Saturday, 2016-05-21

*** tpb has joined #timvideos00:00
*** cyrozap has joined #timvideos00:29
CarlFKmithro: ping - I am having trouble getting my opsis working again00:56
CarlFKread 1487168 bytes from file /tmp/hdmi2usb/opsis_hdmi2usb-hdmi2usbsoc-opsis.bin and flash bank 0 at offset 0x00000000 in 57.619904s (25.205 KiB/s)00:57
CarlFKNumato Opsis in 'jtag' mode at /dev/bus/usb/002/07700:57
CarlFKseems I can get it out of jtag mode00:58
mithroI'll be around in an hour.01:33
*** aom has joined #timvideos02:37
CarlFKmithro: ping02:56
mithroAlmost here, give me 5 more03:00
mithroCarlFK: okay, I'm here now03:02
mithroSo, what is your actual question?03:02
CarlFKmithro: I have a few issues - this one is on my screen now:
mithroCarlFK: and how did you make that happen?03:03
CarlFKmithro: flashed an Atlys with   git describe: v0.0.0-699-gb505665-dirty03:04
CarlFKthat image comes out from output0: [email protected] from pattern03:04
mithroCarlFK: Can you paste the complete version info somewhere?03:04
CarlFKand the stream03:04
tpbTitle: Ubuntu Pastebin (at
mithroCarlFK: can you reproduce it?03:06
CarlFKI think so... sec03:07
CarlFKpowercycle atlys - output0 shows atlys test pattern03:08
CarlFK:~/HDMI2USB-firmware-prebuilt/archive/master/v0.0.0-699-gb505665/atlys/hdmi2usb$ openocd -f board/digilent_atlys.cfg -c "init; pld load 0 atlys_hdmi2usb-hdmi2usbsoc-atlys.bit; exit"03:09
CarlFKfxload -B libusb -D vid=0x16c0,pid=0x06ad -t fx2lp -I hdmi2usb.hex03:09
CarlFKflterm --port /dev/ttyVIZ0 --speed 11520003:09
CarlFK[FLTERM] Starting...03:10
CarlFKBIOS> romboot03:10
mithroIf you need to use "romboot" things are not working already03:10
mithrowhat does memtest show on that?03:10
CarlFKhmm, now something new is happening.  //
CarlFKer, sorry for the missing http03:13
tpbTitle: Ubuntu Pastebin (at
CarlFKit got that far, then stopped.  monitor is "no signal" and went to sleep03:14
CarlFKoh swell.  BIOS> memtest03:17
CarlFKMemtest failed: 532718/532736 words incorrect03:17
mithroCarlFK: that is what I wanted you to look for03:17
CarlFKBIOS> romboot ... Executing booted program.  ...  back to this on the hdmi display
mithroif the memtest doesn't pass straight after reboot, its unlikely to work at all03:21
CarlFKwell, I ran this config for that 1 day conf last week03:26
CarlFKcomplete with fuzzy test pattern at the start03:26
CarlFK>>> 532718 - 53273603:27
CarlFKI am suspicious of that being accurate03:28
CarlFKI think memtest is broken03:28
CarlFKbut.. this is an older firmware rev.  I am going to try current03:30
CarlFKoh wait.. that was current: v0.0.0-69903:31
CarlFKbut, I would rather put the Atlys aside and work on the Opsis03:33
mithroI really need someone who is testing both03:33
CarlFKin two days I leave for a show.  I would like to get the Opsis working and have the Atlys as my backup03:34
CarlFKright now I can't get the Opsis out of jtag (I think thats what is going on)03:36
mithroSo again, what are you trying to do and where is it going on?03:37
CarlFKpowered up opsis,  plug in lcd,  I see the color bars with hdmi2usb/ text in the upper corner  (good)03:40
CarlFKplug in usb, lsusb, Bus 002 Device 010: ID 2a19:544203:41
CarlFKhmm, now in dmesg I see: uvcvideo: Found UVC .. [72621.235056] cdc_acm 2-1:1.2: ttyACM0: USB ACM device03:41
mithrosounds like its working to me?03:42
CarlFK  git describe: v0.0.0-663-g54d888a-dirty03:43
CarlFKI am starting to think 30 seconds, and maybe even a min or two isn't enough time for it to get out of: lsusb ... Bus 002 Device 019: ID 16c0:06ad Van Ooijen Technische Informatica03:44
mithroThe Opsis won't ever come up with that info from stuff on the board03:46
mithroThe only way it gets those IDs is if you load from the PC03:47
CarlFKwell, it seems to retain it without anything (power, usb, hdmi in/out) plugged in03:48
mithroThere are capacitors on the board which could hold charge for a while. However, I think it is more likely you have a udev rule or similar03:51
CarlFKsame box I just plugged it into03:54
mithroThe Opsis will connect first with the 5440 ID, then remunerate to the 5442 ID03:55
mithroYou should see this in the kernel logs03:55
mithroIf you have a udev rule which fires on the first ID, then weird things might happen.03:57
mithroAs it will depend on when you plug it in and when it gets triggered03:58
CarlFKdidn't you say it needed 30 seconds of no power to reset?03:59
CarlFKnew problem - never seen this.  the uvc stream is broken... mplayer tv:// -tv driver=v4l2:device=/dev/video1  .. trying to get a visual ...04:03
mithroYes, as I mentioned the capacitors take a while to discharge..04:04
CarlFKI think that's what was causing it to not go out of 16c0:06ad Van Ooijen...04:06
CarlFKmithro: this is alarming:
CarlFKencoder: 1280x720 @ 12fps (781Mbps) from pattern (q: 50)04:10
CarlFK mplayer tv:// -tv driver=v4l2:device=/dev/video104:10
mithroCarlFK: can you reproduce it?04:10
CarlFKError: Invalid flash boot image length 0xffffffff04:11
CarlFKI see bars on the lcd, but the reboot never finished ..04:13
tpbTitle: Ubuntu Pastebin (at
mithroCarlFK: So, what are the steps to reproduce?04:14
CarlFKim loosing track of problems04:15
CarlFKbut reboot got me here:04:15
mithroCarlFK: yes04:15
CarlFKBus 002 Device 012: ID 16c0:06ad Van Ooijen Technische Informatica04:15
CarlFKeverything unplugged04:18
mithroCarlFK: what do you mean?04:19
CarlFKHDMI2USB>reboot - I did that.  I saw  and lsusb Bus 002 Device 012: ID 16c0:06ad Van Ooijen Technische Informatica04:22
tpbTitle: Ubuntu Pastebin (at
CarlFKand stayed there for over a min.  then I pulled all the cords and it is still sitting04:23
CarlFKgoing to plug it back in now and see what happens04:23
mithroCarlFK: Your explanation makes no sense04:23
CarlFKmithro: I think the amount of time needed for the caps to drain is pretty long04:25
mithroCarlFK: I think you are guessing at things04:25
CarlFK(11:16:17 PM) CarlFK: everything unplugged ... (11:21:11 PM) CarlFK: going to plug it back in now and see what happens04:25
CarlFKthat was about 5 min04:25
CarlFKlsusb shows: Bus 002 Device 016: ID 16c0:06ad Van Ooijen Technische Informatica04:25
mithroCarlFK: You did "reboot" and then got the 06ad means you have a udev rule04:26
mithroCarlFK: nothing to do with caps there04:26
mithroThe 06ad can *only* come from the computer you are connecting to04:26
mithroCarlFK: if you look at the dmesg output, you should be able to see what is going on04:29
CarlFKmithro: the udev rule would have come from a package I instlled form the timvideos ppa - which tumbleweed made but you would have ... ok, let me look04:29
tpbTitle: Ubuntu Pastebin (at
mithroCarlFK: compare that to the following04:31
mithroCarlFK: disconnecting the power, wait 30 seconds, connecting the USB, then connecting the power04:32
mithroCarlFK: and04:32
mithroCarlFK: disconnecting the power, wait 30 seconds, connecting the power, wait 30 seconds, connecting the USB04:32
mithroThen compare it to a machine which doesn't have any packages installed04:38
CarlFKmithro: I think I see the pattern - do you want me to paste dmesg?04:54
mithroCarlFK: if you want to paste it04:55
CarlFK - box with pacakges (so udev rules)04:56
tpbTitle: Ubuntu Pastebin (at
CarlFK - odroid arm, no ppa04:57
tpbTitle: Ubuntu Pastebin (at
mithroThe odroid is doing something weird as it seems to support OTG on that port04:58
CarlFKwhat dir do you think the udev rules are ?04:59
CarlFKnot /etc/udev/rules.d/04:59
mithroCarlFK: you know about locate and find right?05:03
CarlFKmithro: yes - but not sure what to locate and find05:06
CarlFKthere is also some way to monitor what rules fire05:07
mithroWhat does "locate *udev*" find?05:07
CarlFK[email protected]:/usr$ locate *udev*|wc05:09
CarlFK    794     794   2883305:09
CarlFKfound /lib/udev/rules.d05:10
CarlFK85-ixo-usb-jtag.rules # Numato Opsis board - Failsafe Mode (
CarlFKSUBSYSTEM=="usb", ACTION=="add", ATTRS{idVendor}=="2a19", ATTRS{idProduct}=="5440", RUN+="/sbin/fxload -D $tempnode -t fx2lp -I /lib/firmware/ixo-usb-jtag/hw_opsis.hex"05:13
CarlFKmithro: I think this is dependent on the hdmi input
CarlFKthat was with   git describe: v0.0.0-663-g54d888a-dirty05:49
CarlFKI am going to try the most resent05:50
mithroCarlFK: you know what I'm going to ask? :-P05:50
CarlFKmithro: it seems dependent on the input, currently Acer v0.3108 Aspire 1410 - hdmi out05:59
mithroCarlFK: can you reproduce it?05:59
CarlFKif you want steps, step 1, get one of those laptops from ebay ;)06:00
CarlFKor some other device that provides a similar signal06:01
CarlFKI saw it 5 min before last weeks show was about to start06:01
CarlFKon my Atlys #1 (I have 2 now)06:01
CarlFKmithro: I am seeing this noise again:
CarlFKeventually I am going to get the current firmware and find all has been fixed ;)06:17
mithroCarlFK: If you have a procedure that you can make it happen reliably, then we can figure out what is going06:18
CarlFKmithro: these last two are just HDMI2USB>x c 0 106:19
mithroCarlFK: that is not a full procedure06:19
mithroCarlFK: From everything powered off and no cables connected - can you boot everything up and get that behaviour?06:20
CarlFKmithro: I can't seem to get rid of it06:22
CarlFKso sure.  if I plug/unplug cables, it is there.06:23
mithroCarlFK: so, explain to me in excruciating detail of how you got to this..06:23
CarlFKstack 2 i3 desktops on top of each other...06:35
CarlFKpile on a gig swich and a laptop06:35
CarlFKand on the side the exatron vga-hdmi converter set to generate test patterns06:36
CarlFKthat gives me 4 inputs06:36
CarlFKI get the most noise when it is plugged into the laptop06:38
CarlFKlooping an Opsis output1 into input0 seems to cause a problem06:39
mithroCarlFK: that doesn't explain a lot, what are you loading onto what? Are you using an Opsis or an Atlys? What revision? Are you using Input 0 or Input 1?06:39
CarlFKnot sure exactly how that image got there, does it matter?06:40
mithroThis one? ?06:41
tpbTitle: HDMI2USB-firmware-prebuilt/archive/master/v0.0.0-663-g54d888a/opsis/hdmi2usb at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at
CarlFK3' magic cable, input006:42
mithroWhat does the monitor show with the test pattern output?06:43
CarlFKclean, no noise06:45
mithroCarlFK: perfectly clean? Can you take a very close look at the edges of the bars?06:46
CarlFKmithro: looks good.
CarlFKI thin the dots I see are dirt on the screen06:52
mithroCarlFK: Is there normally space between the H and D ?06:52
CarlFKmithro: I don't think so.  it doesn't stand out as something odd, so I think it is normal.06:54
mithroCarlFK: can you capture a full photo?06:55
CarlFK  full06:58
mithroI assume the weirdness around the edges is from the photos not the monitor?06:59
CarlFK  close up of red/blue because it looks like it doesn't line up, but I think that's just how monitors work, like the red dots are above the blue06:59
CarlFKI think so06:59
CarlFKmithro: I should get to bed soon.  any objection to me flashing v0.0.0-699-gb505665 (seems to be current)07:06
mithroCarlFK: before you do07:07
mithroCarlFK: go back to the other input07:07
mithroCarlFK: do you get the same on the encoder?07:07
CarlFKno - encoder is way worse - and mplayer complains about errors07:08
CarlFK  regardless of if the input is hdmi or pattern07:09
CarlFKhuh, now mplayer wont even try07:09
mithroCarlFK: can you use something other than mplayer07:09
mithromplayer seems to get stuck if you have a tty open at the same time07:09
CarlFKgst-launch-1.0 v4l2src device=/dev/video1 ! xvimagesink -v07:13
CarlFKnever plays anything07:13
mithroCarlFK: you need some stuff in between there07:13
CarlFKoh right jpeg..07:13
CarlFKmithro: how does this look gst-launch-1.0 -v v4l2src device=/dev/video1 ! queue ! image/jpeg,width=1280,height=720 ! jpegdec ! videoconvert ! xvimagesink07:16
mithroHow about "v4l2src xxx ! queue ! decodebin ! xvimagesink" ?07:17
CarlFKgst-launch-1.0 -v v4l2src device=/dev/video1 ! queue ! decodebin ! xvimagesink07:18
CarlFKstill nothing07:18
mithroCarlFK: it behaves differently with different inputs?07:19
CarlFKI can't seem to get the uvc stream working07:20
CarlFKthe hdmi is mostly ok, except for the bit of noise and occasional ... this pic:07:20
CarlFKit does that about every 30 seconds for about 1 second07:21
mithroWhat do you mean "occasional"?07:21
mithroCarlFK: this is an Opsis?07:21
CarlFKoccasional = every 30 seconds for about 1 second07:22
mithroCarlFK: When you say "it does that above every 30 seconds for about 1 second" - is that with any input? Or just that one?07:22
CarlFKjust the Aspire 1410 laptop07:23
mithroCarlFK: okay07:23
mithroCarlFK: does the uvc work with the other laptops?07:23
CarlFKif I move the cable to the desktop's hdmi it is fine07:23
CarlFKthe uvc is broken with pattern07:24
mithroCarlFK: can you reset and see if you can get the pattern via UVC with nothing connected?07:26
CarlFKsure.. and... I just moved the input from laptop to extatron for a min, then back.  now it is solid.  no more
mithroCarlFK: can you get it back now?07:28
CarlFKun/plug ... flickering is back.  but less often.  I guess.07:30
mithroCarlFK: can you turn on the debugging for the input then?07:30
CarlFKmithro: whats the command?07:31
mithrodebug input0 on07:31
CarlFKdvisampler0: ph: -17    5    4 // charsync:111 [4 5 5] // WER:  0   0   0 // ch007:33
CarlFKabout 100 then one line of07:33
CarlFKdvisampler0: ph: -16    4    1 // charsync:111 [4 5 5] // WER:1047   0   0 // c007:33
CarlFKand the output flickers07:33
*** sb0 has joined #timvideos07:33
mithroCarlFK: is that consistent?07:36
CarlFKmithro: no  I have only seen 1 set of 3 errors in a row.  it is not doing about 1 error every 10 no errors07:38
mithroCarlFK: We really need deeprave's hdmi thing so we can set up monitoring of those debug values07:39
CarlFKnow a run of ... i lost count.  50 for sure07:39
CarlFKI am running minicom, pretty sure I can turn on logging07:39
mithroCarlFK: yes, but that doesn't help over the course of a whole conference07:40
CarlFKflashing before I fall over ;)07:42
CarlFKI gotta go. I just fell aasleep07:51
*** Bertl_zZ is now known as Bertl08:19
mithroWell, I'm going to go find some dinner08:58
*** aom has quit IRC09:41
*** se6astian|away is now known as se6astian09:46
*** Bertl is now known as Bertl_oO10:36
mithro_florent_: you wouldn't happen to be around?11:19
CarlFKmithro: still up?14:03
mithroCarlFK: just heading to bed14:43
*** sb0 has quit IRC15:28
CarlFKmithro: still up?  v0.0.0-623-g5ac1879-dirty  debug, almost no errors.  for sure nothing over 10, and no spazzing screen15:28
CarlFKhmm, reboot the laptop (input0) and now more errors.  but less errors than before.  (not sure how to measure error count)15:47
*** Bertl_oO is now known as Bertl16:11
CarlFKI only get that noise on the encoder when I plug in the hdmi inputs to Aspire 1410 laptop and a desktop16:32
CarlFKok, so the solution is not to plug in the 2nd hdmi.  'good'16:36
*** se6astian is now known as se6astian|away22:19

Generated by 2.13.1 by Marius Gedminas - find it at!