Friday, 2018-01-19

*** tpb has joined #timvideos00:00
*** ewen has joined #timvideos00:09
ewenKnock! Knock!  Can someone open outer door of hackfest?  Thanks.00:09
*** medicalwei has quit IRC00:40
*** medicalwei has joined #timvideos00:47
mithrocr1901_modern: did I mention that I wanted to support sdcard in SPI mode via bitbanging? (IE a super slow mode?) - the gateware for that should be super simple (IE just a couple of CSR registers right)01:35
*** rohitksingh-demo has joined #timvideos02:39
ewenmithro: rohitksingh-demo: Current status of (Elbert v2 attempt) is that "make firmware" builds a firmware.elf with some ugly #ifdefs, and a linker.ld hack, but "make gateware" fails due to design issues all around clocks/PLL I think.02:44
tpbTitle: WIP: Numato Elbert V2 support by mithro · Pull Request #12 · timvideos/litex-buildenv · GitHub (at
ewenmithro: rohitksingh-demo: All the clocks/PLL bits in elbertv2/ was cargo culted from mimasv2; I'm unclear which bits are actually needed (especially with no SDRAM), and which bits need changes for Spartan 3A versus Spartan 6; pull request has "make gateware" error messages for reference.02:45
*** futarisIRCcloud has joined #timvideos02:54
rohitksingh-demoewen: yup, since we are using 12MHz clock, the PLL parameters are gonna change. I'll look into that02:56
ewenrohitksingh-demo: Thanks!  I did adjust a couple of them for 100 / 12, but was just wildly guessing what should be there.03:02
felix_spartan 3 doesn't have plls03:03
felix_iirc they had those dcm things03:03
felix_so you need to instanciate 03:53 < rohitksingh-demo> ewen: yup, since we are using 12MHz clock, the PLL parameters are gonna change. I'll look into that03:04
rohitksingh-demofleix_: yup, we probably need to use dcm03:04
felix_you need to instantiate a dcm instead of the pll03:05
ewenfelix_: Aha.  That'd explain the build errors.  Unfortunately I'm out of my depth with "instantiate a dcm" so if one of you can figure out what is needed that'd be much appreciated.03:11
xfxfstefanor: changes needed (rails may have already PR'ed them): r6mix/r8mix need hw:1,0 alsa source and 2000ms delay04:18
railsWill be Pring in 10 mins04:19
tpbTitle: GitHub - xfxf/HDMI2USB-misoc-firmware at dvisampler-debug-fix (at
mithroxobs: try this ->
tpbTitle: GitHub - enjoy-digital/transceiver_test: Test of transceivers for board to board communication (at
xobsmithro: I'm missing  Trying to figure out where to get it from.05:46
mithroxobs: Ahh yeah - that is why we statically link -- which causes the other issue...05:46
xfxfmithro: that fix worked btw, made a PR05:46
mithroxfxf: Yeah - just waiting for travis to confirm05:47
*** CarlFK has quit IRC05:52
*** sb0 has quit IRC05:53
xobsmithro: It got further when I added LD_LIBRARY_PATH=$CONDA_DIR/lib:$LD_LIBRARY_PATH to enter_env.sh05:54
xobsNow I'm trying to fix "as: unrecognized option '-mmultiply-enabled'"05:54
mithroxobs: Further means?05:54
mithroxobs: I think that means its using the wrong as05:55
mithroxobs: Because that should be lm32-elf-as or something, right?05:55
xobsThe archive you posted has no as05:58
mithroxobs: conda install binutils-lm32-elf or something?06:09
mithroxfxf: Your patch is merged06:11
*** CarlFK has joined #timvideos06:12
*** ChanServ sets mode: +v CarlFK06:12
mithroha -
tpbTitle: GitHub - azonenberg/protohdl: Streaming FPGA/ASIC code generator for Google Protocol Buffers. (at
*** sb0 has joined #timvideos06:52
*** mcarden has quit IRC06:52
CarlFK[m]name:Marc Merlin: Getting conned into writing IoTuz/ESP32 drivers and example code (while being held prisoner in a share house in Hobart, Tasmania)06:52
CarlFK[m]django.db.utils.DataError: value too long for type character varying(135)06:52
ewenrohitksingh-demo: Papilio Duo schematic:
tpbTitle: Papilio-DUO/Papilio_DUO_Schematic.pdf at master · GadgetFactory/Papilio-DUO · GitHub (at
tpbTitle: GitHub - GadgetFactory/Papilio-DUO: Papilio DUO FPGA board (at
CarlFK[m]name             | character varying(170)   | not null07:11
ewenrohitksingh-demo: This appears to be the full four pages:
tpbTitle: migen-misc/src/migen_misc at master · peteut/migen-misc · GitHub (at
ewenrohitksingh-demo: This seems to be the Papilio Loader:
tpbTitle: GitHub - GadgetFactory/Papilio-Loader: Java and command line apps to load Xilinx Bit files to Papilio FPGA boards. (at
ewenAFAICT the papilio-prog directory has the C/C++ source for the loader (rest seems to be Java GUI wrapper)07:34
ewenrohitksingh-demo: Interestingly it looks like when I try "sudo openocd" it's both disassocating my Wifi and also resetting the Papilio Duo USB (eg, it reports failure to setup, and disconnects).07:38
ewen[Tue Jan 16 12:42:54 2018] usb 1-2: device descriptor read/64, error -7107:38
ewen[Tue Jan 16 12:45:16 2018] usb 1-2: Device not responding to setup address.07:38
tpbTitle: embedded_hal - Rust (at
ewenrohitksingh-demo: Maybe it found it run with sudo?  Info : JTAG tap: xc6s.tap tap/device found: 0x24001093 (mfg: 0x049 (Xilinx), part: 0x4001, ver: 0x2)07:41
ewenBut that seems to be trigger of USB bus resets07:41
ewen(I'm guessing Wifi is also on USB bus.)07:41
ewenrohitksingh-demo: .... relevant bits of openocd file ....07:44
ewen# iManufacturer           1 Gadget Factory07:44
ewen# iProduct                2 Papilio DUO07:44
ewen# iSerial                 3 10000000000007:44
eweninterface ftdi07:44
ewen#ftdi_device_desc "Papilio DUO"07:44
ewenftdi_vid_pid 0x0403 0x7bc007:44
ewen# channel 1 does not have any functionality07:44
ewenftdi_channel 007:44
ewen# just TCK TDI TDO TMS, no reset07:44
ewenftdi_layout_init 0x0008 0x008b07:44
ewenreset_config none07:44
ewenadapter_khz 1000007:44
ewensource [find cpld/xilinx-xc6s.cfg]07:45
ewensource [find cpld/jtagspi.cfg]07:45
ewenrohitksingh-demo: Done, all but some comments07:45
mithrorohitksingh-demo: Did you want me to do the command which does the ident stuff?07:49
rohitksingh-demomithro: yes please! I do not know how to add it07:50
ewenrohitksingh-demo: FYI, these are the URLs I found for board08:01
ewen# Papilio Duo with Xilinx Spartan-6 FPGA08:01
tpbTitle: PapilioDUOHardwareGuide (at
tpbTitle: Papilio DUO-2MB - FPGA - Seeed Studio (at
rohitksingh-demoewen: great! I added the first one08:02
*** ewen has quit IRC08:10
mithrorohitksingh-demo: This is what it looks like ->
tpbTitle: make: Always pass an ident value. by mithro · Pull Request #400 · timvideos/HDMI2USB-litex-firmware · GitHub (at
rohitksingh-demomithro: great! I'll test it in a moment!08:23
*** futarisIRCcloud has quit IRC08:24
*** rqou_ has joined #timvideos08:30
*** puck has quit IRC08:30
*** rqou has quit IRC08:31
*** rqou_ is now known as rqou08:31
*** puck has joined #timvideos08:36
tpbTitle: openocd: Add patch files for neso, saturn, waxwing and papilio-duo by rohitk-singh · Pull Request #25 · timvideos/conda-hdmi2usb-packages · GitHub (at
mithrorohitksingh-demo: Looks good to me!08:44
rohitksingh-demomithro: yay!08:45
rohitksingh-demomithro: and how to submit to openocd upstream?08:45
tpbTitle: OpenOCD: Patch Guidelines (at
mithroTalk to PaulFester in #openocd too08:46
rohitksingh-demomithro: thanks!08:49
*** rqou has quit IRC09:01
cr1901_modernmithro: You did mention that you wanted an SPI mode, but that's not gonna help w/ my current issues, since the command set is basically the same in SPI mode.09:07
cr1901_modernAt this point I would just like something working before the conf for the mimasv2.09:07
cr1901_modernAlso, my sleep schedule has some semblance of normalcy now, yay109:09
mithrocr1901_modern: Well, we are not really using the mimasv2 at the conference...09:19
mithroWe are using the Arty09:20
cr1901_modernOh right, you did mention that.09:20
mithrocr1901_modern: But I don't think you have a pmod->microsd right?09:21
cr1901_modernmithro: No, but I do have a pmod breadboard and sd card breakout board from sparkfun09:22
cr1901_modernmithro: Lemme look something up quickly09:23
rohitksingh-demomithro: any comments? good/bad/horrible?09:24
tpbTitle: Gerrit Code Review (at
mithrorohitksingh-demo: LGTM! You should mention you tested on real hardware09:24
rohitksingh-demomithro: okay09:25
rohitksingh-demohold one09:25
*** rqou has joined #timvideos09:38
mithrorohitksingh-demo: Poke PaulFester in #openocd for review - tell him I sent you09:44
rohitksingh-demomithro: pinged him...doesn't seem to be online now...anyways adding him as reviewer too09:53
CarlFK[m]mithro: kick the AC09:56
CarlFK[m]mithro: kick the ACmithro: kick the AC10:05
CarlFK[m]mithro: kick the AC10:14
mithroCarlFK[m]: done - but why is it so important?10:15
CarlFK[m]My phone kept reminding me to reset the AC10:17
*** aps has joined #timvideos10:21
mithroxobs: you should have build/conda/bin/lm32-elf-as10:22
mithrofrom binutils-lm32-elf10:22
mithroexport PATH=$PWD/conda/bin:$PATH ; conda install <xxx.tar.bz2>10:25
mithroI mean10:25
mithroexport PATH=$PWD/build/conda/bin:$PATH ; conda install <xxx.tar.bz2>10:25
*** sb0 has quit IRC10:51
mithroxobs: libusb-win32 is released just because of the reason you mentioned, to address the OpenOCD/libftdi problems where people want to use the JTAG and Serial port feature at the same time --> which10:56
mithromeans libusb0.sys should only be the driver for the JTAG interface, not the whole device, has the bugs that it does not work well with this option.10:56
cr1901_modernmithro: (Still looking stuff up), I've been having trouble getting my external LA to capture SD card signals in pulseview, but all the issues I've had this past week basically devolve to this:11:07
cr1901_modernThe SD cards I have on hand are too old, and _florent_'s firmware doesn't account for them.11:07
cr1901_modernEmbarrassing, right? Unfortunately, there was no way I would've figured this out on my own w/o delving into the spec11:08
cr1901_modern(embarrassing as in: It took me long enough to figure this out)11:08
mithroxobs: ?11:09
tpbTitle: Automatic USB driver installe (at
mithroxobs: you have to claim the first interface first (interface 0, or channel A).  -- Ref:
tpbTitle: dpaste (at
tpbTitle: Bash On Windows · timvideos/litex-buildenv Wiki · GitHub (at
mithroxobs: You might enjoy this -> :-P11:23
tpbTitle: Notes and Tips · timvideos/litex-buildenv Wiki · GitHub (at
tpbTitle: dpaste (at
mithrocr1901_modern: How does that mimasv2 platform match the one in litex-buildenv?11:24
micolous[m]>>> ','.join((str(x) for x in range(47,95)))11:24
cr1901_modernmithro: I copied and pasted from litex-buildenv and removed the HDMI2USB/litex-buildenv specific constants11:25
CarlFK[m]mithro (IRC): we bothered to create the list of lines you can ignore because it is wonky conerence api data, so you should at least click to see the fruity labor:,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,9411:35
tpbTitle: dpaste (at
tpbTitle: Due to glibc 2.12 limitation, static executables that use time(), cpuinfo() and maybe a few others cannot be run on systems that do not support or use `vsyscall=emulate` · Issue #6747 · conda/conda · GitHub (at
mithrotumbleweed: Try "conda install"11:38
CarlFK[m]micolous: tumbleweed:
tpbTitle: veyepar/ at master · CarlFK/veyepar · GitHub (at
tpbTitle: Adapt (at
*** jea has quit IRC11:47
*** jea has joined #timvideos11:47
micolous[m] all the backend coedz11:48
tpbTitle: GitHub - micolous/streamchicken: backend components for new streaming site (at
tpbTitle: - Non-geek stuff (at
rohitksingh-demomithro: Xilinx Platform Cable too:
tpbTitle: Adapt (at
tumbleweedmithro: your proposal to skip the vsyscall optimisation seems reasonably sensible11:52
mithrotumbleweed: I have yet to figure out how to make it work....11:56
*** TimGremalm has quit IRC12:29
*** TimGremalm has joined #timvideos12:30
*** techman83 has quit IRC12:32
*** techman83 has joined #timvideos12:33
*** ChanServ sets mode: +v techman8312:33
CarlFK[m]mithro (IRC): Warning: API rate limit exceeded for (But here's the good news: Authenticated requests get a higher rate limit. Check out the documentation for more details.)12:58
*** mcarden has joined #timvideos13:16
*** rohitksingh-demo has quit IRC13:21
felix_mithro: oh, the transceiver test repo seems to have the proper reset sequence now. seems like i wasted my time looking at how to do that properly *sigh*13:24
felix_my apporach would have been quite different though13:25
*** rohitksingh-demo has joined #timvideos13:33
*** CarlFK has quit IRC13:34
_florent_felix_: the old code was working but we were bypassing part of the initialization13:52
_florent_felix_: now it should be following xilinx recommandations13:53
felix_yes. i was beginning to implement this all in firmware13:54
felix_small wrapper around the drp and synchronizer for the resets13:54
felix_the issue the workaround is for is btw. that some internal state doesn't get properly reset when the fpga interface is 16 or 32 bit; for 20 or 40 bit the workaround doesn't change things13:56
felix_don't know the details though13:57
_florent_felix_: there is also this:
tpbTitle: misoc/ at master · m-labs/misoc · GitHub (at
tpbTitle: misoc/ at master · m-labs/misoc · GitHub (at
_florent_felix_: i don't think you need phase alignment for sdi, so reusing misoc's core is probably better13:59
felix_i started litegt to properly implement the support fot the gtp14:02
felix_for sdi i don't need much features, but since the gtps seemed to be quite a cause of pain, i wanted to implement the support properly once and be able to reuse that14:03
felix_not having some state machine poking registers, but doing that in firmware, would also eliminate the need for drp arbitration that is needed when the runtime needs to adjust gtp setting14:05
_florent_the initialization is a bit tricky and i think you will still have things to do in the gateware14:07
felix_sure, to get the timing of the reset signals right, some gateware is needed14:09
*** rohitksingh-demo has quit IRC14:23
mithroxobs: non-static libs uploaded14:27
*** hyadez has quit IRC14:42
tpbTitle: Index of /~tim (at
*** CarlFK has joined #timvideos14:57
*** ChanServ sets mode: +v CarlFK14:57
*** sb0 has joined #timvideos15:19
*** hyadez has joined #timvideos15:21
*** aps has quit IRC18:35
*** samsagaz has joined #timvideos19:47
*** CarlFK has quit IRC21:05
*** futarisIRCcloud has joined #timvideos21:51

Generated by 2.13.1 by Marius Gedminas - find it at!