#m-labs on 2018-01-28 — irc logs at freenode.irclog.whitequark.org

2015-03-04 14:45 sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs

00:47 rohitksingh-demo has joined #m-labs

00:53 rohitksingh has joined #m-labs

01:34 rohitksingh-demo has quit [Quit: Leaving.]

01:49 <sb0> whitequark, because all sayma variants use the same rtm gateware.

01:50 <sb0> whitequark, please don't merge them like this

02:49 futarisIRCcloud has joined #m-labs

03:11 rohitksingh has quit [Quit: Connection closed for inactivity]

03:40 <GitHub164> [artiq] sbourdeauducq commented on issue #407: And 68ms with KC705, ARTIQ master and Windows (no AV). https://github.com/m-labs/artiq/issues/407#issuecomment-361035837

04:00 <GitHub173> [artiq] sbourdeauducq commented on issue #883: ``get_rtio_counter_mu`` seems to be working fine. Likely a runtime bug (parallel programming with async RPCs? core device memory corruption when dealing with the large array?) or miscompilation (though compilation is a pretty deterministic process and this bug is intermittent). @whitequark https://github.com/m-labs/artiq/issues/883#issuecomment-361036588

04:03 Ultrasauce has quit [Quit: No Ping reply in 180 seconds.]

04:05 Ultrasauce has joined #m-labs

04:11 <GitHub106> [artiq] sbourdeauducq opened issue #909: regular Urukul DDS support https://github.com/m-labs/artiq/issues/909

04:15 _whitelogger has joined #m-labs

10:29 futarisIRCcloud has quit [Quit: Connection closed for inactivity]

11:55 <whitequark> sb0: I am not convinced

11:56 <sb0> whitequark, what do you propose?

11:56 <whitequark> what I just did

11:56 <whitequark> building the RTM gateware is very quick

11:56 <sb0> ship the same copy of the RTM bitstream in every package?

11:56 <whitequark> yeah

11:56 <sb0> it's quick now because it doesn't do a lot, but it will get bigger

11:57 <sb0> probably will take about the same time as the current kasli drtio satellite target

11:58 <whitequark> well okay but then there is no good way to make a single artiq_flash sayma target.

11:58 <whitequark> so I am not going to do that.

11:58 <sb0> why do you want a single artiq_flash sayma target?

11:59 <whitequark> jbqubit insisted on one...

12:00 <sb0> I can see two alternatives: 1. two commands to flash sayma 2. an ad-hoc option into artiq_flash that runs those same two commands in turn

12:00 <whitequark> you can't specify two --srcbuild options at once

12:00 <sb0> sure, in that case don't use the ad-hoc option

12:00 <sb0> go with the two artiq_flash invokations

12:01 <whitequark> what a pain in the ass

12:01 <whitequark> everything about sayma is user-hostile and you insist on making it worse

12:02 <sb0> adding 15 minutes to every bitstream build for the RTM isn't nice either

12:03 <whitequark> oh, I have an idea

12:03 <sb0> you could disable that build maybe, though

12:03 <whitequark> we can have the RTM package ship three copies of the bitstream

12:03 <sb0> in which case that would be fine

12:03 <whitequark> and put it in the folders for every variant

12:03 <whitequark> that would be functionally equivalent to the current situation but faster

12:04 <sb0> what annoys me the most with your proposal is the increased bitstream build times, not #the minor disk space bloat

12:04 <sb0> having both flashable with a single --srcbuild is nice

12:05 <sb0> and having an option to disable the RTM build would solve the increased compilation time problem in an acceptable way

12:07 <sb0> multiple copies of the bitstream in the RTM package doesn't sound like a good idea

12:09 <whitequark> then the individual AMC packages could copy the RTM bitstream out of the RTM package at build time

12:10 <sb0> we might even not need a RTM package. cache the RTM bitstream outside conda?

12:10 <whitequark> that doesn't really work on buildbot

12:10 <whitequark> not reliably anyway

12:10 <sb0> ok, then fine, we can do that

12:12 <sb0> what in particular with sayma is user-hostile, in your opinion?

12:14 <rqou> i just found out that someone i know somewhat well is also working on a diy mass spectrometer

12:14 <rqou> I'm trying to find out more, maybe there can be some collaboration?

12:15 <sb0> rqou, who's it?

12:16 <rqou> a former classmate

12:18 <rqou> I'll have more info later; I'm still checking e.g. whether this is supposed to be a secret project

12:18 <sb0> is it a diy calutron?

12:21 <whitequark> sb0: moving parts that our tooling doesn't handle

12:21 <sb0> can you be specific?

12:22 <whitequark> for example, the two FPGAs that have to be flashed and loaded separately. but I've just fixed that.

12:22 <whitequark> the conda packages that have to be built in specific precise order on the buildbot or you'll get errors

12:23 <sb0> well sure, loading the RTM FPGA has been an open issue for a while

12:23 <whitequark> I mean with `artiq_flash load` too.

12:24 <sb0> and that's not really an issue with sayma, but with setting up software/infrastructure that no one (before you) had looked into deeply

12:27 <whitequark> you could put it that way but the end result is that sayma is more user-hostile because of its increased complexity

12:41 <GitHub107> [artiq] whitequark commented on issue #883: Possibly a race condition with async RPCs... https://github.com/m-labs/artiq/issues/883#issuecomment-361059964

12:56 <GitHub91> [artiq] sbourdeauducq commented on issue #883: If this is confirmed, it might make sense to switch to a single mailbox. https://github.com/m-labs/artiq/issues/883#issuecomment-361060972

13:36 _whitelogger has joined #m-labs

13:58 <GitHub93> [artiq] whitequark commented on issue #883: Or just fix the race condition. https://github.com/m-labs/artiq/issues/883#issuecomment-361064952

14:00 <GitHub80> [smoltcp] phil-opp commented on issue #117: Done. Sorry for the delay! https://github.com/m-labs/smoltcp/pull/117#issuecomment-361065085

14:03 <GitHub38> [artiq] sbourdeauducq commented on issue #883: I remember you saying that, if another concurrency bug was found in this code, you'd switch to a single mailbox. :) https://github.com/m-labs/artiq/issues/883#issuecomment-361065297

14:13 <GitHub29> [smoltcp] whitequark commented on pull request #117 8dab5b8: To avoid this, `SocketSet::add` should have the signature `fn add<T: Into<Socket>>(socket: T)`. https://github.com/m-labs/smoltcp/pull/117#discussion_r164300521

14:17 futarisIRCcloud has joined #m-labs

14:22 <GitHub64> [smoltcp] phil-opp commented on pull request #117 8dab5b8: Yes, I thought about this, but I wasn't sure if you're okay with the resulting increase in file size. But I assume that it's negligible. https://github.com/m-labs/smoltcp/pull/117#discussion_r164300773

14:25 <GitHub188> [smoltcp] whitequark commented on pull request #117 8dab5b8: It might not be, but space profiling is something that should be done separately and preferably without degrading API ergonomics. https://github.com/m-labs/smoltcp/pull/117#discussion_r164300841

14:27 <GitHub39> [smoltcp] phil-opp commented on issue #117: I made SocketSet::add generic to avoid the `into` calls. https://github.com/m-labs/smoltcp/pull/117#issuecomment-361066972

14:29 <GitHub142> [artiq] whitequark pushed 2 new commits to master: https://github.com/m-labs/artiq/compare/11a8b84355d7...79ea454ec13b

14:29 <GitHub142> artiq/master 79ea454 whitequark: conda: use $SP_DIR instead of $PREFIX/lib/python3.5/site-packages. (#652)...

14:29 <GitHub142> artiq/master 885ab40 whitequark: conda: split RTM and AMC packages back....

14:31 <GitHub185> [smoltcp] phil-opp commented on issue #117: Agreed! https://github.com/m-labs/smoltcp/pull/117#issuecomment-361067270

14:35 <whitequark> bb-m-labs: force build --props=package=artiq-sayma_rtm artiq-board

14:35 <bb-m-labs> build forced [ETA 18m28s]

14:35 <bb-m-labs> I'll give a shout when the build finishes

14:36 <GitHub100> [smoltcp] whitequark closed pull request #117: Return `UdpSocket` from `UdpSocket::new` (instead of `Socket`) (master...patch-1) https://github.com/m-labs/smoltcp/pull/117

14:36 <GitHub21> [smoltcp] whitequark pushed 1 new commit to master: https://github.com/m-labs/smoltcp/commit/c2d18ec071a2cd68aee2724ce13eefa6ccf1f0c1

14:36 <GitHub21> smoltcp/master c2d18ec Philipp Oppermann: Return specific sockets from `new` functions instead of `Socket`....

14:38 <bb-m-labs> build #1163 of artiq-board is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/1163

14:43 <bb-m-labs> build #1164 of artiq-board is complete: Exception [exception interrupted] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/1164 blamelist: whitequark <whitequark@whitequark.org>

14:43 <bb-m-labs> build #2018 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2018 blamelist: whitequark <whitequark@whitequark.org>

14:44 <whitequark> bb-m-labs: force build --props=package=artiq-sayma_rtm artiq-board

14:44 <bb-m-labs> build forced [ETA 18m28s]

14:44 <bb-m-labs> I'll give a shout when the build finishes

14:45 <travis-ci> m-labs/smoltcp#642 (master - c2d18ec : Philipp Oppermann): The build passed.

14:45 <travis-ci> Change view : https://github.com/m-labs/smoltcp/compare/f0c147a5ee16...c2d18ec071a2

14:45 <travis-ci> Build details : https://travis-ci.org/m-labs/smoltcp/builds/334378146

14:51 <bb-m-labs> build #1165 of artiq-board is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/1165

14:53 <GitHub67> [smoltcp] LuoZijun commented on issue #135: @whitequark okay, soon. https://github.com/m-labs/smoltcp/pull/135#issuecomment-361068797

15:19 <GitHub94> [artiq] whitequark pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/9a94482c6ea712865314f6683cbff515e2b7b3cb

15:19 <GitHub94> artiq/master 9a94482 whitequark: conda: fix typo in 885ab409.

15:19 <GitHub184> [artiq] whitequark pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/0edc34a9e5d3790a6692af33e16dff45924e2c5b

15:19 <GitHub184> artiq/master 0edc34a whitequark: artiq_devtool: the proxy artiq_flash action doesn't exist anymore.

15:32 <travis-ci> batonius/smoltcp#125 (master - c2d18ec : Philipp Oppermann): The build passed.

15:32 <travis-ci> Change view : https://github.com/batonius/smoltcp/compare/b0995c13e29e...c2d18ec071a2

15:32 <travis-ci> Build details : https://travis-ci.org/batonius/smoltcp/builds/334389231

15:40 <GitHub119> [smoltcp] LuoZijun opened pull request #136: rebuild ::phy::raw_socket (master...patch-3) https://github.com/m-labs/smoltcp/pull/136

15:47 <bb-m-labs> build #1166 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/1166

15:50 <bb-m-labs> build #2019 of artiq is complete: Failure [failed artiq_corelog] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2019 blamelist: whitequark <whitequark@whitequark.org>

15:56 <GitHub19> [smoltcp] whitequark commented on issue #136: 1. Please remove any changes related to mio. I do not want mio as a dependency of smoltcp, it is too large and complex.... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361073101

16:20 <bb-m-labs> build #1167 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/1167

16:24 <whitequark> bleh, I had to scrap most of the code I wrote in the last few days

16:24 <whitequark> too complex and fragile

16:25 <sb0> in? artiq_flash?

16:26 futarisIRCcloud has quit [Quit: Connection closed for inactivity]

16:26 <GitHub26> [smoltcp] LuoZijun commented on issue #136: @whitequark ... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361075278

16:27 <GitHub13> [smoltcp] LuoZijun commented on issue #136: @whitequark ... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361075278

16:27 <GitHub74> [smoltcp] LuoZijun commented on issue #136: @whitequark ... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361075278

16:28 <bb-m-labs> build #713 of artiq-win64-test is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/713

16:31 <bb-m-labs> build #2020 of artiq is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2020

17:01 <GitHub150> [smoltcp] LuoZijun commented on issue #136: @whitequark ... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361077908

17:02 <GitHub84> [smoltcp] LuoZijun commented on issue #136: @whitequark ... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361077908

17:13 <GitHub20> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305494

17:14 <GitHub74> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305495

17:14 <GitHub51> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305498

17:14 <GitHub188> [smoltcp] hjr3 commented on pull request #125 afcd9ba: yup makes sense. fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305509

17:14 <GitHub131> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305510

17:14 <GitHub83> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305512

17:14 <GitHub12> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305514

17:14 <GitHub7> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305520

17:14 <GitHub111> [smoltcp] hjr3 commented on pull request #125 afcd9ba: `field::DATA` is a function now. https://github.com/m-labs/smoltcp/pull/125#discussion_r164305532

17:15 <GitHub78> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305536

17:15 <GitHub104> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305538

17:15 <GitHub30> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305540

17:15 <GitHub102> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305541

17:15 <GitHub103> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305548

17:15 <GitHub47> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305551

17:15 <GitHub160> [smoltcp] hjr3 commented on pull request #125 afcd9ba: fixed https://github.com/m-labs/smoltcp/pull/125#discussion_r164305555

17:31 <GitHub56> [smoltcp] whitequark commented on issue #136: No. The ioctls are different. Just leave the tap interface as it is. https://github.com/m-labs/smoltcp/pull/136#issuecomment-361080061

17:31 <GitHub79> [smoltcp] whitequark commented on issue #136: > I think RxToken and TxToken implementations have a performance problem... https://github.com/m-labs/smoltcp/pull/136#issuecomment-361080097

20:08 <GitHub129> [sinara] gkasprow pushed 1 new commit to master: https://git.io/vNDNp

20:08 <GitHub129> sinara/master 327ecdd Greg: zotino v1.1 added copper balancing

20:29 mumptai has joined #m-labs

22:37 <rjo> has anyone used openocd to talk to the xadc in 7 series?

22:54 <cr1901_modern> what's the use case?

23:38 <GitHub37> [smoltcp] dlrobertson commented on pull request #125 c3c5988: The `len` used as input to `DATA` should be the result of `data[field::LENGTH]` instead of the buffer length. Also don't make the same mistake I did with ICMPv4 and make sure to check that the buffer is bigger than that before using that field. https://github.com/m-labs/smoltcp/pull/125#discussion_r164317095

23:38 <GitHub165> [smoltcp] dlrobertson commented on pull request #125 c3c5988: `allow(non_snake_case)` if we're sticking to `DATA`. https://github.com/m-labs/smoltcp/pull/125#discussion_r164317124

23:38 <GitHub116> [smoltcp] hjr3 commented on pull request #125 c3c5988: i have this fixed in my latest changes https://github.com/m-labs/smoltcp/pull/125#discussion_r164317145

23:39 <GitHub84> [smoltcp] hjr3 commented on pull request #125 c3c5988: i am fixing this right now https://github.com/m-labs/smoltcp/pull/125#discussion_r164317153

23:53 <GitHub28> [smoltcp] hjr3 commented on pull request #125 839fa3f: this is fixed now. `check_len` is tricky because of the special `Pad1` case and the variable length field. https://github.com/m-labs/smoltcp/pull/125#discussion_r164317726