I wrote a test for 0xAAAA representation and another for 0xAAAAAAAA representation. All those made with 0xAAAA fail and those with 0xAAAAAAAA.

P.N. I used the hex notation for the condition to make the error more explicit. I will convert it to binary.

Signed-off-by: Federico Bruzzone [email protected]

Closes #89

I also added a feature "debug" to put random values in the palette_ram. In can be useful to use this features also to some hard coded test for the render process.

To run clementine in debug mode you just need to run the command: just debug *rom_path/rom.gba*

All the cartridge header implementations have been added to the cartridge_headers.rs file. (#17)

Keep in mind: 1. BIOS must be able to overwrite the boot_mode header, as defined in the extract_boot_mode function comments; 2. BIOS must be able to overwrite the slave_id_number header, as defined in the extract_slave_id_number function comments; 3. If the GBA has been booted by using Joybus transfer mode, we have to put our initialization procedure directly at joybus_mode_entry_point address, or redirect to the actual boot procedure by depositing a "B " opcode in the space used by joybus_mode_entry_point.

Many values are written in hexadecimal format (like 0x0B_00_00_00), and a few others in binary format (like 0b1110_1010_0000_0000_0000_0000_0111_1111). It would be useful to standardize all values in binary, for readability reasons.

However, I think it is better to close #33 first because it would allow the removal of many values.

CartridgeHeader is now in the "Cartridge" file along with the "Cartridge" struct, which is respondible to take raw data and turn it in header fields and ROM (most of the fields are missing at the moment)

Cartridge takes the ROM data under the "rom" field in a reference counted ptr, so the type inside the ARM7TDMI implementation of the CPU was modified

• Logger: Add new logger for clementine
• Everywhere: Use logger in the code with a special arg --log-on-file

Basically the feature is because I would love to analyze the data and istruction post-run of a specific ROM, since now we have enough instruction running but we don't know where are bugs :). This allow us to log on stdout as befor of log on a file saved under /tmp/clementine-{utc::now()}.log with a special arg like ./clementine --log-on-file my_game.gba.

We need to check this path on windows platform.

Now it is possible use some kind of graphic debug tools in the UI. You can easily add new ones and edit existing ones.

N.B. I also refactored all the structure of our workspace

• CpuState was using the inverted logic
• BX was not reading the value from Rn
• Added instruction decode for THUMB mode
• Setting lsb in PC to 0 before fetching

Closes #49

These function complete the bitwise file.

• The first one allow us to clean (set to 0) all bits from msg to idx;
• The second one allow us to clean (set to 0) all bits from idx to 0

Signed-off-by: Federico Bruzzone [email protected]

After executing an instruction that modifies PC, it should not be advanced by 4.

I've added that functions that implement instructions should return a bool representing whether PC should be increased after its execution.

I don't know if this is the best solution, maybe we could use an Option<u32> containing the new PC (returned from the instruction function) if Some or execute advance_program_counter if None. We can discuss it here :)

Tests are broken since they were written taking into account the fact that PC is always increased, when we decide which way to take I can fix the tests

This PR adds LCDRegisters memory region to InternalMemory. It serves as preparation to implement the ppu (hopefully :D).

Structs IORegister and LCDRegisters have been written to model these registers.

Unit tests have been added to InternalMemory.

Added get_byte and set_byte in Bits trait to selectively read/write a specific byte from an u32.

This is a start to a migration on how code is structured in order to have displayed the asm code in one point passing arg for every instruction while decoding.

Documented here: BG_MODE_4

In this PR I also added a macro to simplify lock syntax:

acquire_lock!(lock_var, acquired => { you can use acquired here } );

It was usefull because sometime we have to acquire a lock only in the needed scopes. In this way it is more cleaner to use a mutex more times in the same flow.

For example assume you have two function fn_1(), fn_2()

fn_1() {
  // it needs a lock
  let acquired_var = self.lock_var.lock().unwrap();
  acquired_var.do_something();

 fn_2();
}

fn_2() {
  // it needs a lock
  let acquired_var = self.lock_var.lock().unwrap();
  acquired_var.do_other_stuff();
}

if you call fn_1() a deadlock occurs.

with this macro you can do this without problems:

fn_1() {
  // it needs a lock
  acquire_lock!(self.lock_var, var => { var.do_something(); });
 fn_2();
}

With just test_mode_4 path/rom.gba you can check the correct behavior with the help of PaletteVisualizer.

Added a new UiTool to read values from memory.

I don't like very much how it is implemented but it was the only way (or at least the only one I found) to make it work since we have Gba<T: Cpu> and Cpu trait didn't have any method to access memory and memory is stored into an Rc<RefCell<>>.

Please tell me if you think it could be implemented in a better way (maybe without returning Ref? But I think it's needed since InternalMemory is managed by a RefCell).

It's very simple because I want to use it to debug an issue we may have in the instruction implementations. In the future, we could show a matrix with all the values and so on but for now I think this does the job.

We currently have checks (fmt, test, lint) on PR branches. It would be useful to execute the same set of checks on the "potential" new main branch (PR branch rebased to current main).

This is useful because the PR branch may be not rebased to the main, thus merging the PR could result in broken tests.

Clementine could be run within a Docker container. It should be created a proper Dockerfile, documentation and some github action to build and push the container into a public registry when a PR is merged.