Firmware checkum for DVI3000.

JTAG on DVI3000.

Moderator: elkora

Locked
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Firmware checkum for DVI3000.

Post by usbbdm »

I believe this should apply to DVI3020 too. It is very simple.
Here is the code to calculate it.

Code: Select all

        DWORD nCheckSum = 0;

        unsigned int nCodeLength = (pIn[0x10] << 24) + (pIn[0x11] << 16) + (pIn[0x12] << 8) + (pIn[0x13]);

        for (unsigned int i = 0; i < nCodeLength - 4; i ++)
        {
            nCheckSum += pIn[i + 0x14];
        }

Example
firmware 16.21.
code length is 0x3a413c
checksum is 0x0fd1ef76 at address offset 0x3a414C

Please verify this. I will latter find the place to fake unit address.
haute
Junior Member
Posts: 80
Joined: Mon Jan 09, 2006 9:31 am

Post by haute »

hi, this checksum it was already commented in this post.
http://www.usbjtag.com/bdmjtagbb/viewto ... c&start=30

i think it is a simple checksum for check the uploaded firmware.
I think there are another one, and it can be possible to relationated with nvram.

but I will test but to verify this

thx
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

haute wrote:hi, this checksum it was already commented in this post.
http://www.usbjtag.com/bdmjtagbb/viewto ... c&start=30

i think it is a simple checksum for check the uploaded firmware.
I think there are another one, and it can be possible to relationated with nvram.

but I will test but to verify this

thx
Oh yes I did not read the Spanish one.
haute
Junior Member
Posts: 80
Joined: Mon Jan 09, 2006 9:31 am

Post by haute »

That there is about the table of motorola present in firmware?
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

Nvram at both offset 0x70 and 0xe0 should be the checksum. Checking the firmware now to find out how to calculate checksum.
haute
Junior Member
Posts: 80
Joined: Mon Jan 09, 2006 9:31 am

Post by haute »

hi, Days ago I reduced the information of the NVRAM to the smaller possible number of bytes so that it worked with firmware. The single NVRAM occupied about 60 bytes approximately. Offset 60 and E0 leave them in bytes 0000, and I do not believe that they are checksum. If you want I send that nvram so that you throw a look. It is for 16.25 firmware version. The actual version in use. With only 60 bytes off info it much easy search any checksum.
If you make swap16 to a nvram you can see the firmware version 16.25.

thk for your work
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

Caculate checksum for DVIs are getting close. DVI actually has much more strong checksums than DCT2000/DCT2500. I am about to build small MIPS emulator to emulate the last bulk of check caculation. The first part of checksum caculation is the same as DCT2500.
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

haute
I have the simulator working now. can you go online so we can continue work on the checksum issue?
haute
Junior Member
Posts: 80
Joined: Mon Jan 09, 2006 9:31 am

Post by haute »

hi , i am working out my house, night work. i can go online beteen 4:00 - 6:30 Pm everiday.

You can send the tests to dagra82 to test in in dvi3000.

we can speak abaut ALTERA. I think the counter watchdog is programmed in altera chip. two pines of chip inverter (watchdog) are connect to ALTERA chip.
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

haute wrote:hi , i am working out my house, night work. i can go online beteen 4:00 - 6:30 Pm everiday.

You can send the tests to dagra82 to test in in dvi3000.

we can speak abaut ALTERA. I think the counter watchdog is programmed in altera chip. two pines of chip inverter (watchdog) are connect to ALTERA chip.
haut I am stuck at the checksum, maybe we can first program the altera chip and then I put a break point then we can find the algorithm to compute the checksum.
haute
Junior Member
Posts: 80
Joined: Mon Jan 09, 2006 9:31 am

Post by haute »

Hi usbbdm, as the works of the checksum go?
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

haute wrote:Hi usbbdm, as the works of the checksum go?
Not these days. I think we can live with first modify the boot to bypass the checksum. I am working on HD6412 to provide EJTAG64 support in next release.
husus1
Junior Member
Posts: 57
Joined: Tue Oct 10, 2006 1:20 pm

DVI

Post by husus1 »

please usb,your you are the unical person that can help, tries to remove to the checksum from dvi. thanks us for the obtained thing until, but now can finishing to this project and the next following with others since without you we will not obtain it
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Re: DVI

Post by usbbdm »

husus1 wrote:please usb,your you are the unical person that can help, tries to remove to the checksum from dvi. thanks us for the obtained thing until, but now can finishing to this project and the next following with others since without you we will not obtain it
I worked on it for a while and I found it too difficult. There are two pass of checksum and I have one passed. I will investigate when I have time again. The current work is focus on HDbox (EJTAG64). Sorry for keep you very long on this issue.
usbbdm
Junior Member
Posts: 8981
Joined: Mon Jul 18, 2005 9:33 pm

Post by usbbdm »

Just got the NVRAM from HD6412 and I found that the NVRAM from HD6412 is EXACT the same as DVI3000. I should be able to solve them both at the same time. The first thing I am going to do is to make the box connected. If I do this any box from ebay can be localized.
Locked

Who is online

Users browsing this forum: No registered users and 2 guests