• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 100 subscribers
  • Views 391 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM5748: cryptographic cores documentation

KKF
Expert 4585 points

Part Number: AM5748

Hi,
Where can you find documentation regarding the available cryptographic cores in the AM5748?

Thanks,
KF

  • 0
    Mastermind 8040 points

    Based on PRCM-related info in datasheet and trm, the am574x seems to have the same crypto accelerators as the am572x:

    • 2x omap4 AES accelerator (EIP-39d). There's a linux driver (omap-aes) for a subset of its functionality. I've never seen actual docs from TI for its instances in SoCs, but the same peripheral is also in the Tiva "snowflake" TM4C129 microcontrollers, whose datasheet does document it. There's also an old E2E thread with a fair bit of notes about it.
    • omap4 DES/3DES accelerator. There's a linux driver for it (omap-des).
    • 2x omap5 MD5/SHA1/SHA256/SHA512 accelerator. There's a linux driver (omap-sham) for it and its older siblings. It is very similar to the omap4 MD5/SHA1/SHA256 accelerator (which is documented in aforementioned Tiva snowflake datasheet).
    • omap4 RNG (EIP-75a). There's a linux driver for it (omap-rng). There's also an old E2E thread with a fair bit of notes about it.
    • public-key accelerator (EIP-29). There's no linux driver nor documentation for it. It performs Montgomery modular multiplication and exponentiation.

    Apparently there's also a dedicated dma engine for crypto, although I'm not sure why. I guess on HS devices it might be reserved for secure world while leaving the other DMA engines for the general-purpose OS?

    Given the amount of information already publicly available, in some cases from other TI documentation even, it continues to be a mystery to me why these modules are not simply documented in the TRM.