Mercurial > hg > beaglert
view pru_rtaudio.p @ 39:638bc1ae2500 staging
Improved readibility of the DIGITAL code in the PRU, using register names instead of aliases and expanding some of the macros, removing unused macros. Binaries were not modified
| author | Giulio Moro <giuliomoro@yahoo.it> |
|---|---|
| date | Wed, 13 May 2015 12:18:10 +0100 |
| parents | a9af130097e8 |
| children | 419ce4ebfc4c |
line wrap: on
line source
.origin 0 .entrypoint START #define DBOX_CAPE // Define this to use new cape hardware #define CLOCK_BASE 0x44E00000 #define CLOCK_SPI0 0x4C #define CLOCK_SPI1 0x50 #define CLOCK_L4LS 0x60 #define SPI0_BASE 0x48030100 #define SPI1_BASE 0x481A0100 #define SPI_BASE SPI0_BASE #define SPI_SYSCONFIG 0x10 #define SPI_SYSSTATUS 0x14 #define SPI_MODULCTRL 0x28 #define SPI_CH0CONF 0x2C #define SPI_CH0STAT 0x30 #define SPI_CH0CTRL 0x34 #define SPI_CH0TX 0x38 #define SPI_CH0RX 0x3C #define SPI_CH1CONF 0x40 #define SPI_CH1STAT 0x44 #define SPI_CH1CTRL 0x48 #define SPI_CH1TX 0x4C #define SPI_CH1RX 0x50 #define GPIO0 0x44E07000 #define GPIO1 0x4804C000 #define GPIO_CLEARDATAOUT 0x190 #define GPIO_SETDATAOUT 0x194 #define PRU0_ARM_INTERRUPT 19 #define C_ADC_DAC_MEM C24 // PRU0 mem #ifdef DBOX_CAPE #define DAC_GPIO GPIO0 #define DAC_CS_PIN (1<<5) // GPIO0:5 = P9 pin 17 #else #define DAC_GPIO GPIO1 #define DAC_CS_PIN (1<<16) // GPIO1:16 = P9 pin 15 #endif #define DAC_TRM 0 // SPI transmit and receive #define DAC_WL 32 // Word length #define DAC_CLK_MODE 1 // SPI mode #define DAC_CLK_DIV 1 // Clock divider (48MHz / 2^n) #define DAC_DPE 1 // d0 = receive, d1 = transmit #define AD5668_COMMAND_OFFSET 24 #define AD5668_ADDRESS_OFFSET 20 #define AD5668_DATA_OFFSET 4 #define AD5668_REF_OFFSET 0 #ifdef DBOX_CAPE #define ADC_GPIO GPIO1 #define ADC_CS_PIN (1<<16) // GPIO1:16 = P9 pin 15 #else #define ADC_GPIO GPIO1 #define ADC_CS_PIN (1<<17) // GPIO1:17 = P9 pin 23 #endif #define ADC_TRM 0 // SPI transmit and receive #define ADC_WL 16 // Word length #define ADC_CLK_MODE 0 // SPI mode #define ADC_CLK_DIV 1 // Clock divider (48MHz / 2^n) #define ADC_DPE 1 // d0 = receive, d1 = transmit #define AD7699_CFG_MASK 0xF120 // Mask for config update, unipolar, full BW #define AD7699_CHANNEL_OFFSET 9 // 7 bits offset of a 14-bit left-justified word #define AD7699_SEQ_OFFSET 3 // sequencer (0 = disable, 3 = scan all) #define SHARED_COMM_MEM_BASE 0x00010000 // Location where comm flags are written #define COMM_SHOULD_STOP 0 // Set to be nonzero when loop should stop #define COMM_CURRENT_BUFFER 4 // Which buffer we are on #define COMM_BUFFER_FRAMES 8 // How many frames per buffer #define COMM_SHOULD_SYNC 12 // Whether to synchronise to an external clock #define COMM_SYNC_ADDRESS 16 // Which memory address to find the GPIO on #define COMM_SYNC_PIN_MASK 20 // Which pin to read for the sync #define COMM_LED_ADDRESS 24 // Which memory address to find the status LED on #define COMM_LED_PIN_MASK 28 // Which pin to write to change LED #define COMM_FRAME_COUNT 32 // How many frames have elapse since beginning #define COMM_USE_SPI 36 // Whether or not to use SPI ADC and DAC #define COMM_NUM_CHANNELS 40 // Low 2 bits indicate 8 [0x3], 4 [0x1] or 2 [0x0] channels #define COMM_USE_DIGITAL 44 // Whether or not to use DIGITAL #define MCASP0_BASE 0x48038000 #define MCASP1_BASE 0x4803C000 #define MCASP_PWRIDLESYSCONFIG 0x04 #define MCASP_PFUNC 0x10 #define MCASP_PDIR 0x14 #define MCASP_PDOUT 0x18 #define MCASP_PDSET 0x1C #define MCASP_PDIN 0x1C #define MCASP_PDCLR 0x20 #define MCASP_GBLCTL 0x44 #define MCASP_AMUTE 0x48 #define MCASP_DLBCTL 0x4C #define MCASP_DITCTL 0x50 #define MCASP_RGBLCTL 0x60 #define MCASP_RMASK 0x64 #define MCASP_RFMT 0x68 #define MCASP_AFSRCTL 0x6C #define MCASP_ACLKRCTL 0x70 #define MCASP_AHCLKRCTL 0x74 #define MCASP_RTDM 0x78 #define MCASP_RINTCTL 0x7C #define MCASP_RSTAT 0x80 #define MCASP_RSLOT 0x84 #define MCASP_RCLKCHK 0x88 #define MCASP_REVTCTL 0x8C #define MCASP_XGBLCTL 0xA0 #define MCASP_XMASK 0xA4 #define MCASP_XFMT 0xA8 #define MCASP_AFSXCTL 0xAC #define MCASP_ACLKXCTL 0xB0 #define MCASP_AHCLKXCTL 0xB4 #define MCASP_XTDM 0xB8 #define MCASP_XINTCTL 0xBC #define MCASP_XSTAT 0xC0 #define MCASP_XSLOT 0xC4 #define MCASP_XCLKCHK 0xC8 #define MCASP_XEVTCTL 0xCC #define MCASP_SRCTL0 0x180 #define MCASP_SRCTL1 0x184 #define MCASP_SRCTL2 0x188 #define MCASP_SRCTL3 0x18C #define MCASP_SRCTL4 0x190 #define MCASP_SRCTL5 0x194 #define MCASP_XBUF0 0x200 #define MCASP_XBUF1 0x204 #define MCASP_XBUF2 0x208 #define MCASP_XBUF3 0x20C #define MCASP_XBUF4 0x210 #define MCASP_XBUF5 0x214 #define MCASP_RBUF0 0x280 #define MCASP_RBUF1 0x284 #define MCASP_RBUF2 0x288 #define MCASP_RBUF3 0x28C #define MCASP_RBUF4 0x290 #define MCASP_RBUF5 0x294 #define MCASP_WFIFOCTL 0x1000 #define MCASP_WFIFOSTS 0x1004 #define MCASP_RFIFOCTL 0x1008 #define MCASP_RFIFOSTS 0x100C #define MCASP_XSTAT_XDATA_BIT 5 // Bit to test for transmit ready #define MCASP_RSTAT_RDATA_BIT 5 // Bit to test for receive ready // Constants used for this particular audio setup #define MCASP_BASE MCASP0_BASE #ifdef DBOX_CAPE #define MCASP_SRCTL_X MCASP_SRCTL2 // Ser. 2 is transmitter #define MCASP_SRCTL_R MCASP_SRCTL0 // Ser. 0 is receiver #define MCASP_XBUF MCASP_XBUF2 #define MCASP_RBUF MCASP_RBUF0 #else #define MCASP_SRCTL_X MCASP_SRCTL3 // Ser. 3 is transmitter #define MCASP_SRCTL_R MCASP_SRCTL2 // Ser. 2 is receiver #define MCASP_XBUF MCASP_XBUF3 #define MCASP_RBUF MCASP_RBUF2 #endif #define MCASP_PIN_AFSX (1 << 28) #define MCASP_PIN_AHCLKX (1 << 27) #define MCASP_PIN_ACLKX (1 << 26) #define MCASP_PIN_AMUTE (1 << 25) // Also, 0 to 3 are XFR0 to XFR3 #ifdef DBOX_CAPE #define MCASP_OUTPUT_PINS MCASP_PIN_AHCLKX | (1 << 2) // AHCLKX and AXR2 outputs #else #define MCASP_OUTPUT_PINS (1 << 3) // Which pins are outputs #endif #define MCASP_DATA_MASK 0xFFFF // 16 bit data #define MCASP_DATA_FORMAT 0x807C // MSB first, 0 bit delay, 16 bits, CFG bus, ROR 16bits #define C_MCASP_MEM C28 // Shared PRU mem // Flags for the flags register #define FLAG_BIT_BUFFER1 0 #define FLAG_BIT_USE_SPI 1 #define FLAG_BIT_MCASP_HWORD 2 // Whether we are on the high word for McASP transmission #define FLAG_BIT_USE_DIGITAL 3 // Registers used throughout // r1, r2, r3 are used for temporary storage #define MEM_DIGITAL_BASE 0x11000 //Base address for DIGITAL : Shared RAM + 0x400 #define MEM_DIGITAL_BUFFER1_OFFSET 0x400 //Start pointer to DIGITAL_BUFFER1, which is 256 words after. // 256 is the maximum number of frames allowed #define r27_current r6 // Pointer to current storage location of DIGITAL #define reg_num_channels r9 // Number of SPI ADC/DAC channels to use #define reg_frame_current r10 // Current frame count in SPI ADC/DAC transfer #define reg_frame_total r11 // Total frame count for SPI ADC/DAC #define reg_dac_data r12 // Current dword for SPI DAC #define reg_adc_data r13 // Current dword for SPI ADC #define reg_mcasp_dac_data r14 // Current dword for McASP DAC #define reg_mcasp_adc_data r15 // Current dword for McASP ADC #define reg_dac_buf0 r16 // Start pointer to SPI DAC buffer 0 #define reg_dac_buf1 r17 // Start pointer to SPI DAC buffer 1 #define reg_dac_current r18 // Pointer to current storage location of SPI DAC #define reg_adc_current r19 // Pointer to current storage location of SPI ADC #define reg_mcasp_buf0 r20 // Start pointer to McASP DAC buffer 0 #define reg_mcasp_buf1 r21 // Start pointer to McASP DAC buffer 1 #define reg_mcasp_dac_current r22 // Pointer to current storage location of McASP DAC #define reg_mcasp_adc_current r23 // Pointer to current storage location of McASP ADC #define reg_flags r24 // Buffer ID (0 and 1) and other flags #define reg_comm_addr r25 // Memory address for communicating with ARM #define reg_spi_addr r26 // Base address for SPI // r27, r28 used in macros #define reg_mcasp_addr r29 // Base address for McASP // DIGITAL new code starts here //0 P8_07 36 0x890/090 66 gpio2[2] //1 P8_08 37 0x894/094 67 gpio2[3] //2 P8_09 39 0x89c/09c 69 gpio2[5] //3 P8_10 38 0x898/098 68 gpio2[4] //4 P8_11 13 0x834/034 45 gpio1[13] //5 P8_12 12 0x830/030 44 gpio1[12] //6 P9_12 30 0x878/078 60 gpio1[28] //7 P9_14 18 0x848/048 50 gpio1[18] //8 P8_15 15 0x83c/03c 47 gpio1[15] //9 P8_16 14 0x838/038 46 gpio1[14] //10 P9_16 19 0x84c/04c 51 gpio1[19] //11 P8_18 35 0x88c/08c 65 gpio2[1] //12 P8_27 56 0x8e0/0e0 86 gpio2[22] //13 P8_28 58 0x8e8/0e8 88 gpio2[24] //14 P8_29 57 0x8e4/0e4 87 gpio2[23] //15 P8_30 59 0x8ec/0ec 89 gpio2[25] //TODO during initialization, set the pinmuxers to mode 7, input enable //generic GPIOs constants //#define GPIO1 0x4804c000 #define GPIO2 0x481ac000 //#define GPIO_CLEARDATAOUT 0x190 //SETDATAOUT is CLEARDATAOUT+4 #define GPIO_OE 0x134 #define GPIO_DATAIN 0x138 //gpioX_cleardataout must be adjacent to gpioX_setdataout #define reg_gpio1_oe r2 #define reg_gpio2_oe r3 #define reg_gpio2_cleardataout r4 #define reg_gpio2_setdataout r5 #define reg_gpio1_cleardataout r7 #define reg_gpio1_setdataout r8 #define reg_digital r27 //will first contain digitalOut from render() and digitalIn to render() later //aliases #define reg_gpio1_datain r2 #define reg_gpio2_datain r3 .macro READ_GPIO_BITS .mparam gpio_data, gpio_num_bit, digital_bit, digital QBBC DONE, digital, digital_bit //if the pin is set as an output, nothing to do here QBBC CLEAR, gpio_data, gpio_num_bit SET digital, digital_bit+16 QBA DONE CLEAR: CLR digital, digital_bit+16 QBA DONE DONE: .endm .macro SET_GPIO_BITS .mparam gpio_oe, gpio_setdataout, gpio_cleardataout, gpio_num_bit, digital_bit, digital //sets the bits in GPIO_OE, GPIO_SETDATAOUT and GPIO_CLEARDATAOUT //Remember that the GPIO_OE Output data enable register behaves as follows for each bit: //0 = The corresponding GPIO pin is configured as an output. //1 = The corresponding GPIO pin is configured as an input. QBBS SETINPUT, digital, digital_bit CLR gpio_oe, gpio_num_bit //if it is an output, configure pin as output QBBC CLEARDATAOUT, digital, digital_bit+16 // check the output value. If it is 0, branch SET gpio_setdataout, gpio_num_bit //if it is 1, set output to high QBA DONE CLEARDATAOUT: SET gpio_cleardataout, gpio_num_bit // set output to low QBA DONE SETINPUT: //if it is an input, set the relevant bit SET gpio_oe, gpio_num_bit QBA DONE DONE: .endm QBA START // when first starting, go to START, skipping this section. DIGITAL: //IMPORTANT: do NOT use r28 in this macro, as it contains the return address for JAL //r27 is now the input word passed in render(), one word per frame //[31:16]: data(1=high, 0=low), [15:0]: direction (0=output, 1=input) ) //Preparing the gpio_oe, gpio_cleardataout and gpio_setdataout for each module //r2 will hold GPIO1_OE //load current status of GPIO_OE in r2 MOV r2, GPIO1 | GPIO_OE //it takes 190ns to go through the next instruction LBBO r2, r2, 0, 4 //GPIO1-start //process oe and datain and prepare dataout for GPIO1 //r7 will contain GPIO1_CLEARDATAOUT //r8 will contain GPIO1_SETDATAOUT MOV r8, 0 MOV r7, 0 //map GPIO_ANALOG to gpio1 pins, //r2 is gpio1_oe, r8 is gpio1_setdataout, r7 is gpio1_cleardataout, r27 is the input word //the following operations will read from r27 and update r2,r7,r8 SET_GPIO_BITS r2, r8, r7, 13, 4, r27 SET_GPIO_BITS r2, r8, r7, 12, 5, r27 SET_GPIO_BITS r2, r8, r7, 28, 6, r27 SET_GPIO_BITS r2, r8, r7, 18, 7, r27 SET_GPIO_BITS r2, r8, r7, 15, 8, r27 SET_GPIO_BITS r2, r8, r7, 14, 9, r27 SET_GPIO_BITS r2, r8, r7, 19, 10, r27 //set the output enable register for gpio1. MOV r3, GPIO1 | GPIO_OE //use r3 as a temp register SBBO r2, r3, 0, 4 //takes two cycles (10ns) //GPIO1-end // r2 is now unused //GPIO2-start //r3 will hold GPIO1_OE //load current status of GPIO_OE in r3 MOV r3, GPIO2 | GPIO_OE //it takes 200ns to go through the next instructions LBBO r3, r3, 0, 4 //process oe and datain and prepare dataout for GPIO2 //r4 will contain GPIO2_CLEARDATAOUT //r5 will contain GPIO2_SETDATAOUT MOV r5, 0 MOV r4, 0 //map GPIO_ANALOG to gpio2 pins //r3 is gpio2_oe, r5 is gpio2_setdataout, r4 is gpio2_cleardataout, r27 is the input word //the following operations will read from r27 and update r3,r4,r5 SET_GPIO_BITS r3, r5, r4, 2, 0, r27 SET_GPIO_BITS r3, r5, r4, 3, 1, r27 SET_GPIO_BITS r3, r5, r4, 5, 2, r27 SET_GPIO_BITS r3, r5, r4, 4, 3, r27 SET_GPIO_BITS r3, r5, r4, 1, 11, r27 SET_GPIO_BITS r3, r5, r4, 22, 12, r27 SET_GPIO_BITS r3, r5, r4, 24, 13, r27 SET_GPIO_BITS r3, r5, r4, 23, 14, r27 SET_GPIO_BITS r3, r5, r4, 25, 15, r27 //set the output enable register for gpio2. MOV r2, GPIO2 | GPIO_OE //use r2 as a temp registerp SBBO r3, r2, 0, 4 //takes two cycles (10ns) //GPIO2-end //r3 is now unused //load current inputs in r2, r3 //r2 will contain GPIO1_DATAIN //r3 will contain GPIO2_DATAIN //load the memory locations MOV r2, GPIO1 | GPIO_DATAIN MOV r3, GPIO2 | GPIO_DATAIN //takes 375 nns to go through the next two instructions //read the datain LBBO r2, r2, 0, 4 LBBO r3, r3, 0, 4 //now read from r2 and r3 only the channels that are set as input in the lower word of r27 // and set their value in the high word of r27 //GPIO1 READ_GPIO_BITS r2, 13, 4, r27 READ_GPIO_BITS r2, 12, 5, r27 READ_GPIO_BITS r2, 28, 6, r27 READ_GPIO_BITS r2, 18, 7, r27 READ_GPIO_BITS r2, 15, 8, r27 READ_GPIO_BITS r2, 14, 9, r27 READ_GPIO_BITS r2, 19, 10, r27 //GPIO2 READ_GPIO_BITS r3, 2, 0, r27 READ_GPIO_BITS r3, 3, 1, r27 READ_GPIO_BITS r3, 5, 2, r27 READ_GPIO_BITS r3, 4, 3, r27 READ_GPIO_BITS r3, 1, 11, r27 READ_GPIO_BITS r3, 22, 12, r27 READ_GPIO_BITS r3, 24, 13, r27 READ_GPIO_BITS r3, 23, 14, r27 READ_GPIO_BITS r3, 25, 15, r27 //r2, r3 are now unused //now all the setdataout and cleardataout are ready to be written to the GPIO register. //CLEARDATAOUT and SETDATAOUT are consecutive positions in memory, so we just write 8 bytes to CLEARDATAOUT. //We can do this because we chose cleardata and setdata registers for a given GPIO to be consecutive //load the memory addresses to be written to MOV r2, GPIO1 | GPIO_CLEARDATAOUT //use r2 as a temp register MOV r3, GPIO2 | GPIO_CLEARDATAOUT //use r3 as a temp register //write 8 bytes for each GPIO //takes 30ns in total to go through the following two instructions SBBO r7, r2, 0, 8 //store r7 and r8 in GPIO1_CLEARDATAOUT and GPIO1_SETDATAOUT //takes 145ns to be effective when going low, 185ns when going high SBBO r4, r3, 0, 8 //store r4 and r5 in GPIO2_CLEARDATAOUT and GPIO2_SETDATAOUT //takes 95ns to be effective when going low, 130ns when going high //reversing the order of the two lines above will swap the performances between the GPIO modules //i.e.: the first line will always take 145ns/185ns and the second one will always take 95ns/130ns, //regardless of whether the order is gpio1-gpio2 or gpio2-gpio1 JMP r28.w0 // go back to ADC_WRITE_AND_PROCESS_GPIO .macro HANG //useful for debugging DALOOP: set r30.t14 clr r30.t14 QBA DALOOP .endm // Bring CS line low to write to DAC .macro DAC_CS_ASSERT MOV r27, DAC_CS_PIN MOV r28, DAC_GPIO + GPIO_CLEARDATAOUT SBBO r27, r28, 0, 4 .endm // Bring CS line high at end of DAC transaction .macro DAC_CS_UNASSERT MOV r27, DAC_CS_PIN MOV r28, DAC_GPIO + GPIO_SETDATAOUT SBBO r27, r28, 0, 4 .endm // Write to DAC TX register .macro DAC_TX .mparam data SBBO data, reg_spi_addr, SPI_CH0TX, 4 .endm // Wait for SPI to finish (uses RXS indicator) .macro DAC_WAIT_FOR_FINISH LOOP: LBBO r27, reg_spi_addr, SPI_CH0STAT, 4 QBBC LOOP, r27, 0 .endm // Read the RX word to clear .macro DAC_DISCARD_RX LBBO r27, reg_spi_addr, SPI_CH0RX, 4 .endm // Complete DAC write with chip select .macro DAC_WRITE .mparam reg DAC_CS_ASSERT DAC_TX reg DAC_WAIT_FOR_FINISH DAC_CS_UNASSERT DAC_DISCARD_RX .endm // Bring CS line low to write to ADC .macro ADC_CS_ASSERT MOV r27, ADC_CS_PIN MOV r28, ADC_GPIO + GPIO_CLEARDATAOUT SBBO r27, r28, 0, 4 .endm // Bring CS line high at end of ADC transaction .macro ADC_CS_UNASSERT MOV r27, ADC_CS_PIN MOV r28, ADC_GPIO + GPIO_SETDATAOUT SBBO r27, r28, 0, 4 .endm // Write to ADC TX register .macro ADC_TX .mparam data SBBO data, reg_spi_addr, SPI_CH1TX, 4 .endm // Wait for SPI to finish (uses RXS indicator) .macro ADC_WAIT_FOR_FINISH LOOP: LBBO r27, reg_spi_addr, SPI_CH1STAT, 4 QBBC LOOP, r27, 0 .endm // Read the RX word to clear; store output .macro ADC_RX .mparam data LBBO data, reg_spi_addr, SPI_CH1RX, 4 .endm // Complete ADC write+read with chip select .macro ADC_WRITE .mparam in, out ADC_CS_ASSERT ADC_TX in ADC_WAIT_FOR_FINISH ADC_RX out ADC_CS_UNASSERT .endm // Complete ADC write+read with chip select and also performs IO for digital .macro ADC_WRITE_GPIO .mparam in, out, do_gpio ADC_CS_ASSERT ADC_TX in QBBC GPIO_DONE, reg_flags, FLAG_BIT_USE_DIGITAL //skip if DIGITAL is disabled AND r27, do_gpio, 0x3 // only do a DIGITAL every 2 SPI I/O QBNE GPIO_DONE, r27, 0 //from here to GPIO_DONE takes 1.8us, while usually ADC_WAIT_FOR_FINISH only waits for 1.14us. //TODO: it would be better to split the DIGITAL stuff in two parts: //- one taking place during DAC_WRITE which sets the GPIO_OE //- and the other during ADC_WRITE which actually reads DATAIN and writes CLEAR/SET DATAOUT //r27 is actually r27, so do not use r27 from here to ... LBBO r27, r27_current, 0, 4 JAL r28.w0, DIGITAL // note that this is not called as a macro, but with JAL. r28 will contain the return address SBBO r27, r27_current, 0, 4 //..here you can start using r27 again ADD r27_current, r27_current, 4 //increment pointer GPIO_DONE: ADC_WAIT_FOR_FINISH ADC_RX out ADC_CS_UNASSERT .endm // Write a McASP register .macro MCASP_REG_WRITE .mparam reg, value MOV r27, value SBBO r27, reg_mcasp_addr, reg, 4 .endm // Write a McASP register beyond the 0xFF boundary .macro MCASP_REG_WRITE_EXT .mparam reg, value MOV r27, value MOV r28, reg ADD r28, reg_mcasp_addr, r28 SBBO r27, r28, 0, 4 .endm // Read a McASP register .macro MCASP_REG_READ .mparam reg, value LBBO value, reg_mcasp_addr, reg, 4 .endm // Read a McASP register beyond the 0xFF boundary .macro MCASP_REG_READ_EXT .mparam reg, value MOV r28, reg ADD r28, reg_mcasp_addr, r28 LBBO value, r28, 0, 4 .endm // Set a bit and wait for it to come up .macro MCASP_REG_SET_BIT_AND_POLL .mparam reg, mask MOV r27, mask LBBO r28, reg_mcasp_addr, reg, 4 OR r28, r28, r27 SBBO r28, reg_mcasp_addr, reg, 4 POLL: LBBO r28, reg_mcasp_addr, reg, 4 AND r28, r28, r27 QBEQ POLL, r28, 0 .endm START: // Set up c24 and c25 offsets with CTBIR register // Thus C24 points to start of PRU0 RAM MOV r3, 0x22020 // CTBIR0 MOV r2, 0 SBBO r2, r3, 0, 4 // Set up c28 pointer offset for shared PRU RAM MOV r3, 0x22028 // CTPPR0 MOV r2, 0x00000120 // To get address 0x00012000 SBBO r2, r3, 0, 4 // Load useful registers for addressing SPI MOV reg_comm_addr, SHARED_COMM_MEM_BASE MOV reg_spi_addr, SPI_BASE MOV reg_mcasp_addr, MCASP_BASE // Set ARM such that PRU can write to registers LBCO r0, C4, 4, 4 CLR r0, r0, 4 SBCO r0, C4, 4, 4 // Clear flags MOV reg_flags, 0 // Default number of channels in case SPI disabled LDI reg_num_channels, 8 // Find out whether we should use DIGITAL LBBO r2, reg_comm_addr, COMM_USE_DIGITAL, 4 QBEQ DIGITAL_INIT_DONE, r2, 0 // if we use digital SET reg_flags, reg_flags, FLAG_BIT_USE_DIGITAL /* This block of code is not really needed, as the memory is initialized by ARM before the PRU is started. Will leave it here for future reference DIGITAL_INIT: //set the digital buffer to 0x0000ffff (all inputs), to prevent unwanted high outputs //the loop is unrolled by a factor of four just to take advantage of the speed of SBBO on larger byte bursts, but there is no real need for it MOV r2, 0x0000ffff //value to store. 0x0000ffff means all inputs MOV r3, MEM_DIGITAL_BASE //start of the digital buffer MOV r4, MEM_DIGITAL_BASE+2*MEM_DIGITAL_BUFFER1_OFFSET //end of the digital buffer DIGITAL_INIT_BUFFER_LOOP: SBBO r2, r3, 0, 4 ADD r3, r3, 4 //increment pointer QBGT DIGITAL_INIT_BUFFER_LOOP, r3, r4 //loop until we reach the end of the buffer */ DIGITAL_INIT_DONE: // Find out whether we should use SPI ADC and DAC LBBO r2, reg_comm_addr, COMM_USE_SPI, 4 QBEQ SPI_FLAG_CHECK_DONE, r2, 0 SET reg_flags, reg_flags, FLAG_BIT_USE_SPI SPI_FLAG_CHECK_DONE: // If we don't use SPI, then skip all this init QBBC SPI_INIT_DONE, reg_flags, FLAG_BIT_USE_SPI // Load the number of channels: valid values are 8, 4 or 2 LBBO reg_num_channels, reg_comm_addr, COMM_NUM_CHANNELS, 4 QBGT SPI_NUM_CHANNELS_LT8, reg_num_channels, 8 // 8 > num_channels ? LDI reg_num_channels, 8 // If N >= 8, N = 8 QBA SPI_NUM_CHANNELS_DONE SPI_NUM_CHANNELS_LT8: QBGT SPI_NUM_CHANNELS_LT4, reg_num_channels, 4 // 4 > num_channels ? LDI reg_num_channels, 4 // If N >= 4, N = 4 QBA SPI_NUM_CHANNELS_DONE SPI_NUM_CHANNELS_LT4: LDI reg_num_channels, 2 // else N = 2 SPI_NUM_CHANNELS_DONE: // Init SPI clock MOV r2, 0x02 MOV r3, CLOCK_BASE + CLOCK_SPI0 SBBO r2, r3, 0, 4 // Reset SPI and wait for finish MOV r2, 0x02 SBBO r2, reg_spi_addr, SPI_SYSCONFIG, 4 SPI_WAIT_RESET: LBBO r2, reg_spi_addr, SPI_SYSSTATUS, 4 QBBC SPI_WAIT_RESET, r2, 0 // Turn off SPI channels MOV r2, 0 SBBO r2, reg_spi_addr, SPI_CH0CTRL, 4 SBBO r2, reg_spi_addr, SPI_CH1CTRL, 4 // Set to master; chip select lines enabled (CS0 used for DAC) MOV r2, 0x00 SBBO r2, reg_spi_addr, SPI_MODULCTRL, 4 // Configure CH0 for DAC MOV r2, (3 << 27) | (DAC_DPE << 16) | (DAC_TRM << 12) | ((DAC_WL - 1) << 7) | (DAC_CLK_DIV << 2) | DAC_CLK_MODE | (1 << 6) SBBO r2, reg_spi_addr, SPI_CH0CONF, 4 // Configure CH1 for ADC MOV r2, (3 << 27) | (ADC_DPE << 16) | (ADC_TRM << 12) | ((ADC_WL - 1) << 7) | (ADC_CLK_DIV << 2) | ADC_CLK_MODE SBBO r2, reg_spi_addr, SPI_CH1CONF, 4 // Turn on SPI channels MOV r2, 0x01 SBBO r2, reg_spi_addr, SPI_CH0CTRL, 4 SBBO r2, reg_spi_addr, SPI_CH1CTRL, 4 // DAC power-on reset sequence MOV r2, (0x07 << AD5668_COMMAND_OFFSET) DAC_WRITE r2 // Initialise ADC MOV r2, AD7699_CFG_MASK | (0 << AD7699_CHANNEL_OFFSET) | (0 << AD7699_SEQ_OFFSET) ADC_WRITE r2, r2 // Enable DAC internal reference MOV r2, (0x08 << AD5668_COMMAND_OFFSET) | (0x01 << AD5668_REF_OFFSET) DAC_WRITE r2 // Read ADC ch0 and ch1: result is always 2 samples behind so start here MOV r2, AD7699_CFG_MASK | (0x00 << AD7699_CHANNEL_OFFSET) ADC_WRITE r2, r2 MOV r2, AD7699_CFG_MASK | (0x01 << AD7699_CHANNEL_OFFSET) ADC_WRITE r2, r2 SPI_INIT_DONE: // Prepare McASP0 for audio MCASP_REG_WRITE MCASP_GBLCTL, 0 // Disable McASP MCASP_REG_WRITE_EXT MCASP_SRCTL0, 0 // All serialisers off MCASP_REG_WRITE_EXT MCASP_SRCTL1, 0 MCASP_REG_WRITE_EXT MCASP_SRCTL2, 0 MCASP_REG_WRITE_EXT MCASP_SRCTL3, 0 MCASP_REG_WRITE_EXT MCASP_SRCTL4, 0 MCASP_REG_WRITE_EXT MCASP_SRCTL5, 0 MCASP_REG_WRITE MCASP_PWRIDLESYSCONFIG, 0x02 // Power on MCASP_REG_WRITE MCASP_PFUNC, 0x00 // All pins are McASP MCASP_REG_WRITE MCASP_PDIR, MCASP_OUTPUT_PINS // Set pin direction MCASP_REG_WRITE MCASP_DLBCTL, 0x00 MCASP_REG_WRITE MCASP_DITCTL, 0x00 MCASP_REG_WRITE MCASP_RMASK, MCASP_DATA_MASK // 16 bit data receive MCASP_REG_WRITE MCASP_RFMT, MCASP_DATA_FORMAT // Set data format MCASP_REG_WRITE MCASP_AFSRCTL, 0x100 // I2S mode MCASP_REG_WRITE MCASP_ACLKRCTL, 0x80 // Sample on rising edge MCASP_REG_WRITE MCASP_AHCLKRCTL, 0x8001 // Internal clock, not inv, /2; irrelevant? MCASP_REG_WRITE MCASP_RTDM, 0x03 // Enable TDM slots 0 and 1 MCASP_REG_WRITE MCASP_RINTCTL, 0x00 // No interrupts MCASP_REG_WRITE MCASP_XMASK, MCASP_DATA_MASK // 16 bit data transmit MCASP_REG_WRITE MCASP_XFMT, MCASP_DATA_FORMAT // Set data format MCASP_REG_WRITE MCASP_AFSXCTL, 0x100 // I2S mode MCASP_REG_WRITE MCASP_ACLKXCTL, 0x00 // Transmit on rising edge, sync. xmit and recv MCASP_REG_WRITE MCASP_AHCLKXCTL, 0x8001 // External clock from AHCLKX MCASP_REG_WRITE MCASP_XTDM, 0x03 // Enable TDM slots 0 and 1 MCASP_REG_WRITE MCASP_XINTCTL, 0x00 // No interrupts MCASP_REG_WRITE_EXT MCASP_SRCTL_R, 0x02 // Set up receive serialiser MCASP_REG_WRITE_EXT MCASP_SRCTL_X, 0x01 // Set up transmit serialiser MCASP_REG_WRITE_EXT MCASP_WFIFOCTL, 0x00 // Disable FIFOs MCASP_REG_WRITE_EXT MCASP_RFIFOCTL, 0x00 MCASP_REG_WRITE MCASP_XSTAT, 0xFF // Clear transmit errors MCASP_REG_WRITE MCASP_RSTAT, 0xFF // Clear receive errors MCASP_REG_SET_BIT_AND_POLL MCASP_RGBLCTL, (1 << 1) // Set RHCLKRST MCASP_REG_SET_BIT_AND_POLL MCASP_XGBLCTL, (1 << 9) // Set XHCLKRST // The above write sequence will have temporarily changed the AHCLKX frequency // The PLL needs time to settle or the sample rate will be unstable and possibly // cause an underrun. Give it ~1ms before going on. // 10ns per loop iteration = 10^-8s --> 10^5 iterations needed MOV r2, 1 << 28 MOV r3, GPIO1 + GPIO_SETDATAOUT SBBO r2, r3, 0, 4 MOV r2, 100000 MCASP_INIT_WAIT: SUB r2, r2, 1 QBNE MCASP_INIT_WAIT, r2, 0 MOV r2, 1 << 28 MOV r3, GPIO1 + GPIO_CLEARDATAOUT SBBO r2, r3, 0, 4 MCASP_REG_SET_BIT_AND_POLL MCASP_RGBLCTL, (1 << 0) // Set RCLKRST MCASP_REG_SET_BIT_AND_POLL MCASP_XGBLCTL, (1 << 8) // Set XCLKRST MCASP_REG_SET_BIT_AND_POLL MCASP_RGBLCTL, (1 << 2) // Set RSRCLR MCASP_REG_SET_BIT_AND_POLL MCASP_XGBLCTL, (1 << 10) // Set XSRCLR MCASP_REG_SET_BIT_AND_POLL MCASP_RGBLCTL, (1 << 3) // Set RSMRST MCASP_REG_SET_BIT_AND_POLL MCASP_XGBLCTL, (1 << 11) // Set XSMRST MCASP_REG_WRITE_EXT MCASP_XBUF, 0x00 // Write to the transmit buffer to prevent underflow MCASP_REG_SET_BIT_AND_POLL MCASP_RGBLCTL, (1 << 4) // Set RFRST MCASP_REG_SET_BIT_AND_POLL MCASP_XGBLCTL, (1 << 12) // Set XFRST // Initialisation LBBO reg_frame_total, reg_comm_addr, COMM_BUFFER_FRAMES, 4 // Total frame count (SPI; 0.5x-2x for McASP) MOV reg_dac_buf0, 0 // DAC buffer 0 start pointer LSL reg_dac_buf1, reg_frame_total, 1 // DAC buffer 1 start pointer = N[ch]*2[bytes]*bufsize LMBD r2, reg_num_channels, 1 // Returns 1, 2 or 3 depending on the number of channels LSL reg_dac_buf1, reg_dac_buf1, r2 // Multiply by 2, 4 or 8 to get the N[ch] scaling above MOV reg_mcasp_buf0, 0 // McASP DAC buffer 0 start pointer LSL reg_mcasp_buf1, reg_frame_total, r2 // McASP DAC buffer 1 start pointer = 2[ch]*2[bytes]*(N/4)[samples/spi]*bufsize CLR reg_flags, reg_flags, FLAG_BIT_BUFFER1 // Bit 0 holds which buffer we are on MOV r2, 0 SBBO r2, reg_comm_addr, COMM_FRAME_COUNT, 4 // Start with frame count of 0 /* This block of code is not really needed, as the memory is initialized by ARM before the PRU is started. Will leave it here for future reference //Initialise all SPI and audio buffers (DAC0, DAC1, ADC0, ADC1) to zero. //This is useful for analog outs so they do not have spikes during the first buffer. //This is not very useful for audio, as you still hear the initial "tumpf" when the converter starts //and each sample in the DAC buffer is reset to 0 after it is written to the DAC. QBBC SPI_INIT_BUFFER_DONE, reg_flags, FLAG_BIT_USE_SPI //Initialize SPI buffers //compute the memory offset of the end of the audio buffer and store it in r4 SUB r4, reg_dac_buf1, reg_dac_buf0 // length of the buffer, assumes reg_dac_buf1>ref_dac_buf0 LSL r4, r4, 2 //length of four buffers (DAC0, DAC1, ADC0, ADC1) ADD r4, reg_dac_buf0, r4 //total offset MOV r2, 0// value to store MOV r3, 0 // offset counter SPI_INIT_BUFFER_LOOP: SBCO r2, C_ADC_DAC_MEM, r3, 4 ADD r3, r3, 4 QBGT SPI_INIT_BUFFER_LOOP, r3, r4 SPI_INIT_BUFFER_DONE: //Initialize audio buffers //compute the memory offset of the end of the audio buffer and store it in r4 SUB r4, reg_mcasp_buf1, reg_mcasp_buf0 // length of the buffer, assumes reg_mcasp_buf1>ref_mcasp_buf0 LSL r4, r4, 2 //length of four buffers (DAC0, DAC1, ADC0, ADC1) ADD r4, reg_mcasp_buf0, r4 //total offset MOV r2, 0 // value to store MOV r3, 0 // offset counter MCASP_INIT_BUFFER_LOOP: SBCO r2, C_MCASP_MEM, r3, 4 ADD r3, r3, 4 QBGT MCASP_INIT_BUFFER_LOOP, r3, r4 */ // Here we are out of sync by one TDM slot since the 0 word transmitted above will have occupied // the first output slot. Send one more word before jumping into the loop. MCASP_DAC_WAIT_BEFORE_LOOP: LBBO r2, reg_mcasp_addr, MCASP_XSTAT, 4 QBBC MCASP_DAC_WAIT_BEFORE_LOOP, r2, MCASP_XSTAT_XDATA_BIT MCASP_REG_WRITE_EXT MCASP_XBUF, 0x00 // Likewise, read and discard the first sample we get back from the ADC. This keeps the DAC and ADC // in sync in terms of which TDM slot we are reading (empirically found that we should throw this away // rather than keep it and invert the phase) MCASP_ADC_WAIT_BEFORE_LOOP: LBBO r2, reg_mcasp_addr, MCASP_RSTAT, 4 QBBC MCASP_ADC_WAIT_BEFORE_LOOP, r2, MCASP_RSTAT_RDATA_BIT MCASP_REG_READ_EXT MCASP_RBUF, r2 WRITE_ONE_BUFFER: // Write a single buffer of DAC samples and read a buffer of ADC samples // Load starting positions MOV reg_dac_current, reg_dac_buf0 // DAC: reg_dac_current is current pointer LMBD r2, reg_num_channels, 1 // 1, 2 or 3 for 2, 4 or 8 channels LSL reg_adc_current, reg_frame_total, r2 LSL reg_adc_current, reg_adc_current, 2 // N * 2 * 2 * bufsize ADD reg_adc_current, reg_adc_current, reg_dac_current // ADC: starts N * 2 * 2 * bufsize beyond DAC MOV reg_mcasp_dac_current, reg_mcasp_buf0 // McASP: set current DAC pointer LSL reg_mcasp_adc_current, reg_frame_total, r2 // McASP ADC: starts (N/2)*2*2*bufsize beyond DAC LSL reg_mcasp_adc_current, reg_mcasp_adc_current, 1 ADC reg_mcasp_adc_current, reg_mcasp_adc_current, reg_mcasp_dac_current MOV reg_frame_current, 0 QBBS DIGITAL_BASE_CHECK_SET, reg_flags, FLAG_BIT_BUFFER1 //check which buffer we are using for DIGITAL // if we are here, we are using buffer0 MOV r27_current, MEM_DIGITAL_BASE QBA DIGITAL_BASE_CHECK_DONE DIGITAL_BASE_CHECK_SET: //if we are here, we are using buffer1 MOV r27_current, MEM_DIGITAL_BASE+MEM_DIGITAL_BUFFER1_OFFSET //so adjust offset appropriately DIGITAL_BASE_CHECK_DONE: WRITE_LOOP: // Write N channels to DAC from successive values in memory // At the same time, read N channels from ADC // Unrolled by a factor of 2 to get high and low words MOV r1, 0 ADC_DAC_LOOP: QBBC SPI_DAC_LOAD_DONE, reg_flags, FLAG_BIT_USE_SPI // Load next 2 SPI DAC samples and store zero in their place LBCO reg_dac_data, C_ADC_DAC_MEM, reg_dac_current, 4 MOV r2, 0 SBCO r2, C_ADC_DAC_MEM, reg_dac_current, 4 ADD reg_dac_current, reg_dac_current, 4 SPI_DAC_LOAD_DONE: // On even iterations, load two more samples and choose the first one // On odd iterations, transmit the second of the samples already loaded // QBBS MCASP_DAC_HIGH_WORD, r1, 1 QBBS MCASP_DAC_HIGH_WORD, reg_flags, FLAG_BIT_MCASP_HWORD MCASP_DAC_LOW_WORD: // Load next 2 Audio DAC samples and store zero in their place LBCO reg_mcasp_dac_data, C_MCASP_MEM, reg_mcasp_dac_current, 4 MOV r2, 0 SBCO r2, C_MCASP_MEM, reg_mcasp_dac_current, 4 ADD reg_mcasp_dac_current, reg_mcasp_dac_current, 4 // Mask out the low word (first in little endian) MOV r2, 0xFFFF AND r7, reg_mcasp_dac_data, r2 QBA MCASP_WAIT_XSTAT MCASP_DAC_HIGH_WORD: // Take the high word of the previously loaded data LSR r7, reg_mcasp_dac_data, 16 // Every 2 channels we send one audio sample; this loop already // sends exactly two SPI channels. // Wait for McASP XSTAT[XDATA] to set indicating we can write more data MCASP_WAIT_XSTAT: LBBO r2, reg_mcasp_addr, MCASP_XSTAT, 4 QBBC MCASP_WAIT_XSTAT, r2, MCASP_XSTAT_XDATA_BIT MCASP_REG_WRITE_EXT MCASP_XBUF, r7 // Same idea with ADC: even iterations, load the sample into the low word, odd // iterations, load the sample into the high word and store // QBBS MCASP_ADC_HIGH_WORD, r1, 1 QBBS MCASP_ADC_HIGH_WORD, reg_flags, FLAG_BIT_MCASP_HWORD MCASP_ADC_LOW_WORD: // Start ADC data at 0 LDI reg_mcasp_adc_data, 0 // Now wait for a received word to become available from the audio ADC MCASP_WAIT_RSTAT_LOW: LBBO r2, reg_mcasp_addr, MCASP_RSTAT, 4 QBBC MCASP_WAIT_RSTAT_LOW, r2, MCASP_RSTAT_RDATA_BIT // Mask low word and store in ADC data register MCASP_REG_READ_EXT MCASP_RBUF, r3 MOV r2, 0xFFFF AND reg_mcasp_adc_data, r3, r2 QBA MCASP_ADC_DONE MCASP_ADC_HIGH_WORD: // Wait for a received word to become available from the audio ADC MCASP_WAIT_RSTAT_HIGH: LBBO r2, reg_mcasp_addr, MCASP_RSTAT, 4 QBBC MCASP_WAIT_RSTAT_HIGH, r2, MCASP_RSTAT_RDATA_BIT // Read data and shift 16 bits to the left (into the high word) MCASP_REG_READ_EXT MCASP_RBUF, r3 LSL r3, r3, 16 OR reg_mcasp_adc_data, reg_mcasp_adc_data, r3 // Now store the result and increment the pointer SBCO reg_mcasp_adc_data, C_MCASP_MEM, reg_mcasp_adc_current, 4 ADD reg_mcasp_adc_current, reg_mcasp_adc_current, 4 MCASP_ADC_DONE: QBBC SPI_SKIP_WRITE, reg_flags, FLAG_BIT_USE_SPI // DAC: transmit low word (first in little endian) MOV r2, 0xFFFF AND r7, reg_dac_data, r2 LSL r7, r7, AD5668_DATA_OFFSET MOV r8, (0x03 << AD5668_COMMAND_OFFSET) OR r7, r7, r8 LSL r8, r1, AD5668_ADDRESS_OFFSET OR r7, r7, r8 DAC_WRITE r7 // Read ADC channels: result is always 2 commands behind // Start by reading channel 2 (result is channel 0) and go // to N+2, but masking the channel number to be between 0 and N-1 LDI reg_adc_data, 0 ADD r8, r1, 2 SUB r7, reg_num_channels, 1 AND r8, r8, r7 LSL r8, r8, AD7699_CHANNEL_OFFSET MOV r7, AD7699_CFG_MASK OR r7, r7, r8 //ssssssssssssssssssssssssssss ADC_WRITE_GPIO r7, r7, r1 // Mask out only the relevant 16 bits and store in reg_adc_data MOV r2, 0xFFFF AND reg_adc_data, r7, r2 // Increment channel index ADD r1, r1, 1 // DAC: transmit high word (second in little endian) LSR r7, reg_dac_data, 16 LSL r7, r7, AD5668_DATA_OFFSET MOV r8, (0x03 << AD5668_COMMAND_OFFSET) OR r7, r7, r8 LSL r8, r1, AD5668_ADDRESS_OFFSET OR r7, r7, r8 DAC_WRITE r7 // Read ADC channels: result is always 2 commands behind // Start by reading channel 2 (result is channel 0) and go // to N+2, but masking the channel number to be between 0 and N-1 ADD r8, r1, 2 SUB r7, reg_num_channels, 1 AND r8, r8, r7 LSL r8, r8, AD7699_CHANNEL_OFFSET MOV r7, AD7699_CFG_MASK OR r7, r7, r8 ADC_WRITE r7, r7 // Move this result up to the 16 high bits LSL r7, r7, 16 OR reg_adc_data, reg_adc_data, r7 // Store 2 ADC words in memory SBCO reg_adc_data, C_ADC_DAC_MEM, reg_adc_current, 4 ADD reg_adc_current, reg_adc_current, 4 // Toggle the high/low word for McASP control (since we send one word out of // 32 bits for each pair of SPI channels) XOR reg_flags, reg_flags, (1 << FLAG_BIT_MCASP_HWORD) // Repeat 4 times for 8 channels (2 samples per loop, r1 += 1 already happened) // For 4 or 2 channels, repeat 2 or 1 times, according to flags ADD r1, r1, 1 QBNE ADC_DAC_LOOP, r1, reg_num_channels QBA ADC_DAC_LOOP_DONE SPI_SKIP_WRITE: // We get here only if the SPI ADC and DAC are disabled // Just keep the loop going for McASP // Toggle the high/low word for McASP control (since we send one word out of // 32 bits for each pair of SPI channels) XOR reg_flags, reg_flags, (1 << FLAG_BIT_MCASP_HWORD) ADD r1, r1, 2 QBNE ADC_DAC_LOOP, r1, reg_num_channels ADC_DAC_LOOP_DONE: // Increment number of frames, see if we have more to write ADD reg_frame_current, reg_frame_current, 1 QBNE WRITE_LOOP, reg_frame_current, reg_frame_total WRITE_LOOP_DONE: // Now done, swap the buffers and do the next one // Use r2 as a temp register MOV r2, reg_dac_buf0 MOV reg_dac_buf0, reg_dac_buf1 MOV reg_dac_buf1, r2 MOV r2, reg_mcasp_buf0 MOV reg_mcasp_buf0, reg_mcasp_buf1 MOV reg_mcasp_buf1, r2 XOR reg_flags, reg_flags, (1 << FLAG_BIT_BUFFER1) //flip the buffer flag // Notify ARM of buffer swap AND r2, reg_flags, (1 << FLAG_BIT_BUFFER1) // Mask out every but low bit SBBO r2, reg_comm_addr, COMM_CURRENT_BUFFER, 4 // Increment the frame count in the comm buffer (for status monitoring) LBBO r2, reg_comm_addr, COMM_FRAME_COUNT, 4 ADD r2, r2, reg_frame_total SBBO r2, reg_comm_addr, COMM_FRAME_COUNT, 4 // If LED blink enabled, toggle every 4096 frames LBBO r3, reg_comm_addr, COMM_LED_ADDRESS, 4 QBEQ LED_BLINK_DONE, r3, 0 MOV r1, 0x1000 AND r2, r2, r1 // Test (frame count & 4096) QBEQ LED_BLINK_OFF, r2, 0 LBBO r2, reg_comm_addr, COMM_LED_PIN_MASK, 4 MOV r1, GPIO_SETDATAOUT ADD r3, r3, r1 // Address for GPIO set register SBBO r2, r3, 0, 4 // Set GPIO pin QBA LED_BLINK_DONE LED_BLINK_OFF: LBBO r2, reg_comm_addr, COMM_LED_PIN_MASK, 4 MOV r1, GPIO_CLEARDATAOUT ADD r3, r3, r1 // Address for GPIO clear register SBBO r2, r3, 0, 4 // Clear GPIO pin LED_BLINK_DONE: // Check if we should finish: flag is zero as long as it should run LBBO r2, reg_comm_addr, COMM_SHOULD_STOP, 4 QBEQ WRITE_ONE_BUFFER, r2, 0 CLEANUP: MCASP_REG_WRITE MCASP_GBLCTL, 0x00 // Turn off McASP // Turn off SPI if enabled QBBC SPI_CLEANUP_DONE, reg_flags, FLAG_BIT_USE_SPI MOV r3, SPI_BASE + SPI_CH0CONF LBBO r2, r3, 0, 4 CLR r2, r2, 13 CLR r2, r2, 27 SBBO r2, r3, 0, 4 MOV r3, SPI_BASE + SPI_CH0CTRL LBBO r2, r3, 0, 4 CLR r2, r2, 1 SBBO r2, r3, 0, 4 SPI_CLEANUP_DONE: // Signal the ARM that we have finished MOV R31.b0, PRU0_ARM_INTERRUPT + 16 HALT