Sunday, 2016-08-07

*** tpb has joined #timvideos00:00
*** Bertl_oO is now known as Bertl_zZ02:14
mithropring cr1901 / cr1901_modern102:24
mithros/pring/ping/02:24
cr1901mithro: Hi! Yes, I'm currently petsitting for a week, and I couldn't bring my minispartan with me. I'll be making a trip home tomorrow tho so I can further debug from last time02:25
mithrocr1901: I'd like to give your counter stuff a go on my atlys/opsis boards here02:26
cr1901mithro: The counter itself should work fine. I can make a patch (sans the minispartan stuff) and show you how to add it into your code02:27
mithrocr1901: That would be awesome02:28
cr1901mithro: Shall I continue looking into minispartan woes as well? I mean, I'm going to need to get it working anyway02:28
mithrocr1901: I think we need to get on _florent_ getting the litedram stuff merged02:29
mithrocr1901: I'm pretty sure https://github.com/enjoy-digital/scarab-soc hasn't been updated to his latest stuff, right?02:29
tpbTitle: GitHub - enjoy-digital/scarab-soc: ScarabHardware miniSpartan6+ SoC based on Migen/MiSoC (at github.com)02:29
cr1901mithro: That may be true as well, but there is DEF some EDID woes as well02:29
mithrocr1901: I disable the ability to write the lm32 firmware onto the SPI flash for the moment - it was breaking the opsis/atlys boards because it loads the firmware contents into the dram and tries to run it from there02:30
cr1901mithro: Okay, I'll take a look. Something important tho: flterm doesn't really play nicely with Windoze, so I've always been trying to flash onto the ROM.02:32
cr1901mithro: My copy of target/minispartan.py in hdmi2usb vs scarab-soc is *not* that drastically different as of this moment02:32
mithrocr1901: Hrm? You mean you have been flashing to the SPI flash?02:32
cr1901mithro: Yes, this whole time02:32
* cr1901 backs up slowly and runs02:35
mithrocr1901: Nah, that is fine02:37
mithrocr1901: Although I'd like to know why flterm isn't working for you02:38
cr1901mithro: I'll have to check when I get home tomorrow (I'll only be there from about 10 to 5)02:38
cr1901I don't remember the details02:38
mithrocr1901: Okay02:39
mithrocr1901: Are you able to get the pull request bit done before tomorrow?02:39
cr1901mithro: I have access to the repo remotely. Let's see what I can do...02:41
mithrocr1901: Thanks!02:47
cr1901mithro: This is gonna have to wait till morning thanks to line ending bullshit thinking EVERY file has been modified03:31
cr1901I only set up (S)FTP access to my machine :/. Lesson learned03:32
mithrocr1901: Okay03:32
mithroCarlFK: are you going to be in a position to test VGA capture over the next couple of days?03:54
CarlFKmithro: right now I only have an opsis and the vga-atlys board.  my atlys are all being shipped back, should arrive in 2 or 3 days (mon or tues)03:56
mithroCarlFK: okay, so that is no then :P03:58
mithroCarlFK: well, atleast until later in the week03:59
CarlFKmithro: oh wait.. I take that back.   I brought an atlys with me.  so yeah, I can do something in about 12 hours (11p tonight, im beat04:00
mithroCarlFK: well, don't have anything to test right this second04:00
mithroCarlFK: but might in about 12 hours04:01
mithroCarlFK: depends on if Rohit gets a chance to do hacking today or not04:01
xfxfmithro: i'll message you here instead of disturbing you constantly, reply when you're free05:04
xfxfthe gamma edid patch is still in, but i still think the gamma is wrong05:04
xfxfi'll have a deeper look, but do you remember the thing we had that was causing crashes?05:04
xfxfi.e. how would i reproduce?05:05
xfxf(my laptop runs gnome 3 and it's working fine)05:05
mithroxfxf: I'm pretty sure it was only older versions of gnome 305:06
mithroxfxf: Maybe take a look at http://www.edidreader.com/ and https://cgit.freedesktop.org/xorg/app/edid-decode05:10
tpbTitle: Web Based EDID Reader (at www.edidreader.com)05:10
*** rohitksingh has joined #timvideos06:59
rohitksinghmithro: ping07:08
mithrorohitksingh: pong!07:08
rohitksinghmithro: this build failed because timvideos:master doesn't have gateware merged yet. https://travis-ci.org/timvideos/HDMI2USB-misoc-firmware/builds/15032548507:09
tpbTitle: Travis CI - Test and Deploy Your Code with Confidence (at travis-ci.org)07:09
mithrorohitksingh: That is saying you are trying to merge something which has dependencies no in that pull request07:10
rohitksinghmithro: and I think we can try your fix...we can run OVDD from 3.3V rails directly and run other sensitive PLL supplies etc from 5V07:11
mithrorohitksingh: yeah07:11
mithrorohitksingh: but lets hold off on that07:11
mithrorohitksingh: one thing at a time :P07:11
rohitksinghmithro: yes I missed the fact that travis will fail07:11
mithrorohitksingh: You only have one pull request open at the moment?07:12
rohitksinghmithro: 2. This one also https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/215 Can you see if this can be merged or not. Doesn't have any dependency, and have fixed everything discussed in the comments thread. Its gateware only code07:13
tpbTitle: Initial gateware code for VGA capture by rohit91 · Pull Request #215 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)07:13
mithrorohitksingh: Oh - I missed it because it is below all the other WIP requests07:15
rohitksinghmithro: oh. can you take a look at it now. I will fix any issues on those 2 commits.07:18
mithrorohitksingh: Is there any reason not to have #279 as part of #215?07:28
*** cr1901 has quit IRC07:28
rohitksinghmithro: Initially I did that itself, but I thought we might need some discussion on atlys_video target, like we are having now about 3.3V tolerant Ethernet. that's why :p07:30
rohitksinghmithro: I can add it to #215 if you want07:31
mithrorohitksingh: I think that is the right move here07:32
rohitksinghmithro: okhay i'll add it to #215 after lunch. having lunch right now.07:33
*** cr1901 has joined #timvideos07:43
mithrorohitksingh: Could you take a look at https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/280 and tell me what you think? I'm just trying to make the buttons clearable bit now08:11
tpbTitle: Support LEDs, Switches and Buttons on Atlys (plus improve VCCB2 support). by mithro · Pull Request #280 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)08:11
rohitksinghmithro: okay08:11
rohitksinghmithro: Excellent. Its much flexible now that we can configure individual banks' IOStandards.08:30
rohitksinghmithro: why did we remove encoder's leds though?08:31
xfxfCarlFK: what do you do in production, run dvsmon or screen?09:08
xfxf(in terms of auto-restarting ingest.py things)09:08
mithrorohitksingh: It seems to fail if you have the leds in two modules09:14
rohitksinghmithro: yeah. it will fail. I was just unsure of why not just have only 6 higher/lower bits of leds controllable from registers, unless of course we need all 8 bits...09:18
mithrorohitksingh: we can look at reenabling that functionality in the future09:21
rohitksinghmithro: yeah we can do that09:22
mithrorohitksingh: I'm going to fix this so we have two seperate pull requests here09:23
mithrorohitksingh: Can you take a careful look over 280 now? It just contains the IO standard fixes now (Check that banks and stuff match). If you think it looks good, can you comment on it?09:38
rohitksinghmithro: sure09:41
mithrorohitksingh: Then I think we should rebase your VGA stuff on top of that IO standard fixes09:42
rohitksinghmithro: yeah. only platform file will get modified09:43
*** ssk1328 has joined #timvideos09:43
ssk1328mithro: In the conversation, we had here, https://logs.timvideos.us/%23timvideos/%23timvideos.2016-08-05.log.html#t2016-08-05T08:39:06, I said no, but I am actually combining two sinks into a single source in the updated VTG module09:45
ssk1328mithro: Here https://github.com/ssk1328/HDMI2USB-misoc-firmware/blob/float-arithmetic/gateware/hdmi_out/format.py#L7809:46
tpbTitle: HDMI2USB-misoc-firmware/format.py at float-arithmetic · ssk1328/HDMI2USB-misoc-firmware · GitHub (at github.com)09:46
ssk1328mithro: Or the other way would be to combine the two two streams into a single layout even before they arrive at VTG module09:48
mithrossk1328: So this code -> https://github.com/ssk1328/HDMI2USB-misoc-firmware/blob/float-arithmetic/gateware/hdmi_out/format.py#L93 should join the two streams together?09:49
tpbTitle: HDMI2USB-misoc-firmware/format.py at float-arithmetic · ssk1328/HDMI2USB-misoc-firmware · GitHub (at github.com)09:49
ssk1328mithro: Yes09:49
ssk1328mithro: Payload from two sinks is getting connected to payload of one source09:50
mithrossk1328: I think you should extract the piece of code which has two sinks and one source - then you can do a testbench for that module09:50
mithrossk1328: Looking at that code, I'm not certain the stb/acks are being combined / asserted in the right way09:50
ssk1328mithro: You mean here https://github.com/ssk1328/HDMI2USB-misoc-firmware/blob/float-arithmetic/gateware/hdmi_out/format.py#L14609:51
tpbTitle: HDMI2USB-misoc-firmware/format.py at float-arithmetic · ssk1328/HDMI2USB-misoc-firmware · GitHub (at github.com)09:51
mithrossk1328: Yeah - I don't think what that is doing is right?09:52
ssk1328mithro: The default is doing something like this https://github.com/timvideos/HDMI2USB-misoc-firmware/blob/master/gateware/hdmi_out/format.py#L14209:53
tpbTitle: HDMI2USB-misoc-firmware/format.py at master · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)09:53
mithrossk1328: yeah - I think that is fine when you only have one source09:53
mithrossk1328: but when you have two sources you are pulling from, things are more complicated09:54
ssk1328mithro: You mean a different logic at the if condition or some more changes elsewhere?09:54
mithrossk1328: both I think09:54
mithrossk1328: but I'm unsure09:54
ssk1328mithro: Maybe I can use a something to pack the tokens before arriving at VTG module09:55
mithrossk1328: Once you have a two sources -> one sink combinder object, you should be able to have minimal changes to the VTG module09:55
ssk1328mithro: So I won't have to play with VTG09:55
mithrossk1328: Yeah09:55
ssk1328mithro: Yeah, excatly09:55
ssk1328mithro: https://migen.readthedocs.io/en/latest/dataflow.html#pack09:55
tpbTitle: Dataflow Migen X documentation (at migen.readthedocs.io)09:55
ssk1328mithro: I can maybe use a pack before VTG09:56
mithrossk1328: I'm unsure from the description - but I think that only operates from one sink09:57
ssk1328mithro: Oh09:57
mithrossk1328: IE it takes an 8bit @ 100 MHz and converts it to 32bit @ 25MHz type thing09:58
ssk1328mithro: Yeah, its different09:58
mithroYou want something which takes 2 separate 8bit@100MHz streams and generates a 16bit@100MHz stream09:59
ssk1328mithro: Yes, that I understood10:00
rohitksinghssk1328: I think you should use this one https://migen.readthedocs.io/en/latest/dataflow.html#combinator? this seems to be the one which combines tokens10:00
tpbTitle: Dataflow Migen X documentation (at migen.readthedocs.io)10:00
ssk1328rohitksingh: Seems correct for the use here10:02
ssk1328rohitksingh: I will check this out10:02
rohitksinghssk1328: great10:03
rohitksinghmithro: we can map banks directly with pins no?10:10
rohitksinghmithro: https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/280/files#diff-e951b052fef48179c79f7ca67755519fR5910:10
tpbTitle: Rework IO standards for Atlys board by mithro · Pull Request #280 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)10:10
rohitksinghmithro: for example, i'm not sure of whether pin L15 is actually on bank 1 unless I check in datasheet10:11
mithrorohitksingh: there are comments above it10:25
rohitksinghmithro: oh. diff wasn't showing them :p with comments directly placed above, it is okay then10:27
mithrossk1328: The Combinator the rohit pointed too does seem like the right thing to be using here11:01
mithrorohitksingh: Looking at the VGA pull request, there isn't any lm32 firmware code to use this yet?11:15
rohitksinghmithro: yeah this pull request is only for gateware. This one is easier to merge, and has no conflicts....I'll create new pull request for firmware...it will require much close review thats why11:17
mithrorohitksingh: Okay11:18
rohitksinghmithro: you told me to review cas stuff...I can't find it anywhere now except this -> https://github.com/timvideos/HDMI2USB-misoc-firmware/commit/e39f2fb and there is an invalid literal "10ms" which github has also flagged11:22
tpbTitle: Finishing the TODO that make buttons set and then clearable. · timvideos/HDMI2USB-misoc-firmware@e39f2fb · GitHub (at github.com)11:22
mithrorohitksingh: Already fixed -> https://github.com/mithro/HDMI2USB-misoc-firmware/blob/control-n-status/gateware/cas.py#L6111:23
tpbTitle: HDMI2USB-misoc-firmware/cas.py at control-n-status · mithro/HDMI2USB-misoc-firmware · GitHub (at github.com)11:23
mithrorohitksingh: So, it would be good to put a few extra comments into your VGA pull request11:25
mithrorohitksingh: IE Some links to the vmodvga repository and stuff would be good11:25
rohitksinghmithro: sure11:26
mithrossk1328: any luck?11:27
*** andi-m has quit IRC11:39
rohitksinghmithro: now that we are using 3.3 iostandard for ethernet, I think I should change back atlys_video target to inherit MiniSoC again ?11:42
mithrorohitksingh: Maybe?11:44
mithrorohitksingh: I merged #28011:44
*** andi-m has joined #timvideos11:44
*** cr1901_modern has joined #timvideos11:47
rohitksinghmithro: awesome. I'll fix #215 then11:47
*** cr1901_modern1 has quit IRC11:49
mithrorohitksingh: Here is the LED/Switch/Push Button stuff in a seperate pull request -> https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/282/files11:49
tpbTitle: Adding a module to allow firmware access to the LEDs, Push Buttons and Switches (on the Atlys) by mithro · Pull Request #282 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)11:49
*** andi-m has quit IRC11:53
*** cr1901 has quit IRC11:55
*** Bertl_zZ is now known as Bertl12:00
*** andi-m has joined #timvideos12:04
*** cr1901 has joined #timvideos12:09
*** andi-m has quit IRC12:19
*** andi-m has joined #timvideos12:24
*** shenki has quit IRC12:25
*** shenki has joined #timvideos12:25
*** ChanServ sets mode: +v shenki12:25
*** andi-m has quit IRC12:28
*** andi-m has joined #timvideos12:33
ssk1328mithro: gateware build in progress12:41
mithrossk1328: Can you commit and push what you have so far so I can take a look at it?12:41
ssk1328mithro: Except the changes since last few hours, everything else is updated, I will push the latest code as well12:42
mithrossk1328: I want to see how you are using the Combinator thingy12:43
ssk1328mithro: Yes12:43
*** andi-m has quit IRC12:46
*** andi-m has joined #timvideos12:47
ssk1328mithro: I have defined a module myself https://github.com/ssk1328/HDMI2USB-misoc-firmware/blob/float-arithmetic/gateware/hdmi_out/format.py#L71]12:48
tpbTitle: HDMI2USB-misoc-firmware/format.py at float-arithmetic · ssk1328/HDMI2USB-misoc-firmware · GitHub (at github.com)12:48
ssk1328mithro: Which is similar  to Combinator defined here https://github.com/timvideos/migen/blob/ad57f636eb95c0c0ba539fe9048266120f3f5c54/migen/flow/plumbing.py#L1912:48
tpbTitle: migen/plumbing.py at ad57f636eb95c0c0ba539fe9048266120f3f5c54 · timvideos/migen · GitHub (at github.com)12:48
mithrossk1328: Why can't you use that one?12:49
ssk1328mithro: So instead of this I was looking at CombinatorialActor class, which I didn't was appropriate, so I wrote the class myself, and later while scanning through found this and understood this was the Combinator class rohitksingh was referring to12:50
rohitksinghmithro: shall I add vga signals in platform file to _io or _io_vccb2_3v3? currently it should go to later one12:53
mithrorohitksingh: lets go with the second one for now12:54
rohitksinghmithro: great. meanwhile can you take a look at the docstring and see if is sufficient or needs something more?12:55
rohitksinghmithro: here https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/215/commits/49dec2397a1445938060653a657937378a9af7c5#diff-03a5e3b5ec77c130e9b6342b134584f3R712:55
tpbTitle: Initial gateware code for VGA capture by rohit91 · Pull Request #215 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)12:55
mithrorohitksingh: Yeah, the comment looks like the type of stuff I was looking for - your formatting is a bit weird12:57
mithrorohitksingh: See https://github.com/m-labs/migen/blob/master/migen/genlib/coding.py#L912:57
tpbTitle: migen/coding.py at master · m-labs/migen · GitHub (at github.com)12:57
mithrorohitksingh: And https://github.com/m-labs/migen/blob/master/migen/genlib/sort.py12:58
tpbTitle: migen/sort.py at master · m-labs/migen · GitHub (at github.com)12:58
rohitksinghmithro: ohkay. hadn't seen how migen does that. I'll fix it12:58
mithrorohitksingh: Format seems to be describe a bit here -> http://hplgit.github.io/teamods/sphinx_api/html/sphinx_api.html13:01
tpbTitle: Simple Formatting Rules (at hplgit.github.io)13:01
mithroIt's not one of the formatting systems I'm use to - I normally use the Google one for obvious reasons13:02
mithrorohitksingh: But consistency with the existing stuff is the best way to go13:02
mithrorohitksingh: I don't want you to do this now, but I think we could probably replace your current counter stuff with the VTG stuff13:04
mithrorohitksingh: some more links would be good13:04
ssk1328mithro: Finally static mixing is working! :)13:04
rohitksinghmithro: VTG is  video timing generator, right?13:05
mithrossk1328: OMG! Lets see13:05
mithrorohitksingh: yes13:05
rohitksinghmithro: okay I'll do that...btw pads are which types? lists or some migen class?13:05
mithrorohitksingh: Not sure :-P13:06
*** andi-m has quit IRC13:06
mithrorohitksingh: I'm unsure how it would need to be modified - but it already has a bunch of stuff around the data valid stuff13:06
mithrorohitksingh / ssk1328: I've been working on some ideas around how to rework a lot of the pixel / video code to make it easier to understand and have better structure ->   https://docs.google.com/document/d/1L8lz7u2uj6MrzSQv4b1Vk6Rmic26okyRklOju5IWLYA/edit#13:07
tpbTitle: MiSoC HDMI refactoring - Google Docs (at docs.google.com)13:07
mithrobasically it would be nice if we had some better primitives like "pixels" and "pixel streams" and stuff13:08
*** andi-m has joined #timvideos13:09
mithrolove your comments - still very much a WIP13:09
ssk1328mithro: This seems nice, I have always wanted to understand more of this13:10
rohitksinghmithro: yeah. i'll read it tonight...13:10
ssk1328mithro: https://photos.google.com/share/AF1QipOJ0tE656GhYkF4sU3jD3PExuONIAmLrsQplsSLk78PwWPl-T7Tiq2Zrjfm-xqUYw/photo/AF1QipOSkGlF0MK-cVYDG3Q8m8qGLIuCLB8sMg35pEL5?key=RGVxV2tleUdWQVdEV0k2cTZHWGFMUk10Sy1vODB313:10
mithrossk1328: YAY! It's all lined up too13:10
ssk1328mithro: pattern mixed with input from my laptop screen extended13:10
ssk1328mithro: The brightness less, just have to change a number in gateware code13:11
rohitksinghssk1328: great job! :013:11
rohitksingh* :)13:11
ssk1328mithro: Took longer than expected! But feels good to see this :)13:11
mithrossk1328: Yeah, we really should have started down this path first rather than worrying about the mask stuff first13:11
rohitksinghbrb...have to go outside13:12
ssk1328mithro: I guess next important task is 1. to do this for all outputs; 2. Firmware for dynamic mixing13:14
mithrossk1328: I added a diagram to your Mixing doc13:19
*** rohitksingh has quit IRC13:27
mithrossk1328: Does that type of structure make sense?13:28
mithrossk1328: Basically, trying to separate out bits into separate blocks13:28
ssk1328mithro: I have a PPT to attend right now, I will be back in sometime13:31
mithrossk1328: Okay13:34
mithrossk1328: Added some more diagram which might be helpful....13:34
mithrossk1328: You should also see if you can replace your own custom thing with the inbuilt one as less code is always good13:40
*** Bertl is now known as Bertl_oO13:55
mithrowell, bed time for me14:53
CarlFKxfxf: dvsmon14:54
*** sb0 has quit IRC17:00
*** rohitksingh has joined #timvideos18:40
*** rohitksingh has quit IRC18:50
*** springermac has quit IRC18:55
*** springermac has joined #timvideos18:55
*** springermac has quit IRC19:01
*** springermac has joined #timvideos19:02
*** ssk1328 has quit IRC19:45
cr1901_modernmithro: https://github.com/timvideos/HDMI2USB-misoc-firmware/pull/28320:49
tpbTitle: Frequency Counter for Debugging by cr1901 · Pull Request #283 · timvideos/HDMI2USB-misoc-firmware · GitHub (at github.com)20:49
*** cr1901 has quit IRC23:55

Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!