uartWireTiming()

ChipsetTiming fl::uartWireTiming	(	const ChipsetTimingConfig &	timing,
		u32	max_baud_rate )

The wire timing the UART encoder actually produces for a chipset.

The encoder quantises T0H/T1H onto a grid of period / pulses_per_bit, so what appears on the wire is not the chipset's nominal timing. A receiver decoding a UART-encoded frame has to classify against the quantised values, not the datasheet ones – otherwise it is matching a waveform nobody sent.

Derived through the same fitUartWave() rules that built the LUT, so the decode windows cannot drift from the encoding.

Parameters

timing	Chipset timing configuration
max_baud_rate	Backend baud ceiling (selects the frame geometry)

Returns

Quantised wire timing, or a zeroed ChipsetTiming if infeasible

Definition at line 196 of file uart_wave_encoder.cpp.hpp.

                                                             {
    const Wave10Lut lut = buildWave10LutForMaxBaud(timing, max_baud_rate);
    if (lut.pulses_per_bit == 0) {
        return ChipsetTiming{0, 0, 0, timing.reset_us, "uart_infeasible"};
    }
    // Re-run the fit at the geometry the LUT settled on. With P fixed the
    // result is independent of max_baud_rate (that argument only gates
    // feasibility), so this reproduces exactly the pulse counts encoded.
    const UartWaveFit fit = fitUartWave(timing, lut.pulses_per_bit, max_baud_rate);
    if (!fit.ok) {
        return ChipsetTiming{0, 0, 0, timing.reset_us, "uart_infeasible"};
    }
    const u32 period_ns = timing.total_period_ns();
    const u32 pulse_width_ns = period_ns / lut.pulses_per_bit;
    const u32 t0h = fit.pulses_0 * pulse_width_ns;
    const u32 t1h = fit.pulses_1 * pulse_width_ns;
    return ChipsetTiming{t0h, t1h - t0h, period_ns - t1h, timing.reset_us,
                         "uart_wire"};
}

References buildWave10LutForMaxBaud(), FL_NO_EXCEPT, and fl::Wave10Lut::pulses_per_bit.

Here is the call graph for this function: