01:12 <sb0> mithro, yes, read it at every clock cycle in a loop...

03:43 <GitHub28> [artiq] sbourdeauducq pushed 1 new commit to master: https://git.io/v174M

03:43 <GitHub28> artiq/master db5957a Sebastien Bourdeauducq: firmware: use bsp crate in ad9154

03:46 <sb0> rjo, how was ad9156_reg autogenerated?

03:46 <sb0> *ad9516

04:16 <GitHub56> [smoltcp] whitequark pushed 2 new commits to master: https://git.io/v17Bu

04:16 <GitHub56> smoltcp/master 59dae01 whitequark: Rethink the buffering strategy with Managed<T>.

04:16 <GitHub56> smoltcp/master 8ff9b17 whitequark: Significantly simplify buffering.

04:21 <bb-m-labs> build #275 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/275

04:45 <bb-m-labs> build #1174 of artiq is complete: Failure [failed python_unittest_1] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1174 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>

05:15 <whitequark> sb0: 11:58 < whitequark> as for network, I just finished some more necessary refactoring, so now the application side of the stack works fine

05:15 <whitequark> done

05:15 <GitHub88> [smoltcp] whitequark pushed 2 new commits to master: https://git.io/v170Y

05:15 <GitHub88> smoltcp/master 7d838b6 whitequark: I was going to license this under 0-clause BSD.

05:15 <GitHub88> smoltcp/master 8b89de2 whitequark: Working UDP loopback.

05:15 <whitequark> that's basically all the high-risk parts gone

05:27 <sb0> TCP isn't high-risk?

05:57 <whitequark> I'm not implementing it from scratch

05:58 <whitequark> whereas no one that I know of has quite did this sort of use case of Rust's features (AsMut, BorrowMut, lifetimes, combination thereof) before and it wasn't obvious that it will work

05:58 <whitequark> for that matter it took a fair amount of fiddling before it did.

05:59 <whitequark> the benefit is of course that the stack has zero unsafe code and I don't see any reason it will gain any.

06:00 <whitequark> well, *maybe* field accesses, if someone figures out a way to make that robust. but probably not me.

07:27 <GitHub190> [smoltcp] whitequark pushed 2 new commits to master: https://git.io/v17zM

07:27 <GitHub190> smoltcp/master 37f565a whitequark: Fix lifetime variance.

07:27 <GitHub190> smoltcp/master ef04295 whitequark: Fix a Cargo warning.

10:10 <GitHub14> [artiq] jordens reopened issue #582: plan ARTIQ users group meeting https://git.io/v17rt

10:27 <sb0> memory safety (under the assumption that everything else - dma engines, c code, etc. is working properly) is one thing, protocol snafu is another

10:29 <sb0> what is rust-managed?

10:32 <sb0> if smoltcp depends on it, can you put it under m-labs?

10:32 <sb0> on github

12:07 <whitequark> right now it's just a placeholder for code I haven't written

12:07 <whitequark> and may not even write depending on feedback

12:08 <whitequark> also I don't put things under m-labs right away because it means I can't edit the repo description and URL

12:08 <whitequark> very annoying

12:14 <whitequark> sb0: well the problem that we have now is memory safety, and I hope to not encounter it again

12:14 <whitequark> apart from stack overflows and the binutils bugs I don't recall any memory safety issues with the new runtime

12:14 <whitequark> looking forward to when lld finally gains MEMORY{} support and we don't have to touch binutils ever

12:34 <mithro> sb0: https://github.com/m-labs/misoc/pull/47 -- Sorry it has taken so long, still needs a little more cleanup but its getting there...

22:50 <GitHub50> [artiq] cjbe opened issue #646: Passing kernel function pointers between classes https://git.io/v15kX