code protection

[Linuxbuilders]

Hi,

chip 18f4620
ccs 4.114

[Ttelmah]

Now, I have no idea what programmer you are using (you don't tell us), or which bit actually gets set by what you are trying... For what the fuses do, _read the data sheet_. There are perhaps a dozen fuses controlling individual 'parts' of the code protection. For example, you can specify that other parts of your _code_ can read a block, but it is protected from reads by the external programmer. There are several fuses allowing this to be done for individual blocks of memory. This is a _vital_ ability, since _all_ constant data, and the values used to initialise variables are stored in the ROM, and the code _must_ be able to access this. I suspect you have turned on the 'read protection' bit for the main code area, but not the 'boot block' (the low 800 addresses selectable separately, for use by a bootloader) - (which will result on '0' being read), but have not enabled the code to be able to read the block you have protected, hence it won't work....

Assuming you haven't got a bootloader, you need to set EBTR0 to EBTR3 to '1', EBTRB to '1', CP0 to CP3 to '0', CPB to '0', WRT0 to WRT3 to '0' (assuming you are not using program memory to store values yourself from the code), WRTB to '0', and WRTC to '0'.
If you are writing values to program memory from the code, then you will need to enable the WRT bit corresponding to the area involved. I have also ignored the EEPROM at this point.

Now, the 'danger' of allowing table reads, and writes, comes if you have a bootloader present, and this is not itself using an encryption check. What people can do (since the bootloader _must_ be able to write other blocks), is write a small program into the code somewhere, which then reads the rest of the code back. The route round this, is to not allow table _reads_ from the bootloader area (prevents it from being able to verify the code - but also prevents this route 'in'....).

Best Wishes

[Linuxbuilders]

Thnx, pickit2 I use. I do not use boot loader, I plan in the future but not at this stage. Thanx for help. I need to do some digging now and work out how to protect entire program then. I do not self write into program memory because I have EEPROM for configuration. I would like to protect it too if it is possible.

As per datasheet I need to set register 23-6 and 23-7 and then 23-10 and 23-11 and then ID bits, etc. How do I set it manually? I could not find where CCS is defining those registers under "protect fuse", it is not in the device file. Could someone please let me know how to manually set registers then? I need to set READ PROTECTION bits then, I want to be able to override program and eeprom if required or wipe the chip out and load new soft on it. I just need to make impossible to read the program from it and from the EEPROM.

Thank you for help.
_________________
Help "d" others and then you shell receive some help from "d" others.

[Ttelmah]

The fuses are all there in CCS, just with names (supposedly) designed to make them easier to understand. So you have 'CPB' - code protect bootblock, 'PROTECT' - protects the main code block, then the write protect ones, as 'WRTx' where 'x' is the area to protect, and 'NOWRTx' disable write protection on area 'x', 'TR' - table reads allowed (this is the one I suspect you need 'on'), etc. etc..
You can always look at the end of the LST file to see exactly what bits are set/cleared by the CCS fuses, and the .h file has a list of the ones for your processor.

Best Wishes

[Linuxbuilders]

thnx for that, will play with it and let you know how it goes.
_________________
Help "d" others and then you shell receive some help from "d" others.

[Linuxbuilders]

Hi, (Icannot put whole progam here because nobody have weeks to read it and I cannot do it for commercial reasons too so please try to understand what I am writing here)

[I am not using boot loader]

done my testing and:

[temtronic]

Assuming you're using MPLAB to compile ...

MPLAB has options for the 'fuses' that might be overwriting the CCS C code 'options'

or

your programming software might be doing that(don't know which programmer you're using).

You should cut a really,really small program that shows the problem that others can cut/paste/test.probably the quickest way to figure out this head scratcher.

[Linuxbuilders]

I am using native CCS PCWHD + IDE 4.114 and pk2cmd.exe app plus PICkit2 for programming because I am running W7 64 bit.

Very hard to paste small program because I have dual processor PCB with so much code that it would be a killer. Small program may not show the problem too because of its size.

What I suspect is:
1. problem is somehow related to bootloader reserved space which I use for normal program because I do not use boot loader at all at this stage
2. ccs bug most probably

a fact that I cannot protect EEPROM from reading even when FUSE is on would suggest some dodgy thing happenig during compilation.

I do not know assembler very well so I am not able to debug it this way.
Maybe someone could point me for what to look in the assembly code so I could find it and post here.

thnx
_________________
Help "d" others and then you shell receive some help from "d" others.

[dyeatman]

TTelmah mentioned the fuses breakout at the end of the listing. How about showing us what you have?

Example:

[Mike Walne]

You're not giving us anything to work on so that we can try to help.

A really short section of compilable code,which shows the problem, is the way forward.

Mike

[Linuxbuilders]