| *** tpb has joined #timvideos | 00:00 | |
| mithro | Currently watching the recording from the video last night.... | 00:28 |
|---|---|---|
| mithro | https://twit.tv/shows/floss-weekly/episodes/359?autostart=false | 00:29 |
| tpb | Title: FLOSS Weekly 359 | TWiT (at twit.tv) | 00:29 |
| CarlFK | mithro: good morning - can you give me the time that you talk about presenters having trouble hooking up and how we only give them 1 choice ? | 00:38 |
| mithro | CarlFK: When I find it, I'll give you the timecode | 00:40 |
| CarlFK | thanks | 00:40 |
| shenki | good morning | 00:57 |
| shenki | who is our makestuff expert? | 00:57 |
| shenki | https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/120 | 00:57 |
| tpb | Title: Bootstrap fails · Issue #120 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 00:57 |
| mithro | shenki: It seems like liberror is getting pulled by `make deps` | 01:00 |
| shenki | mithro: getting pulled? | 01:01 |
| mithro | shenki: downloaded | 01:01 |
| mithro | shenki: you could try using openocd as the programmer instead | 01:02 |
| shenki | mithro: okay. i kind of wanted to build the enitre show and get a feel for it (my urgency is decreased now that i couldn't get it going last night) | 01:02 |
| shenki | mithro: are you planning on moving to openocd? | 01:02 |
| mithro | shenki: I think we are going to end up moving to openocd | 01:02 |
| mithro | shenki: It already works with the Opsis | 01:03 |
| shenki | mithro: okay | 01:04 |
| shenki | mithro: so openocd has a firmware for the fx2? | 01:04 |
| shenki | (how does the fpga get programmed) | 01:04 |
| mithro | shenki: No, that is provided by the ixo-usb-jtag | 01:04 |
| mithro | shenki: but there is a PPA package for that which has udev rules to load it | 01:05 |
| shenki | mithro: cool | 01:05 |
| mithro | shenki: https://launchpad.net/~timvideos/+archive/ubuntu/fpga-support | 01:06 |
| tpb | Title: Support packages for FPGA development : “TimVideos.Us - Live event streaming” team (at launchpad.net) | 01:06 |
| mithro | CarlFK: the discussion about the problem starts around 41:12 | 01:08 |
| mithro | Heading to lunch | 01:08 |
| mithro | bblr | 01:08 |
| mithro | shenki: you'll need my "everything" branch of openocd | 01:08 |
| mithro | shenki: hopefully the changes will get into upstream openocd soon, they have a +2 already | 01:09 |
| CarlFK | mithro: thanks | 01:12 |
| CarlFK | shenki: i put tim's openocd everything branch on my ppa for trusty | 01:13 |
| shenki | mithro: nice | 01:22 |
| shenki | CarlFK: cool | 01:23 |
| CarlFK | shenki: http://paste.ubuntu.com/12882007/ | 01:24 |
| tpb | Title: Ubuntu Pastebin (at paste.ubuntu.com) | 01:24 |
| mithro | If you install fxload from CarlFK's repo it is likely to break a couple of other things | 02:05 |
| mithro | shenki: if you have my openocd in the path, just do a "PROG=openocd make load-gateware" | 02:06 |
| mithro | Bertl: A couple of the riser cables I ordered turned up today | 03:02 |
| *** CarlFK has quit IRC | 03:14 | |
| *** CarlFK has joined #timvideos | 03:31 | |
| *** ChanServ sets mode: +v CarlFK | 03:31 | |
| *** Bertl is now known as Bertl_zZ | 06:26 | |
| *** se6astian|away is now known as se6astian | 08:36 | |
| G33KatWork | mithro: sorry, i was busy over the last two days. I didn't try the fx2 yet. I played around with migen and my atlys over the weekend and basically read a bit of code and tried things using adept (and needed to fix my docker container to use USB devices which took quite a while). | 09:43 |
| G33KatWork | i'm just building the video mixer bitstream for the opsis and will try the fx2 | 09:43 |
| G33KatWork | now that my env seems to work :) | 09:43 |
| G33KatWork | oh and I just tested it, the HDMI2USB bitstream on the atlys seems to cause problems during routing as well | 09:44 |
| G33KatWork | any suggestions what to do for debugging why it's not able to route it? | 09:44 |
| seaLne | won't route or generates errors? i've seen it complaining but still producing a bit file for the opsis | 09:50 |
| G33KatWork | it generates a bitfile with unrouted signals | 09:52 |
| G33KatWork | after running for ages | 09:52 |
| G33KatWork | like a few hours | 09:53 |
| G33KatWork | mithro: i suppose i should use your opsis-fixes branch for the opsis, right? | 09:55 |
| seaLne | yes | 09:57 |
| seaLne | though i've not had much luck so far | 09:57 |
| seaLne | so getting another guinea pig would be good | 09:58 |
| G33KatWork | k, i'll try my luck. give me a few minutes | 10:04 |
| seaLne | i'm using mithro openocd fork https://github.com/mithro/openocd for jtag | 10:08 |
| tpb | Title: mithro/openocd · GitHub (at github.com) | 10:08 |
| seaLne | which has the fixes to support tho opsis | 10:08 |
| G33KatWork | >usb-jtag - Altera USB Blaster Emulation with a FX2 | 10:20 |
| G33KatWork | lol | 10:20 |
| G33KatWork | using an altera usb blaster for a xilinx fpga :> | 10:20 |
| G33KatWork | hm, how do i actually load the gateware on the opsis? | 10:28 |
| seaLne | make load-gateware | 10:29 |
| G33KatWork | >ERROR:iMPACT:1075 - setCable must be set before this operation. | 10:30 |
| seaLne | if using openocd PROG=openocd make load-gateware | 10:30 |
| G33KatWork | my fx2 is still in unconfigured mode according to its usb PID | 10:30 |
| seaLne | its loaded with make load-fx2 | 10:30 |
| seaLne | currently i can't connect with make connect-lm32 after make load-fx2 | 10:32 |
| G33KatWork | the fx2 loaded, but i still can't load the gateware | 10:34 |
| G33KatWork | but okay, it's trying to do it via impact | 10:34 |
| G33KatWork | how does openocd connect to the fpga? using the fx2? | 10:35 |
| seaLne | i had to tell the openocd configure to add support for the blaster | 10:35 |
| G33KatWork | if i use that, i need to load another firmware in the fx2, right? | 10:35 |
| seaLne | don't think so | 10:35 |
| seaLne | turn opsis on; dmesg should show JTAG; PROG=openocd make load-gateware; make load-fx2 | 10:36 |
| G33KatWork | http://opsis.hdmi2usb.tv/getting-started/usb-ids.html#unconfigured-mode | 10:37 |
| tpb | Title: Numato Opsis : USB IDs (at opsis.hdmi2usb.tv) | 10:37 |
| G33KatWork | this is the mode i get | 10:37 |
| seaLne | not sure sorry | 10:38 |
| G33KatWork | not a problem | 10:38 |
| G33KatWork | i slowly understand how it is supposed to work | 10:39 |
| G33KatWork | my fpga is unconfigured and doesn't supply the fx2 with a firmware | 10:39 |
| G33KatWork | so it comes up as device 2a19:5440 | 10:39 |
| G33KatWork | yours seems to be able to do that, so it comes up as 2a19:5441 which provides usb-jtag and uart according to the table | 10:40 |
| mithro | G33KatWork: almost | 10:40 |
| G33KatWork | ah yeah | 10:41 |
| G33KatWork | make load-fx2 does that | 10:41 |
| G33KatWork | in our cases | 10:41 |
| G33KatWork | it now comes up as 5442 | 10:41 |
| mithro | G33KatWork: seaLne has the ixo-usb-jtag package | 10:41 |
| G33KatWork | in the mode switcher? | 10:41 |
| mithro | Which has a udev rule which takes the Opsis from unconfigured to JTAG mode | 10:42 |
| G33KatWork | does the fx2 firmware in the misoc-git also provide jtag access? | 10:42 |
| G33KatWork | alright, got it | 10:42 |
| seaLne | ah, i'd forgoten that step | 10:42 |
| mithro | No | 10:42 |
| mithro | Well, the make stuff programmer provides JTAG firmware for the fx2 on the Atlys but doesn't work with the Opsis | 10:43 |
| G33KatWork | k, so i use the firmware from the HDMI2USB-mode-switch git? | 10:43 |
| mithro | Ixo-usb-jtag works with both the Opsis and Atlys | 10:43 |
| mithro | I think the mode switch has a slightly older ixo-usb-jtag firmware in it | 10:44 |
| mithro | Need to update it | 10:45 |
| G33KatWork | i just tried to compile it | 10:45 |
| G33KatWork | cp: cannot stat 'examples/cdc/to-uart/build/cdc-acm-to-uart.ihx': No such file or directory | 10:45 |
| mithro | My branch is needed btw | 10:46 |
| G33KatWork | yeah, got that one | 10:46 |
| mithro | mode-switch? | 10:46 |
| G33KatWork | //github.com/mithro/HDMI2USB-mode-switch.git | 10:46 |
| G33KatWork | yeah | 10:46 |
| mithro | Okay, I haven't tested the build firmware script in the last couple of days | 10:47 |
| G33KatWork | might be my weird setup again ;) | 10:47 |
| mithro | Nah, just likely haven't updated it for other changes | 10:48 |
| G33KatWork | ixo-usb-jtag builds fine | 10:50 |
| mithro | The one which failed is the serial UART emulation | 10:50 |
| G33KatWork | yeah | 10:50 |
| mithro | When I finish making the FPGA emulate an EEPROM so it provides firmware to the fx2, this should all go away | 10:51 |
| G33KatWork | well, i need to understand it anyway if i want to customize stuff later | 10:52 |
| mithro | G33KatWork: BTW do you want a little starter project to get going with our firmware? | 10:52 |
| mithro | Got plenty of options :-) | 10:53 |
| mithro | BTW i'm currently on my phone | 10:53 |
| G33KatWork | somethings borked with the paths in the build script | 10:53 |
| G33KatWork | if i build it from hand it seems to work | 10:54 |
| G33KatWork | what do you have in mind? | 10:54 |
| G33KatWork | project-wise | 10:54 |
| mithro | Depends on what you want to work on | 10:54 |
| mithro | IE it would be good to add CRC calculation for incoming frame data and then checking on the output | 10:55 |
| mithro | Or a "flipped" mode were we invert the image | 10:56 |
| *** Bertl_zZ is now known as Bertl | 10:56 | |
| mithro | Or https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/94 | 10:56 |
| tpb | Title: Have frequency (and hsync/vsync) information on the inputs in the status message · Issue #94 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 10:57 |
| G33KatWork | for another project i'm currently implementing a stereo vision system on a zynq. this weekend i finished the PCB for my small custom camera. I thought about porting that stuff over to the opsis once i'm done with all that as a case-study | 10:58 |
| G33KatWork | just read a few papers on how to do rectification on an fpga | 10:58 |
| mithro | Or https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/35 | 10:58 |
| tpb | Title: Dump EDID from connected monitor · Issue #35 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 10:58 |
| G33KatWork | yeah, cool. i'll have a look at a few issues once i get everything running | 10:59 |
| mithro | These are probably a couple of days / hours at most projects | 10:59 |
| mithro | Just little tasters :-) | 10:59 |
| G33KatWork | i thought about implementing an i2c master core using migen | 11:00 |
| G33KatWork | because that doesn't exist yet | 11:00 |
| G33KatWork | it seems | 11:00 |
| mithro | I have a bigger project to create some pixel shaders for HDMI2USB scaling | 11:00 |
| G33KatWork | even you guys do bitbanging | 11:00 |
| G33KatWork | haha, pixel shaders. that's a bit too crazy for me :D | 11:01 |
| G33KatWork | but it sounds awesome | 11:01 |
| mithro | It's surprisingly easy | 11:01 |
| mithro | But I estimate a 3 month project for a good PhD student | 11:01 |
| G33KatWork | might be usable for my stereo vision suff as well to perform the rectification | 11:01 |
| mithro | yeah | 11:02 |
| G33KatWork | i found a nice paper where they use lookup-tables and a fancy compression scheme to make them fit inside of BRAMs in the fpga | 11:02 |
| mithro | If you do create an i2c core, you should do something similar to what the ardunios have :-) | 11:02 |
| G33KatWork | you mean avrs? | 11:03 |
| mithro | Yeah | 11:03 |
| G33KatWork | (hipstermode: when i used microcontrollers first, arduinos weren't a thing yet) | 11:03 |
| G33KatWork | and i'm not even that old :> | 11:03 |
| mithro | I need EEPROM emulation for the fx2 if you want to work on that :-) | 11:04 |
| mithro | I used PICs | 11:04 |
| mithro | They are ****so**** much better then avrs | 11:04 |
| G33KatWork | i started with AVRs, because they are were easily programmable | 11:04 |
| G33KatWork | parallel port and a few wires | 11:04 |
| mithro | But for some reason the avrs won despite PICs being better in every way | 11:04 |
| G33KatWork | and the open source toolchain | 11:05 |
| mithro | I'll be hacking on stuff on Sunday | 11:05 |
| G33KatWork | and they don't have weird register widths :D | 11:05 |
| G33KatWork | 18 bit... wtf? | 11:05 |
| mithro | Have work offsite for the next 2 days | 11:06 |
| G33KatWork | and there is no freaking disassembler for PIC24s out there | 11:06 |
| G33KatWork | not even IDA | 11:06 |
| mithro | If you are going that level just go arm | 11:07 |
| G33KatWork | yeah, already did that | 11:07 |
| mithro | PIC12 or PIC18 | 11:07 |
| G33KatWork | these stm32f4 discovery boards are so cheap | 11:07 |
| mithro | 12 cent micro :-) | 11:07 |
| G33KatWork | we built a small gaming console using these with a bit-banged vga port | 11:08 |
| G33KatWork | for a youth project at our hackerspace | 11:08 |
| mithro | Cool | 11:09 |
| G33KatWork | >raise NotImplemented("Not yet finished...") | 11:09 |
| G33KatWork | meh ;) | 11:09 |
| G33KatWork | hm, how do i load the gateware now using openocd? | 11:18 |
| G33KatWork | ah, PROG, not PROGRAMMER | 11:19 |
| G33KatWork | mithro: what configure flags do i need for the jtag interface? | 11:24 |
| G33KatWork | i already have --with-ftdi and usb-blaster-2, but it still whines about a missing lowlevel driver | 11:25 |
| G33KatWork | --enable, not with | 11:25 |
| mithro | :-) | 11:26 |
| mithro | Heading to bed | 11:26 |
| G33KatWork | hm, it still doesn't like it | 11:27 |
| G33KatWork | mithro: gn8 | 11:27 |
| G33KatWork | --enable-usb_blaster_libftdi did the trick | 11:32 |
| G33KatWork | okay, gateware is loaded | 11:32 |
| G33KatWork | loaded the fx2 serial firmware | 11:35 |
| G33KatWork | tried to connect using flterm | 11:35 |
| G33KatWork | i get a bunch of usb errors | 11:35 |
| G33KatWork | and the serial port disappears | 11:35 |
| G33KatWork | *sadpanda* | 11:35 |
| CarlFK | G33KatWork: hi | 14:10 |
| CarlFK | what board are you using? | 14:10 |
| G33KatWork | Opsis | 14:24 |
| CarlFK | k - seems that is different enough from the Atlys that I can't really help | 14:43 |
| *** se6astian is now known as se6astian|away | 15:07 | |
| *** CarlFK has quit IRC | 16:28 | |
| *** CarlFK has joined #timvideos | 16:29 | |
| *** ChanServ sets mode: +v CarlFK | 16:29 | |
| *** se6astian|away is now known as se6astian | 16:41 | |
| *** GChriss has joined #timvideos | 17:16 | |
| *** se6astian is now known as se6astian|away | 22:34 | |
| *** Bertl is now known as Bertl_zZ | 22:42 | |
| *** CarlFK has quit IRC | 22:44 | |
| *** CarlFK has joined #timvideos | 23:38 | |
| *** ChanServ sets mode: +v CarlFK | 23:38 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!