window from 10 us to 21.7 us. The receive queue goes to 8192 bytes for the same reason -- with no flow control, whatever does not fit is lost rather than deferred. The RadioURL on the border router host must set the matching uart-baudrate=460800.
53 lines
2.2 KiB
Plaintext
53 lines
2.2 KiB
Plaintext
# OpenThread RCP on the Seeed XIAO nRF54L15.
|
|
|
|
# The XIAO has no external SPI flash populated.
|
|
CONFIG_SPI_NOR=n
|
|
|
|
# Nordic uses this on the nRF54L15 DK to keep the main stack clear of CRACEN
|
|
# (the hardware crypto engine) stack usage.
|
|
CONFIG_MAIN_STACK_SIZE=2048
|
|
|
|
# Nothing may write to uart20 except spinel. The boot banner and the UART log
|
|
# backend would both land mid-stream and desynchronise otbr-agent's HDLC
|
|
# decoder. Logs ride in-band over spinel instead, where otbr-agent decodes them.
|
|
CONFIG_BOOT_BANNER=n
|
|
CONFIG_LOG=y
|
|
CONFIG_LOG_MAX_LEVEL=1
|
|
CONFIG_LOG_BACKEND_UART=n
|
|
CONFIG_LOG_BACKEND_SPINEL=y
|
|
CONFIG_LOG_PROCESS_THREAD_STACK_SIZE=2048
|
|
|
|
# Transmit at the nRF54L15's full +8 dBm.
|
|
#
|
|
# Zephyr's default is 0 dBm and the coprocessor sample never overrides it, while
|
|
# Nordic's Matter samples set 8 -- so an RCP built from this repo was running 8 dB
|
|
# weaker than the Matter devices talking to it. That asymmetry hurts the
|
|
# border-router-to-device direction specifically, which is the one that matters
|
|
# for beacon responses during a joiner's active scan.
|
|
CONFIG_OPENTHREAD_DEFAULT_TX_POWER=8
|
|
|
|
# The XIAO needs GPIO, unlike the DKs this sample targets.
|
|
#
|
|
# The upstream coprocessor prj.conf sets CONFIG_GPIO=n ("Disable GPIO"), which
|
|
# is fine on a DK whose RCP has no GPIO-controlled RF hardware. This board
|
|
# selects its antenna path with one: P2.03 powers the RF switch and P2.05
|
|
# selects ceramic vs external, both declared as regulator-fixed nodes with
|
|
# regulator-boot-on in the board DTS.
|
|
#
|
|
# With GPIO off, REGULATOR_FIXED cannot build (it depends on GPIO), so those
|
|
# nodes get no driver and neither pin is ever driven -- the RF switch stays
|
|
# unpowered and the border router runs with a crippled antenna path. Measured
|
|
# -93 dBm at close range before this was found, against a -101 dBm sensitivity
|
|
# floor.
|
|
CONFIG_GPIO=y
|
|
CONFIG_REGULATOR=y
|
|
CONFIG_REGULATOR_FIXED=y
|
|
|
|
# Give the spinel link a deeper receive queue.
|
|
#
|
|
# The default 2048 bytes is sized for a link with RTS/CTS, where the host stops
|
|
# sending once the buffer fills. This one has no flow control, so anything that
|
|
# does not fit is simply lost. 8192 covers a burst of full-size frames arriving
|
|
# while the OpenThread thread is busy.
|
|
CONFIG_OPENTHREAD_COPROCESSOR_UART_RING_BUFFER_SIZE=8192
|