add better demo code

hv.m

@@ -157,7 +157,7 @@ hv_return_t hv_vcpu_create(hv_vcpu_t* vcpu, hv_vcpu_exit_t** exit, hv_vcpu_confi
   }
   printf("vcpu_zone = %p\n", args.output_vcpu_zone);
   if (args.output_vcpu_zone->ro.ver != kHvVcpuMagic) {
-    printf("Invalid magic! expected %lx, got %lx\n", kHvVcpuMagic, args.output_vcpu_zone->ro.ver);
+    printf("Invalid magic! expected %llx, got %llx\n", kHvVcpuMagic, args.output_vcpu_zone->ro.ver);
     const bool yolo = true;
     if (!yolo) {
       hv_trap(HV_CALL_VCPU_DESTROY, nil);
@@ -291,13 +291,62 @@ hv_return_t hv_vcpus_exit(hv_vcpu_t* vcpus, uint32_t vcpu_count) {
   return hv_trap(HV_CALL_VCPU_RUN_CANCEL, (void*)mask);
 }
 
+const char kVMCode[] = {
+    // Compute ((2 + 2) - 1)
+    0x40, 0x00, 0x80, 0xD2,  // mov x0, #2
+    0x00, 0x08, 0x00, 0x91,  // add x0, x0, #2
+    0x00, 0x04, 0x00, 0xD1,  // sub x0, x0, #1
+    // Write it to memory pointed by x1
+    0x20, 0x00, 0x00, 0xF9,  // str x0, [x1]
+    // Reboot the computer with PSCI/SMCCC
+    // 0x84000009 is PSCI SYSTEM_RESET using SMC32 calling convention
+    0x20, 0x01, 0x80, 0xd2,  // mov x0, 0x0009
+    0x00, 0x80, 0xb0, 0xf2,  // movk x0, 0x8400, lsl #16
+    0x02, 0x00, 0x00, 0xD4,  // hvc #0
+    // Infinite loop
+    0x00, 0x00, 0x00, 0x14,  // b .
+};
+
 int main() {
+  // https://github.com/zhuowei/FakeHVF/blob/main/simplevm.c
+  // https://gist.github.com/imbushuo/51b09e61ecd7b7ac063853ad65cedf34
+  // allocate 1MB of RAM, aligned to page size, for the VM
+  const uint64_t kMemSize = 0x100000;
+  const uint64_t kMemStart = 0x69420000;
+  const uint64_t kMemResultOutOffset = 0x100;
+  const uint64_t kMemGuestResultOut = kMemStart + kMemResultOutOffset;
+
+  char* vm_ram = mmap(NULL, kMemSize, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+  if (vm_ram == MAP_FAILED) {
+    abort();
+  }
+  // copy our code into the VM's RAM
+  memcpy(vm_ram, kVMCode, sizeof(kVMCode));
+
   hv_return_t err = hv_vm_create(nil);
   printf("vm create %x\n", err);
-  hv_vcpu_t cpu = 0;
+
+  err = hv_vm_map(vm_ram, kMemStart, kMemSize, HV_MEMORY_READ | HV_MEMORY_WRITE | HV_MEMORY_EXEC);
+  printf("hv_vm_map %x\n", err);
+
+  hv_vcpu_t vcpu = 0;
   hv_vcpu_exit_t* exit = nil;
-  err = hv_vcpu_create(&cpu, &exit, nil);
+  err = hv_vcpu_create(&vcpu, &exit, nil);
   printf("vcpu create %x\n", err);
-  err = hv_vcpu_run(cpu);
+
+  // Set the CPU's PC to execute code from our RAM
+  err = hv_vcpu_set_reg(vcpu, HV_REG_PC, kMemStart);
+  printf("vcpu set reg pc %x\n", err);
+  // Set CPU's x1 register to point to where we want to read the result
+  err = hv_vcpu_set_reg(vcpu, HV_REG_X1, kMemGuestResultOut);
+  printf("vcpu set reg x1 %x\n", err);
+  // Set CPU's CPSR to EL1
+  err = hv_vcpu_set_reg(vcpu, HV_REG_CPSR, 0x3c4);
+  printf("vcpu set reg cpsr %x\n", err);
+
+  err = hv_vcpu_run(vcpu);
   printf("run %x\n", err);
+
+  uint64_t result_out = *((uint64_t*)(vm_ram + kMemResultOutOffset));
+  printf("The result of (2 + 2) - 1 is %d\n", (int)result_out);
 }

userclient_hv_trap.m

@@ -60,6 +60,6 @@ kern_return_t hv_trap(unsigned int hv_call, void* hv_arg) {
     return 0x1337;
   }
   uint64_t signed_code_ptr = gCodePtrs[hv_call];
-  return IOConnectTrap6(gUserClient, /*index=*/0, /*arg1*/ hv_arg, /*arg2*/ 0, signed_code_ptr,
-                        gDiscriminant, 0, 0);
+  return IOConnectTrap6(gUserClient, /*index=*/0, /*arg1*/ (uintptr_t)hv_arg, /*arg2*/ 0,
+                        signed_code_ptr, gDiscriminant, 0, 0);
 }