| *** tpb has joined #timvideos | 00:00 | |
| *** FelixVi has quit IRC | 01:24 | |
| *** rohitksingh_work has joined #timvideos | 03:45 | |
| shenki | shorne_: thanks. i saw them, will respond soon | 04:15 |
|---|---|---|
| shenki | shorne_: i have a strong preference for email patch review :) I should have sent them out to the list | 04:15 |
| shenki | shorne_: i'll do that once i've fixed them up a bit | 04:15 |
| shenki | _florent_: thanks, i'll give it a go | 04:15 |
| shenki | _florent_, mithro: do we have shadow GPRs enabled in the mor1k design? | 04:16 |
| shenki | _florent_, mithro: my understanding from reading the verilog was that we do not | 04:16 |
| mithro | _florent_: I'm pretty sure we do? | 04:16 |
| shenki | parameter OPTION_RF_NUM_SHADOW_GPR = 0, | 04:17 |
| shenki | litex/litex/soc/cores/cpu/mor1kx/verilog/rtl/verilog/mor1kx.v | 04:17 |
| shenki | line 62 | 04:17 |
| mithro | oh | 04:17 |
| mithro | shadow GPRs, not shadow CSRs | 04:17 |
| shenki | yup | 04:18 |
| rohitksingh_work | mithro: Received the letters last night! Thanks! :) | 04:47 |
| shorne_ | shenki: yeah, I like email patch review too, I just wanted to get your some quick feedback | 05:02 |
| shorne_ | I would recomment enableding at least 1 extra set of GPRs | 05:03 |
| shenki | okay. i'll do a build now and give it a spin tonight. i don't have my board with me today | 05:03 |
| shorne_ | But... another thought is having an extra stack for exception handling | 05:03 |
| shenki | that's essentialy what we're doing with tim's hacks, yeah? using the bottom of memory as an emergency stack | 05:04 |
| shorne_ | storing stuff to random spots in memory is not smp safe | 05:04 |
| shenki | ah. make it per-cpu? | 05:04 |
| shenki | we should turn on a smp in the cpu i'm using, so i can test your smp support out | 05:05 |
| shenki | and find more bugs in my driver no doubt | 05:05 |
| shorne_ | yeah, I would like to really use the r1 (stack pointer) and some kind of per CPU stack to manage the exceptions | 05:07 |
| shorne_ | Currently for lockdep support we require frame pointers for exceptions, so I have to create the stack frame before calling into the lockdep on/off functions | 05:08 |
| shorne_ | I havent really looked into it, but using the SPR's has been the way settled on by Stefan W and Stefan K | 05:09 |
| shenki | okay | 05:09 |
| shenki | what do other arches do to solve this problem? | 05:09 |
| shorne_ | thats what I mean, when I havent really looked into it | 05:10 |
| shenki | okay | 05:10 |
| shorne_ | I think they have explicit stack pointers for exceptions | 05:10 |
| shorne_ | maybe in hardware? doesnt x86 have something like that? | 05:10 |
| shorne_ | got to go | 05:11 |
| *** CarlFK has quit IRC | 08:58 | |
| *** anant3110 has joined #timvideos | 10:09 | |
| *** aps has joined #timvideos | 11:22 | |
| anant3110 | Hey, This is Anant, I made contact with you a month ago and got busy with my exams thereafter and hence couldn't contribute. I am hoping to make some impact this time. So, as a recap, I know beginner to intermediate level of Verilog and have experience in working with FPGA boards like Basys3, Zedboard and Zynq. I would want to start making some contributions. Can you help me out? | 11:27 |
| *** tsglove has joined #timvideos | 11:48 | |
| *** rohitksingh_work has quit IRC | 12:48 | |
| *** anant3110 has quit IRC | 12:48 | |
| *** anant3110 has joined #timvideos | 13:04 | |
| anant3110 | I am interested in the HDMI2USB project specially. Please give me some details about setting up my environment and start solving some issues. | 13:16 |
| *** rohitksingh has joined #timvideos | 13:37 | |
| *** aps has quit IRC | 13:42 | |
| *** CarlFK has joined #timvideos | 14:50 | |
| *** ChanServ sets mode: +v CarlFK | 14:50 | |
| CarlFK[m] | anant3110 (IRC): do you have any fpga boards? | 14:57 |
| anant3110 | My institute has FPGA boards. I will have to check if they have the ones that you people work on. | 14:59 |
| CarlFK[m] | do you have a linux box you can setup the build environment on? | 15:15 |
| anant3110 | No, I will have to check with that. | 15:25 |
| anant3110 | Oh wait I do. | 15:26 |
| CarlFK[m] | if you are comfortable with Windows, it would be good to see if you can build on windows... | 15:27 |
| anant3110 | I am more comfortable on linux only, the thing is I had arch before this and I really didn't want to work on that. So, I wasn't sure if I had removed it to install Ubuntu. | 15:29 |
| CarlFK[m] | I think Ubuntu 16.04 (xenial) will work the best. even though this page says 14.04 https://github.com/timvideos/HDMI2USB-litex-firmware/blob/master/getting-started.md | 15:32 |
| tpb | Title: HDMI2USB-litex-firmware/getting-started.md at master · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com) | 15:32 |
| CarlFK[m] | I am only testing builds, not doing real fpga dev work, and I am trying to write up docs on it | 15:33 |
| anant3110 | I have Ubuntu 17.04. I am guessing that will work too? | 15:41 |
| cr1901_modern | CarlFK[m]: FINALLY, the JTAG appears to be behaving. Nice idea to connect it to a relay | 15:43 |
| CarlFK[m] | anant3110 (IRC): yes , or if you run into problems we can figure it out | 15:44 |
| CarlFK[m] | cr1901_modern (IRC): yay!!! | 15:44 |
| CarlFK[m] | cr1901_modern (IRC): any idea why it wasn't working at first ? | 15:44 |
| cr1901_modern | Prob needed to be power cycled. But AFAICT I did the same thing as yesterday | 15:45 |
| cr1901_modern | I did need to mode swap from jtag to serial and back tho | 15:45 |
| CarlFK[m] | mithro gets some credit for the relay board - he asked me to set that up | 15:46 |
| CarlFK[m] | and that relay board is what fell into the Opsis and knocked off some components | 15:47 |
| cr1901_modern | bahahaha | 15:47 |
| cr1901_modern | how much does the relay board cost? | 15:47 |
| cr1901_modern | I could prob use one of those tbh | 15:47 |
| CarlFK[m] | $30 | 15:51 |
| CarlFK[m] | er, 35. whatever. | 15:51 |
| CarlFK[m] | lets see if I can find the page that lists it | 15:52 |
| anant3110 | Hey, So just tell me about the model of boards or any type of hardware you would require. I will get them issued from my institute. | 15:53 |
| CarlFK[m] | anant3110 (IRC): this is 'best' https://store.digilentinc.com/atlys-spartan-6-fpga-trainer-board-limited-time-see-nexys-video/ | 15:55 |
| tpb | Title: Atlys Spartan-6 FPGA Trainer Board (LIMITED TIME) - Digilent (at store.digilentinc.com) | 15:55 |
| cr1901_modern | CarlFK[m]: Okay cool, I'm satisfied that firmware upload works | 16:17 |
| CarlFK | cr1901_modern: neat - want me to wire up an eithernet connection? | 16:17 |
| cr1901_modern | CarlFK: Not necessary, but thanks | 16:18 |
| CarlFK | k | 16:18 |
| cr1901_modern | firmware upload currently only works via serial | 16:18 |
| CarlFK | cr1901_modern: I can take over the box now right? | 16:21 |
| anant3110 | I am not very sure if Spartan-6 is available in the institute. Would be great if you could suggest some alternatives as well. They definitely have Zedboard Zynq. But I think you would require boards which have HDMI ports? | 16:21 |
| cr1901_modern | CarlFK: Yes I'm satisfied | 16:22 |
| CarlFK | anant3110: https://github.com/timvideos/HDMI2USB-litex-firmware#other-boards | 16:27 |
| tpb | Title: GitHub - timvideos/HDMI2USB-litex-firmware: A version of the HDMI2USB firmware based around LiteX tools produced by @Enjoy-Digital (based on misoc+migen created by @M-Labs) (at github.com) | 16:27 |
| CarlFK | cr1901_modern: juser@cnt3:/media/sda1/HDMI2USB-firmware-prebuilt-master$ find |wc | 16:30 |
| CarlFK | 25908 25908 1960748 | 16:30 |
| CarlFK | 25,908 files ! | 16:30 |
| anant3110 | Thanks Carl. I ll get back when I have my inventory up and my environment set up :) | 16:33 |
| *** tsglove has quit IRC | 16:39 | |
| *** tsglove has joined #timvideos | 16:39 | |
| CarlFK[m] | cr1901_modern (IRC): can you walk me what I can paste into the promt to flash? | 16:42 |
| CarlFK[m] | juser@cnt3:/media/sda1/HDMI2USB-firmware-prebuilt-master/archive/master/v0.0.3-1054-gbbede92/opsis/hdmi2usb/lm32$ hdmi2usb-mode-switch --flash-gateware image-gateware+bios+firmware.bin | 16:43 |
| CarlFK[m] | like do I need to sudo that? | 16:43 |
| *** anant3110 has quit IRC | 17:21 | |
| *** abhas has joined #timvideos | 17:25 | |
| abhas | I had a question regarding capturing audio. Is there a way to capture audio from a presenter's laptop? For the upcoming conference, there might be some speakers who would play animations etc with audio. Would there be a way to capture that using the Opsis board? Also does audio work in the passthrough mode on whatever is connected to the output of the opsis board? | 17:30 |
| CarlFK[m] | abhas (IRC): currently the Opsis ignores hdmi audio | 17:42 |
| CarlFK[m] | I would use the laptop headphone jack into a DI box into the audio mixer that the presenter's mic goes to | 17:43 |
| abhas | CarlFK[m]: So then the audio gets recorded separately and you would mix it on the video during encoding or video editing? | 17:50 |
| CarlFK[m] | no. ignore the laptop audio for a moment. how do you plan on capturing the presenter's voice? | 17:52 |
| abhas | There is going to be a video camera to capture the presenter and the presenter's mic would go into the video camera's audio input. | 17:54 |
| abhas | So while the video is being recorded on the camera, I think I will be able to additonally capture it on the opsis board. | 17:54 |
| abhas | At the conference, there are 3 conference rooms. And so we just got 3 opsis boards so we can try to do both recording and streaming. | 17:55 |
| CarlFK[m] | presenter's voice goes into video camera. how does it get out of the camera? | 17:58 |
| abhas | It doesn't. It would get recorded into the video camera. | 17:58 |
| CarlFK[m] | oh gosh. | 17:58 |
| CarlFK[m] | um.. some how it has to get out... what is your plan? | 17:59 |
| CarlFK[m] | I am guessing record to SD cards, and copy files off of them to a computer after the conference? | 18:00 |
| abhas | I am actually yet to reach the venue. I will reach there on Tuesday. So if I can get a live audio out from the camera, I will feed that into the recording computer and then we can record both the audio and video together | 18:00 |
| abhas | CarlFK[m]: THat is one plan. But that will require too much time to get the videos online. Isn't it? | 18:01 |
| CarlFK[m] | well, too much work | 18:01 |
| CarlFK[m] | do you know about voctomix? | 18:01 |
| abhas | Yeah. You had mentioned it. I am going to try it out tonight | 18:02 |
| CarlFK[m] | when is the conference? | 18:05 |
| abhas | Wednesday / Thursday (Dec 20/21) | 18:05 |
| abhas | http://swatantra.net.in | 18:06 |
| tpb | Title: Swatantra Celebrating Freedom (at swatantra.net.in) | 18:06 |
| *** rohitksingh has quit IRC | 18:11 | |
| CarlFK[m] | next week. I don't think you have time for vocotmix | 18:12 |
| CarlFK[m] | you should have started weeks ago. Oh well. use SD cards and put in many hours of work after the event. | 18:12 |
| CarlFK[m] | or, if you think that won't happen (it generally does not work out) then try to use voctomix, but know that you may fail. | 18:13 |
| abhas | I will try out voctomix now. Lets see how it fares. Hopefully it will work out. | 18:15 |
| CarlFK[m] | do you have a box you can install linux on? | 18:16 |
| abhas | oh yeah... I just did a fresh install of Debian. | 18:17 |
| abhas | I have an issue with the HDMI out when I enable both the HDMI inputs. The display (both on the HDMI out and encoder) fluctuates between both the HDMI inputs. If I disable one input (input1 off), then it works fine. | 18:18 |
| abhas | Is this a known issue? | 18:18 |
| CarlFK | yes | 18:22 |
| abhas | so is there a way to have 2 inputs and switch between them? | 18:26 |
| CarlFK | "yes" - fix the firmware | 18:26 |
| CarlFK | otherwise, maybe change the refresh rate to 50 or something. but I would not recommend it | 18:27 |
| CarlFK | I would not be experimenting any more - use what is known to work | 18:28 |
| CarlFK | also, do you have 3 other people that will operate the equipment during the event? | 18:29 |
| CarlFK | hmm, before we spend time on this, do you want A) less work, more chance of fail or B) lots of work (SD cards, edit for days/weeks ), but less chance of fail ? | 18:31 |
| *** abhas_ has joined #timvideos | 18:31 | |
| *** abhas has quit IRC | 18:32 | |
| *** abhas has joined #timvideos | 18:32 | |
| abhas | CarlFK: Sure. I understand. | 18:32 |
| CarlFK | abhas: If you will get fired for not making videos, then do SD cards and spend many many hours. like 2 hours per talk - so it will be weeks before you are done | 18:40 |
| CarlFK | if you don't like the sound of that, then try voctomix | 18:41 |
| abhas | :) -- Well -- using Opsis is an experiment. I am sure there are other things that we can fallback to. but nothing like trying to get this to work. | 18:43 |
| abhas | I have voctomix running on a newly setup Debian computer now. | 18:44 |
| abhas | CarlFK: Both voctocore and voctogui are running. Will try to see how to get the encoder feed into it | 18:45 |
| CarlFK[m] | abhas (IRC): https://github.com/CarlFK/voctomix-outcasts#getting-started--demo | 18:46 |
| tpb | Title: GitHub - CarlFK/voctomix-outcasts (at github.com) | 18:46 |
| CarlFK[m] | that will get you started with some simple ini files | 18:47 |
| abhas | okay -- so I am able to run the test1.sh script and it shows the pattern from the opsis board now. | 18:51 |
| abhas | sorry -- I misspoke.. It shows the inbuilt test patterns. I can see the opsis pattern using the show_hdmi2usb.sh script | 18:53 |
| CarlFK[m] | k, so 'we' all use systemd scripts to do as much as possible | 19:00 |
| CarlFK[m] | but it makes it very hard to see what is going on | 19:00 |
| CarlFK[m] | https://github.com/CarlFK/dvsmon | 19:01 |
| tpb | Title: GitHub - CarlFK/dvsmon: DVswitch monitor (at github.com) | 19:01 |
| CarlFK[m] | I wrote that to manage dvswitch, and added support for vocto | 19:01 |
| CarlFK[m] | https://github.com/CarlFK/dvsmon/blob/master/prod.sh | 19:01 |
| tpb | Title: dvsmon/prod.sh at master · CarlFK/dvsmon · GitHub (at github.com) | 19:01 |
| CarlFK[m] | ./dvs-mon.py -c vocto-prod1.py | 19:02 |
| CarlFK[m] | vocto-prod1.py holds a bunch of voctomix commands that make something useful | 19:02 |
| abhas | sure. let me try this out | 19:02 |
| CarlFK[m] | dvs-mon.py gives each command a run/kill button and you can see stdout/stderr | 19:03 |
| CarlFK[m] | I haven't used in quite a while, there is likely a wx-python dependency that needs fixing | 19:03 |
| CarlFK[m] | for sure you should use it to learn vocto, and maybe use it next week | 19:04 |
| CarlFK[m] | it is better than the 5-8 shell windows, one for each command | 19:04 |
| abhas | what would be the best resource to learn vocto? | 19:05 |
| mithro | abhas: Talking to Carl :-P | 19:05 |
| CarlFK[m] | playing with it | 19:05 |
| CarlFK[m] | :) | 19:05 |
| CarlFK[m] | vocto assumes you can get live video from the camera into vocto. what hardware do you have that can do that? | 19:06 |
| abhas | other than the opsis board? I have the laptop's built in webcan on /dev/video0 | 19:07 |
| CarlFK[m] | you mentioned a camera that you would hook audio into - what's that? | 19:08 |
| abhas | I don't know the exact model yet. | 19:10 |
| CarlFK[m] | what interface to you have on the computer? | 19:11 |
| abhas | no video capture interface... :( I assumed (wrongly) that I would be able to feed the camera output to the opsis and the switch between the inputs on the fly | 19:13 |
| abhas | I guess some sort of firewire interface might be required for this (I saw that the debconf people use firewire to capture the video) | 19:14 |
| CarlFK | debconf and the rest of us stopped using firewire about 2 years ago | 19:15 |
| CarlFK | it works, but it is harder now because less devices have support for it | 19:16 |
| abhas | hmm... Then I must've been reading an old account on the wiki | 19:16 |
| CarlFK | yep | 19:16 |
| abhas | so what sort of capture interface or card do you use? | 19:17 |
| CarlFK | https://www.blackmagicdesign.com/products/decklink/techspecs/W-DLK-06#W-DLK-06 DeckLink Mini Recorder | 19:18 |
| tpb | Title: Blackmagic Design: DeckLink Tech Specs (at www.blackmagicdesign.com) | 19:18 |
| CarlFK | audio into the camera, hdmi video+audio into the DeckLink card, ingest.py to feed that into voctomix | 19:18 |
| abhas | understood | 19:19 |
| CarlFK | mithro: do you know if the Opsis can do 2 inputs and no hdmi outputs? | 19:20 |
| mithro | CarlFK: should be able too, but haven't tested | 19:20 |
| abhas | okay -- so suppose I use the Opsis just to record (or stream) the presenter's laptop, can I do it using voctomix? Or even do the recording using mplayer or gstreamer? | 19:22 |
| CarlFK[m] | I would go with Opsis for the presenters laptop, usb webcam for a camera, some sort of linux audio device (either sound card or usb audio) to capture audio - feel all of this into vocto | 19:23 |
| CarlFK[m] | forget about live streaming. things are rushed enough | 19:24 |
| CarlFK[m] | the more you try to do the less you will actually do ;) | 19:24 |
| abhas | :) | 19:25 |
| CarlFK[m] | do you know if there will be a PA system (mic and loudspeakers) in the room for the audience ? | 19:26 |
| abhas | yup | 19:26 |
| abhas | Those will be there | 19:26 |
| CarlFK[m] | good - get audio form that. feed laptop analog audio into it. | 19:26 |
| abhas | I think there conference room will have a mixer to feed the audio to the speakers over loopback | 19:26 |
| abhas | so the mixer will give one composite (mixed) audio stream -- with the microphone + laptop audio together | 19:27 |
| CarlFK[m] | I am not sure what you mean by loopback but I don't care either. | 19:27 |
| CarlFK[m] | right | 19:28 |
| abhas | so what I will try to do now is learn how to configure voctomix to use 2 input video streams (webcam + opsis) and audio input from the laptop audio interface | 19:30 |
| CarlFK[m] | http://koipond.org.uk/Debconf/DebconfRoom3.svg | 19:34 |
| CarlFK[m] | make a diagarm like that | 19:34 |
| abhas | sure | 19:35 |
| CarlFK[m] | (webcam + opsis) and audio input - that is 2 usb video and 1 audio | 19:35 |
| CarlFK[m] | all vocto streams expect a video and audio track | 19:35 |
| CarlFK[m] | so make 2 inputs: webcam+pa audio, opsis+test audio tone (default in injest.py) | 19:36 |
| abhas | I was reading the ingest.py scripts and there are places where the video-source is specified as "blackmagic". Is this a reserved word? I initially thought, there must be a corresponding section in the ini file which maps the actual source device to this name, but the ini file does not mention it. | 19:36 |
| CarlFK[m] | use ingest.py for both - othewise if you have any probems the fist thin I will say is "use ingest" | 19:36 |
| CarlFK[m] | blackmagic is brand name of the DecLink card | 19:36 |
| abhas | yeah i gathered... so does it show up as a v4l device? | 19:37 |
| CarlFK[m] | ingest.py sends to a port - 10000 or 10001 - voctomix says "cam is on port 10000" | 19:37 |
| abhas | basically -- I am trying to understand how to name the inputs | 19:37 |
| CarlFK[m] | no, there is a gstreamer module to talk to the blackmangic driver. | 19:37 |
| CarlFK[m] | cam and grabber | 19:38 |
| abhas | oh okay | 19:38 |
| CarlFK[m] | https://github.com/CarlFK/video-stack-deploy/blob/nbpy/roles/voctomix/templates/voctocore.ini.j2 | 19:40 |
| tpb | Title: video-stack-deploy/voctocore.ini.j2 at nbpy · CarlFK/video-stack-deploy · GitHub (at github.com) | 19:40 |
| CarlFK[m] | I have to go to lunch now - be back in an hour ror so | 19:40 |
| CarlFK[m] | play around with things now at the prompt | 19:40 |
| CarlFK[m] | you need some comfort about constructing things with the simple commands | 19:41 |
| abhas | CarlFK: Sure. Thanks a lot for the pointers. Let me read this up | 19:41 |
| CarlFK[m] | if you want to try and use ansible to setup a box... https://github.com/CarlFK/video-stack-deploy/tree/nbpy/scripts#how-to-use | 19:42 |
| tpb | Title: video-stack-deploy/scripts at nbpy · CarlFK/video-stack-deploy · GitHub (at github.com) | 19:42 |
| CarlFK[m] | but would play first | 19:42 |
| abhas | I thought I'd do that. But it also sets up a lot of other stuff. So I was planning to read the scripts and undestand the process first and try to replicate it by hand before trying to automate it | 19:43 |
| CarlFK[m] | mithro (IRC): assuming modeswitch udev rules and everything is setup/installed, it *should* find and program the board regardless of what state the board is in, right? | 19:51 |
| CarlFK[m] | and if it doesn't, then look to the problem/solution section | 19:51 |
| * CarlFK[m] goes off to lunch now - skay says so. | 19:51 | |
| *** abhas has quit IRC | 19:56 | |
| mithro | CarlFK[m]: Correct -- it should even find the board *without* the udev rules installed | 20:06 |
| mithro | CarlFK[m]: It just might not be able to do anything to the board without the udev rules | 20:06 |
| *** abhas has joined #timvideos | 20:06 | |
| *** abhas has joined #timvideos | 20:09 | |
| abhas | so here is what I've been able to do quickly: 1) run voctocore and voctogui 2) run 3 separate ingest.py streams (opsis, webcam, test pattern) on the corresponding ports so that they show up correctly in voctogui | 20:12 |
| mithro | https://github.com/riscv/riscv-wiki/wiki/RISC-V-Cores-and-SoCs | 20:36 |
| tpb | Title: RISC V Cores and SoCs · riscv/riscv-wiki Wiki · GitHub (at github.com) | 20:36 |
| CarlFK[m] | back | 20:58 |
| CarlFK[m] | abhas (IRC): well done! | 20:59 |
| CarlFK | mithro: "Ubuntu 18.04 - Bionic Beaver" really!? | 21:58 |
| CarlFK | https://linux.conf.au/programme/miniconfs/fpga/ "A computer running one of the following 64bit Linux Distro, ..." | 21:59 |
| tpb | Title: linux.conf.au 2018 - Sydney 22-26 January 2018 (at linux.conf.au) | 21:59 |
| skay | mithro: is that a typo? what do you run? | 22:00 |
| skay | mithro: should I try and convince someone to try out things on bionic????? | 22:00 |
| * skay is running artful for regular stuff and there are still annoyances and broken things with it | 22:01 | |
| mithro | ? | 22:42 |
| mithro | Hrm... I just copied and pasted from somewhere... | 22:43 |
| mithro | skay: Have you seen Probot? https://changelog.com/podcast/264 | 22:43 |
| CarlFK[m] | mithro (IRC): I;m pretty sure you don't want Bionic - stuff will break on you | 22:47 |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!
