Back to Tock

Tock OT Notes 5/21/2020

doc/wg/opentitan/notes/2020-05-21.md

latest2.2 KB
Original Source

Tock OT Notes 5/21/2020

Present:

  • Brad Campbell (Chair), University of Virginia
  • Jon Flatley, Google
  • Silvestrs Timofejevs, lowRISC
  • Johnathan Van Why, Google
  • Alistair Francis, Western Digital
  • Alphan Ulusoy, Google
  • Vadim Sukhomlinov, Google

Updates

  • Jon: Have PRs ready for internal host-side emulation and testing for internal review. Working on roadmap towards v1.

  • Plan to share with Tock in the next week or two.

  • Alphan: PR for more efficient MPU config approved.

  • Debug optional for processes generated more discussion. Will iterate.

  • Also have question about disabling in the MPU.

  • Silvestrs: PR for device configuration in progress.

  • Brad: Fixed addresses TLV merged.

  • Vadim: Working on a minimal process syscall shim, allowing for LTO of apps.

  • Working on a system for storing known data about apps. No need to parse TBF headers every time.

  • One goal is to make a very lightweight libtock-c app. Push more into the kernel or somehow share code.

  • Jon: This may extend to process testing host side, if there are new load processes and create() functions.

  • Brad: We used to do app setup in the kernel. But, we weren't sure if what we had implemented was general or not. Once Amit demonstrated we could do it all app side we switched to that since then we didn't have to answer the generality question.

  • Vadim: Worried about code duplication.

  • Also, we want to support vendor provided libraries in apps that access HW, are blocking, and written in C.

  • Brad: We have discussed letting apps access hardware, but it never went anywhere. Difficult to provide useful fine-grained access control.

  • Alistair: There is an I/O PMP separate from the main PMP in RISC-V. Just covers the address access space. Might be interesting to see how that could help in this case.

Chomebook Use Case

  • See slides.
  • Vadim: If there is overlap or ongoing projects in these areas, let me know.

PMP

  • Alistair: Disable PMP not needed (is a no-op), want to support enhanced PMP that would apply to kernel as well.
  • We are not currently checking the lock bits.
  • Alphan: We also want to use some regions for the boot code.
  • Alistair: Those are different lock bits, also need to check locked regions.
  • Brad: Will continue discussion next week.