How was the fourth parameter determined in TIMER_PIN_MAP?

[lida2003]

I have checked the code and STM32 spec & config files (Googled), but still don't have any idea about the fourth parameter.

Can any one help me out?

/*
Details of the TIMER_PIN_MAP macro:
    i => is the index in the PG timer IO config array,
    p => is the hardware pin to be mapped,
    o => is the hardware pin occurrence (1 based index) within timerHardware (full) that should be used
                e.g. 1 = first occurence of the pin, 2 = second occurence, etc.
    d => the configured dma opt for the pin
*/
#define TIMER_PIN_MAP(i, p, o, d)  \
        { config[i].ioTag = IO_TAG(p); config[i].index = o; config[i].dmaopt = d; }

[haslinghuis]

4th column is simply the dma pin option assigned (using -1 if there is none).

[lida2003]

From my perspective,
a) bdshot need dma
b) dshot doesn't need dma
c) led doesn't need dma

But the opt value for dma is quite randomly assigned from my current view.

So any specific rules apply to assign the values for dma.

e.g. NEUTRONRCF435/config.h

Why led strip pin use opt value 7, not 12?

#define TIMER_PIN_MAPPING               TIMER_PIN_MAP( 1, LED_STRIP_PIN, 2, 7 ) \
                                        TIMER_PIN_MAP( 2, UART2_RX_PIN, 1,  6 ) \
                                        TIMER_PIN_MAP( 3, SERVO1_PIN, 1,  11 ) \
                                        TIMER_PIN_MAP( 4, SERVO2_PIN, 1,  10 ) \
                                        TIMER_PIN_MAP( 5, SERVO3_PIN, 2,  8 ) \
                                        TIMER_PIN_MAP( 6, SERVO4_PIN, 2,  9 ) \
                                        TIMER_PIN_MAP( 7, MOTOR1_PIN, 2,  0 ) \
                                        TIMER_PIN_MAP( 8, MOTOR2_PIN, 2,  2 ) \
                                        TIMER_PIN_MAP( 9, MOTOR3_PIN, 1,  1 ) \
                                        TIMER_PIN_MAP(10, MOTOR4_PIN, 1,  3 )

[haslinghuis]

For ledstrip using 7 would use DMA2, there is no 12 here
dma pin a01 list should give an overview

# dma pin a01 list
# 0: DMA1 Channel 0 Channel 73
# 1: DMA1 Channel 0 Channel 73
# 2: DMA1 Channel 0 Channel 73
# 3: DMA1 Channel 0 Channel 73
# 4: DMA1 Channel 0 Channel 73
# 5: DMA1 Channel 0 Channel 73
# 6: DMA1 Channel 0 Channel 73
# 7: DMA2 Channel 0 Channel 73

[lida2003]

It seems dma pin a01 list doesn't work now. there is no output.

#define RX_PPM_PIN                      PA3
#define LED_STRIP_PIN                   PA1

#define MOTOR1_PIN                      PC9
#define MOTOR2_PIN                      PC8
#define MOTOR3_PIN                      PC7
#define MOTOR4_PIN                      PC6

#define SERVO1_PIN                      PB6
#define SERVO2_PIN                      PB7
#define SERVO3_PIN                      PB1
#define SERVO4_PIN                      PB0

[haslinghuis]

It seems CLI is freezing (reflashed NEUTRONRCF435 to master)

[lida2003]

I have reflashed latest build eee0cb5 for AT32F435G target.

I also something like "CLI freezing". I will build eee0cb5 for NEUTRONRCF435 target, see what happens.

# version
# Betaflight / AT32F435G (A435) 4.5.0 Oct 23 2023 / 07:04:36 (eee0cb546) MSP API: 1.46

[lida2003]

Not good.
a) sometimes "CLI freezing"
b) sometimes just die "oops, no responding", need to reboot.

# version
# Betaflight / AT32F435G (A435) 4.5.0 Oct 23 2023 / 09:42:04 (eee0cb546) MSP API: 1.46
# config rev: 6fa0501
# board: manufacturer_id: NERC, board_name: NEUTRONRCF435

--- Edit ---

Very much like: #11486

[haslinghuis]

Just tested AIRBOTF435 and it does have same issue.
To reproduce type dma pin {PIN} list to halt operation.

[lida2003]

@haslinghuis Thanks for confirmation.

BTW dma pin a01 list, it shows 0 ~7.
Is that means I can choose any number of 0 to 7, right?

For ledstrip using 7 would use DMA2, there is no 12 here dma pin a01 list should give an overview

# dma pin a01 list
# 0: DMA1 Channel 0 Channel 73
# 1: DMA1 Channel 0 Channel 73
# 2: DMA1 Channel 0 Channel 73
# 3: DMA1 Channel 0 Channel 73
# 4: DMA1 Channel 0 Channel 73
# 5: DMA1 Channel 0 Channel 73
# 6: DMA1 Channel 0 Channel 73
# 7: DMA2 Channel 0 Channel 73

[haslinghuis]

Correct, just tried on another session but the list is cut off due to the bug present.