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
<GitHub35> [artiq] sbourdeauducq commented on issue #854: It's EMS, never tried it to Poland. https://github.com/m-labs/artiq/issues/854#issuecomment-363272717
<GitHub111> [artiq] sbourdeauducq commented on issue #915: Why are the automatic tests not catching that? https://github.com/m-labs/artiq/issues/915#issuecomment-363282476
<GitHub18> [artiq] sbourdeauducq commented on issue #854: Also @gkasprow can you try it concurrently on your board? Or have you done that already and ethernet is working with no problem at all? https://github.com/m-labs/artiq/issues/854#issuecomment-363282974
<GitHub110> [artiq] sbourdeauducq commented on issue #854: That board I sent was for the 1.8V bug. I don't know how Ethernet behaves on it with the latest MMC firmware. It may or may not exhibit the 'no link after rgmii setup' bug. https://github.com/m-labs/artiq/issues/854#issuecomment-363283389
kristianpaul has quit [Ping timeout: 240 seconds]
<GitHub127> [artiq] jbqubit commented on issue #910: Please just fix the docs and close. Somebody else can pay for the Vivado check. https://github.com/m-labs/artiq/issues/910#issuecomment-363296157
rohitksingh_work has joined #m-labs
<GitHub155> [artiq] whitequark commented on issue #910: > ARTIQ documentation does not specify supported version of Vivado. cf #908... https://github.com/m-labs/artiq/issues/910#issuecomment-363315583
<GitHub77> [artiq] sbourdeauducq commented on issue #910: Yes, it's a rather complicated situation. https://github.com/m-labs/artiq/issues/910#issuecomment-363316376
<GitHub180> [artiq] whitequark commented on issue #915: Catching what? https://github.com/m-labs/artiq/issues/915#issuecomment-363316735
<GitHub163> [artiq] sbourdeauducq commented on issue #908: @hartytp Are you looking into this?... https://github.com/m-labs/artiq/issues/908#issuecomment-363318585
<GitHub41> [artiq] sbourdeauducq commented on issue #908: @hartytp Are you looking into this?... https://github.com/m-labs/artiq/issues/908#issuecomment-363318585
sb0 has quit [Quit: Leaving]
sb0 has joined #m-labs
<GitHub169> [artiq] sbourdeauducq commented on issue #915: The ARTIQ conda package not installing correctly on Windows, either on the release-3 branch or master; both specify binutils-or1k-linux >=2.27. https://github.com/m-labs/artiq/issues/915#issuecomment-363324526
<GitHub95> [misoc] whitequark pushed 1 new commit to master: https://github.com/m-labs/misoc/commit/2e2b78387b0740bf287dcdaf6ad3297ff147adfe
<GitHub95> misoc/master 2e2b783 whitequark: integration: if --variant exists, build in that subfolder of --output-dir.
<bb-m-labs> build #379 of misoc is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/misoc/builds/379
<GitHub16> [smoltcp] canndrew opened pull request #153: Relax type constraints on payload_mut methods (master...relax-payload_mut-constraints) https://github.com/m-labs/smoltcp/pull/153
<GitHub196> [smoltcp] m-labs-homu commented on issue #153: :pushpin: Commit abbaa37 has been approved by `whitequark`
<GitHub83> [smoltcp] whitequark commented on issue #153: @m-labs-homu r+ https://github.com/m-labs/smoltcp/pull/153#issuecomment-363344425
<GitHub60> [smoltcp] m-labs-homu force-pushed auto from f06af16 to 5fdd447: https://github.com/m-labs/smoltcp/commits/auto
<GitHub60> smoltcp/auto 5fdd447 Andrew Cann: Relax type constraints on payload_mut methods...
<GitHub73> [smoltcp] m-labs-homu commented on issue #153: :hourglass: Testing commit abbaa37be4df7f86b8864d1d81ac69b88a5377c2 with merge 5fdd44757ad8ed0151d23feab8a16be5b08b425f... https://github.com/m-labs/smoltcp/pull/153#issuecomment-363344469
<sb0> whitequark, can you update the windows binutils package? nist cannot use artiq-3 until this is fixed.
<sb0> well at least that lab who uses windows...
<whitequark> ok
<GitHub45> [artiq] whitequark closed issue #912: Sayma output folder: use different folder for different variant https://github.com/m-labs/artiq/issues/912
<GitHub74> [artiq] whitequark pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/61c64a76bea0aece490daa89f43b93f07e820c9a
<GitHub74> artiq/master 61c64a7 whitequark: gateware: use a per-variant subfolder in --output-dir. (fixes #912)...
<whitequark> sb0: no, I can't do that
<whitequark> I need VNC access to buildbot and I only have a mobile ISP right now, which is too slow to be usable.
<sb0> rjo, kasli-1 reworked, the xadc works
<sb0> letting the fpga heat up a bit...
<GitHub83> [smoltcp] m-labs-homu merged auto into master: https://github.com/m-labs/smoltcp/compare/89b39b7c19d9...5fdd44757ad8
<GitHub84> [smoltcp] m-labs-homu commented on issue #153: :sunny: Test successful - [status-travis](https://travis-ci.org/m-labs/smoltcp/builds/337901594?utm_source=github_status&utm_medium=notification)
<travis-ci> m-labs/smoltcp#714 (auto - 5fdd447 : Andrew Cann): The build passed.
<GitHub194> [smoltcp] m-labs-homu closed pull request #153: Relax type constraints on payload_mut methods (master...relax-payload_mut-constraints) https://github.com/m-labs/smoltcp/pull/153
<sb0> rjo, seems to be around 47C using some si5324 test design (including cpu) that florent had flashed into it
<sb0> rjo, by the way, I have the 6mm heatsinks (6.5K/W) instead of the 10mm ones (6K/W)
<sb0> hm still climbing a bit, now 49C
<travis-ci> m-labs/smoltcp#715 (master - 5fdd447 : Andrew Cann): The build passed.
<sb0> whitequark, when will you be able to do it?
<rjo> sb0: ok. i have a couple bigger heat sinks on order.
<bb-m-labs> build #1187 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/1187
<whitequark> sb0: well I fly to HK today
<rjo> whitequark: and it might be worth checking whether http://ddietze.github.io/Py-Hardware-Support/picam.html is useful. afaik that was only tested on windows through.
<whitequark> WTF, this comes with a KERNEL MODULE?
<whitequark> are they nuts?
<rjo> whitequark: i haven't figured out what's that for. but afaict we don't need it.
<rjo> if you browse around at ftp://ftp.princetoninstruments.com/Public/Software/Official/PICam there is a manual as well ftp://ftp.piacton.com/Public/Manuals/Princeton%20Instruments/PICam%205.x%20Programmers%20Manual.pdf
<whitequark> why does this garbage install its files with permissions 0700...
<rjo> there are a bunch of things in there. let's maximize functionality and work around the dumb things (like the kernel module) if we can. i am happy to have a windows machine do the firmware upgrade if that's needed.
<whitequark> there is a linux firmware updater on that ftp
<bb-m-labs> build #717 of artiq-win64-test is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/717
<whitequark> wait
<whitequark> "FW_for_Linux.exe" ?
<rjo> whitequark: yes. but it might not be that seamless. anyway. let's work around the things you don't like pragmatically without too much complaining. ;)
<whitequark> but I like complaining.
<bb-m-labs> build #2033 of artiq is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2033
<rjo> whitequark: haha ;) well.
<whitequark> I've looked at the libraries.
<whitequark> 27 megabytes of C++ written in the best traditions of AbstractSingletonProxyFactoryBean
<rjo> whitequark: FYI for target level of functionality, that picam python library is sufficient.
<whitequark> DependentRelevance<50462767, 50593816, TrivializeOnValues<50593816>, DependentRelevanceStrategy<50462767, TrivializeOnValues<50593816>, 50593816, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1>::DependentTrigger>::DoGetDependency() const
<rjo> whitequark: yeah. if you are really convinced you can reverse engineer the protocol without **any** risk to the camera, we can consider that.
<whitequark> no, scratch that. all of the enumerations are mapped to their actual protocol values via this C++ horror
<whitequark> even if I reverse-engineer the protocol structure I'm not going to extract the mappings from the library
<GitHub157> [artiq] jordens commented on issue #915: In that build build 520 was installed on windows while the parent build was 521. https://github.com/m-labs/artiq/issues/915#issuecomment-363358461
<rjo> whitequark: but we should prioritize #915 over that camera driver.
<GitHub167> [artiq] hartytp commented on issue #908: > @hartytp Are you looking into this?... https://github.com/m-labs/artiq/issues/908#issuecomment-363362537
<sb0> 55C now
<GitHub166> [artiq] hartytp commented on issue #908: Anyway to be clear, in case I do find time to look into this, your plan is basically to dig through the git history, building various versions of Sayma gateware/firmware with SAWG (at a few hours per build) until we find the point where it stopped working? IIRC, that's a bit complicated by the fact that the tools to build Sayma have changed a bit over time, so it's not always
<GitHub98> [artiq] sbourdeauducq commented on issue #908: > Doesn't sound like fun.... https://github.com/m-labs/artiq/issues/908#issuecomment-363363547
<GitHub180> [artiq] hartytp commented on issue #908: Anyway to be clear, in case I do find time to look into this, your plan is basically to dig through the git history, building various versions of Sayma gateware/firmware with SAWG (at a few hours per build) until we find the point where it stopped working? IIRC, that's a bit complicated by the fact that the tools to build Sayma have changed a bit over time, so it's not always
<GitHub19> [artiq] sbourdeauducq commented on issue #908: > Doesn't sound like fun.... https://github.com/m-labs/artiq/issues/908#issuecomment-363363547
<GitHub56> [artiq] hartytp commented on issue #908: Well, as I said, as this seems like standard yak shaving for getting a board up and running, rather than a particular hardware/design issue with Sayma. So, do you mind taking a look at it first -- it's likely to be quicker for you since you've probably kept a closer eye on the changes that have been made to ARTIQ over the past weeks. https://github.com/m-labs/artiq/issues/9
<GitHub191> [artiq] sbourdeauducq commented on issue #908: > Doesn't sound like fun.... https://github.com/m-labs/artiq/issues/908#issuecomment-363363547
<sb0> there is a kernel module for talking to the camera over ethernet?
<whitequark> sb0: yes.
<whitequark> it includes a chunk of some RTOS.
<whitequark> oh, no, it's worse
<sb0> this reminds me of the xilinx trash for talking to the jtag cables a couple years ago (in a rare stroke of genius, xilinx replaced that with libusb since then)
<whitequark> they distribute a proprietary .a and an open-source RTOS-like API that delegates to Linux
<whitequark> clearly derived from some windows trash
<whitequark> OS_VOID NET_CopyToMemory( OS_VOID* aSource, OS_UINT32 aSourceOffset, OS_VOID* aDestination, OS_UINT32 aAmountOfDataToCopy )
<whitequark> I can't believe they did #define OS_VOID void.
<whitequark> this has got to be the most moronic thing I've ever seen in vendor code.
<GitHub61> [smoltcp] canndrew opened issue #154: Add an EthernetRepr type https://github.com/m-labs/smoltcp/issues/154
<GitHub95> [smoltcp] whitequark commented on issue #154: Doesn't seem very useful other than for completeness. https://github.com/m-labs/smoltcp/issues/154#issuecomment-363367162
<whitequark> sb0: are you using sayma-3?
<GitHub173> [smoltcp] canndrew commented on issue #154: It would be useful for me with the code I'm writing right now. I'll submit a PR for it. https://github.com/m-labs/smoltcp/issues/154#issuecomment-363370795
<whitequark> nevermind, it was my own openocd
<whitequark> sb0: openocd on sayma-3 hangs with "Info : clock speed 5000 kHz"
<whitequark> oh, and after power-cycle sayma-1 has a broken scan chain somehow.
<whitequark> xc7 is there, xcu is not
<bb-m-labs> build #200 of conda-win64 is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/conda-win64/builds/200
attie has quit [Read error: Connection reset by peer]
attie has joined #m-labs
<GitHub115> [artiq] gkasprow commented on issue #854: On our board we observe that ARTIQ gets rx preamble errors... https://github.com/m-labs/artiq/issues/854#issuecomment-363381365
<GitHub137> [artiq] gkasprow commented on issue #854: The link LED is on. I observe nice Rx and Tx clock signals. https://github.com/m-labs/artiq/issues/854#issuecomment-363383257
<GitHub89> [artiq] gkasprow commented on issue #854: @marmeladapk tried with different Rx delay setting https://github.com/m-labs/artiq/issues/854#issuecomment-363384166
<GitHub164> [artiq] gkasprow commented on issue #854: @marmeladapk tried with different Tx delay setting https://github.com/m-labs/artiq/issues/854#issuecomment-363384166
<GitHub83> [artiq] gkasprow commented on issue #854: @sbourdeauducq How to adjust Rx clock delay? https://github.com/m-labs/artiq/issues/854#issuecomment-363385673
<sb0> whitequark, not using saymas
<sb0> should I replug the USB connectors?
<whitequark> sb0: I'm not sure if this is USB
<whitequark> try it
<GitHub146> [artiq] sbourdeauducq commented on issue #854: @whitequark Can you help @gkasprow transmit some packets (assuming broken RX)? https://github.com/m-labs/artiq/issues/854#issuecomment-363389354
<sb0> whitequark, replugged it
attie has quit [Ping timeout: 256 seconds]
attie has joined #m-labs
<sb0> 54C after running for a couple hours
<sb0> not too bad...
<sb0> whitequark, xcu missing on the JTAG chain is likely the 1v8 rail failing at startup
<whitequark> ah
<GitHub117> [artiq] gkasprow commented on issue #854: @sbourdeauducq you experimented with Rx clock delay. What values make sense? https://github.com/m-labs/artiq/issues/854#issuecomment-363394381
<GitHub170> [artiq] sbourdeauducq commented on issue #854: Those around what I committed to the repository (135).... https://github.com/m-labs/artiq/issues/854#issuecomment-363395301
<GitHub198> [artiq] gkasprow commented on issue #854: OK, we will try with @marmeladapk.... https://github.com/m-labs/artiq/issues/854#issuecomment-363395880
<GitHub89> [artiq] sbourdeauducq commented on issue #854: You can also use microscope with MiSoC/ARTIQ. http://github.com/m-labs/microscope... https://github.com/m-labs/artiq/issues/854#issuecomment-363396753
<GitHub144> [artiq] gkasprow commented on issue #854: I want to be able to observe the RGMII signals on rx and tx side. But this I do with my existing design, without MAC but for single packets. I'd like to run it with higher data rate.... https://github.com/m-labs/artiq/issues/854#issuecomment-363398120
<GitHub132> [artiq] marmeladapk commented on issue #854: @sbourdeauducq Well, kudos for description of microscope, we had a good laugh in our lab. :) https://github.com/m-labs/artiq/issues/854#issuecomment-363398578
<GitHub143> [artiq] sbourdeauducq commented on issue #854: I did receive correct packets using my design, and I dumped them after the whole chain, the contents and CRC were correct. Maybe you did the rework a bit differently than me (did you cut the RX clock trace or just added the wire?) and your timing is different. A RX phase adjustement and/or verifying that our reworks are the same (I cut the trace, but maybe I did not do
<GitHub173> [artiq] gkasprow commented on issue #854: @sbourdeauducq I cut the trace and added shortest possible wire. I documented it with photos in dedicated [issue](https://github.com/m-labs/sinara/issues/484).... https://github.com/m-labs/artiq/issues/854#issuecomment-363400764
<GitHub30> [artiq] sbourdeauducq commented on issue #854: Also, it is impossible for a RGMII design to "simply work", other than by chance, or if they have some sort of automatic calibration which sounds impractical. The timing windows are small and this needs phase calibration for every board design.... https://github.com/m-labs/artiq/issues/854#issuecomment-363401223
<GitHub83> [artiq] sbourdeauducq commented on issue #854: Also, it is impossible for a RGMII design to "simply work", other than by chance, or if they have some sort of automatic calibration which sounds impractical. The timing windows are small, the PHY timings and PCB layouts are varied, and this needs phase calibration for every board design.... https://github.com/m-labs/artiq/issues/854#issuecomment-363401223
<GitHub40> [artiq] sbourdeauducq commented on issue #854: Also, it is impossible for a RGMII design to "simply work", other than by chance, or if they have some sort of automatic calibration which sounds impractical. The timing windows are small, the PHY timings and PCB layouts are varied, and this needs phase calibration for every board design. As I remember, you had to do some hacks (e.g. invert the clock on IDDR) to get your
<GitHub108> [artiq] gkasprow commented on issue #854: @sbourdeauducq I can route the clock to same bank using some of TLCKC/TCLKD clock pairs. They won't be used in the future. The clock trace delay would be much longer, but at least it won't depend on FPGA silicon PTV variation. https://github.com/m-labs/artiq/issues/854#issuecomment-363401949
<GitHub127> [artiq] gkasprow commented on issue #854: OK, let's fix the Rx path first. https://github.com/m-labs/artiq/issues/854#issuecomment-363402152
<GitHub2> [artiq] sbourdeauducq commented on issue #854: > The clock trace delay would be much longer, but at least it won't depend on FPGA silicon PTV variation.... https://github.com/m-labs/artiq/issues/854#issuecomment-363402681
<GitHub157> [artiq] gkasprow commented on issue #854: OK, my wire is roughly 2cm shorter. Cannot we add some auto-tuning capability? It can be done via UART-controlled fine delay that can be setup by packet loss rate. It's just a few lines of Python code. https://github.com/m-labs/artiq/issues/854#issuecomment-363403189
<GitHub79> [artiq] sbourdeauducq commented on issue #854: We can, but:... https://github.com/m-labs/artiq/issues/854#issuecomment-363404114
<GitHub70> [artiq] gkasprow commented on issue #854: Yes, but this would be problematic for other users of Sayma rev 1.1.... https://github.com/m-labs/artiq/issues/854#issuecomment-363404591
<GitHub92> [artiq] sbourdeauducq commented on issue #854: Let's focus on the board issues so that they can be effectively fixed in the next PCB revision. For actual users, there is the Ethernet-FMC escape route. https://github.com/m-labs/artiq/issues/854#issuecomment-363405027
<GitHub43> [artiq] gkasprow commented on issue #854: Did you get the FMC-Ethernet board? I think my piece can be switched to 1.8V so could verify of this works, providing that you get yours from same vendor. https://github.com/m-labs/artiq/issues/854#issuecomment-363405980
attie has quit [Ping timeout: 260 seconds]
attie has joined #m-labs
<GitHub161> [artiq] sbourdeauducq commented on issue #854: I did not get it yet. https://github.com/m-labs/artiq/issues/854#issuecomment-363406515
<GitHub6> [artiq] gkasprow commented on issue #854: To use same clock bank, one would need to move the clock wire to another board edge:... https://github.com/m-labs/artiq/issues/854#issuecomment-363406730
sb0 has quit [Quit: Leaving]
<GitHub159> [artiq] gkasprow commented on issue #854: One would also need to get rid of the coupling capacitor... https://github.com/m-labs/artiq/issues/854#issuecomment-363406940
<GitHub153> [artiq] gkasprow commented on issue #854: One would also need to get rid of the coupling capacitor... https://github.com/m-labs/artiq/issues/854#issuecomment-363406940
<GitHub40> [artiq] gkasprow commented on issue #854: One would also need to get rid of the coupling capacitor... https://github.com/m-labs/artiq/issues/854#issuecomment-363406940
<GitHub104> [artiq] gkasprow commented on issue #854: And the clock pin (TCLKC_D_P) is M25 https://github.com/m-labs/artiq/issues/854#issuecomment-363407694
rohitksingh_work has quit [Read error: Connection reset by peer]
sb0 has joined #m-labs
attie has quit [Remote host closed the connection]
attie has joined #m-labs
kristianpaul has joined #m-labs
attie has quit [Ping timeout: 268 seconds]
attie has joined #m-labs
<GitHub73> [smoltcp] dlrobertson commented on pull request #139 bf04ed3: Yeah, I wasn't sure if this `Repr` would be more usable if it was like the ICMP, TCP, and UDP reprs where the `data` is in the `Repr`, or like like IPv6 and IPv4 where the `data` is not in the `Repr`s. https://github.com/m-labs/smoltcp/pull/139#discussion_r166328078
rohitksingh has joined #m-labs
rohitksingh_ has joined #m-labs
rohitksingh has quit [Ping timeout: 260 seconds]
rohitksingh_ is now known as rohitksingh
rohitksingh1 has joined #m-labs
rohitksingh2 has joined #m-labs
rohitksingh1 has quit [Quit: Leaving.]
rohitksingh2 has quit [Ping timeout: 240 seconds]
rohitksingh1 has joined #m-labs
rohitksingh2 has joined #m-labs
rohitksingh2 has quit [Client Quit]
FabM has quit [Ping timeout: 256 seconds]
rohitksingh1 has quit [Quit: Leaving.]
rohitksingh1 has joined #m-labs
FabM has joined #m-labs
rohitksingh1 has quit [Quit: Leaving.]
attie has quit [Ping timeout: 268 seconds]
attie has joined #m-labs
cr1901_modern1 has joined #m-labs
cr1901_modern has quit [Ping timeout: 248 seconds]
rohitksingh has quit [Quit: Connection closed for inactivity]
cr1901_modern1 has quit [Quit: Leaving.]
cr1901_modern has joined #m-labs
mumptai has joined #m-labs
<GitHub154> [smoltcp] dlrobertson commented on issue #141: @whitequark would you like all the changes to the TCP timestamps etc to be in this PR too? If so, I'm assuming they should be in a separate commit https://github.com/m-labs/smoltcp/pull/141#issuecomment-363552444
attie has quit [Ping timeout: 260 seconds]
attie has joined #m-labs
mumptai has quit [Remote host closed the connection]