Wednesday, 2015-10-07

*** tpb has joined #timvideos00:00
*** _florent_ has joined #timvideos00:18
*** springermac has quit IRC01:17
*** travis-ci has joined #timvideos01:17
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#206] (f13c17a): The build has errored. (
*** travis-ci has left #timvideos01:17
*** springermac has joined #timvideos01:45
*** CarlFK has joined #timvideos01:48
*** ChanServ sets mode: +v CarlFK01:48
*** Bertl_zZ is now known as Bertl03:15
*** travis-ci has joined #timvideos06:19
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#206] (f13c17a): The build has errored. (
*** travis-ci has left #timvideos06:19
*** travis-ci has joined #timvideos06:39
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#18] (cd873c2): The build is still failing. (
*** travis-ci has left #timvideos06:39
seaLnemithro: is there actually any thing to copy if the target is base?06:39
seaLneah there seems to be06:49
*** travis-ci has joined #timvideos07:00
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#19] (baf0140): The build was canceled. (
*** travis-ci has left #timvideos07:00
mithroseaLne: just the .bit file07:01
mithroReally we want the xsvf file rather than the .bit but the only way to generate that currently is via setting the programmer to fpgalink and attempting to program07:02
seaLneso no bin?07:06
seaLnei'm not really sure what all of the files do yet :)07:07
seaLnemithro: does the .bit include the lm32 stuff?07:09
seaLnemithro: i'm wondering if putting the .bit in a firmware directory is confusing as we talk about firmware and gateware being different when building normally07:11
seaLneto me gateware might be a clearer name for the directory?07:11
*** se6astian|away is now known as se6astian07:18
mithroseaLne: well the .hex file is for the Cypress and that is firmware07:18
mithroseaLne: I also logged
tpbTitle: `make gateware` should generate the xsvf file · Issue #91 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
*** travis-ci has joined #timvideos07:19
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#20] (bb35cbb): The build was fixed. (
*** travis-ci has left #timvideos07:19
mithroseaLne: see the diagram at
tpbTitle: HDMI2USB-firmware-prebuilt/atlys at master · mithro/HDMI2USB-firmware-prebuilt · GitHub (at
seaLnemithro: so do we want the .hex file as well?07:21
mithroseaLne: yes but the fx2 .hex file rather than .hex for the lm3207:22
seaLnemithro: where is it built? and is it different between boards?07:24
mithromake firmware-fx2 produces07:33
seaLneis it supposed to be putting stuff in /firmware/fx2 rather than third party?07:34
seaLnemithro: am i right that travis isn't currently building fx2? i might as well add it?07:44
*** travis-ci has joined #timvideos08:11
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#22] (3c3370b): The build was broken. (
*** travis-ci has left #timvideos08:11
mithroseaLne: The firmware target builds the fx2 and lm32 firmware08:26
mithroseaLne: the fx2 firmware ends up in firmware/fx208:27
seaLneso it should be building it just now?08:27
*** CarlFK has quit IRC08:30
mithroseaLne: it only builds on HDMI2USB target though08:30
mithroseaLne: I think setting PYTHONHASHSEED=0 has made the 2 hour timeout thing reliably reproducible...08:32
mithroseaLne: you can see it building here ->
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at
seaLnehmm did i push something wrong by accident08:56
mithroseaLne: ?08:56
seaLneah no got confused by the flash on talking about my recently pushed copy-build branch and thought i'd pushed to upstream somehow08:57
tpbTitle: timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroseaLne: nope! :P08:57
seaLnepanic over08:57
*** travis-ci has joined #timvideos09:12
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#25] (15ace3e): The build passed. (
*** travis-ci has left #timvideos09:12
mithroseaLne: Hows the pushing going?09:53
mithroseaLne: It would probably be a good idea to save the log output of the build too09:54
seaLnei'd wondered about that for the log09:55
tpbTitle: Pastebin | IRCCloud (at
mithroseaLne: Looks like you have it mostly working?09:55
seaLnei think i broke my repo when i tried to fixup :(09:55
mithroseaLne: There is a line just before the non-shallow bit which says "Raw output saved in '/home/travis/build/sealne/HDMI2USB-misoc-firmware/build/output.2015_10_07.log'09:56
mithroseaLne: I wonder if just capturing that log would be good enough?09:56
mithroseaLne: How did you break it?09:57
seaLnei don't seem to have any build-copy stuff anymore09:57
seaLnei set the intitial commit to r, f the following ones and moved the clead picks to before the first one09:58
mithroso, unless you did a "git gc" the commits will still be there10:01
seaLnegitk doesn't show them10:01
mithroseaLne: git has a thing called the "reflog" for precisely things like this10:01
mithroseaLne: It'll let you get back to the old state10:01
seaLnei haven't pushed so origin is still fine10:01
mithroseaLne: btw you can use "gitk --all" to see everything at once10:03
mithroit can be a bit confusing - but it is my preferred mode10:03
seaLnegit reset --hard 02f845098502f0904374f1ab4f53f5a78033ff88 got me back. is that a reasonable thing to do?10:03
mithroseaLne: yes10:04
seaLnenot moving things about just rewording the first commit and fixup all up until the merge with master still leaves me with
tpbTitle: KDE Paste (at
mithroseaLne: you are using "git rebase -i HEAD~30" again - this is most certainly not what you want10:07
mithroyou should always use a known tag for the rebase10:08
seaLneso the tag that you want to start from?10:08
mithroseaLne: It is "git rebase -i <location to put the rebased changes onto>"10:09
seaLnenot sure waht you mean by location there?10:09
seaLneas in origin/copy-build?10:10
mithroseaLne: the commit (which you can use a symbolic name like a branch name instead)10:10
mithroseaLne: "git rebase -i HEAD~30" says "take all commits which are not in HEAD~30 and put them on top of HEAD~30"10:13
mithroseaLne: "git rebase -i origin/copy-build" says "take all the commits which are not in origin/copy-build and put them on top of origin/copy-build"10:13
seaLneexcept "git rebase -i origin/build-copy" just has "noop" in the file10:18
mithroseaLne: that means there is nothing in your branch which isn't already in origin/build-copy10:19
seaLneRebase 02f8450..02f8450 onto 02f8450 would be the problem?10:19
mithroseaLne: you probably want upstream/master10:19
seaLneyeah i want to fixup my millions of commits10:19
mithroseaLne: so I recommend adding an "upstream" remote so you can do things like "git rebase -i upstream/master"10:19
mithroseaLne: you can do that with "git remote add upstream && git fetch upstream"10:20
tpbTitle: timvideos/HDMI2USB-misoc-firmware · GitHub (at
*** springermac has quit IRC10:20
seaLneaha yeah that looks like it10:20
*** springermac has joined #timvideos10:21
seaLnemithro: for the log could i just create a readme with a link to travis?10:27
mithroseaLne: Travis output doesn't get the full log output10:28
seaLnehow do we get it then?10:31
mithro8:54 PM <@mithro> seaLne: There is a line just before the non-shallow bit which says "Raw output saved in '/home/travis/build/sealne/HDMI2USB-misoc-firmware/build/output.2015_10_07.log'10:32
mithro8:54 PM <@mithro> seaLne: I wonder if just capturing that log would be good enough?10:32
seaLneah sorry missed that10:33
mithroWhat is COPY_REPO_OWNER ?10:44
seaLnesealne or timvideos10:47
seaLnecopying the log seems fine
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at
seaLnemithro: how can i un-change .travis.yml? i guess something broke when i was removing using the token in there10:53
mithroseaLne: "git checkout upstream/master .travis.yml"10:54
mithroseaLne: despite all my comments this looks really good!10:59
mithroseaLne: Thanks for this...11:00
mithroseaLne: btw we have two options, we can merge with the minor fixes required and then you can send another pull request with the stuff moved into its own reusable script, or we can wait until you do it11:02
seaLnemithro: sorry i'm a bit unsure what you are meaning by "FYI - This script is actually also used outside travis for building the Xilinx package...."?11:06
mithroseaLne: FYI == For your information11:07
seaLnethe lack of GITHUB_TOKEN should stop anything else trying to upload the files11:07
seaLnei ment the rest of that comment11:07
mithroseaLne: Yeah - I think your changes are fine with that, just wanted to make sure you knew we used it outside of travis despite it being in a .travis directory11:08
seaLnethat should fix not displaying error11:12
tpbTitle: Only display token error if being run in travis · sealne/[email protected] · GitHub (at
mithroseaLne: I wonder if we are doing a "stable release" if we might want to use this script?11:14
seaLnewill we not just symlink stable to the build?11:14
seaLnemithro: hmm where did this come from
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at
mithroseaLne: good question11:17
mithroscripts/ line 4: realpath: command not found11:17
mithroseaLne: if you look above travis had a bit of a screw up11:18
mithroFailed to fetch  Unable to connect to [IP: 80]11:18
mithroE: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?\11:18
seaLneah yeah. thought it was weird that would break there11:19
*** travis-ci has joined #timvideos11:21
travis-ci[sealne/HDMI2USB-misoc-firmware/master#26] (f13c17a): The build has errored. (
*** travis-ci has left #timvideos11:21
seaLnegetting too used to going straight to the bottom of the log to look at my stuff :)11:23
mithroseaLne: yeah11:23
seaLnemithro: what do you want to do about the token for travis in timvideos? should it do this as "timvideos"?11:24
mithroseaLne: yeah11:25
mithroseaLne: take a look at the travis build for some of the websites?11:25
seaLnenone seem to say what "timvideos" email address is? i can change COPY_REPO_OWNER and the in the branch and then you can add the token in travis? needs updated first though11:29
tpbTitle: timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
mithroseaLne: actually can we reuse the information from the person who pushed the commit?11:29
seaLnethats an idea. assuming github doesn't mind that being "forged" not sure how to find out though11:30
mithroseaLne: yeah11:30
mithroseaLne: The TimVideos robot would still be the committer - the other person would be the author?11:32
seaLnerobot? it would be whosever token it was that would be the committer11:33
seaLneie we set some how to whoever made the last commit and hope that is displayed when viewing the file?11:34
tpbTitle: Commits · timvideos/planet · GitHub (at
tpbTitle: timvideos-robot ( Robot) · GitHub (at
seaLneso you mean use a token from timvideos-robot?11:35
mithroseaLne: yes11:37
seaLnebtw who is doing the Low speed I/O” TOFE board design?11:38
seaLneshame presumably the IO board can't be used at the same time as VGA11:40
mithroseaLne: correct11:40
seaLneright so what do you think the plan is now to finish the build-copy?11:41
mithroseaLne: How long do you think it'll take to do the changes I asked in the pull request?11:42
seaLnewell i can't add the token or tidy up prebuilt repo11:43
mithroseaLne: I've done the tidy up and can add the token11:44
seaLne doesn't look fixed?11:46
tpbTitle: timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
mithro <-- that looks right?11:46
tpbTitle: mithro/HDMI2USB-firmware-prebuilt · GitHub (at
seaLnestable and testing symlinks shouldn't exist till we have something to point them at11:47
mithroseaLne: I want to link to the firmware revision that carl uses as testing for now11:48
seaLnejust looking at ways to get committer name/email11:50
xfxfmithro: i reported a month or two back (and created github tickets) for two of my monitors that would not accept any output from the HDMI2USB11:53
xfxfit's still happening on the latest firmware11:53
xfxfneither monitor detects output11:53
xfxfand yes, tried setting to DVI input, doesn't work11:53
mithroxfxf: hrm - we have fixed every other monitor11:53
mithroxfxf: first, can you check what resolutions the monitor actually supports?11:54
xfxfboth work with my xbox 360 at 720p / 1080p as i've used them for gaming before11:54
xfxfbut will dig up specs11:54
xfxfnot finding any 'all resolutions supported' info11:57
xfxfthe monitors are:11:57
xfxfdell u2713hm, BenQ XL2411T11:57
mithroxfxf: read the edid11:57
xfxfi'm 99% positive i've used 720p50 with both with a gaming console before11:57
mithroxfxf: I'm not after hearsay - tell me what the EDID of the monitor says11:58
seaLnemithro: i think that should now use the last committer for uploading12:01
xfxfsure, once i figure out how to do that on a mac12:01
mithroxfxf: stop using crappy operating systems ;)12:01
seaLnemithro: err no it won't need to get git stuff from right repo for setting this duh12:03
seaLnewill this work assuming its define earlier git config "$ORIG_COMMITTER_EMAIL"12:05
seaLnewill bash and git play happily there?12:06
mithroseaLne: not sure12:06
xfxfmithro: those work?12:07
seaLnelooks like it will12:08
mithroxfxf: It does look like it supports 720p @ 60Hz and 720p @ 50Hz12:10
xfxfyep, no output from the hdmi2usb though12:10
mithroxfxf: next step is to plug in a laptop to the monitor that you can configure to a give resolution12:10
xfxfneither detect output12:10
xfxfyep, sure, did that to get you those edid dumps12:10
mithroxfxf: and then force the resolution to 720p60 and see what happens12:11
xfxfdone, getting output on the dell12:11
mithroand I don't mean using a Mac which might be lying to you12:11
xfxfi'm using SwitchResX to force the resolution so personally actually trust it's doing what it says12:13
xfxfdon't have any linux boxes here with HDMI out12:13
*** springermac has quit IRC12:13
xfxfwill in a few days but trying to get this tested+working for another meetup tomorrow night12:13
*** springermac has joined #timvideos12:14
xfxfand tried both on a windows machine using nvidia control panel/tool to force to [email protected], both work12:16
mithroxfxf: okay12:16
xfxfwait, never mind, now it's working12:18
mithroxfxf: have you got the jumpers right?12:18
xfxfi unplugged my second input (camera)12:19
xfxfi'll investigate to see what was happening12:19
xfxfand yes was using video_mode and video_matrix to set/connect things12:19
xfxfi turn camera on12:19
xfxfHDMI2USB>dvisampler1: FIFO overflow12:19
xfxfit's outputting a wierd resolution though12:20
mithroxfxf: if you aren't getting it 100s of times then it is fine12:20
mithroxfxf: did you forget to turn on the outputs?12:20
xfxfno, they were on12:20
mithroxfxf: I mean in the firmware12:21
xfxfi know that's what you mean - yes, turned output0 on12:21
xfxfhmm, now i'm getting hanging at [FLTERM] Starting12:30
xfxfpower cycled board + ran 'make load' three times now, ends up there12:30
mithroxfxf: define hang?12:30
mithroxfxf: did you try hitting enter?12:30
xfxfof course12:30
xfxfby hang nothing happens12:30
xfxfno response to any input12:31
mithroxfxf: I can't look over your shoulder - can you paste an example of what your screen looks like?12:31
xfxfthat's turning off board, turning on board, 'make load'12:31
xfxfnormally i'd push enter there and get a console, nothing12:32
mithroxfxf: if you didn't flash the board correctly then flterm won't show anything12:32
mithroxfxf: I think the error is further up12:32
tpbTitle: Pastebin | IRCCloud (at
mithroyou didn't set "PROG=fpgalink"12:34
mithroIt's interesting that it didn't fail at that step12:34
xfxfok, these instructions need updating i think12:34
xfxfthey refer to some out of date stuff so it's not clear on what is or isn't right anymore12:34
mithroData width read from the bitstream file = 1.12:34
mithroINFO:iMPACT:501 - '1': Added Device xc6slx45 successfully.12:34
mithro>ERROR:iMPACT:1075 - setCable must be set before this operation.12:34
mithroxfxf: can you log a bug about that error not causing `make load` to stop?12:35
mithroxfxf: can you also fix that bug? :P12:35
mithroxfxf: we should probably make PROG=fpgalink the default these days12:35
xfxfmore than happy + want to hack on this, just trying to remain focused to get a repeatable working process at the moment so i can attempt recording12:36
xfxfand yeah, agreed12:36
*** travis-ci has joined #timvideos12:40
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#30] (11d4635): The build has errored. (
*** travis-ci has left #timvideos12:40
seaLnemithro: i think the build-copy is ready?12:44
mithroseaLne: okay - I need to push my branch into prebuilt first, right?12:44
seaLnethen add the token and make sure the robot has commit to prebuilt12:45
mithroseaLne: done!12:46
mithroseaLne: adding token now12:46
xfxfmithro: input1 doesn't seem to be working properly for me - is there a jumper i might need to pull or something?12:47
mithroxfxf: its a bit complicated12:48
mithroxfxf: actually!12:48
mithroxfxf: take a look a the matrix help12:48
mithroit describes how the
tpbTitle: HDMI2USB-misoc-firmware/ at master · timvideos/HDMI2USB-misoc-firmware · GitHub (at
seaLnemithro: rebuild once you've done that maybe before accepting pull request?12:49
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at
mithroseaLne: pull requests can't build the full firmware12:50
seaLneno access to XILINX... ?12:51
mithroseaLne: yeah12:51
xfxfta, pins were wrong12:51
xfxftesting again12:51
seaLnei guess makes sense12:51
mithroseaLne: I'm going to merge as is12:53
* seaLne crosses fingers12:54
xfxfsorry, it's input0 which isn't working properly12:54
xfxfinput1 works fine with multiple inputs12:54
tpbTitle: Pastebin | IRCCloud (at
xfxfif i swap input0 + input1 connections, exactly the same (as in, input0 still not working, but the opposite device now outputs to hdmi2usb fine)12:55
xfxfmithro: any idea?12:57
mithroxfxf: you checked all the jumpers? Is the device seeing the board?12:57
xfxfyes checked all jumpers according to that matrix help12:59
xfxfnot sure what you mean by latter question12:59
xfxfoh, sorry, right12:59
xfxfno, it's not12:59
mithroxfxf: can you try setting video_mode while the device is still plugged in?13:00
seaLneany idea roughly how many people have atlys boards they are using?13:00
mithroseaLne: 2-3 people I think13:00
xfxfmithro: i did, didn't fix it13:01
seaLnewe should swap over default build target at some point13:01
mithroxfxf: what is the device?13:01
mithroseaLne: Yes probably13:01
xfxfdevice(s).  one is my mac, one is a canon XA20 set to 720p50 out13:01
xfxfboth work perfectly on input113:01
xfxfneither work on input013:01
xfxfi think this was the original problem i had when i said 'it wasn't working, oh wait, now it is'13:02
xfxfi probably swapped inputs during diagnosis13:02
mithroxfxf: okay, we haven't seen that one before13:02
xfxfwhat should i do?13:07
mithroxfxf: I don't know - try and see if you can get anything detected on input0?13:08
mithroxfxf: you only have one Atlys board right?13:08
xfxfonly one yes13:08
xfxfso can't easily rule HW issue13:08
mithroseaLne: hrm - ?13:09
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at
seaLneits called GITHUB_TOKEN ?13:10
mithroseaLne: oh - everything else uses GH_TOKEN13:10
seaLneother timvideos travis stuff you mean already uses one?13:11
mithroxfxf: we effectively run the same "gateware" on both inputs13:12
xfxfi can't get anything on input013:12
mithroseaLne: yeah - many of the ruby tools use that name13:12
mithroxfxf: do you have two outputs connected?13:12
xfxfonly one13:13
mithroxfxf: to output0 ?13:13
mithroxfxf: and you are sure you are connecting to the right connectors?13:14
mithroxfxf: using the same cable?13:14
xfxfyes and yes but tried different cables too13:14
xfxfinput top right is one not working at all13:18
xfxfer sorry13:18
xfxflet me verify13:18
xfxfbottom left rather13:19
xfxfnot working at all13:19
xfxfsame source, same cable, plug into the other hdmi on on top right, works fine13:19
mithro* JP4 has a jumper (it connects / disconnects 5V to HDMI pin 18).13:19
mithrois not done in your photo13:20
xfxfi know13:20
xfxfit's fixed13:20
xfxfi disconnected it before to test13:20
xfxfconnected again, same problem13:20
xfxfsorry, should have altered before taking photo13:20
mithroSCL and SDA / JP2 have the jumpers correct right?13:21
xfxfi believe so, followed instructions13:21
mithroxfxf: the next step would be to use the edid_debug target to figure out if you are seeing anything.13:24
mithroxfxf: but _florent_ hasn't written up any instructions on how to use that13:24
tpbTitle: HDMI2USB-misoc-firmware/ at master · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroxfxf: I would try and get CarlFK to see if he can reproduce13:25
xfxfwhat is required to get my mac to recognise the hdmi2usb as a capture device?13:28
xfxfit doesn't seem to be13:28
mithroxfxf: that is a good question - I do not know13:28
mithroxfxf: Need to figure out how our UVC device differs from a UVC device it likes13:29
xfxfhmm interesting side effect with my XA20 too13:30
xfxfhdmi2usb doesn't see any input unless i go into the camera, change the resolution to something else, then back to 720p13:30
xfxfif i power cycle the hdmi2usb or camera, doesn't work until i do the above13:30
xfxfunsure if hdmi2usb or xa20 bug13:30
xfxfxa20 works fine on my atem gear though13:31
xfxfi gotta go to bed13:32
xfxfi might postpone this recording tomorrow, i can't even get this working locally13:33
xfxfta for help, talk tomorrow13:33
mithroxfxf: you mean run "video_mode 9"?13:33
xfxfmithro: no, change the modes on the camera itself13:33
mithroxfxf: interesting13:33
xfxfon linux13:34
xfxfit seems random that /dev/video0 actually shows up13:34
xfxfi'm yet to actually get capture from it13:34
xfxfsomething is flakey as hell13:35
xfxfthat said it might be USB passthrough on my VM13:35
xfxfhence ordering a dedicated laptop13:35
mithroxfxf: what do you mean "seems random that /dev/video0 shows up"?13:35
tpbTitle: Pastebin | IRCCloud (at
xfxfit's either there or it isn't13:35
xfxfit goes away randomly13:35
xfxfi don't know how to articulate it any more than that13:35
mithroxfxf: that doesn't mean much unless you tell me what happens inbetween13:35
xfxfright now it's not there but i am connected to the HDMI2USB console perfectly13:35
xfxfnothing happens in between that i'm aware of13:36
xfxfit's there13:36
xfxfnext it ins't13:36
xfxfdmesg doesn't indicate any USB connects/disconnects13:36
mithroxfxf: something is obviously happening :)13:36
xfxfsure, but i can't see what13:36
tpbTitle: Pastebin | IRCCloud (at
xfxfis earlier from dmesg, but fyi13:37
xfxfhmm also13:37
tpbTitle: Pastebin | IRCCloud (at
tpbTitle: Pastebin | IRCCloud (at
xfxfi have no idea why they're reporting differently13:37
xfxfboth are with 'make load' after a power cycle and connecting to the hdmi2usb board perfectly13:38
mithroxfxf: the second one is when the FX2 is in JTAG programmer mode13:38
xfxfsomething is flaky - i can be more specific if you tell me what i should be looking at13:38
mithrothe first one is when operating in HDMI2USB mode13:38
xfxfOK, but remember i'm not explicitly calling things myself, i'm just doing 'make load' in an attempt to get it to capture13:38
mithromake load does 3 steps13:39
mithromake load-gateware13:39
mithromake load-lm3213:39
mithromake load-fx213:39
mithromake load-gateware with PROG=fpgalink will cause the FX2 to go into the Makestuff mode13:39
xfxfyeah, sure, i get that13:39
xfxfoh, wait, sorry13:39
xfxfabove is probably hapening during the 3 steps of make load13:39
mithromake load-fx2 will cause the FX2 to into the Product: HDMI2USB mode13:40
xfxfi was trying to imply that i was doing the 'same thing' and getting different results13:40
xfxfbut, right13:40
xfxfi'm not clear at all on what i'm supposed to do to actually be able to capture from it then13:40
xfxfi figured 'make load', get to the console, setup the right modes/matrix/output, then i'll have a /dev/video0 device to capture from13:40
xfxfi don't get the latter13:40
mithroevery time either of those things happen, the FX2 will disconnect and reconnect to the USB bus, so you will see "USB disconnect" and "new high speed device"13:41
mithromake load-fx2 doesn't happen until after make load-lm32 finishes - IE you close the console13:41
tpbTitle: `make load-lm32` should actually load the firmware and then exit · Issue #90 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
xfxfok, so how do i capture from the device at the same time as being able to change inputs?13:41
xfxfbecause that is my goal13:41
mithromake load-gateware; make load-fx2; make connect-lm3213:42
xfxfis that a different order than what 'make load' does, and if so, why>13:42
mithroxfxf: yes13:42
mithroxfxf: make load does a "make load-gateware; make load-lm32; make load-fx2"13:43
mithrowhich would make sense if Issue #90 was fixed13:43
xfxfnot working13:44
xfxfi'm going to go to bed, i'll work on this again on the weekend13:44
mithroxfxf: this works for me "PROG=fpgalink make load-gateware; make load-fx2; make connect-lm32"13:46
xfxfthat's exactly what i tried13:48
xfxfi just get the FLTERM starting prompt without the console actually starting13:48
mithrodid make load-gateware succeed?13:49
mithroreplace ; with &&13:49
*** travis-ci has joined #timvideos13:50
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#214] (3b0f648): The build failed. (
*** travis-ci has left #timvideos13:50
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at
xfxfokay, getting capture now, but the USB passthrough is not coping with the bandwidth13:54
xfxfgetting epic MPEG corruption13:54
xfxfi zink i'll wait until i have a dedicated laptop for this13:55
xfxfunless i can figure out how to get the mac to see it as a valid UVC device13:55
xfxfactually, it recognises it out of the box13:59
xfxfagain, USB passthrough wierdness was occuring13:59
xfxfnot getting video data from it thoug14:00
mithrodid you turn the encoder on?14:00
mithrowhat does status show?14:00
tpbTitle: Pastebin | IRCCloud (at
mithroxfxf: I'm unsure if we advertise the right resolution in the usb firmware14:01
mithro@ 0 fps means something isn't reading the encoder data14:01
xfxfcorrect, OS X doesn't seem to actually read it14:01
xfxfgives the hdmi2usb as a valid capture device but won't capture from it14:02
xfxflivestream producer shows 'no data', vlc just defaults to the next capture card14:02
mithroxfxf: this isn't something I can debug remotely, we will have to investigate another time14:05
mithroxfxf: I do know Linux is probably the most forgiving14:05
mithroseaLne: there are a couple of small things which need to be done before the Opsis is the preferred device for doing HDMI2USB recording14:08
*** travis-ci has joined #timvideos14:19
travis-ci[sealne/HDMI2USB-misoc-firmware/build-copy#35] (0dcb924): The build has errored. (
*** travis-ci has left #timvideos14:19
mithroseaLne: It is unclear to me which command failed?14:36
*** rohitksingh has joined #timvideos14:41
seaLnemithro: its the push i think. just to confirm things are as i think they are, there is the travis variable GITHUB_TOKEN with the token from timvideos-robot? and timvideos-robot has commit privs for HDMI2USB-firmware-prebuilt?15:33
mithroseaLne: yes I added that15:33
mithroOh wait15:33
mithroseaLne: Need to check is the robot has push access15:34
mithroseaLne: any idea why there isn't an error message?15:34
seaLneprobably the redirect to /dev/null to not show the token15:34
seaLnethe other possibility is that github doesn't like setting to something other than the account that is commiting?15:40
seaLnei think that is less likely though15:40
mithroseaLne: you can use the method the XILINX_PASSPHRASE uses to make sure you don't echo it15:41
seaLnethe things i read suggested the token would be mentioned when you pushed15:41
seaLnei wanted to be very sure that didn't happen15:42
mithroseaLne: ? the secure values never end up even getting into a pull request environment15:42
seaLneyou could be right, i wanted to err on the safe side15:43
seaLnedoes the robot have commit rights?15:45
mithroseaLne: Otherwise I could just send a pull request with "echo $XXXX" in the .travis.yml file :)15:45
mithroseaLne: they did not15:46
*** se6astian is now known as se6astian|away15:46
mithroseaLne: Trying with pushing access enabled :)15:48
*** travis-ci has joined #timvideos15:56
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#214] (3b0f648): The build failed. (
*** travis-ci has left #timvideos15:56
seaLnew00t :)15:57
*** CarlFK has joined #timvideos15:57
*** ChanServ sets mode: +v CarlFK15:57
tpbTitle: HDMI2USB-firmware-prebuilt/archive/v0.0.0-396-g3b0f648/atlys/base at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
mithroseaLne: so - what happens if two builds complete in a similar amount of time?16:01
mithroseaLne: IE You can't push because another push has happened while you where committing....16:01
seaLnehmm maybe there should be a pull at the start of each target build?16:02
seaLneno that wouldn't quite cover all of that16:03
seaLnei'll have a think16:03
mithroseaLne: I think if you might want to just wrap the "git push --quiet origin master > /dev/null 2>&1" stuff in a loop which does a "git pull; git push;" a couple of times?16:03
seaLneyou still might get some of the files of one of the other builds?16:04
mithroor even just retry the whole uploading upload step?16:04
seaLneor do the checkout to a suffix directory?16:05
mithroseaLne: the builds should all end up in separate directories right?16:05
seaLneHDMI2USB-firmware-prebuilt-atlys-base for example16:05
seaLnethey are uploaded to different places16:06
mithroseaLne: yeah16:06
mithroseaLne: so the only problem is the fact that git push will fail because commits exist that are not local16:06
seaLneyou could still plausibly get something pushed in between a pull and add,commit,push16:07
mithroseaLne: yeah16:08
seaLnewhich is about as close as you could get the pull16:08
mithroseaLne: it's pretty unlikely - but not impossible16:08
mithroseaLne: I wonder if the archive also needs a "travis build ID" or something?16:09
mithroseaLne: If I click rebuild on travis, it'll overwrite the old version for that git ID right?16:09
seaLnethere should definitly be another pull before the add, we can see if we actually end up with things happening after that16:09
seaLnebut the travis build id would still be the same16:10
mithroI guess it will still appear in the git history16:10
seaLnethis is intended to be a simple way for people to get the latest builds not to deploy a phone network on so i'd be tempted to see how it goes in real life :)16:11
mithroseaLne: I'm planning on using it in the tool16:12
tpbTitle: Create a tool to manage the download and flashing of the HDMI2USB firmware · Issue #82 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroseaLne: oh! We should generate a md5sum file for the content16:13
CarlFKmithro: any chance of that being ready in about 28 hours?16:13
mithroCarlFK: not a chance in hell unless you are planning on working on it16:13
CarlFKI am planning on recording another show in about 40 hours.16:14
seaLneCarlFK: i think the build in is probably good16:14
tpbTitle: HDMI2USB-firmware-prebuilt/archive/v0.0.0-416-g02f8450/atlys/hdmi2usb at master · sealne/HDMI2USB-firmware-prebuilt · GitHub (at
CarlFKI am heading the gate - boarding starts in 15 min16:16
CarlFKthen 17 hours in the air, 2 hour layover, 2 more hours and I am back in chicago.16:16
seaLnemithro: you could click the rebuild on 214.2 ?16:16
mithroCarlFK: I can find the board on the system in the various states reasonably reliably16:16
seaLneah they are already running now16:17
mithroCarlFK: I can also detect if drivers have been bound to the USB device and unbind them16:17
mithroCarlFK: things left todo are16:17
mithroCarlFK: (a) Get the loading of stuff onto the FPGA16:18
seaLnemithro: so for fixes to a pull request that have been applied what is your prefered workflow?16:18
mithroCarlFK: (b) Get the "flashing" part working16:18
mithroseaLne: send new pull request16:18
mithroseaLne: Not really anything else you can do...16:18
CarlFKmithro: my plan once I get to chicago is try to flash the "probably good" files using the make commands I have been using16:18
seaLnecan i still use my build-copy branch or should i create a new branch?16:19
CarlFKand then whip though the few hdmi sources I have around the house16:19
mithroCarlFK: "flash" is different from "load into ram"16:19
*** Bertl is now known as Bertl_oO16:19
CarlFKwhatever it is I have been doing16:19
CarlFKnot enough time for any R&D for another 4 or 5 days16:20
mithroCarlFK: flashing is highly dangerous because if you flash bad contents the only way to recover the board is to use an external JTAG programmer16:20
CarlFKyeah, none of that for now16:20
seaLnedoes the opsis in theory support loading gateware from an EEPROM somehere? i haven't looked closely enough to see if it has one16:21
mithroseaLne: the Opsis has SPI flash which the gateware can be flashed into16:22
CarlFKbb in either 5 min or 30 hours... bye!16:22
mithroseaLne: the Atlys does too16:22
*** CarlFK has quit IRC16:22
seaLneyeah, wasn't sure about the opsis16:22
mithroseaLne: the Atlys also has an EEPROM the .hex file can be burnt into16:22
mithroseaLne: the Opsis doesn't have that EEPROM, but the FPGA is able to pretend to be the EEPROM16:23
mithroseaLne: this allows us to put multiple .hex files on the board and have the FX2 appear as different devices depending on if the power button was held down while booting16:23
mithroseaLne: but that doesn't quite work yet :P16:24
*** se6astian|away is now known as se6astian16:25
mithroseaLne: It is on my TODO list :)16:27
*** travis-ci has joined #timvideos16:33
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#214] (3b0f648): The build passed. (
*** travis-ci has left #timvideos16:33
mithroseaLne: I created a couple of github issues for the remaining things about the prebuilt stuff16:37
seaLnei'll probably not get to them till tomorrow (UTC+1)16:39
mithroseaLne: that is perfectly fine!16:40
mithroseaLne: already gotten it much further then it was previously!16:40
mithroseaLne: if you were to disappear now, those github issues would remind me of what needs to be finished16:42
*** rohitksingh has quit IRC17:22
*** rohitksingh has joined #timvideos17:38
*** travis-ci has joined #timvideos18:37
travis-ci[sealne/HDMI2USB-misoc-firmware/master#36] (3b0f648): The build has errored. (
*** travis-ci has left #timvideos18:37
*** se6astian is now known as se6astian|away19:21
*** se6astian|away is now known as se6astian19:24
*** rohitksingh has quit IRC19:49
*** se6astian is now known as se6astian|away20:11
xfxfyay, i7 arrived23:14

Generated by 2.13.1 by Marius Gedminas - find it at!