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/app_init.cc

@@ -0,0 +1,23 @@
+#include <cstdint>
+
+extern "C" int main();
+extern uint32_t _bss_begin, _bss_end, _initial_stack_pointer;
+
+__attribute__((section(".app_init")))
+void AppInit() {
+    *(uint32_t*)(0x40000000) = 0;
+    asm ("mov sp, %0"
+            :
+            : "r" (&_initial_stack_pointer)
+            :
+        );
+
+    // clear .bss
+    for (uint32_t* ptr = &_bss_begin; ptr < &_bss_end; ptr++) {
+        *ptr = 0;
+    }
+
+    main();
+
+    while(true) {}
+}