arm: working bootloader, example app

'make' will produce two outputs:
- bootloader.elf to be loaded into the bitstream
- app.bin to be loaded through the programmer

Loading app.bin is as simple as:
- reset the board
- `python3 prog.py app.bin`

arm/main.cc

@@ -1,30 +1,12 @@
-#include <cstdint>
-
-#include "xuartlite.h"
-
-struct Gpio {
-    volatile uint32_t data;
-};
-
-#define gpio0 ((Gpio*) 0x40000000)
+#include "gpio.h"
 
 namespace {
 
-constexpr uintptr_t kUart0BaseAddress = 0x40001000;
-XUartLite uart0;
-XUartLite_Config uart0_config = {
-    .DeviceId = 0,
-    .RegBaseAddr = kUart0BaseAddress,
-    .BaudRate = 115200,
-    .UseParity = false,
-    .DataBits = 8,
-};
-
 [[maybe_unused]]
 void sleep(int ms) {
     for (int i = 0; i < ms; i++) {
         // sleep for 1 ms
-        for (int j = 0; j < 50000; j++) {
+        for (int j = 0; j < 22000; j++) {
             asm("");
         }
     }
@@ -32,24 +14,19 @@ void sleep(int ms) {
 
 }  // namespace
 
-int main() {
-    int recv = 0;
-    gpio0->data = 37;
+extern "C" int main() {
+    uint8_t cnt = 0;
 
-    XUartLite_CfgInitialize(&uart0, &uart0_config, uart0_config.RegBaseAddr);
+    for (int i = 0; i < 256; i++) {
+        gpio0->data = cnt;
+        cnt++;
 
-    while (true) {
-        uint8_t c;
-        int received_bytes = XUartLite_Recv(&uart0, &c, 1);
-        if (received_bytes < 1) {
-            continue;
-        }
-        recv++;
-        gpio0->data = recv & 0xff;
-        XUartLite_Send(&uart0, &c, 1);
-        if (c == '\r') {
-            c = '\n';
-            XUartLite_Send(&uart0, &c, 1);
-        }
+        sleep(125);
     }
-}
+
+    auto* airc = reinterpret_cast<volatile uint32_t*>(0xe000ed0c);
+    uint32_t a = *airc;
+    *airc = a | 0x04;
+
+    while (1) {}
+}