~ruther/qmk_firmware

302b35c2a0ac90208e523944e8cc4b44a793d8d5 — Takeshi ISHII 4 years ago 017aa59
fix matrix_io_delay() timing in quantum/matrix.c (#9603)

* fix matrix_io_delay() timing in quantum/matrix.c

* Updated comments explaining the need for matrix_io_delay() in quantum/matrix.c

* fix matrix_io_delay() timing in quantum/split_common/matrix.c

* Update quantum/matrix.c

Co-authored-by: Ryan <fauxpark@gmail.com>

* Update quantum/split_common/matrix.c

Co-authored-by: Ryan <fauxpark@gmail.com>

* Update quantum/matrix.c

Co-authored-by: Ryan <fauxpark@gmail.com>

* Update quantum/split_common/matrix.c

Co-authored-by: Ryan <fauxpark@gmail.com>

* add waitOutputPinValid() and wait_cpuclock() into quantum/quantum.h and tmk_core/common/wait.h

* add matrix_output_select_delay() and matrix_output_unselect_delay()

* fix quantum/matrix_common.c, tmk_core/common/matrix.h

* fix tmk_core/common/wait.h

* fix quantum/quantum.h, tmk_core/common/wait.h

* waitOutputPinValid() rename to waitInputPinDelay() in quantum/quantum.h.

* waitOutputPinValid() rename to waitInputPinDelay() in quantum/matrix_common.c

* update tmk_core/common/wait.h

* update comment in quantum/matrix.c, quantum/split_common/matrix.c

* update quantum/quantum.h: Make more margin in the GPIO_INPUT_PIN_DELAY default value.

Co-authored-by: Ryan <fauxpark@gmail.com>
M quantum/matrix.c => quantum/matrix.c +10 -4
@@ 101,9 101,9 @@ static bool read_cols_on_row(matrix_row_t current_matrix[], uint8_t current_row)
    // Start with a clear matrix row
    matrix_row_t current_row_value = 0;

    // Select row and wait for row selecton to stabilize
    // Select row
    select_row(current_row);
    matrix_io_delay();
    matrix_output_select_delay();

    // For each col...
    for (uint8_t col_index = 0; col_index < MATRIX_COLS; col_index++) {


@@ 116,6 116,9 @@ static bool read_cols_on_row(matrix_row_t current_matrix[], uint8_t current_row)

    // Unselect row
    unselect_row(current_row);
    if (current_row + 1 < MATRIX_ROWS) {
        matrix_output_unselect_delay();  // wait for row signal to go HIGH
    }

    // If the row has changed, store the row and return the changed flag.
    if (current_matrix[current_row] != current_row_value) {


@@ 147,9 150,9 @@ static void init_pins(void) {
static bool read_rows_on_col(matrix_row_t current_matrix[], uint8_t current_col) {
    bool matrix_changed = false;

    // Select col and wait for col selecton to stabilize
    // Select col
    select_col(current_col);
    matrix_io_delay();
    matrix_output_select_delay();

    // For each row...
    for (uint8_t row_index = 0; row_index < MATRIX_ROWS; row_index++) {


@@ 175,6 178,9 @@ static bool read_rows_on_col(matrix_row_t current_matrix[], uint8_t current_col)

    // Unselect col
    unselect_col(current_col);
    if (current_col + 1 < MATRIX_COLS) {
        matrix_output_unselect_delay();  // wait for col signal to go HIGH
    }

    return matrix_changed;
}

M quantum/matrix_common.c => quantum/matrix_common.c +5 -0
@@ 1,3 1,4 @@
#include "quantum.h"
#include "matrix.h"
#include "debounce.h"
#include "wait.h"


@@ 83,8 84,12 @@ uint8_t matrix_key_count(void) {
    return count;
}

/* `matrix_io_delay ()` exists for backwards compatibility. From now on, use matrix_output_unselect_delay(). */
__attribute__((weak)) void matrix_io_delay(void) { wait_us(MATRIX_IO_DELAY); }

__attribute__((weak)) void matrix_output_select_delay(void) { waitInputPinDelay(); }
__attribute__((weak)) void matrix_output_unselect_delay(void) { matrix_io_delay(); }

// CUSTOM MATRIX 'LITE'
__attribute__((weak)) void matrix_init_custom(void) {}


M quantum/quantum.h => quantum/quantum.h +29 -0
@@ 210,6 210,13 @@ typedef uint8_t pin_t;

#    define togglePin(pin) (PORTx_ADDRESS(pin) ^= _BV((pin)&0xF))

/*   The AVR series GPIOs have a one clock read delay for changes in the digital input signal.
 *   But here's more margin to make it two clocks. */
#    if !defined(GPIO_INPUT_PIN_DELAY)
#        define GPIO_INPUT_PIN_DELAY 2
#    endif
#    define waitInputPinDelay() wait_cpuclock(GPIO_INPUT_PIN_DELAY)

#elif defined(PROTOCOL_CHIBIOS)
typedef ioline_t pin_t;



@@ 225,6 232,28 @@ typedef ioline_t pin_t;
#    define readPin(pin) palReadLine(pin)

#    define togglePin(pin) palToggleLine(pin)

#endif

#if defined(__ARMEL__) || defined(__ARMEB__)
/* For GPIOs on ARM-based MCUs, the input pins are sampled by the clock of the bus
 * to which the GPIO is connected.
 * The connected buses differ depending on the various series of MCUs.
 * And since the instruction execution clock of the CPU and the bus clock of GPIO are different,
 * there is a delay of several clocks to read the change of the input signal.
 *
 * Define this delay with the GPIO_INPUT_PIN_DELAY macro.
 * If the GPIO_INPUT_PIN_DELAY macro is not defined, the following default values will be used.
 * (A fairly large value of 0.25 microseconds is set.)
 */
#    if !defined(GPIO_INPUT_PIN_DELAY)
#        if defined(STM32_SYSCLK)
#            define GPIO_INPUT_PIN_DELAY (STM32_SYSCLK/1000000L / 4)
#        elif defined(KINETIS_SYSCLK_FREQUENCY)
#            define GPIO_INPUT_PIN_DELAY (KINETIS_SYSCLK_FREQUENCY/1000000L / 4)
#        endif
#    endif
#    define waitInputPinDelay() wait_cpuclock(GPIO_INPUT_PIN_DELAY)
#endif

// Atomic macro to help make GPIO and other controls atomic.

M quantum/split_common/matrix.c => quantum/split_common/matrix.c +10 -4
@@ 114,9 114,9 @@ static bool read_cols_on_row(matrix_row_t current_matrix[], uint8_t current_row)
    // Start with a clear matrix row
    matrix_row_t current_row_value = 0;

    // Select row and wait for row selecton to stabilize
    // Select row
    select_row(current_row);
    matrix_io_delay();
    matrix_output_select_delay();

    // For each col...
    for (uint8_t col_index = 0; col_index < MATRIX_COLS; col_index++) {


@@ 129,6 129,9 @@ static bool read_cols_on_row(matrix_row_t current_matrix[], uint8_t current_row)

    // Unselect row
    unselect_row(current_row);
    if (current_row + 1 < MATRIX_ROWS) {
        matrix_output_unselect_delay();  // wait for row signal to go HIGH
    }

    // If the row has changed, store the row and return the changed flag.
    if (current_matrix[current_row] != current_row_value) {


@@ 160,9 163,9 @@ static void init_pins(void) {
static bool read_rows_on_col(matrix_row_t current_matrix[], uint8_t current_col) {
    bool matrix_changed = false;

    // Select col and wait for col selecton to stabilize
    // Select col
    select_col(current_col);
    matrix_io_delay();
    matrix_output_select_delay();

    // For each row...
    for (uint8_t row_index = 0; row_index < ROWS_PER_HAND; row_index++) {


@@ 188,6 191,9 @@ static bool read_rows_on_col(matrix_row_t current_matrix[], uint8_t current_col)

    // Unselect col
    unselect_col(current_col);
    if (current_col + 1 < MATRIX_COLS) {
        matrix_output_unselect_delay();  // wait for col signal to go HIGH
    }

    return matrix_changed;
}

M tmk_core/common/matrix.h => tmk_core/common/matrix.h +3 -0
@@ 55,6 55,9 @@ matrix_row_t matrix_get_row(uint8_t row);
/* print matrix for debug */
void matrix_print(void);
/* delay between changing matrix pin state and reading values */
void matrix_output_select_delay(void);
void matrix_output_unselect_delay(void);
/* only for backwards compatibility. delay between changing matrix pin state and reading values */
void matrix_io_delay(void);

/* power control */

M tmk_core/common/wait.h => tmk_core/common/wait.h +52 -0
@@ 6,10 6,62 @@
extern "C" {
#endif

#if defined(__ARMEL__) || defined(__ARMEB__)
#    ifndef __OPTIMIZE__
#        pragma message "Compiler optimizations disabled; wait_cpuclock() won't work as designed"
#    endif

#    define wait_cpuclock(x) wait_cpuclock_allnop(x)

#    define CLOCK_DELAY_NOP8 "nop\n\t nop\n\t nop\n\t nop\n\t   nop\n\t nop\n\t nop\n\t nop\n\t"

__attribute__((always_inline))
static inline void wait_cpuclock_allnop(unsigned int n) { /* n: 1..135 */
    /* The argument n must be a constant expression.
     * That way, compiler optimization will remove unnecessary code. */
    if (n < 1) { return; }
    if (n > 8) {
        unsigned int n8 = n/8;
        n = n - n8*8;
        switch (n8) {
        case 16: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case 15: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case 14: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case 13: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case 12: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case 11: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case 10: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  9: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  8: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  7: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  6: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  5: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  4: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  3: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  2: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  1: asm volatile (CLOCK_DELAY_NOP8::: "memory");
        case  0: break;
        }
    }
    switch (n) {
    case 8: asm volatile ("nop"::: "memory");
    case 7: asm volatile ("nop"::: "memory");
    case 6: asm volatile ("nop"::: "memory");
    case 5: asm volatile ("nop"::: "memory");
    case 4: asm volatile ("nop"::: "memory");
    case 3: asm volatile ("nop"::: "memory");
    case 2: asm volatile ("nop"::: "memory");
    case 1: asm volatile ("nop"::: "memory");
    case 0: break;
    }
}
#endif

#if defined(__AVR__)
#    include <util/delay.h>
#    define wait_ms(ms) _delay_ms(ms)
#    define wait_us(us) _delay_us(us)
#    define wait_cpuclock(x) __builtin_avr_delay_cycles(x)
#elif defined PROTOCOL_CHIBIOS
#    include <ch.h>
#    define wait_ms(ms)                     \