Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

pipistrello runtime does not compile [PPP] #658

Closed
jordens opened this issue Jan 23, 2017 · 3 comments
Closed

pipistrello runtime does not compile [PPP] #658

jordens opened this issue Jan 23, 2017 · 3 comments

Comments

@jordens
Copy link
Member

jordens commented Jan 23, 2017

smoltcp doesn't support PPP/SLIP or something like those.

@whitequark
Copy link
Contributor

whitequark commented Jan 31, 2017

This needs some sort of serial encapsulation, since we don't have a PPP implementation anymore.

@sbourdeauducq suggested SLIP. The argument for SLIP was that it's available on Linux and Windows, but that's not really the case; the last version of Windows that shipped with SLIP support was XP. Also the Linux tooling looks horrible from the quick look I made at it.

What I would prefer to do instead is to use a simple TAP interface on Linux, a framing with fixed maximum overhead, and a bespoke splitter on the host system, using the code already in smoltcp. This gets us several things:

  1. Simpler code on the core device. No need to implement a SLIP interface type alongside the Ethernet interface type in smoltcp. Sharing more code paths between pipistrello and kc705, preventing bit rot.
  2. Being able to see console messages at the same time as forwarding network packets.
  3. This implementation is trivially portable to any version of Windows using the OpenVPN tap interface drivers and a few DeviceIoControl calls.

@sbourdeauducq
Copy link
Member

I would make the case for dropping Pipistrello from ARTIQ 3.

  • Smoltcp is already taking a lot of time, with issues like low throughput still not resolved.
  • Only one user survey respondent seems to have a serious interest in it. I have contacted him to find out more.
  • If we want to attract beginners, then Pipistrello should have good support and testing. This includes Windows support. Doing all this development and testing takes time. If the goal is to make physics more accessible, then I'd rather spend resources on ionpak or a good laser servo.
  • Even with this effort, it would still be substandard due to the low speed of the serial interface, and the confusion that pppd (or equivalent) causes to beginners.
  • Spartan-6 FPGAs are getting long in the tooth.
  • The Kasli board ought to be reasonably priced and a much better alternative - easier to use, faster, and more powerful.

@jbqubit
Copy link
Contributor

jbqubit commented May 19, 2017

Several prospective ARTIQ users had poor experience with Pipistrello due to it's second-class support. Agreed that it's better to drop support and focus on supporting Kasli in the future.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

4 participants