// AVR gamecube controller interface
// Copyright (C) 2014 Florian Jung
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see .
/* Setup:
lfuse: 0xe0
hfuse: 0xd9
attach a 12MHz quartz to the atmega8, using the appropriate XTAL pins.
supply 3.3V to the 3.3V line of the gamecube controller.
supply 5V to the 5V line of the gamecube controller.
connect the GND and SHIELD lines of the controller to the common GND.
connect the DATA line of the gamecube controller to the C5 pin.
!!! ensure that the PORTC has bit 0 cleared at any time. !!!
!!! otherwise, +5V is supplied to the pin; the gamecube !!!
!!! controller, however, may shortcut the line down to !!!
!!! GND at any time, leading to HARDWARE DAMAGE. !!!
*/
#include
#include
static char buffer[300]; // contains the exploded bits, i.e. each bit sent/received occupies one byte here.
#define OUT0 "sbi %2, 5 \n" /* pull line to GND */
#define OUT1 "cbi %2, 5 \n" /* tristate the line */
#define WAIT10 "ldi r16, 1 \n rcall delay_loop \n"
#define WAIT34 "ldi r16, 9 \n rcall delay_loop \n"
#define WAIT10MINUS10 ""
#define WAIT34MINUS10 "ldi r16,5 \n rcall delay_loop \n nop \n nop \n"
/* send the first "len" bytes stored in "bytes" to the controller,
* then receive the reply and store it into the global "buffer". */
int send_recv_gc(char* bytes, int len)
{
/* Phase 1: Send the data.
* Phase 1.1: Explode the bits into buffer. buffer[0..7] will contain the bits of bytes[0],
* buffer[8..15] contains the bits of bytes[1], and so on. MSB first.
* Phase 1.2: Actually send the data.
*
* intermediate Phase: wait for data line to become high again.
*
* Phase 2: Receive the reply.
* Phase 2.1: Busy-loop until the line is pulled down for the first time.
* Phase 2.2: Actually receive now:
* A counter is set to 0x80.
* Busy-loop until the line becomes high. Increment the counter in each iteration.
* (The line is now high.)
* Busy-loop until the line becomes low again. Decrement the counter each iteration.
* (Done receiving the bit)
* Write out the counter to buffer[], and proceed with the next bit.
*
* If the counter over- or underflows, stop receiving, because the line
* seems to be idle again (i.e. data transfer is finished). That's a timeout
* of ca. 53 us. (when running at 12 MHz)
*
* buffer[] now contains the counter values of the bits.
* if (buffer[42] > 0x80), i.e. (buffer[42] & 0x80), then the line was longer LOW than HIGH -> bit42 = 0
* otherwise, the line was longer HIGH than LOW -> bit42 = 1.
*/
// The NOPs are there because of symmetry reasons.
// the "// 2" comments after the assembly directives are the number of cycles this
// instruction will take to execute.
char* buf_ptr = buffer;
/****** SEND PART ******/
int k=0;
for (int i=0; i>1, k++)
buffer[k] = ((bytes[i] & j)!=0)?1:0 ;
len=len*8+1;
asm volatile(
"push r31 ; save Z\n"
"push r30 \n\n"
";;;;;;;; SEND PART ;;;;;;;;\n\n"
"send_loop: \n"
"sbiw %3, 1 \n" // 2
"breq send_done \n" // 1 if not done, 2 if done
"ld r16, z+ \n" // 2
"tst r16 \n" // 1
"brne send_one \n\n" // 1 if zero, 2 if nonzero
"; otherwise, send zero \n"
"nop \n" // 1
OUT0
WAIT34
OUT1
WAIT10MINUS10
"rjmp send_loop \n\n" // 2
"send_one: \n"
OUT0
WAIT10
OUT1
WAIT34MINUS10
"rjmp send_loop \n\n"
"delay_loop:\n"
"; this costs 7 + 3*(r16) cycles \n"
"dec r16\n"
"brne delay_loop\n"
"ret\n\n"
"send_done: \n"
"; now send the stop bit and release the line \n"
"nop \n nop \n nop \n"
OUT0
"; instead of WAIT10, do sensible work \n"
"pop r30 ; restore Z \n" // 2
"pop r31 \n" // 2
"clr r16 \n" // 1
"nop \n nop \n nop \n nop \n nop \n" // 5
OUT1
"; done :) \n\n\n"
"; now the final thing is to wait for DATA become high again (should be immediately anyway) \n"
"send_final_loop: \n"
"inc r16 \n"
"breq timeout \n"
"sbis %1, 5 \n"
"rjmp send_final_loop \n\n\n"
";;;;;;;; RECEIVE PART ;;;;;;;;\n\n"
"clr r16 \n"
"recv_wait_low_initial_loop: \n"
"inc r16 \n"
"breq timeout \n"
"sbic %1, 5 \n"
"rjmp recv_wait_low_initial_loop \n" // from low to the start of counting: 6 cycles = 0.5us
"nop \n"
"nop \n"
"nop \n"
"nop \n"
"recv_loop: \n"
"recv_low_begin: \n"
"ldi r16, 128 \n" // 1
"recv_low_loop: \n"
"inc r16 \n" // 1
"breq timeout \n" // 1 if no timeout
"sbis %1, 5 \n" // 1 // von high auf dec: 6
"rjmp recv_low_loop \n\n" // 2 if executed, 1 if skipped.
"nop \n" // to account for the rjmp recv_loop below.
"nop \n"
"nop \n"
"nop \n"
"recv_high_begin: \n"
"nop \n" // to account for the ldi in recv_low_begin.
"recv_high_loop: \n"
"dec r16 \n" // 1
"breq timeout \n" // 1 if no timeout
"sbic %1, 5 \n" // 1 // von low auf inc: 6
"rjmp recv_high_loop \n\n" // 2 if executed, 1 if skipped
"st z+, r16 \n" // 2
"rjmp recv_loop \n\n" // 2
"timeout: \n"
: "+z" ((unsigned char volatile*) buf_ptr)
: "I" (_SFR_IO_ADDR(PINC)),
"I" (_SFR_IO_ADDR(DDRC)),
"w" (len)
: "r16", "memory"
);
return buf_ptr-buffer;
// a value of >=128 means "0", <127 means "1" bit.
}
int main (void)
{
DDRC=0x00;
PORTC=0x00;
DDRB=0xFF;
DDRD=0x00;
PORTB=0x55;
int temp=0;
int n_received;
while(1)
{
temp++;
if (!(PIND & 0x08)) // check if uC is hung up
{
//PORTB=~(1 << ((temp>>9)%6));
PORTB=~temp;
}
else if (!(PIND & 0x20)) // debug num_received
{
PORTB=~n_received>>3;
}
else if (!(PIND & 0x40)) // clear debug output
{
PORTB=~0x00;
}
else
{
// decode "buffer" and write button states to PORTB
//PORTB=~buffer[4];
unsigned char tmp2=0;
for (int i=0;i<8;i++)
tmp2|= ( (buffer[55-i]&0x80)?0:(1< tmp2 )
tmp|=1;
PORTB=~tmp;
}
_delay_ms(0.3);
char foo[] = { 0x40, 0x03, 0x02 };
if (!(PIND & 0x10))
foo[2]=0x03;
n_received=send_recv_gc(foo, 3);
}
return 0; // never reached
}