[sw] add support for newlib's system calls

[stnolting]

This PR add support for RISC-V's newlib system calls by providing simple stubs for "OS" functions like fork() and write() and even printf() and scanf(). A new library file sw/lib/source/syscalls.c is added to provide these stubs.

Furthermore, this PR updates the processor's linker script (sw/common/neorv32.ld) by providing definitions for setting up the heap. Application software can now use standard dynamic memory allocation functions like malloc() and free().

---

📚 A new section in the data sheet ("RAM Layout") illustrates the default RAM usage.

⚠️ The concepts from this PR are still experimental - even though all tests using functions provided by syscalls.c seem to work fine. Any feedback/help is highly appreciated!

©️ The stubs for the system calls were copied (and modified) from @openhwgroup's https://github.com/openhwgroup/cv32e40p/blob/master/example_tb/core/custom/syscalls.c (project license: SOLDERPAD HARDWARE LICENSE version 0.51) and @newaetech's newaetech/chipwhisperer@df8e2fa

Further references on handling newlib system calls: https://interrupt.memfault.com/blog/boostrapping-libc-with-newlib#implementing-newlib

[stnolting]

✔️🚀 A minimal test program using dynamic memory allocation and printf:

#include <neorv32.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <unistd.h>

int main() {

  // initialize the neorv32 runtime environment
  // this will take care of handling all CPU traps (interrupts and exceptions)
  neorv32_rte_setup();

  // setup UART0 at default baud rate, no parity bits, no hw flow control
  neorv32_uart0_setup(BAUD_RATE, PARITY_NONE, FLOW_CONTROL_RTSCTS);

  // output via "STDOUT" (UART0.TX)
  printf("printf test\n");

  // allocate 4 bytes of memory
  char *char_buffer;
  char_buffer = (char *) malloc(4 * sizeof(char));

  // get 4 chars from "STDIN" (UART0.RX)
  read((int)stdin, char_buffer, 4 * sizeof(char));

  // output via UART0.TX
  neorv32_uart0_printf("%c %c %c %c\n", char_buffer[0], char_buffer[1], char_buffer[2], char_buffer[3]);

  free(char_buffer);

  return 0;
}

And the resulting output (when entering 1, 2, 3, 4 via the UART console):

printf test
1 2 3 4

⚠️ Dynamic memory should be used with caution to prevent collisions with the stack! The stack starts at the end of the RAM (like DMEM) and grows downwards. The heap start at the end of the initialized and un-initialized data sections (.data and .bss) and grows upwards. The default maximum heap size is 1/4 of the RAM (DMEM). This definition can be altered by explicitly setting the __heap_size symbol.

⚠️ printf() and scanf() are quite powerful functions and require a lot of memory. Furthermore, they use UART0 only. It is highly recommended to use the NEORV32-optimized functions like neorv32_uart[0/1]_printf().