mirror of
https://github.com/mcgurk/Arduino-USB-HID-RetroJoystickAdapter
synced 2024-11-25 10:32:19 -05:00
80 lines
4.0 KiB
Markdown
80 lines
4.0 KiB
Markdown
- https://www.protovision.games/hardw/4_player.php?language=en
|
|
|
|
```
|
|
& 'C:\Program Files (x86)\Arduino\hardware\tools\avr\bin\avr-objdump.exe' -S "C:\\Users\\lehti\\AppData\\Local\\Temp\\arduino_build_208947/c64_4joystick-adapter.ino.elf" > c:\temp\koe.txt
|
|
```
|
|
|
|
```
|
|
cli // 1 clock
|
|
PORTB = *ptr; // is this atomic? probably, because ptr is 6-bit pointer. nope...
|
|
f98: e0 91 26 01 lds r30, 0x0126 ; 0x800126 <__data_end> // 2 clocks
|
|
f9c: f0 91 27 01 lds r31, 0x0127 ; 0x800127 <__data_end+0x1> // 2 clocks
|
|
fa0: 80 81 ld r24, Z // 1 clock
|
|
fa2: 85 b9 out 0x05, r24 ; 5 // 1 clock
|
|
sei // 1 clock
|
|
```
|
|
500ns
|
|
|
|
```
|
|
cli
|
|
mov r31, r1
|
|
mov r30, gpior0
|
|
ld r24, Z
|
|
out 0x05, r24
|
|
sei
|
|
```
|
|
375ns
|
|
|
|
## avr
|
|
- https://ww1.microchip.com/downloads/en/DeviceDoc/Atmel-7766-8-bit-AVR-ATmega16U4-32U4_Datasheet.pdf
|
|
- https://ww1.microchip.com/downloads/en/DeviceDoc/Atmel-7810-Automotive-Microcontrollers-ATmega328P_Datasheet.pdf
|
|
- http://ww1.microchip.com/downloads/en/devicedoc/atmel-0856-avr-instruction-set-manual.pdf
|
|
|
|
## using data/i/o space
|
|
- lds can load from io (if +0x20 added to address) or data space
|
|
- sts can store from register to io (if +0x20 added to address) or data space
|
|
- in can load from io to register
|
|
- out can write from register to io
|
|
- only one io/data address space access within one instruction?
|
|
- lds/sts takes 2 cycles and in/out takes 1 cycle? ldi takes 1 cycle?
|
|
- lds/sts: "M" \_SFR_MEM_ADDR(GPIOR2), in/out: "I" \_SFR_IO_ADDR(GPIOR0)
|
|
|
|
## interrupts
|
|
- https://billgrundmann.wordpress.com/2009/03/02/the-overhead-of-arduino-interrupts/
|
|
- https://forum.arduino.cc/t/how-fast-can-i-interrupt/25884/5
|
|
```
|
|
This is from the datasheet for the AT90USB82 processor; things in parenthesis are from me...
|
|
|
|
-The interrupt execution response for all the enabled AVR interrupts is five clock cycles minimum (the processor is fixin' to execute the interrupt)
|
|
-The vector is normally a jump to the interrupt routine, and this jump takes three clock cycles (the processor jumps to the ISR)
|
|
-SREG must be saved and restored (the processor doesn't do this for us and SREG is important)
|
|
-A return from an interrupt handling routine takes three clock cycles
|
|
-When the AVR exits from an interrupt, it will always return to the main program and execute one more instruction before any pending interrupt is served
|
|
|
|
Those are the things necessary just to get the ISR called. We have not yet added the application stuff (incrementing an unsigned long in BetterSense's case).
|
|
|
|
Adding those up gives us 5+3+2+3+1 = 14. The absolute maximum number of interrupts per second that can be handled by the AT90USB82 is 16 million instructions per second / 14 instructions per interrupt = 1,142,857 interrupts per second.
|
|
```
|
|
|
|
## avr asm
|
|
- http://www.nongnu.org/avr-libc/user-manual/inline_asm.html
|
|
- https://ucexperiment.wordpress.com/2016/03/04/arduino-inline-assembly-tutorial-1/
|
|
- https://ucexperiment.wordpress.com/2016/03/11/arduino-inline-assembly-tutorial-5-2/
|
|
|
|
## atomic
|
|
- https://groups.google.com/a/arduino.cc/g/developers/c/cmu0Qy32zxY
|
|
- http://www.gammon.com.au/forum/?id=11488
|
|
- https://home.csulb.edu/~hill/ee346/Lectures/10%20ATmega32U4%20Interrupts.pdf
|
|
|
|
## register as variable
|
|
- https://forum.arduino.cc/index.php?topic=43760.5
|
|
- http://www.nongnu.org/avr-libc/user-manual/FAQ.html#faq_regbind
|
|
- volatile register unsigned char my_register asm("r2");
|
|
- https://www.avrfreaks.net/forum/binding-variable-register
|
|
|
|
## running code from ram
|
|
- https://forum.arduino.cc/index.php?topic=425962.0
|
|
- https://forum.arduino.cc/index.php?topic=470631.0
|
|
- "AVRs are Harvard architecture CPUs, so they CANNOT run code out of RAM, so that directive can't possibly do anything of any value whatsoever."
|
|
- "Yes, Harvard IS the reason. A Harvard CPU, by definition, has separate code and data memory spaces. They execute code from one memory, and fetch data from a different memory. What you're seeing is most likely the compiler pretending the directive can do what you want, but the linker doing the only thing it can do - putting that code in FLASH."
|