| *** tpb has joined #timvideos | 00:00 | |
| shenki | hello | 00:01 |
|---|---|---|
| shenki | mithro: badmouthing me to the community i see | 00:01 |
| shenki | ;) | 00:01 |
| shenki | Bertl: hello | 00:01 |
| shenki | Bertl: where do i know you from? | 00:01 |
| Bertl | OLPC | 00:01 |
| shenki | ohh, yes, i recongise the nick now | 00:01 |
| Bertl | how are you? everything fine? | 00:02 |
| shenki | yeah, doing well. i work on server firmware these days | 00:02 |
| Bertl | I heard from mithro | 00:02 |
| shenki | lots of FOSS, all out in the open. it's good | 00:03 |
| shenki | how about yourself? | 00:03 |
| Bertl | that's the way it is supposed to be :) | 00:03 |
| Bertl | my current focus is on the Axiom camera | 00:03 |
| Bertl | and I'm fine except for an ugly cold at the moment | 00:03 |
| shenki | i just searched for the axiom; looks impressive! | 00:06 |
| shenki | sorry to hear about your cold; get well soon! | 00:07 |
| shenki | Bertl: what are you doing on the axiom? | 00:07 |
| Bertl | a lot :) | 00:07 |
| Bertl | hardware, software, firmware | 00:08 |
| mithro | shenki: I've sent Bertl one of the preproduction Opsis boards | 00:13 |
| shenki | mithro: cool | 00:21 |
| shenki | Bertl: nice. what does it run? | 00:21 |
| Bertl | arch linux | 00:26 |
| shenki | on x86? | 00:46 |
| Bertl | no, the Zynq has two arm cores | 00:46 |
| shenki | ah yep | 00:46 |
| shenki | how do you find the zynq? | 00:46 |
| Bertl | well, it is overrated, but we started with a zedboard, and we got stuck with the fast bringup and easy way to control FPGA stuff | 00:47 |
| Bertl | the very good prices for Zynq development boards (compared to most other FPGAs) played a role too | 00:48 |
| shenki | ok | 00:50 |
| xfxf | oh, rad, I was looking at the Axiom camera's the other day, seem like really neat kit | 01:43 |
| Bertl | glad you like it | 01:45 |
| *** Bertl is now known as Bertl_oO | 01:45 | |
| *** Niharika has quit IRC | 02:30 | |
| *** xfxf has quit IRC | 02:31 | |
| *** CARAM__ has quit IRC | 02:34 | |
| mithro | Hey xfxf. | 02:51 |
| mithro | Going to do some hacking / testing today? | 02:51 |
| *** xfxf has joined #timvideos | 03:04 | |
| *** Niharika has joined #timvideos | 03:55 | |
| *** CARAM__ has joined #timvideos | 04:01 | |
| CarlFK | mithro: is your "find board and upload bytes" .py app checked in yet? | 06:05 |
| mithro | CarlFK: its not at that stage yet - but yes | 06:06 |
| mithro | https://github.com/mithro/HDMI2USB-find-board | 06:06 |
| tpb | Title: mithro/HDMI2USB-find-board · GitHub (at github.com) | 06:06 |
| CarlFK | mithro: what did you want to use instead of makestuff ? | 06:10 |
| mithro | CarlFK: openocd seems to be the best option | 06:10 |
| mithro | Bertl_oO: any chance you can get me a populated 1xHDMI Plugin Module? | 06:14 |
| CarlFK | mithro: looks like it is already packaged: https://packages.debian.org/jessie/openocd | 06:19 |
| tpb | Title: Debian -- Details of package openocd in jessie (at packages.debian.org) | 06:19 |
| mithro | CarlFK: yes - but it needs patches | 06:20 |
| mithro | CarlFK: http://openocd.zylin.com/#/q/owner:mithro%2540mithis.com+status:open | 06:20 |
| tpb | Title: Gerrit Code Review (at openocd.zylin.com) | 06:20 |
| mithro | Bertl_oO: Do you have a schematic for that 1x HDMI board? | 06:26 |
| CarlFK | mithro: is your patched version somewhere I can check out? | 06:31 |
| mithro | CarlFK: nope | 06:31 |
| mithro | CarlFK: let me upload it somewhere | 06:32 |
| Bertl_oO | mithro: http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_plugin_module_1xHDMI_v0.5.sch.pdf | 06:32 |
| mithro | Bertl_oO: thanks! | 06:32 |
| CarlFK | thanks.. I'll see about rolling the patches into the packaging | 06:32 |
| mithro | CarlFK: They'll be upstream shortly | 06:32 |
| CarlFK | mithro: shortly = hours or days? | 06:33 |
| Bertl_oO | mithro: and yes, I can send you one once we start populating them for production (which should be soon) | 06:33 |
| mithro | CarlFK: days | 06:33 |
| mithro | CarlFK: but not weeks | 06:33 |
| mithro | CarlFK: my stuff will be at https://github.com/mithro/openocd soon | 06:34 |
| tpb | Title: mithro/openocd · GitHub (at github.com) | 06:34 |
| CarlFK | mithro: k. now that I think about it it will be days before I can spend any time on it . i am 1/2 asleep right now and may not be thinking to well :p | 06:35 |
| mithro | CarlFK: I'd work on getting a package from their git repo going | 06:36 |
| CarlFK | mithro: btw - 9 more talks done using the Atlys. | 06:36 |
| mithro | CarlFK: great! That doesn't give me much information though | 06:38 |
| CarlFK | mithro: once I got it up, it just worked. | 06:39 |
| mithro | CarlFK: Can you start recording information about speakers laptops in some way? | 06:40 |
| CarlFK | mithro: I have been thinking of trying to collect speakers laptop model number before the event. "what do you plan on using?" | 06:41 |
| CarlFK | trying to get that at the event... not enough time | 06:41 |
| mithro | CarlFK: can you snap a photo of it? | 06:41 |
| CarlFK | mithro: not easily. for the 9 talks I was in the back of the room. 1/2 the time the next talk was being announced before the presenter had put their laptop on the podium | 06:44 |
| mithro | CarlFK: It's annoying that we can't figure out what type of laptop it is via the board itself | 06:46 |
| mithro | and it is generally frowned upon to exploit their machine via the display adapter to run a script to upload it :P | 06:48 |
| *** xfxf1 has joined #timvideos | 07:03 | |
| xfxf1 | mithro: compiled up that older firmware, input0 still not working | 07:04 |
| mithro | xfxf1: which one did you compile? | 07:04 |
| xfxf1 | the commit you linked me to - one from 9th sept afaik | 07:04 |
| mithro | xfxf1: can you try a couple of versions older then that one too? | 07:05 |
| xfxf1 | sure | 07:05 |
| xfxf1 | oddly this 2nd gen i7 takes exponentially longer to compile than my macbook retina i7 inside a VM | 07:06 |
| xfxf1 | am observing a lot of the compile toolchain seems to be single threaded? | 07:07 |
| xfxf1 | i suspect turbo boost isn't working on the i7 properly, but any tricks to get things happening faster than the defaults? | 07:07 |
| *** rohitksingh has joined #timvideos | 07:07 | |
| mithro | xfxf1: the compile toolchains are single threaded | 07:08 |
| mithro | Bertl_oO: what are the VCC, VCCIO, V_I2C and PWRGOOD pins in your spec? | 07:08 |
| xfxf1 | any progress on making your travis CI stuff build firmware daily? | 07:09 |
| xfxf1 | because that's be super rad for being able to step back to older firmwares for testing regressions rapidly | 07:09 |
| xfxf1 | assuming you'd keep all historical firmware | 07:09 |
| mithro | xfxf1: it'll build every commit from now on | 07:09 |
| xfxf1 | right, but does it make available publically each firmware, or only the latest? | 07:10 |
| mithro | xfxf1: each firmware | 07:10 |
| xfxf1 | oh, neat, good stuff | 07:10 |
| mithro | xfxf1: it just updates a symlink | 07:10 |
| mithro | to tell you which one is the "current unstable" | 07:10 |
| xfxf1 | how hard/feasible would it to make whatever CI hook scripts you're using step through the last two months or so of commits? | 07:10 |
| xfxf1 | what was the right sequence for events (make targets) to flash the gateware, firmware, the connect to the device? | 07:14 |
| xfxf1 | i'll update flash-hdmi2usb.sh as it's currently broken | 07:14 |
| xfxf1 | alternatively i note you have a 'firmware' Makefile target, i can just shove it in there | 07:16 |
| Bertl_oO | mithro: VCC is a configureable voltage (typically 3.3V though), VCCIO is the reference voltage for the LVDS pairs | 07:16 |
| mithro | Bertl_oO: so in your case that is 2.5V? | 07:17 |
| Bertl_oO | V_I2C is a separate supply usually for reading the EEPROM before the card is activated | 07:17 |
| Bertl_oO | in our case either 1.8V or 2.5V yes | 07:17 |
| mithro | Bertl_oO: PWRGOOD is some type of power good? | 07:30 |
| Bertl_oO | yeah, but we haven't decided yet what to do with it :) | 07:31 |
| mithro | Bertl_oO: I'd recommend changing it to a "RST" pin | 07:32 |
| Bertl_oO | good idea, will see if we can do that | 07:32 |
| mithro | Bertl_oO: basically holding low (or high?) causes the board to stay in reset state, releasing makes it start operating -- basically the same to PWRGOOD I guess | 07:33 |
| Bertl_oO | yeah, but I have to check if we have I/Os left over for this | 07:34 |
| Bertl_oO | no point of a reset when it cannot be asserted | 07:35 |
| Bertl_oO | hmm, is a little tricky, definitely won't happen in the Early Betas | 07:38 |
| Bertl_oO | i.e. please ignore for now or assign an I/O whatever you like | 07:39 |
| Bertl_oO | it is not used anywhere as far as I can see | 07:39 |
| Bertl_oO | but we can label it PWRGD/#RST | 07:41 |
| Bertl_oO | off for a nap ... bbl | 07:44 |
| *** Bertl_oO is now known as Bertl_zZ | 07:45 | |
| xfxf1 | mithro: stepped back to a commit on 8th sept about 4 before the otehr, input0 not working, stepping back to about a week earlier now | 07:50 |
| xfxf1 | hopefully will be able to indicate the commit where the regression occurs assuming no build issues | 07:50 |
| mithro | I'm wondering if there is something we are missing. | 07:51 |
| xfxf1 | i'm having problems building older firmware, something to do with liteeth | 08:09 |
| xfxf1 | not sure if it's worth me persisting, comments mithro? | 08:09 |
| xfxf1 | my goal here is to get you info to get the board fixed asap | 08:10 |
| mithro | xfxf1: yeah the older firmware is probably not going to build because it didn't use submodules so there isn't any way to really go back to that date | 08:10 |
| xfxf1 | i want to use it but currently can't, have now sat through two user groups now with it not working | 08:10 |
| xfxf1 | ok | 08:10 |
| xfxf1 | so what can i do to help? | 08:10 |
| mithro | xfxf1: You have one input working right? | 08:12 |
| xfxf1 | yes but i want to use both | 08:12 |
| xfxf1 | and i'm yet to get reliable USB capture | 08:13 |
| mithro | xfxf1: work on reliable USB capture | 08:13 |
| xfxf1 | /dev/videoX doesn't reliably appear | 08:13 |
| xfxf1 | sure, i don't have that | 08:13 |
| xfxf1 | as per issues posted the other day | 08:13 |
| mithro | xfxf1: Which issues? | 08:13 |
| xfxf1 | the video device didn't appear? | 08:14 |
| xfxf1 | and it was unusable under OS X, but i've got a dedicated linux laptop now | 08:14 |
| xfxf1 | i ran the commands you told me (which i have forgotten; as per above, can you let me know the right sequence of make commands to get a board working that has usb capture working + is controllable? i will fix the flash script) | 08:14 |
| xfxf1 | and it wasn't producing reliable results | 08:14 |
| xfxf1 | i found i had to reset the board and try again | 08:14 |
| xfxf1 | it was failing enough that i just gave up trying to capture with it, it was holding up presenters | 08:15 |
| mithro | xfxf1: you do know that the IRC channel has logs? | 08:16 |
| xfxf1 | sure, why is that an issue? this is a board with a non-stable firmware, stuff like this is expected, no? | 08:16 |
| mithro | xfxf1: I me regarding the "which I have forgotten part" | 08:17 |
| mithro | s/me/mean/ | 08:17 |
| mithro | make load-gateware; make load-fx2; make connect-lm32 | 08:17 |
| xfxf1 | oh - right, nod, will go back then. i hadn't done that because irccloud seems broken at the moment (hence logging on with another client) so my logs weren't accessible | 08:17 |
| xfxf1 | ta | 08:17 |
| mithro | xfxf1: we have publicly available logs with search :P | 08:19 |
| mithro | xfxf1: logs.timvideos.us/%23timvideos/search?q=xfxf | 08:19 |
| *** mithro changes topic to "Website - http://code.timvideos.us | Code - http://github.com/timvideos | Logs at http://logs.timvideos.us/%23timvideos/ | Mailing lists: https://groups.google.com/group/timvideos" | 08:21 | |
| *** mithro changes topic to "Website - http://code.timvideos.us | Code - http://github.com/timvideos | Logs - http://logs.timvideos.us/%23timvideos/ | Mailing lists - https://groups.google.com/group/timvideos" | 08:21 | |
| mithro | xfxf1: "make load" should work once https://github.com/timvideos/HDMI2USB-misoc-firmware/issues/90 is fixed | 08:27 |
| tpb | Title: `make load-lm32` should actually load the firmware and then exit · Issue #90 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 08:27 |
| *** CARAM__ has quit IRC | 08:38 | |
| *** Niharika has quit IRC | 08:38 | |
| *** xfxf has quit IRC | 08:38 | |
| mithro | xfxf1: you still here? | 08:43 |
| xfxf1 | mithro, having dinner, then will do above | 08:45 |
| mithro | xfxf1: I need someone to +1 the openocd change by checking it works on their hardware too | 08:45 |
| xfxf1 | sure, tell me what i need to do | 08:46 |
| mithro | xfxf1: you have a Ubuntu trusty system right? | 08:47 |
| xfxf1 | correct | 08:47 |
| mithro | xfxf1: Install the ixo-usb-jtag from our FPGA support PPA | 08:48 |
| mithro | xfxf1: then clone my version of openocd from github | 08:48 |
| mithro | xfxf1: once you get to that, I'll explain what to do | 08:48 |
| *** CARAM__ has joined #timvideos | 08:52 | |
| xfxf1 | will do, but best you give me a list of instructions - balancing doing this with family duties so i'll be coming back and forth from my computer | 08:53 |
| xfxf1 | ok, both instructions above done | 08:55 |
| xfxf1 | and literally; it's cloned, i haven't compiled it | 08:56 |
| *** Niharika has joined #timvideos | 08:57 | |
| xfxf1 | mithro, ping | 09:03 |
| mithro | xfxf1: checkout master | 09:04 |
| mithro | xfxf1: compile it | 09:04 |
| mithro | xfxf1: ./configure --prefix=/opt/openocd | 09:04 |
| mithro | Oh | 09:04 |
| mithro | you might need to run ./bootstrap first | 09:04 |
| xfxf1 | if you're intending on scripting this, had to install libtool/autoconf too | 09:06 |
| mithro | xfxf1: not really | 09:07 |
| mithro | xfxf1: if that all compiles and then you can run /opt/openocd/bin/openocd --help | 09:08 |
| mithro | xfxf1: then you can test with my patches | 09:08 |
| xfxf1 | yep, working | 09:09 |
| xfxf1 | assume you don't want any output from --help, rather to check the binary works? | 09:09 |
| mithro | xfxf1: yeah | 09:12 |
| mithro | great! | 09:12 |
| mithro | Did you install the ixo-usb-jtag package from the PPA? | 09:13 |
| xfxf1 | yes | 09:13 |
| mithro | xfxf1: okay | 09:13 |
| mithro | xfxf1: plug in your Atlys board | 09:13 |
| xfxf1 | already plugged in | 09:13 |
| mithro | xfxf1: power cycle your Atlys board then | 09:14 |
| xfxf1 | same :) | 09:14 |
| mithro | xfxf1: In the lsusb output you shouldn't see the Atlys board, you should see something with the vid_pid of 0x16C0 0x06AD | 09:15 |
| mithro | Bertl_zZ: when you get up, I have a mapping for the TOFE->apertus interface I'd like you to look over | 09:16 |
| *** rohitksingh has quit IRC | 09:16 | |
| mithro | might be labelled something like | 09:18 |
| mithro | Van Ooijen Technische Informatica | 09:18 |
| *** xfxf has joined #timvideos | 09:28 | |
| xfxf1 | mithro, nothing with that vid/pid | 09:34 |
| mithro | xfxf1: what does lsusb show? | 09:34 |
| xfxf1 | 04e2:1410 exar corp | 09:35 |
| mithro | xfxf1: That is the uart | 09:35 |
| mithro | xfxf1: There are two micro usb ports on the Atlys | 09:35 |
| mithro | xfxf1: you need both of them plugged in | 09:35 |
| xfxf1 | yes, i have both connected | 09:35 |
| xfxf1 | i'm only seeing that | 09:35 |
| mithro | xfxf1: can you show me the full output of lsusb? | 09:36 |
| xfxf1 | ugh, dmesg is showing 'unable to enumerate USB device on port 1' | 09:36 |
| xfxf1 | let me try another cable | 09:36 |
| *** Niharika has quit IRC | 09:39 | |
| *** Niharika has joined #timvideos | 09:39 | |
| xfxf1 | nope, doing it with different cables | 09:41 |
| xfxf1 | rebooting machine | 09:41 |
| mithro | xfxf1: hrm - I'm getting it here too - trying to figure out what is going on | 09:41 |
| xfxf1 | has your package changed anything to do with that port? | 09:41 |
| xfxf1 | ah | 09:41 |
| xfxf1 | glad i'm of service as a rubber duck at least ;) | 09:43 |
| mithro | xfxf1: can you uninstall the package | 09:46 |
| mithro | xfxf1: and clone my ixo-usb-jtag package | 09:46 |
| xfxf1 | sure, i assume i can dpkg-buildpackage it or something? | 09:47 |
| mithro | yeah | 09:47 |
| mithro | xfxf1: run fakeroot debian/rules get-orig-source first | 09:48 |
| xfxf1 | ok | 09:48 |
| xfxf1 | typo in debian/rules, you want https on your repo, not http | 09:49 |
| xfxf1 | installed package, same problem, unsure if i need to do anything else | 09:50 |
| mithro | Hrm... | 09:50 |
| mithro | What does "sdcc --version" on your machine say? | 09:50 |
| mithro | xfxf1: you uninstalled the package and power cycled the Atlys right? | 09:51 |
| xfxf1 | root@capcore:~/src/ixo-usb-jtag# sdcc --version | 09:51 |
| xfxf1 | SDCC : mcs51/gbz80/z80/z180/r2k/r3ka/ds390/pic16/pic14/TININative/ds400/hc08/s08 3.3.0 #8604 (Dec 30 2013) (Linux) | 09:51 |
| xfxf1 | and yes to both | 09:51 |
| mithro | hrm.... | 09:52 |
| mithro | What does | 09:52 |
| mithro | md5sum output/hw_nexys.hex | 09:52 |
| mithro | give? | 09:52 |
| xfxf1 | 92bf118bc3eb07aad7f8c09f453a587c output/hw_nexys.hex | 09:52 |
| mithro | Hrm | 09:54 |
| mithro | xfxf1: can you try | 09:54 |
| mithro | rm -rf output; fakeroot debian/rules clean; fakeroot debian/rules binary; md5sum output/hw_nexys.hex | 09:54 |
| xfxf1 | same result as above | 09:55 |
| mithro | same md5sum? | 09:55 |
| xfxf1 | btw if it makes things easier i can easily enough give you remote ssh access to this box, equally happy to follow instructions | 09:55 |
| mithro | xfxf1: I don't quite understand why you are getting different build output | 09:56 |
| mithro | xfxf1: but every time you build you get the same md5sum? | 09:56 |
| xfxf1 | correct | 09:57 |
| xfxf1 | just did it again | 09:57 |
| mithro | xfxf1: what does "git log" show? | 09:57 |
| xfxf1 | dpkg-deb: building package `ixo-usb-jtag' in `../ixo-usb-jtag_0.0.0-85-g2df5d2f-1_all.deb'. | 09:57 |
| xfxf1 | 92bf118bc3eb07aad7f8c09f453a587c output/hw_nexys.hex | 09:57 |
| mithro | Can you send me your output/hw_nexys.hex? | 09:58 |
| mithro | xfxf1: can you give me remote access? I still need someone to reset the Atlys though | 10:01 |
| xfxf1 | mithro: are you getting my messages? can't recall if freenode allows private messages when not authenticated | 10:04 |
| mithro | Yes | 10:04 |
| xfxf1 | good | 10:04 |
| *** se6astian has joined #timvideos | 10:04 | |
| xfxf1 | atlys reset, ping me if you need me to reset it again at any point | 10:04 |
| xfxf1 | feel free to apt-get or install whatever on it too, it'll end up getting reloaded at some point once i'm confident in stability of use of the device and i can roll more of a specialised 'turn on and record' install for it | 10:05 |
| mithro | xfxf1: can you restart the atlys | 10:06 |
| xfxf1 | done | 10:06 |
| mithro | so... the firmware I build on my computer works | 10:06 |
| xfxf1 | wierd | 10:07 |
| xfxf1 | feel free to look at what's installed on mine - followed your instructions | 10:07 |
| xfxf1 | ubuntu 14.04 LTS 64-bit install, i did install gnome desktop instead of unity but i doubt that'll make any difference whatsoever to this build | 10:07 |
| mithro | xfxf1: can you reboot the atlys again? | 10:08 |
| xfxf1 | done | 10:08 |
| mithro | xfxf1: reboot again | 10:08 |
| xfxf1 | done | 10:08 |
| mithro | GAH | 10:11 |
| xfxf1 | is that an instruction? :P | 10:12 |
| mithro | okay - WTF debian | 10:13 |
| mithro | ?ASlink-Warning-Undefined Global '_dscr_attrpow' referenced by module 'eeprom' | 10:14 |
| mithro | make[1]: *** [usbjtag.hex] Error 1 | 10:14 |
| mithro | make[1]: Leaving directory `/home/videoteam/src/ixo-usb-jtag' | 10:14 |
| xfxf1 | assume it's behaving differently on yours | 10:14 |
| xfxf1 | is doing something as crude as an strace of the build on yours+mine then diffing them useful, even if overly verbose? | 10:15 |
| mithro | xfxf1: I'm pretty sure I have found the cause | 10:15 |
| mithro | xfxf1: what I don't understand is why the rules didn't error | 10:16 |
| xfxf1 | need me to do anything? | 10:19 |
| mithro | not yet | 10:19 |
| mithro | xfxf1: you are getting the same md5sum as me now | 10:21 |
| mithro | xfxf1: new package pushed to the ppa - will take 5 minutes to build | 10:23 |
| mithro | xfxf1: how goods your python? :P | 10:23 |
| xfxf1 | guess we'll find out! | 10:24 |
| Bertl_zZ | ~. | 10:24 |
| mithro | xfxf1: I need a pure-python version of https://github.com/libusb/libusb/blob/master/examples/ezusb.c | 10:24 |
| tpb | Title: libusb/ezusb.c at master · libusb/libusb · GitHub (at github.com) | 10:24 |
| *** Bertl_zZ is now known as Bertl | 10:25 | |
| Bertl | back now ... | 10:25 |
| Bertl | mithro: url? | 10:25 |
| mithro | xfxf1: The other option is a C version of https://github.com/mithro/HDMI2USB-find-board/blob/master/hdmi2usb-mode-switch.py | 10:25 |
| tpb | Title: HDMI2USB-find-board/hdmi2usb-mode-switch.py at master · mithro/HDMI2USB-find-board · GitHub (at github.com) | 10:25 |
| xfxf1 | my python won't be the issue there, my c will be though | 10:25 |
| mithro | https://docs.google.com/spreadsheets/d/15rivtOcSG4kqKf0AUK_OwBfs67XQtW8byr8kc5FUTgE/edit#gid=1654125645 | 10:25 |
| tpb | Title: Tims Open FPGA Expansion (TOFE) board connector interface - Google Sheets (at docs.google.com) | 10:25 |
| xfxf1 | ah, you need both in the same lang? why? | 10:26 |
| mithro | xfxf1: well - once it is in C, we'd then have to port it to windows/mac | 10:26 |
| mithro | Bertl: so - one thing I notice that is missing from your "spec" is that LVDS_3A/LVDS_3B are used as clock pins | 10:27 |
| mithro | Bertl: so you really want them to map into the FPGA GCLK, right? | 10:28 |
| *** se6astian is now known as se6astian|away | 10:29 | |
| Bertl | not sure how you got to that conclusion, please elaborate? | 10:29 |
| mithro | Bertl: Looking at the schematic for the HDMI board | 10:30 |
| mithro | Bertl: this one -> http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_plugin_module_1xHDMI_v0.5.sch.pdf | 10:30 |
| Bertl | okay? | 10:30 |
| Bertl | the HDMI clock is produced identical to the data channels | 10:30 |
| mithro | Bertl: the LVDS_3_P and LVDS_3_N are mapped to the Clock+/Clock- signals of the HDMI port | 10:30 |
| Bertl | i.e. with the serializer | 10:31 |
| mithro | Bertl: oh - I forget, you are output only :) | 10:32 |
| Bertl | yup | 10:32 |
| Bertl | there is one clock input capable pair routed to North PCIE (LVDS_5) | 10:33 |
| Bertl | but that is more coincidence than planning | 10:34 |
| mithro | Bertl: well that is fortunate :P - My LVDS_5 pairs are also on clock pins | 10:34 |
| mithro | Bertl: so you label the connectors as "North" and "South"? | 10:35 |
| Bertl | yes | 10:35 |
| Bertl | but doesn't the cyan block denote a clock input? | 10:35 |
| Bertl | because it looks like you mapped them to LVDS_3 now | 10:35 |
| mithro | Bertl: GCLK in the TOFE spec just means it is mapped to a GCLK capable pin | 10:37 |
| Bertl | ah, got it | 10:37 |
| mithro | think of them as "clock capable" rather then "clock required" | 10:37 |
| Bertl | so which one of the yellow ones are clock capable? | 10:38 |
| mithro | Hrm? | 10:39 |
| Bertl | sorry, I'm a little slow, but I guess I understand now what the table means | 10:39 |
| mithro | Cyan == IO pins connected to the GCLK pins | 10:40 |
| Bertl | with 38.5°C thinking gets a little slow ... | 10:40 |
| Bertl | the middle column is the mapping for the adapter | 10:40 |
| Bertl | the left column is the TOFE | 10:40 |
| Bertl | the right column are the plugin header :) | 10:40 |
| mithro | Bertl: 38.5C in Europe? | 10:40 |
| Bertl | body temperature unfortunately .... | 10:41 |
| Bertl | I got a cold and fever now as well ... | 10:41 |
| mithro | Bertl: you should go back to bed! | 10:42 |
| Bertl | yeah, well, I got up because I couldn't sleep ... will probably return soon | 10:42 |
| mithro | Bertl: I had another crazy idea for the Opsis V2 :P | 10:43 |
| Bertl | let's hear ... | 10:44 |
| mithro | Bertl: See the new picture in the doc | 10:44 |
| Bertl | picture where? | 10:45 |
| Bertl | the PCIe slots? | 10:47 |
| Bertl | not sure how that is supposed to work though (the right angle) | 10:47 |
| mithro | Bertl: yeah - if we map 4 sets of GTP to TOFE-4x connectors | 10:47 |
| Bertl | but there are right angle PCIe connectors | 10:48 |
| mithro | Bertl: Yes, you can get 4x right angle / left angle adapters cheaply | 10:48 |
| Bertl | well, cheaply is relative | 10:49 |
| Bertl | but they usually do not have much space to the PCB | 10:49 |
| mithro | Bertl: They also cost ~$5 USD to make in individual quantities | 10:50 |
| Bertl | http://vserver.13thfloor.at/Stuff/AXIOM/BETA/axiom_beta_plugin_module_3xPMOD_v1.1.brd.pdf | 10:53 |
| Bertl | thought that might help to get North and South orientation right :) | 10:54 |
| mithro | Bertl: thanks! | 10:56 |
| Bertl | np | 10:56 |
| mithro | Bertl: the kind of cool thing about this setup is the TOFE-4x connectors would actually PCI-Express Gen3 compatible | 10:57 |
| *** travis-ci has joined #timvideos | 11:04 | |
| travis-ci | [timvideos/HDMI2USB-misoc-firmware/master#232] (658511b): The build passed. (https://travis-ci.org/timvideos/HDMI2USB-misoc-firmware/builds/84760710) | 11:04 |
| *** travis-ci has left #timvideos | 11:04 | |
| mithro | xfxf1: you still around? | 11:08 |
| *** rohitksingh has joined #timvideos | 11:08 | |
| xfxf1 | yes | 11:08 |
| mithro | xfxf1: okay - new package -> https://launchpad.net/~timvideos/+archive/ubuntu/fpga-support/+build/8108081 | 11:10 |
| tpb | Title: i386 build of ixo-usb-jtag 0.0.0-88-gd9e335b-1 : Support packages for FPGA development : “TimVideos.Us - Live event streaming” team (at launchpad.net) | 11:10 |
| mithro | xfxf1: is the Atlys now appears correctly? | 11:19 |
| xfxf1 | i'm not clear on what you want me to do | 11:19 |
| xfxf1 | install that package, ensure USB sees the device, then? | 11:20 |
| mithro | xfxf1: get to that stage | 11:20 |
| mithro | xfxf1: power cycling the Atlys should get the new stuff | 11:20 |
| mithro | xfxf1: any luck? | 11:22 |
| xfxf1 | Oct 11 22:22:16 capcore kernel: [ 6172.280616] usb 2-1.2: Product: USB-JTAG-IF | 11:24 |
| xfxf1 | Oct 11 22:22:16 capcore kernel: [ 6172.280621] usb 2-1.2: Manufacturer: ixo.de | 11:24 |
| xfxf1 | Oct 11 22:22:16 capcore kernel: [ 6172.280626] usb 2-1.2: SerialNumber: hw_nexy | 11:24 |
| xfxf1 | assume that's bettre? | 11:24 |
| xfxf1 | s/bettre/better | 11:24 |
| mithro | xfxf1: yeah | 11:26 |
| mithro | xfxf1: happening every time you unplug / replug and power cycle? | 11:26 |
| xfxf1 | yep seems to be | 11:26 |
| mithro | Great! | 11:26 |
| mithro | xfxf1: now you won't be able to use the misoc firmware with that package installed currently | 11:27 |
| xfxf1 | ok, what is the purpose of this package? | 11:27 |
| mithro | xfxf1: so remember to uninstalled it and power cycle the Atlys board before you try and use the "make xxx" stuff | 11:27 |
| xfxf1 | sure | 11:27 |
| mithro | xfxf1: The board has a reconfigurable USB chip on it, the "fx2" | 11:28 |
| xfxf1 | yep, i recall the fx2 + the fpgalink method of flashing it currently | 11:28 |
| mithro | xfxf1: This package automatically configures the Atlys with a firmware suitable for jtag programming | 11:29 |
| mithro | xfxf1: fpgalink does a similar thing | 11:29 |
| mithro | xfxf1: now we have the fx2 in this mode, we can use compatible jtag programming tools to load stuff onto the FPGA | 11:29 |
| mithro | xfxf1: openocd is one of these tools | 11:29 |
| mithro | xfxf1: so, if you go and fetch the latest version of openocd from my repo | 11:30 |
| mithro | xfxf1: and checkout the "everything" branch | 11:30 |
| mithro | xfxf1: then configure / make / make install | 11:30 |
| xfxf1 | ok, already compiled master, will git pull / checkout everything andreinstall | 11:30 |
| mithro | xfxf1: we should in theory be able to talk to the FPGA on the Atlys board using usb-jtag | 11:31 |
| xfxf1 | done | 11:32 |
| mithro | bin/openocd -f board/digilent_atlys.cfg -c "init; xc6s_print_dna xc6s.tap; exit" | 11:33 |
| xfxf1 | jtagspi_program | 11:34 |
| xfxf1 | Error: no lowlevel driver found for ftdi or lowlevel driver opening error | 11:34 |
| xfxf1 | where is the cfg in -f looking relative to? | 11:34 |
| mithro | xfxf1: oh, you need to install libftdi-dev and recompile openocd | 11:34 |
| xfxf1 | ack | 11:35 |
| mithro | xfxf1: then you might want | 11:35 |
| mithro | --enable-usb-blaster-2 --enable-usb_blaster_libftdi | 11:35 |
| mithro | as ./configure will complain if you didn't install the deps correctly | 11:36 |
| xfxf1 | to configure or openocd? | 11:36 |
| xfxf1 | Altera USB-Blaster II Compatible yes (auto) | 11:37 |
| xfxf1 | seems to not explicitly need it? | 11:37 |
| mithro | xfxf1: Not sure | 11:37 |
| mithro | we can investigate later when I'm not falling asleep | 11:37 |
| xfxf1 | jtagspi_program | 11:39 |
| xfxf1 | Info : usb blaster interface using libftdi | 11:39 |
| xfxf1 | Info : This adapter doesn't support configurable speed | 11:39 |
| xfxf1 | Info : JTAG tap: xc6s.tap tap/device found: 0x34008093 (mfg: 0x049, part: 0x4008, ver: 0x3) | 11:39 |
| xfxf1 | DNA = 100110101111010000101001011111101001000110011101101111011 (0x135e852fd233b7b) | 11:39 |
| mithro | xfxf1: \o/ | 11:39 |
| xfxf1 | if you're tired btw we can do this tomorrow | 11:40 |
| mithro | xfxf1: I want you to verify it works, so you can +1 the codereview, so it can get merged | 11:40 |
| mithro | then I'll go home | 11:40 |
| xfxf1 | oh wow, still at the office? | 11:41 |
| mithro | xfxf1: I have a better setup in the office then at home | 11:41 |
| xfxf1 | air enough, and not bad offices to be at anyway i guess! | 11:41 |
| mithro | xfxf1: okay | 11:41 |
| mithro | xfxf1: btw - that is the "Device DNA" or the serial number of the FPGA in your board | 11:42 |
| mithro | xfxf1: so do you have a HDMI2USB .bit file somewhere? | 11:43 |
| xfxf1 | probably, these are generated from compiling up the firmware? | 11:43 |
| mithro | xfxf1: yeah - otherwise grab it from the prebuilt repo | 11:44 |
| xfxf1 | hmm where does it stick the built firmware? | 11:44 |
| mithro | xfxf1: third_party/misoc/build/ I believe | 11:45 |
| xfxf1 | ah. yes, have it | 11:45 |
| xfxf1 | jtagspi_program | 11:45 |
| xfxf1 | Info : usb blaster interface using libftdi | 11:45 |
| xfxf1 | Info : This adapter doesn't support configurable speed | 11:45 |
| xfxf1 | Info : JTAG tap: xc6s.tap tap/device found: 0x34008093 (mfg: 0x049, part: 0x4008, ver: 0x3) | 11:45 |
| xfxf1 | DNA = 100110101111010000101001011111101001000110011101101111011 (0x135e852fd233b7b) | 11:45 |
| xfxf1 | re | 11:46 |
| xfxf1 | er | 11:46 |
| xfxf1 | atlys_hdmi2usb-hdmi2usbsoc-atlys.bit | 11:46 |
| xfxf1 | rather | 11:46 |
| mithro | https://github.com/m-labs/migen/blob/master/mibuild/openocd.py | 11:46 |
| tpb | Title: migen/openocd.py at master · m-labs/migen · GitHub (at github.com) | 11:46 |
| mithro | xfxf1: we are going to do the "load_bitstream" bit | 11:46 |
| mithro | ../bin/openocd -f board/digilent_atlys.cfg -c "init; pld load 0 <bitfilename>; exit" | 11:48 |
| mithro | I think | 11:48 |
| mithro | That worked for me | 11:50 |
| xfxf1 | loaded file atlys_hdmi2usb-hdmi2usbsoc-atlys.bit to pld device 0 in 14s 162788us | 11:50 |
| mithro | xfxf1: mine only took 13 seconds :P | 11:50 |
| mithro | xfxf1: you should be able to do a "make connect-lm32" in the misoc-firmware | 11:51 |
| mithro | xfxf1: then hit enter and you should have the HDMI2USB> promt | 11:51 |
| xfxf1 | yep | 11:51 |
| xfxf1 | ido | 11:51 |
| xfxf1 | i assume this hasn't loaded the fx2 firmware, so no usb capture? | 11:52 |
| mithro | xfxf1: correct! | 11:52 |
| xfxf1 | can we use openocd to do that | 11:52 |
| xfxf1 | ? | 11:52 |
| mithro | xfxf1: and as the USB ID won't be what it expects you'll have to load the firmware manually | 11:52 |
| mithro | openocd is for talking to the FPGA not the FX2 | 11:53 |
| xfxf1 | ah, right | 11:53 |
| mithro | it talks to the FPGA *via* the FX2 | 11:53 |
| mithro | xfxf1: take a look at the | 11:53 |
| mithro | firmware/fx2/download.sh script | 11:53 |
| xfxf1 | i'm keen on shoving something into a udev rule so the device is all setup when it's plugged in / turned on | 11:53 |
| xfxf1 | right, so lsusb to find whatever usb id's are being used, and run fxload by hand? | 11:54 |
| mithro | xfxf1: fxload takes a /dev/bus/usb/ path | 11:54 |
| mithro | so you need to figure that out from lsusb | 11:54 |
| mithro | /sbin/fxload -D /dev/bus/usb/003/059 -t fx2lp -I firmware/fx2/hdmi2usb.hex | 11:54 |
| mithro | because I had | 11:55 |
| mithro | Bus 003 Device 059: ID 16c0:06ad Van Ooijen Technische Informatica | 11:55 |
| xfxf1 | hmm that hex file isn't in that dir for me | 11:56 |
| mithro | xfxf1: did you make it first? | 11:56 |
| mithro | xfxf1: "make firmware-fx2" | 11:56 |
| xfxf1 | i thought i already had | 11:56 |
| xfxf1 | t i'll try again | 11:56 |
| xfxf1 | ok done | 11:57 |
| xfxf1 | now have an OpenMoko device in lsusb | 11:57 |
| xfxf1 | [ 8133.310677] uvcvideo: Found UVC 1.00 device HDMI2USB (1d50:60b7) | 11:57 |
| xfxf1 | [ 8133.311019] uvcvideo: UVC non compliance - GET_DEF(PROBE) not supported. Enabling workaround. | 11:57 |
| mithro | xfxf1: and you should have a /dev/videoX | 11:58 |
| xfxf1 | yep i do | 11:58 |
| xfxf1 | trying to capture from it now | 11:58 |
| mithro | xfxf1: if you get something like "timeouts" it means you haven't turned the encoder on | 11:58 |
| xfxf1 | ok cool, appear to be getting capture | 12:01 |
| mithro | xfxf1: so a couple of things to note | 12:03 |
| mithro | the "connect-lm32" is just using flterm to connect to the hardware UART | 12:04 |
| xfxf1 | ack, knew that | 12:04 |
| mithro | which uses the vizzino module which is crap | 12:06 |
| mithro | so make sure you have that module installed if you are having issues using connect-lm32 | 12:12 |
| xfxf1 | sure | 12:13 |
| mithro | There isn't really any way to tell if you have loaded the FPGA correctly, except looking for the HDMI2USB> prompt on the serial port | 12:13 |
| mithro | xfxf1: Go to http://openocd.zylin.com/#/q/owner:mithro%2540mithis.com+status:open | 12:13 |
| tpb | Title: Gerrit Code Review (at openocd.zylin.com) | 12:13 |
| mithro | xfxf1: create an account using your github | 12:13 |
| mithro | xfxf1: take a look at the patches | 12:13 |
| mithro | xfxf1: add +1 saying you tested they worked with your board | 12:13 |
| xfxf1 | sure | 12:14 |
| xfxf | 2961? | 12:14 |
| xfxf | ha confusing, using multiple irc clients now | 12:15 |
| mithro | xfxf1: you obviously can't say that for the Numato Opsis boards | 12:15 |
| mithro | xfxf1: but you can confirm it for the other ones | 12:15 |
| mithro | xfxf1: btw - you should know now how to flash the board manually | 12:16 |
| xfxf | i mean which one am i approving? there's 4 listed | 12:16 |
| xfxf | and nod, good exercise | 12:16 |
| mithro | xfxf1: they all do different things | 12:16 |
| xfxf | okay, all four there affect what you got me to test? | 12:16 |
| mithro | xfxf1: one fixes the ixo-usb-jtag | 12:16 |
| mithro | xfxf1: one adds the Digilent Atlys support | 12:17 |
| mithro | xfxf1: one adds support for printing that serial number | 12:17 |
| xfxf | also, "Strategy: Cherry Pick" | 12:17 |
| mithro | xfxf1: one adds the Opsis support | 12:17 |
| xfxf | that's strange to me, I'm used to a PR usually representing gitflow and a request to merge back to develop | 12:17 |
| mithro | xfxf1: ignore that | 12:17 |
| xfxf | how does this work, you grab a bunch of commits and ask for those to be approved? | 12:17 |
| xfxf | ah | 12:17 |
| mithro | They are all tcl, which is almost as bad as perl | 12:18 |
| xfxf | the http to https typo i spotted before, has that been fixed / is that part of this patch? | 12:19 |
| xfxf | this UI is confusing, it's taking me a bit to grok the info on it | 12:19 |
| mithro | xfxf1: that was for ixo-usb-jtag not openocd | 12:19 |
| xfxf | ah | 12:19 |
| mithro | xfxf1: send a pull request for that one | 12:19 |
| xfxf | assume on all four up the top, click reply, hit +1, and put in "Works for me on Diligent Atlys board." ? | 12:19 |
| mithro | xfxf1: well, not the opsis one | 12:20 |
| xfxf | sure, done for three | 12:20 |
| mithro | xfxf1: Then take a look at https://github.com/mithro/HDMI2USB-find-board/blob/master/hdmi2usb-mode-switch.py to take a gander at how you figure out where the atlys board is and what state it is in | 12:21 |
| tpb | Title: HDMI2USB-find-board/hdmi2usb-mode-switch.py at master · mithro/HDMI2USB-find-board · GitHub (at github.com) | 12:21 |
| mithro | xfxf1: home time for me | 12:23 |
| xfxf1 | np, ta | 12:23 |
| xfxf1 | nite | 12:23 |
| mithro | xfxf1: maybe be back later, maybe not | 12:23 |
| mithro | xfxf1: still have to eat some food | 12:23 |
| mithro | xfxf1: pretty much any issue on https://github.com/timvideos/HDMI2USB-misoc-firmware/issues is open | 12:24 |
| tpb | Title: Issues · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com) | 12:24 |
| mithro | xfxf1: Figuring out how to "make load-lm32" do something useful would be good | 12:24 |
| CarlFK | hey gang | 12:32 |
| CarlFK | xfxf1: (06:53:00 AM) mithro: /sbin/fxload -D /dev/bus/usb/003/059 -t fx2lp -I firmware/fx2/hdmi2usb.hex | 12:49 |
| CarlFK | where did you get fxload from? | 12:49 |
| xfxf1 | there's a package for it that the setup scripts install | 13:09 |
| xfxf1 | ubuntu package 'fxload' | 13:09 |
| CarlFK | k - I thought maybe it was a patched version that was built from who knows what | 13:20 |
| *** Bertl is now known as Bertl_zZ | 13:25 | |
| *** se6astian|away is now known as se6astian | 14:12 | |
| CarlFK | xfxf1: fxload loads the .hex file (I guess) do you know what loads the .bin? | 14:18 |
| CarlFK | other than make | 14:19 |
| *** rohitksingh has quit IRC | 15:49 | |
| *** rohitksingh has joined #timvideos | 16:05 | |
| *** Bertl_zZ is now known as Bertl | 17:51 | |
| *** Bertl is now known as Bertl_oO | 19:02 | |
| *** rohitksingh has quit IRC | 19:51 | |
| *** se6astian is now known as se6astian|away | 21:59 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!