Enforce stricter validation

[Lucas-Morais-Freire]

Hi. First of all, thanks for the continued work!

I was testing the simulator with some addresses listed in the "invalid" section of the configuration json (with shared memory set to False). I had a bit register on address 0 and one invalid register on address 1. I tried reading coils starting on address 15 with count 16. Of course, coil 15 is part of register 0 and that's okay, but coils 16 - 30 should be illegal. The current way the range for bit access is computed is like this:

def validate(self, func_code, address, count=1):
        """Check to see if the request is in range.

        :meta private:
        """
        if func_code in self._bits_func_code:
            # Bit count, correct to register count
            count = int((count + WORD_SIZE - 1) / WORD_SIZE)
            address = int(address / 16)

        real_address = self.fc_offset[func_code] + address
        if real_address < 0 or real_address > self.register_count:
            return False

        fx_write = func_code in self._write_func_code
        return self.loop_validate(real_address, real_address + count, fx_write)

I was wondering if it would be desirable to change it to this:

def validate(self, func_code, address, count=1):
        """Check to see if the request is in range.

        :meta private:
        """
        if func_code in self._bits_func_code:
            # Bit count, correct to register count
            count = (addr + count - 1)//WORD_SIZE - addr//WORD_SIZE + 1 # distance between start register and end register + 1
            address = address // WORD_SIZE

        real_address = self.fc_offset[func_code] + address
        if real_address < 0 or real_address > self.register_count:
            return False

        fx_write = func_code in self._write_func_code
        return self.loop_validate(real_address, real_address + count, fx_write)

this way, the count takes into account the starting address as well so that you can catch cases like the one I describe above.

[janiversen]

Pull requests are welcome.

[janiversen]

Please be aware that the simulator datastore is changing on dev, we move towards a much simpler configuration.

[Lucas-Morais-Freire]

Yeah, I'm aware. Though i have been working on this important application for a while and having to change to the new version might break stuff. So I'd rather work on the current version for now. But I hope we can use the new one sometime!