Security vulnerability with MAX CPLD's !!

Question

I'm currently using an EPM3256 CPLD in a prototype to prove a design and was thinking of using the same CPLD in the final revision until a colleague alerted me to possible security vulnerabilities with these devices as well as many other vendor parts.

Scratching around the internet I found a dude that has done a thesis on the subject and another mob that will actually circumvent the security on these devices and read back the design information so it can be copied by unscrupulous operators !!

http://www.mcu-reverse.com/altera/epm3256a (http://www.mcu-reverse.com/altera/epm3256a)

http://www.cl.cam.ac.uk/~sps32/mcu_lock.html (http://www.cl.cam.ac.uk/~sps32/mcu_lock.html)

Until now I was always under the impression that once the security bit was set the security of these devices was bullet proof. I don't feel happy expending lots of development time only to know that the investment in time and money could easily be given away to some unscrupulous third party :(

The question to Altera is have they improved the situation in later devices and what have they done to make this type of invasive chip probing impossible or impractical ??

regards

altera_forum · Answer

Even high-grade parts are hackable. Either Altera or Xilinx afaik, so nothing to expect for CPLDs...

altera_forum · Answer

You have the Cyclone III LS family which is more targeted on security and protection against tempering and cloning.

But nothing is bullet proof. You can make it more difficult to reverse engineer a system, but to someone with unlimited resources anything is crackable. You just need to make it hard enough so that it isn't economically viable.

As for the Max V... I didn't find in the datasheet how the configuration is transferred from the internal flash to the SRAM registers during power up. If this transfer is done serially, then in theory it could be possible to get the configuration stream back by carefully monitoring the power supply. That said I rather think that it is a parallel interface, and in that case getting the stream back is near impossible in this way. The only way to get the configuration would then be to open the chip and do a more invasive attack, which would require extensive knowledge and equipment.

altera_forum · Answer

Sorry I didn't see that you were using a Max 3000A and not a Max V. In that case it makes things a lot harder, as the Max 3000 doesn't use a separate configuration memory but has EEPROM cells directly integrated in the LABs. In this case there is no way to access the configuration by probing the power supply or trying to trigger errors that way, as there in no configuration data loading. The only solution I see to get back the data in that case is either through a magical undocumented JTAG command (which I'm not aware of, and I don't think Altera would have done this) or by opening the chip and trying to make that bit change in some way. So this doesn't change what I said, while it isn't impossible it requires lots of knowledge and equipment, and probably isn't cheap.

altera_forum · Answer

--- Quote Start ---

Sorry I didn't see that you were using a Max 3000A and not a Max V. In that case it makes things a lot harder, as the Max 3000 doesn't use a separate configuration memory but has EEPROM cells directly integrated in the LABs. In this case there is no way to access the configuration by probing the power supply or trying to trigger errors that way, as there in no configuration data loading. The only solution I see to get back the data in that case is either through a magical undocumented JTAG command (which I'm not aware of, and I don't think Altera would have done this) or by opening the chip and trying to make that bit change in some way. So this doesn't change what I said, while it isn't impossible it requires lots of knowledge and equipment, and probably isn't cheap.

--- Quote End ---

Thanks for the reply

The thing is, the site I posted the link for is offering that service. These dudes in china don't care about IP protection or patents :( Is there anything in the MAX II or V parts that would make an invasive or semi-invasive attack difficult or is it just a matter of time before they can reverse engineer those ? Maybe someone from Altera can tell us.

regards

http://www.mcu-reverse.com/contact-us (http://www.mcu-reverse.com/contact-us)

--- Quote Start ---

Email: mcureverse@gmail.com

Tel: 86-0755-25895856

FAX: 86-0755-2589681

Skype: simendi0755

Address for IC samples shipping:

Room 2110, Changhong Building, No. 3013 Sun Gang East Road, Luohu District, Shenzhen, Guangdong Province, China. (Zip code: 518010)

How to get started?

A. Please send an inquiry by skype, email or call with description of protected device you have.

B. Our staff will response quickest with quotation and delivery time and conditions for service.

C. Ship us the protected device, along with at least two or more not-programmed samples.

D. Before extracting the code, you need to pay 50% deposit by Paypal.

E. After extracting the code, we will send you back two programmed devices for you to test.

F. after making sure we had done our job with success, you will need to make final payment to receive the source code( .bin file or .hex file).

--- Quote End ---

altera_forum · Answer

i recall similar discussions on microcontroller forums. since the MAX II/V is flash based, you might take a look at something like avrfreaks.net to see the status of reading out flash bits from the die itself

i would focus on innovation and/or customer service instead of combating cloning

Forum Discussion

Security vulnerability with MAX CPLD's !!

10 Replies

Recent Discussions

Download links not working

Agilex 5 Power

Agilex 5 Sulfur Partial Write Issue on F2H ACE‑Lite I/F (256‑bit) with AXI Master of 128‑bit

Cyclone 10 LP's Extended Industrial parts

Quartus Prime Pro 25.1 fatal error during fitter: Windows "Efficiency mode" required