Saturday, 2016-06-25

*** tpb has joined #timvideos		00:00
*** aravind_ has joined #timvideos		00:20
aravind_	Hello	00:20
CarlFK	hi	00:20
aravind_	Hi Carl	00:20
aravind_	im new here	00:20
CarlFK	hello - what brings you here ?	00:26
aravind_	im an electronics engineer.. looking for pcb design work..	00:27
aravind_	what about you?	00:27
aravind_	@mithro suggested some projects to work on	00:27
CarlFK	I'm mostly into software and going to events to record talks	00:34
CarlFK	and running down the street to get a pizza. brb	00:34
aravind_	cool. Nice to meet you	00:35
mithro	aravind_: I'll be around in a couple of hours	00:37
mithro	Currently out at brunch	00:37
aravind_	Sure no prob.. i will be here till my blankets pull me in	00:38
aravind_	:)	00:38
mithro	If you are still around then, we can chat about possible projects	00:38
mithro	I'll also respond to your email	00:38
aravind_	Great :)	00:38
mithro	tumbleweed / CarlFK: the vizzini driver should be replaced with the stuff Ben (The debian dvswitch guy?) and shenki working on	00:40
* mithro pokes shenki with a pointy stick - you were going to get that upstreamed		00:40
*** aom_ has quit IRC		00:47
CarlFK	mithro: it may have happened, but now it needs to trickle down into the distros.. and the odroid is pretty far behind - like on kernel 3.14	00:51
mithro	CarlFK: I'm 100% sure it hasn't happened	01:30
mithro	I stopped bugging shenki long enough he has probably forgotten all about it	01:31
mithro	aravind_: I'll be around in about 10-15 minutes, you still awake?	01:32
aravind_	Sure @mithro	01:33
mithro	Great, talk to you then	01:34
mithro	aravind_: couple of questions while I walk	01:34
aravind_	sure	01:35
mithro	aravind_: what OS do you generally use?	01:35
mithro	aravind_: Have you used KiCad before?	01:35
aravind_	mithro i use OS X	01:35
aravind_	yes I have used KiCad before	01:35
mithro	What is your level of experience in this area?	01:36
aravind_	circuit design --> been in industry for last 10 years	01:37
mithro	What are you hoping to get out of this? Practice? Something for your portfolio? See something you created be used? Cool friends to hang out with? Want to give back to the community?	01:38
mithro	Need something interesting to do in your spare time?	01:38
aravind_	Well all of it + i plan to start design services , a small step to be my own boss :)	01:39
mithro	Ah cool!	01:40
aravind_	:)	01:40
aravind_	as of now, my plan is to work when I get spare time (few hours a week)	01:41
mithro	So, I guess your more looking to come up with a good workflow where you can do designs quickly and efficiently?	01:41
mithro	KiCad is probably the least tested on Mac, but at least you have things like proper scripting environment	01:43
aravind_	I have used KiCad for some designs in my mac (not very complex though)	01:44
mithro	Almost at a computer now	01:44
aravind_	okie dokie	01:44
mithro	On the experience side, do you know about things like differential traces, length matching and impedance matching (even if you have not done them in KiCad)?	01:46
mithro	Btw what have you used professionally? Altium	01:47
mithro	?	01:47
aravind_	yes	01:47
aravind_	professionaly mostly PADS and CR2000	01:47
aravind_	i was a circuit designer mostly , layout being done by layout engineer.	01:48
mithro	Finally at a computer now! :)	01:49
mithro	Typing is much easier than doing it on the phone while walking ;)	01:49
mithro	CarlFK: ping? You still around?	01:50
CarlFK	mithro: here	01:50
aravind_	i concur :)	01:50
mithro	aravind_: so, are you just interested in the PCB side of things, or want to do the full circuit design too?	01:51
aravind_	happy to be on full circuit design	01:51
aravind_	spec + schematics + layout + sims	01:52
mithro	aravind_: okay cool	01:52
aravind_	@mithro I understand your projects are version controlled using git	01:53
mithro	aravind_: as I mentioned, the two projects that I think would be good to start with are, the TOFE-VGA board and the TOFE-HiLink board	01:53
aravind_	Okay.	01:53
aravind_	@mithro I will go through them in detail. Anyone else working on those at the moment?	01:54
mithro	aravind_: the TOFE-VGA was being worked on by seaLne - but he ran into issues with RSI, so had to stop	01:57
mithro	I believe it was at the stage of the schematic being pretty much done and he was starting the PCB layout / design	01:57
aravind_	@mithro any detailed doc for the proj?	01:58
mithro	aravind_: Sure, let me give you a bit of background	02:01
aravind_	@mithro Thanks	02:01
mithro	aravind_: The HDMI2USB project originally started using the Digilent Atlys boards for developing our firmware	02:01
mithro	aravind_: The problem is that the Digilent Atlys boards don't have any VGA capability and VGA is probably going to be around until the heat death of the universe	02:02
aravind_	:)	02:02
aravind_	thats true, people still use VGA for something or other	02:02
mithro	aravind_: The Atlys boards have an expansion connected - it looks like the old SCSI ones and is called the VHDCI (Very High Density Connector Interface)	02:03
aravind_	okay	02:03
mithro	aravind_: so, in 2014 we had a Google Summer of Code student develop an expansion board which connects to the VHDCI interface and allows VGA capture	02:06
aravind_	okay	02:06
mithro	This board works okay and there are about 20-30 floating around the world	02:07
mithro	aravind_: You can find that board design at https://github.com/timvideos/HDMI2USB-vmodvga	02:07
tpb	Title: GitHub - timvideos/HDMI2USB-vmodvga: VGA capture expansion board for HDMI2USB (or Digilent Atlys prototype board). (at github.com)	02:07
mithro	aravind_: with a bunch of description from Rohit	02:08
aravind_	okay	02:08
aravind_	still getting my grip around git :)	02:08
mithro	aravind_: GitHub has quite a lot of good documentation about using Git+GitHub on Mac	02:09
mithro	aravind_: They have a bunch of tools which are Mac only but have nice UI and stuff	02:09
CarlFK	aravind_: I have been using git for 5 years. I know about 1%: I commit, I push, and sometimes I clone to some other box.	02:09
CarlFK	that's all you really "need" to know to get going.	02:09
aravind_	@mithro you are right :). I am following the tutorial now	02:09
mithro	aravind_: The Atlys board however is a development board, which isn't really all that suitable for what we want to do - which is record conferences	02:10
aravind_	okay	02:10
aravind_	got it	02:10
mithro	aravind_: so we created our own board which is pretty similar called the Numato Opsis	02:10
aravind_	so a dedicated board is needed	02:10
mithro	aravind_: the VHDCI connector however is super expensive - like $4 to $8 each	02:11
mithro	aravind_: so we didn't use it on the Opsis	02:11
aravind_	yes i think its a dense connector for I/O expansion	02:12
mithro	aravind_: instead we decided to reuse PCI-Express connectors instead - which means that the expansion boards don't even need a connector	02:12
aravind_	ah okay	02:12
mithro	aravind_: The usage of PCI-Express connectors in this manner is a "open standard" I'm trying to promote which is called TOFE (Tim's Open FPGA Expansion)	02:13
aravind_	got it	02:13
aravind_	@mithro so the PCIe pinouts dont match to the standard	02:13
mithro	aravind_: I've tried to keep it as close to PCI-Express as possible, we just allow the signals to be slower than PCI-Express uses	02:14
aravind_	okay. whats the max signal rate in our case?	02:14
mithro	aravind_: All the power pins and data pins are in the same locations, so you can plug in a PCI-Express card and it won't blow up	02:14
aravind_	ah okay	02:15
mithro	aravind_: I'm trying to do as minimal amount of work as possible ;)	02:15
aravind_	I understand :)	02:15
mithro	aravind_: The signal speeds are DC -> ~1.5Gbit/s	02:15
mithro	aravind_: PCI-Express starts at about 1.5Gbit/s and goes up to 16Gbit/s (possible 32Gbit/s?)	02:16
aravind_	is this per lane speed? or overall?	02:16
mithro	aravind_: per lane	02:16
*** sb0 has joined #timvideos		02:16
aravind_	okay cool	02:16
aravind_	any post layout sims? to check sig integrity?	02:17
mithro	aravind_: So, at 1.5Gbit/s we do care about things like impedance matching and differential traces, etc	02:17
mithro	aravind_: Nope, we just kind of pray and hope most of the time :P	02:17
aravind_	hows the board stackup done?	02:17
aravind_	ha ha	02:17
mithro	aravind_: We have two existing TOFE boards - the "LowSpeedIO" and the "Milkymist Compat" boards - both operate in the 100kbit/s to 1Megabit/s type range where it doesn't really matter	02:18
aravind_	okay understood	02:19
mithro	aravind_: another cool advantage of being mostly compatible with the PCI-Express standard is that we can buy adapters and stuff intended for usage inside a computer and use them	02:19
aravind_	yes	02:19
aravind_	who will be end users?	02:19
mithro	aravind_: Things like this -> https://www.amazon.com/gp/product/B00KZIAHW8/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1 and https://www.amazon.com/Tenext-Express-Adapter-Flexible-Extension/dp/B00EY7BW0C/ref=pd_sim_sbs_147_1?ie=UTF8&dpID=41C0qqno%2B1L&dpSrc=sims&preST=_AC_UL160_SR160%2C160_&refRID=0QMEHFW8SHXMFZH1900D	02:20
CarlFK	me :)	02:20
mithro	aravind_: so to finally get to the point	02:21
aravind_	CarlFK :)	02:21
aravind_	okay	02:21
mithro	aravind_: the TOFE-VGA board takes the design used on the vmodvga and changes it to instead fit the TOFE connector + form factor	02:21
aravind_	okay	02:21
aravind_	mech specs are defined?	02:22
mithro	aravind_: yeah - PCI-Express + a couple of extra mounting holes	02:22
aravind_	great	02:22
aravind_	I saw in your google drive that these boards are manufactured with 2 vendors	02:22
mithro	aravind_: I've also been trying to build up things like templates	02:23
aravind_	Any specific rules from them?	02:23
mithro	aravind_: so it is easy for people to create new TOFE boards and such	02:23
aravind_	ah great	02:23
mithro	aravind_: The aim would be that boards we create should be producible by both hackvana and ohs park	02:24
aravind_	okay	02:24
mithro	aravind_: neither do controlled impedance stuff	02:24
aravind_	oh :(	02:24
mithro	aravind_: so again we just do our best	02:25
aravind_	then we should really pray	02:25
aravind_	:)	02:25
mithro	aravind_: The apertus/AXIOM guys (see Bertl_zZ and se6astian\|away) have gotten okay results with OHS Park using the rules I put in that document	02:25
aravind_	okay. I will follow the same rules	02:26
mithro	aravind_: yeah	02:26
*** sb0 has quit IRC		02:26
*** sb0 has joined #timvideos		02:26
mithro	aravind_: so, the current work on the board VGA board can be found at https://github.com/timvideos/HDMI2USB-TOFE-VGA	02:26
tpb	Title: GitHub - timvideos/HDMI2USB-TOFE-VGA: A VGA capture board for use with TOFE-8x compatible boards like the Numato Opsis. (at github.com)	02:26
mithro	aravind_: I'll download it this afternoon and check what the current status of it is	02:27
aravind_	i will checkout from git and have a look around.	02:27
aravind_	ah okay	02:27
aravind_	then i will checkout after you verify	02:27
mithro	aravind_: feel free to check it out and have a look at it too	02:27
aravind_	okay no prob.	02:27
aravind_	Any documentation?	02:27
mithro	aravind_: you can see stuff around the TOFE boards at https://github.com/timvideos/?utf8=%E2%9C%93&query=TOFE	02:28
tpb	Title: Tim Videos Conference Recording Systems · GitHub (at github.com)	02:28
mithro	aravind_: I'm collecting TOFE related documentation at https://tofe.io too	02:28
aravind_	mithro: thanks	02:28
aravind_	:) lots of things to have a look	02:28
mithro	aravind_: yeah, the TOFE-LowSpeedIO and TOFE-Milkymist boards where both done with quite an old version of KiCad	02:29
aravind_	do we have a repo of kicad libs?	02:29
aravind_	footprints and symbols?	02:29
mithro	https://github.com/timvideos/HDMI2USB-TOFE-kicad-template	02:29
tpb	Title: GitHub - timvideos/HDMI2USB-TOFE-kicad-template: KiCad template for creating TOFE compatible expansion boards. Fork the repository, rename it and create your own board! (at github.com)	02:29
aravind_	Thank you	02:30
mithro	aravind_: The original approach was you forked that repo and ran the script to start your own board	02:30
aravind_	ah okay	02:30
mithro	aravind_: however, that turned out to be really hard to merge when library parts where updated / fixed	02:31
aravind_	okay	02:31
mithro	aravind_: so I think I'm going to change to a git-submodule based approach	02:31
aravind_	okay	02:31
mithro	aravind_: but yeah, we are still figuring these things out	02:31
aravind_	i understand :)	02:31
aravind_	i have to do the same as well	02:32
mithro	aravind_: It's surprising how little people have already done in this area	02:32
aravind_	why is it like that?	02:32
aravind_	what could be the reason?	02:32
mithro	aravind_: I think lots of small reasons	02:33
aravind_	okay	02:33
mithro	aravind_: Most Electrical Engineers don't use git	02:33
aravind_	yea its the first time for me	02:33
mithro	aravind_: KiCad has only just gotten things like differential pairs, length matching, etc	02:33
aravind_	i have used SVN a lot	02:33
mithro	aravind_: Many EEs don't even use SVN :)	02:34
aravind_	we were forced to use SVN ;)	02:34
mithro	aravind_: The most common thing I've seen is folders with dates in them	02:34
aravind_	nowadays at work I use something custom developed	02:34
mithro	aravind_: GitHub actually allows you to access git repositories using SVN	02:34
aravind_	ic	02:34
mithro	aravind_: as someone who knows how that would work I find it really scarey :P	02:35
aravind_	ha ha	02:35
aravind_	its scares me everytime	02:35
mithro	aravind_: The board I have been working on most recently is the https://github.com/mithro/HDMI2USB-TOFE-AXIOM-Adapter	02:35
tpb	Title: GitHub - mithro/HDMI2USB-TOFE-AXIOM-adapter: Adapter which allows the expansion cards from the AXIOM be used on a TOFE compatible board. (at github.com)	02:35
aravind_	okay	02:35
mithro	aravind_: One thing KiCad lacks is the ability to set DRC for things like length matching	02:36
aravind_	yes. i saw it has an autorouter, but i dont trust them	02:36
mithro	aravind_: I doubt the autorouter does diffpairs	02:36
aravind_	i hope in future versions there will be DRC for these features and also we could add production specific rules	02:36
mithro	aravind_: Actually, I've never gotten it to work because it use to be this stupid web service thing	02:37
aravind_	but KiCad seems to be good for complex boards, seen some examples	02:37
mithro	aravind_: To work around this, projectgus wrote an external Python script which uses the KiCad scripting interface to do it	02:37
aravind_	mithro: apologies, time for me to go	02:38
mithro	aravind_: That is one great thing about KiCad is that you can script it :)	02:38
aravind_	its 3.30 am here and its post BREXIT :D	02:38
mithro	aravind_: Length matching checks -> https://github.com/mithro/kicad-length-matching-checks	02:38
tpb	Title: GitHub - mithro/kicad-length-matching-checks: Couple of scripts which checks length matching of your KiCad design. (at github.com)	02:38
mithro	aravind_: No worries	02:38
mithro	aravind_: Feel free to hang out in this IRC channel and chat with people	02:38
aravind_	I will go through the links you sent me	02:38
aravind_	and set up the repo + KiCad and scratch around	02:39
aravind_	okie dokie	02:39
aravind_	take care @mithro	02:39
aravind_	have a nice day	02:39
mithro	aravind_: Sorry about the pretty huge information dump - you can review the logs here -> https://logs.timvideos.us/%23timvideos/%23timvideos.2016-06-25.log.html	02:39
*** aravind_ has left #timvideos		02:39
mithro	CarlFK: so, you want to chat about something?	02:43
CarlFK	mithro: 2 things - 1 easy - I need cases this week, and the one you picked out is out of stock	02:44
CarlFK	I don't need rack mount... so what do you think of this one:	02:45
mithro	CarlFK: any ITX compatible case should work	02:45
CarlFK	Link: https://amzn.com/B01B575EMA	02:45
mithro	CarlFK: I even going to probably 3d print a full size backplate this weekend	02:45
CarlFK	k - i figured it was that easy	02:46
CarlFK	next.. have you ever written python to read data from a google spread sheet ?	02:46
mithro	CarlFK: I think you'll be shocked at how much bigger that case is then the one I used	02:46
CarlFK	it is?	02:46
CarlFK	Case Dimensions: 228.6mm x 195.6mm x 63.5mm, 9" x 7.7" x 2.5"	02:47
mithro	CarlFK: It will be at least twice the height probably about 33% wider?	02:47
CarlFK	Hmm.. I think I am ok with that. I hope to mount it on the back of a 22" lcd and stuff an odrod sbc in there too, so a little extra hight may work out fine	02:49
CarlFK	and maybe stuff the power supply inside the case too.	02:50
mithro	CarlFK: yeah, I'm not seeing anything else on amazon within a similar price range	02:50
CarlFK	super cool would be if I can figure out how to use the LCD's power supply, or at least nice power cord management	02:51
mithro	CarlFK: maybe something like https://smile.amazon.com/Silverstone-Aluminum-Mini-ITX-Center-PT13B/dp/B00DL4BMMC/ref=sr_1_8?s=pc&ie=UTF8&qid=1466822809&sr=1-8&keywords=VESA+ITX+case	02:51
*** sab_123 has joined #timvideos		02:56
mithro	Morning sab_123	02:57
sab_123	mithro, Hi	02:57
sab_123	mithro, pushed some changes	02:58
mithro	sab_123: Yes, I'll reply to your comments shortly	02:58
sab_123	mithro, once these 2 get merged in , I can push the timer_uart(I kind of have it working)	02:59
sab_123	mithro, then I can push the i2c stuff(I have gotten it working)	02:59
sab_123	with a 24C256	02:59
sab_123	mithro, then will push the FTDI_serial	02:59
mithro	sab_123: The USB loopback FTDI serial stuff should be a pull request independent of everything else	03:00
mithro	sab_123: But before you send that	03:00
mithro	sab_123: make sure you reread my email and go look at the code I linked	03:00
sab_123	the UVC stuff	03:01
sab_123	yep	03:01
mithro	sab_123: pay special attention to the things I talk about in the email	03:01
mithro	sab_123: I really don't want to have to repeat myself	03:02
mithro	CarlFK: a lot of "signage LCDs" have a 12V output on them	03:03
CarlFK	hmm	03:03
mithro	CarlFK: but I doubt you want to pay the premium those cost	03:04
CarlFK	yeah, I was guessing they wouldn't be cheep.	03:04
mithro	CarlFK: so, it sounds like you are building a combined "capture+confidence screen" thingy?	03:31
CarlFK	mithro: correct	03:31
mithro	CarlFK: What is your setup going to look like?	03:32
mithro	sab_123: Got a moment to discuss the uart stuff	03:33
sab_123	mithro, yes	03:33
CarlFK	mithro: um.. physically or .. cables? not sure what you mean ... oh, talk uart.. please :)	03:33
mithro	CarlFK: nothing to do with your uart stuff	03:33
mithro	CarlFK: What parts are you using?	03:34
mithro	CarlFK: LCD Monitor+Opsis+Odroid?	03:34
mithro	sab_123: Have you programmed much in OOP languages like Java, Python, C++, etc?	03:34
sab_123	mithro, Java yes	03:35
mithro	sab_123: so you understand the difference between a "class definition" and the "instance" of a class?	03:35
sab_123	yes	03:36
mithro	sab_123: Okay, so here we want to do something like OOP in C	03:36
sab_123	mithro, ok	03:37
mithro	sab_123: But C doesn't make this particularly easy :P	03:37
sab_123	mithro, ok	03:37
mithro	sab_123: In our case, we have a "Fast UART class" which can be connected to pretty much any pin in the FX2	03:38
mithro	sab_123: A specific instance of the Fast UART is connected to a specific pin	03:39
sab_123	But you can create objects to access the same functions	03:39
sab_123	cant	03:39
sab_123	SO there is only a couple of ways out	03:39
sab_123	Use those macros	03:39
mithro	sab_123: Stop	03:39
mithro	sab_123: we are discussing the high level at the moment, we'll get to the implimentation details in a minute	03:40
mithro	sab_123: So, we might have one (instance) of Fast UART on PA2 which we are using as a debug console , and another (instance) of Fast UART on PB4 which we are using to send commands to the FPGA	03:41
sab_123	mithro, ok	03:41
mithro	sab_123: this usage is going to be different on different boards	03:42
sab_123	mithro, ok	03:42
mithro	sab_123: IE on one board the debug console might be on PA2, on another it might be on PB4	03:42
mithro	sab_123: on many of the boards the setup is going to be pretty static (IE On the Opsis, the debug pin will always be X)	03:43
mithro	sab_123: while some boards like the little FX2 dev board you have can be much more dynamic	03:44
mithro	Generally however, this is not going to be dynamically changed at runtime - the pin usage is probably going to be decided while developing some type of firmware	03:45
mithro	sab_123: IE the UVC firmware might put the debug console on pin PA2 and the CDC firmware might put the debug console on PB5 - but it is probably not going to change while the firmware is running...	03:46
mithro	sab_123: Does that make sense?	03:46
sab_123	mithro, yes	03:46
mithro	sab_123: Lots of constraints + flexibility	03:47
mithro	sab_123: To make things even more complicated we have different types of UARTs, we have the fast transmit only, the timer/interrupt one, the hardware peripheral based one, etc	03:47
mithro	sab_123: in Java, we would define a simple UART interface and then each of the different types of UARTs would impliment the interface	03:48
mithro	sab_123: right?	03:49
sab_123	mithro, yes	03:49
mithro	sab_123: So if you were doing something like this on a "big computer", you might create a structure which had a bunch of function pointers	03:51
mithro	sab_123: The function pointers would be set so they matched the right implementation	03:51
mithro	sab_123: Using function pointers to make C OOP like is a fairly common pattern	03:52
mithro	sab_123: You can look up more details on how to do this later if you want	03:53
sab_123	mithro, ok	03:53
mithro	sab_123: There are a couple of downsides of this approach however	03:53
mithro	sab_123: The function pointers lead to a level of indirection that compiler can't really optimise around	03:54
mithro	sab_123: For example, It can be hard for it to do dead code removal, IE Remove the implementations you didn't end up using	03:55
mithro	sab_123: Pointers are also generally a very costly construct on small systems	03:55
mithro	sab_123: And you need quite a few of them (~one per function in your interface)	03:56
mithro	sab_123: So, this approach generally doesn't really work for something like the FX2 :(	03:56
mithro	sab_123: So we have to find a different approach	03:57
mithro	sab_123: does all that make sense so far?	03:57
sab_123	I have to read up on why function pointers are expensive, but I will do that later	03:58
sab_123	The rest make sense	03:58
mithro	sab_123: pointers in general are pretty expensive	03:58
mithro	sab_123: On "big" 32bit platforms, pointers are the same size as your data you normally operate on	03:59
mithro	sab_123: On the FX2, your data size is a byte - however pointers generally need to be 16bits	04:00
mithro	sab_123: 16bits != 8bits :(	04:00
sab_123	mithro, agreed. But lests say I want to pass a struct	04:00
sab_123	lets	04:00
sab_123	The struct is pretty huge	04:00
sab_123	Dont pointers help then?	04:00
mithro	sab_123: yes, a single pointer to a struct is going to be better then a large structure	04:01
mithro	sab_123: but you should be avoiding the large structure if possible	04:01
mithro	sab_123: BTW This doesn't apply if the structure is at a static location in memory	04:01
mithro	sab_123: because then the pointer is actually a constant	04:02
mithro	sab_123: Generally, on something like the FX2 you want everything statically, preallocated	04:03
mithro	sab_123: Most types of dynamism are really expensive on 8bit systems	04:04
mithro	sab_123: luckily, as I pointed out - we only really need dynamism at compile time, not run time	04:05
sab_123	mithro, ok	04:06
mithro	sab_123: so, what we end up with - is we want a system where we can say something like "Create a set of functions for using a UART on PA1"	04:09
sab_123	mithro, I still dont get how you plan to use it.	04:09
sab_123	mithro, can i talk about implementation now?	04:10
mithro	sab_123: So, I do something like "CREATE_SOFT_UART(<name>, <soft uart configuration>)" and then end up with a bunch of functions like "<name>_rx" and "<name>_init" which I can call in my code	04:11
mithro	I can also do something like	04:11
mithro	#ifdef BOARD_OPSIS; CREATE_HARD_UART(uart0, <config>); #elif BOARD_ATLYS; CREATE_SOFT_UART(uart0, <config>); #else .....	04:12
sab_123	mithro, so there must be different macros	04:12
sab_123	one for init	04:12
sab_123	one for the tx	04:12
mithro	As the CREATE_HARD_UART and CREATE_SOFT_UART both generate the same set of functions with the same signature, the code doesn't need to know which style of uart is actually in use	04:13
mithro	sab_123: no, there should be one macro which defines all the functions in the API	04:13
sab_123	The API I get	04:13
mithro	sab_123: CREATE_HARD_UART defines the complete set, _init, _rx, _tx, etc	04:13
sab_123	We can use one macro	04:13
sab_123	But the firmware needs to know which function is present	04:14
sab_123	lets say a macro dumps the contents into api.h	04:14
mithro	sab_123: what do you mean?	04:15
sab_123	This is what you are looking for right	04:15
sab_123	an autogenerated api.h	04:15
sab_123	Once we do this	04:15
sab_123	The uart_main.c	04:15
sab_123	needs to know what function to call in putchar(c)	04:15
sab_123	Is it uart0_tx(c)	04:15
sab_123	Is it uart1_tx(c)	04:15
mithro	sab_123: That is the firmware's job	04:16
sab_123	So if the firmware is setting the pins	04:16
sab_123	and calling the functions	04:16
sab_123	generated in api.h	04:16
mithro	sab_123: If the firmware might be using the UART for printf/gets	04:16
mithro	sab_123: Or it might be using it for something totally different	04:16
sab_123	SO basically you want to auto-generate the api.h file	04:17
sab_123	Am I correct?	04:17
sab_123	or off track?	04:17
mithro	sab_123: So the firmware will define the putchar/getchar to do what it prefers	04:17
mithro	sab_123: define "auto-generate"	04:17
sab_123	the uartx_init which we defined yesterday	04:18
sab_123	Should be made uart1_init	04:18
sab_123	uart2_init	04:18
sab_123	etc	04:18
sab_123	Just by calling the macro	04:18
mithro	sab_123: yes	04:19
sab_123	That is going to be rather hard to do	04:19
sab_123	What is going to be in the api.h file then?	04:19
sab_123	How do we tell the macro	04:19
sab_123	to put its contents into a api.h file	04:19
sab_123	Or will you call this macro	04:20
sab_123	from api.h	04:20
mithro	sab_123: api.h is purely for documentation	04:20
sab_123	so these function definitions	04:20
sab_123	will be in the example code	04:21
sab_123	and we need to implement uart0_init	04:21
sab_123	in soft_uart.c	04:21
mithro	sab_123: in soft_uart there will be a macro which creates a whole set of functions	04:21
mithro	sab_123: Which have the same "shape" as the functions define in api.h	04:22
mithro	sab_123: but all start with the given prefix name	04:22
sab_123	What about their implementations?	04:23
sab_123	Should these be hard coded	04:23
sab_123	or auto generated	04:24
mithro	sab_123: what do you mean?	04:24
sab_123	https://github.com/RacingTornado/fx2lib/blob/0803a717617b9909e568f9c9ea8c1e64de2954af/examples/uart_main/uart_main.c#L25	04:25
tpb	Title: fx2lib/uart_main.c at 0803a717617b9909e568f9c9ea8c1e64de2954af · RacingTornado/fx2lib · GitHub (at github.com)	04:25
mithro	sab_123: that code should be calling the macros - not defining them	04:26
sab_123	What I am asking is this	04:26
sab_123	Do you want the macro to just define the function name	04:26
sab_123	Like what I have done	04:26
sab_123	Or do you want the functionality also to be autogenerated?	04:27
sab_123	For examples	04:27
sab_123	in soft_uart.c	04:27
sab_123	we will still have functions called uart0_init()	04:27
sab_123	uart1_init()	04:28
sab_123	and uart2_init()	04:28
sab_123	assuming the user wont pass a pin number greater than 3	04:28
mithro	sab_123: No	04:28
mithro	sab_123: The user can pass any name for the uart	04:29
mithro	sab_123: and any pin	04:29
sab_123	and who implemets the functionality?	04:29
sab_123	The macro?	04:29
mithro	sab_123: So, there is no way you can just have every single combination in that file	04:29
sab_123	exactly	04:30
mithro	sab_123: Depending on how it gets implemented, the macro might generate all the code	04:30
sab_123	the entir fast_uart	04:30
mithro	sab_123: Or it might just generate wrapping functions which call internal functions	04:30
sab_123	in a macro	04:30
sab_123	wow	04:31
sab_123	okay	04:31
mithro	sab_123: yes, it might end up that way	04:31
sab_123	Thats not going to be very readble	04:31
mithro	sab_123: you might get away with using inline function plus constants	04:31
sab_123	I dont understand that	04:32
mithro	sab_123: There are a number of ways to make it readable	04:32
sab_123	An inline function is just like any other, except its body is pasted	04:32
sab_123	How would that help?	04:32
sab_123	I agree with the wrapper stuff	04:32
sab_123	Thats one way to proceed	04:32
sab_123	https://github.com/mithro/HDMI2USB-misoc-firmware/blob/fx2-refactor/firmware/fx2/uvc.c#L981	04:33
tpb	Title: HDMI2USB-misoc-firmware/uvc.c at fx2-refactor · mithro/HDMI2USB-misoc-firmware · GitHub (at github.com)	04:33
mithro	sab_123: You create an function like "inline fast_uart_rx(const BYTE pin_location, char c) {" then your macro does a "void NAME##_rx(char c) { fast_uart_rx(PINLOC, c); }"	04:34
sab_123	ah	04:35
sab_123	okay	04:35
mithro	sab_123: but that requires the compile to notice that PINLOC is actually a constant and optimise it out	04:35
sab_123	That way	04:35
mithro	s/compile/compiler/	04:35
sab_123	For the pinloc	04:36
sab_123	we need a case right	04:36
sab_123	there is no easier way to do it	04:36
mithro	sab_123: No	04:36
sab_123	macro?	04:36
sab_123	for the .equ?	04:37
mithro	sab_123: just make the person pass in the right values you need for things like the pin bit address	04:37
sab_123	and create the .equ in the macro itself?	04:38
sab_123	because you need to do clr name	04:38
sab_123	so name needs to point to PA2 or PA3	04:38
mithro	sab_123: yes	04:38
sab_123	okay	04:38
sab_123	I will try to get started on this	04:38
mithro	sab_123: Understanding now?	04:38
sab_123	Yes	04:38
sab_123	I havent done anything like this before	04:39
sab_123	So its quite interesting	04:39
sab_123	But I am pretty sure I will run across a lot of problems.	04:39
mithro	sab_123: So - you can start by ignoring the "genericisation / dynasim" by hard coding the name to be "uart0_" and the pin names just #define at the top of the file	04:39
sab_123	ok	04:40
sab_123	Is there any repo where you have something like this done(just for reference)	04:40
mithro	sab_123: once you have that working, you just need to figure out a way to replace the "uart0_" and pin names	04:40
CarlFK	mithro: confidence monitor/capture : 1 short hdmi going to the podium, power, cat5 and hdmi away to me and projector .	04:42
mithro	CarlFK: hrm?	04:43
CarlFK	mithro: you were asking what my setup was going to look like	04:43
mithro	CarlFK: I think you mean, 1 short HDMI coming from the presenters laptop on the podium -- long power, cat5 running to "carl" and one long HDMI cable running to the projector ?	04:44
CarlFK	right	04:45
mithro	CarlFK: your not running HDMI to you right? the HDMI for the presenters laptop goes directly into the Opsis on the back of the monitor?	04:45
CarlFK	correct	04:46
mithro	CarlFK: great, you almost gave me a heart attack with the "power, cat5 and hdmi away to me" bit ;)	04:46
CarlFK	lol	04:46
mithro	CarlFK: Couple of things I have been thinking about	04:47
mithro	CarlFK: Power over Ethernet is generally 48V - it wouldn't take much to do a little board which has ethernet pass through and extracts the power needed for the Opsis	04:48
mithro	CarlFK: I'm pretty sure that we'll get capture over Ethernet by the end of the year too	04:48
mithro	CarlFK: but that might not make sense if you need to run power for the monitor anyway....	04:49
mithro	CarlFK: I assume you are just getting cheap PC monitors?	04:50
CarlFK	correct	04:50
CarlFK	$100 ish	04:50
mithro	CarlFK: do they use a power brick? Or just take a kettle plug?	04:52
mithro	CarlFK: If you could find ones which use a 12V power brick, there are some interesting possibilities :P	04:52
CarlFK	mithro: I haven't picked one out yet. if brick, then it would be nice if I could split that	04:53
CarlFK	yeah	04:53
mithro	CarlFK: yeah	04:53
mithro	LCD monitors are surprisingly low wattage these days	04:53
mithro	CarlFK: I was looking at it the other day and thought they had the decimal in the wrong location...	04:54
CarlFK	heh	04:54
CarlFK	mithro: modeswitch - is this the repo I should be using https://github.com/mithro/HDMI2USB-mode-switch.git	04:56
tpb	Title: GitHub - mithro/HDMI2USB-mode-switch: Tool for switching boards supported by HDMI2USB firmware between multiple different modes (programming, webcam, etc). (at github.com)	04:56
mithro	CarlFK: it was like 8 Watts or something	04:56
mithro	CarlFK: you can use the upstream one now	04:56
mithro	CarlFK: I was sure it was suppose to be 80W	04:56
CarlFK	mithro: yeah.. i guess 80w light bulb is mostly heat	04:57
mithro	CarlFK: Will merge the latest commits upstream now	04:58
CarlFK	mithro: upstream has less docs	04:58
CarlFK	lol	04:58
mithro	CarlFK: I was sure I had pushed them....	04:58
mithro	CarlFK: just making sure the tests pass before hand :)	04:59
mithro	CarlFK: merged	05:02
sab_123	mithro, Will get started then. Thanks for the help	05:03
*** sb0 has quit IRC		05:04
mithro	sab_123: I'd like to merge https://github.com/mithro/fx2lib/pull/17 but there are still a couple of small outstanding things	05:07
tpb	Title: Created a UART API definition for allowing proper abstraction by RacingTornado · Pull Request #17 · mithro/fx2lib · GitHub (at github.com)	05:07
sab_123	mithro, ok	05:08
*** mobydikc has joined #timvideos		05:10
mobydikc	CarlFK: does prod1.sh do the hdmi capture?	05:11
CarlFK	mithro do the board have real eproms? like where is the firmware getting loaded when it doesn't go away on power cycle?	05:11
*** mobydikc has left #timvideos		05:12
CarlFK	mobydikc: um.. where is that?	05:12
CarlFK	lol	05:12
mithro	CarlFK: The Atlys has SPI flash (for the FPGA gateware) and an EEPROM (for the FX2 firmware)	05:12
mithro	CarlFK: The Opsis has SPI flash (for the FPGA gateware + FX2 firmware) and a tiny EEPROM which just stores the MAC address and code to load the FX2 firmware from the FPGA	05:13
CarlFK	ah, got it.	05:13
CarlFK	how do I load v0.0.0-639-g956a8b3 into the Atlys spi/eeporm? (and do you call that flashing or burning or what?	05:14
mithro	CarlFK: which is why the Opsis can change how the FX2 behaves	05:14
mithro	CarlFK: I'm happy to flash the Atlys's SPI	05:15
mithro	CarlFK: but I've never been game to flash the Atlys's EEPROM	05:15
mithro	CarlFK: If you screw the second up, it is a bit tricky to fix	05:15
mithro	CarlFK: I think we should probably use the "burn" terminology because flashing is so overloaded?	05:16
CarlFK	mithro: agreed	05:16
mithro	Now I have to remeber what I was actually going to do today....	05:17
CarlFK	so how do I use modeswitch ?	05:17
mithro	CarlFK: In theory or in practice?	05:17
CarlFK	practice	05:17
CarlFK	I want to setup my atlys so it powers up in a functional state	05:17
CarlFK	make conda ... did that.	05:18
mithro	CarlFK: "make conda" sets up a self contained development environment, probably not what you want in "production" but lets start with that	05:21
mithro	CarlFK: Basic operation is "hdmi2usb-mode-switch --mode=XXXX"	05:23
CarlFK	um.. it isn't on my path	05:23
mithro	CarlFK: it won't be unless you install it	05:23
mithro	CarlFK: set your path to PATH=$PWD/conda/bin:$PATH	05:24
mithro	or run ./conda/bin/hdmi2usb-mode-switch	05:26
mithro	I'm hoping someone at debconf will put it in a deb package	05:26
*** mobydikc has joined #timvideos		05:28
*** sab_123 has quit IRC		05:28
CarlFK	mithro: first i want : --mode jtag ... right?	05:28
mobydikc	CarlFK: I think I missed that. Set mode in what?	05:28
CarlFK	mobydikc: um.. there are 2 hdmi captures - laptop/projector and camera	05:29
mobydikc	I assume grabber is laptop	05:29
mobydikc	I have cam1 on injgest now	05:29
mobydikc	How do I make it the camera?	05:30
CarlFK	https://github.com/CarlFK/dvsmon/blob/master/vocto-prod1.py#L13 ingest --video-source blackmagichdmi --audio-source blackmagichdmi	05:30
tpb	Title: dvsmon/vocto-prod1.py at master · CarlFK/dvsmon · GitHub (at github.com)	05:30
CarlFK	I think thats what you want?	05:30
mobydikc	I'm using the PC and camera you used at nodepdx. I have the camera's HDMI out going to the PC's D+HDMI capture card	05:31
mobydikc	I shouldn't need to download any code to recreate what you had running on this machine last weekend	05:31
CarlFK	correct	05:32
mithro	CarlFK: It depends on what you are trying to do	05:32
CarlFK	mobydikc: im just showing you where the commands are stored	05:32
CarlFK	mithro: I want to burn v0.0.0-639-g956a8b3 to Atlys spi	05:33
mithro	CarlFK: If you want to run openocd commands against your board "--mode=jtag", such as using the make commands in HDMI2USB-misoc-firmware, then that is what you want	05:33
mithro	CarlFK: I haven't tested it recently, but you should in theory be able to just do "hdmi2usb-mode-switch --flash-gateware=XXXXX.bin"	05:34
mithro	--load-gateware LOAD_GATEWARE	05:34
mithro	Load gateware onto the FPGA.	05:34
mithro	--flash-gateware FLASH_GATEWARE	05:34
mithro	Flash gateware onto the SPI flash which the FPGA boots	05:34
mithro	from.	05:34
CarlFK	mithro: is XXXX the full path and file name?	05:34
mithro	CarlFK: yes	05:35
CarlFK	hmm, lets forget about burn and stick with load and flash so that we don't have to change the help	05:35
mithro	CarlFK: in theory, it should find an Atlys board in any mode, switch to jtag mode, and call openocd correctly	05:36
mobydikc	CarlFK: my irc thingy reset, so I'm not sure what exact commands you said I need	05:36
mobydikc	If I want cam1 to be the actual camera should I still run ingest?	05:36
mithro	CarlFK: that initial switch bit might be broken....	05:36
mithro	CarlFK: if you add "--verbose" you can see the commands it is running	05:37
CarlFK	mobydikc: ingest is what feeds both sources: cam and grabber	05:37
mobydikc	ok, so I get the bouncy ball when I need camera	05:38
CarlFK	mithro: ^^^ mike has all my equipment from NodePDX and is trying to set it up.. yay!	05:38
mithro	CarlFK: If, hdmi2usb-mode-switch --flash-gateware=XXXX.bin doesn't work we should fix it so that it does	05:38
CarlFK	mobydikc: bouncy ball is the default video. --video-source blackmagichdmi	05:38
mobydikc	do I adjust that in dvs-mon?	05:38
CarlFK	mobydikc: today, no. because today you need the special gst libs... so ... in a new term:	05:39
CarlFK	gst-master	05:39
mobydikc	Ok, that's where I can do set mode -i	05:39
mobydikc	?	05:40
CarlFK	umm. what is set mode -i	05:40
mobydikc	i thought that's the last thing you told me before I rebooted. OK forget that	05:41
mobydikc	I did gst-master	05:41
mobydikc	in the shell	05:41
CarlFK	mobydikc: cd voctomix-outcasts	05:41
CarlFK	./ingest.py --video-source blackmagichdmi --audio-source blackmagichdmi	05:41
mobydikc	do I run that while dvs-mon is running (so so so sorry for dumb questions)	05:42
CarlFK	well... ingest sends a stream to the vocto core (the mixer)	05:43
CarlFK	it needs to be running. either open a shell and run it, or use dvsmon to run it	05:44
mithro	CarlFK: so......	05:45
CarlFK	mithro: PermissionError: [Errno 13] Permission denied: '/sys/bus/usb/devices/1-1.5:1.0/driver/unbind	05:45
CarlFK	mithro: for testing, should I just sudo, or put the udeve rules somewhere ?	05:46
mithro	CarlFK: you need to "make unbind-helper" or run as root	05:46
mobydikc	i'm rebooting. what's the optimal order to start dvs-mon and run that blackmagic command?	05:46
CarlFK	I'll start with runing as root.	05:46
mithro	CarlFK: the unbind helper lets me disconnect linux kernel drivers from the device without being root	05:47
mithro	CarlFK: I recommend the unbind-helper	05:47
mithro	CarlFK: because that is what I test with	05:47
CarlFK	mobydikc: dvs-mon, click the run for voctocore and vocto gui. then open a new shell, gst-master, cd ~/voctomix-outcasts; ./ingest.py --video-source blackmagichdmi --audio-source blackmagichdmi	05:48
CarlFK	unbind-helper: command not found	05:48
CarlFK	unbind-helper.c ?	05:48
mithro	CarlFK: Did you do a "make unbind-helper" in the directory?	05:48
CarlFK	mithro: doing that now.	05:49
CarlFK	mithro: sudo chmod 755 bin/unbind-helper	05:49
mithro	CarlFK: what?	05:50
CarlFK	why does it need sudo for that ?	05:50
CarlFK	mithro: http://paste.ubuntu.com/17841465/	05:51
tpb	Title: Ubuntu Pastebin (at paste.ubuntu.com)	05:51
mithro	CarlFK: I guess it doesn't for that line - it does for the following two	05:51
mithro	CarlFK: the unbind-helper tool is a setuid binary running as root	05:52
CarlFK	mithro: bin needs to be added to the path too	05:54
mithro	CarlFK: hrm?	05:55
CarlFK	mithro: I am adding notes to the readme, like PATH=$PWD/conda/bin:$PWD/bin:$PATH	05:55
mithro	CarlFK: I'm pretty sure the code should find the unbind-helper without you adding it to the bin	05:56
CarlFK	mithro: oh, I thought you ran it	05:56
mithro	CarlFK: no	05:56
mithro	https://github.com/timvideos/HDMI2USB-mode-switch/blob/master/hdmi2usb/modeswitch/lsusb.py#L21	05:56
tpb	Title: HDMI2USB-mode-switch/lsusb.py at master · timvideos/HDMI2USB-mode-switch · GitHub (at github.com)	05:56
CarlFK	mithro: Write to /sys/bus/usb/drivers/uvcvideo/unbind failed http://paste.ubuntu.com/17841615/	05:58
CarlFK		05:58
tpb	Title: Ubuntu Pastebin (at paste.ubuntu.com)	05:58
mithro	CarlFK: does "/sys/bus/usb/devices/1-1.5:1.0/driver/unbind" exist?	06:00
mithro	CarlFK: and can you do a "ls -l bin/unbind-helper" ?	06:00
CarlFK	--w------- 1 root root 4096 Jun 25 00:37 /sys/bus/usb/devices/1-1.5:1.0/driver/unbind	06:00
CarlFK	-rwxr-xr-x 1 juser juser 9056 Jun 25 00:46 bin/unbind-helper	06:00
mithro	CarlFK: did you cancel the "make unbind-helper" command?	06:01
*** Bertl_zZ is now known as Bertl		06:01
CarlFK	mithro: maybe. I hit enter until the sudo failed, then $ chmod 755 bin/unbind-helper	06:01
CarlFK	then re-ran make	06:02
mithro	CarlFK: so you do see that there are two very important commands after that one in the makefile?	06:02
*** sb0 has joined #timvideos		06:04
CarlFK	mithro: makefiles that call sudo is... not going to go over well with the debian folk	06:05
mithro	CarlFK: They wouldn't be using the Makefile	06:05
mithro	CarlFK: they would be doing the equivalent in the debian/rules file and the binary would be installed as being owned by root when the package is installed	06:05
CarlFK	well.. I suspect the way you have this setup is going to make it hard for them. but I am not sure what would be better ... so ... never mind.	06:08
CarlFK	-rwsr-xr-x 1 root root 9056 Jun 25 00:46 bin/unbind-helper	06:09
mithro	CarlFK: The makefile sets up a developer environment - they are going to be doing a "python setup.py install" like any other python module	06:09
CarlFK	mithro: Running 'fxload -t fx2lp -D /dev/bus/usb/001/011 -I /home/juser/HDMI2USB-mode-switch/hdmi2usb/firmware/fx2/atlys/ixo-usb-jtag.hex -v'	06:11
CarlFK	b'Permission denied : /dev/bus/usb/001/011\n' http://paste.ubuntu.com/17841879/	06:11
tpb	Title: Ubuntu Pastebin (at paste.ubuntu.com)	06:11
mithro	CarlFK: they'll probably want to rewrite/rework the unbind helper anyway	06:11
mithro	CarlFK: oh... That would be fixed by the udev rules	06:12
mithro	Hrm, I wonder if the udev rules could also fix the unbind permissions....	06:13
mithro	CarlFK: okay, maybe you should be running this as root if you haven't installed the udev rules	06:14
CarlFK	LOL sudo: hdmi2usb-mode-switch: command not found	06:14
CarlFK	it does't get the users path	06:15
mithro	CarlFK: yeah	06:15
CarlFK	mithro: this installs, right? sudo setup.py install	06:16
mithro	CarlFK: Yes, but I've never tested that :P	06:16
CarlFK	lol	06:16
mithro	CarlFK: I only have so much time	06:16
CarlFK	from setuptools import setup	06:16
CarlFK	ImportError: No module named 'setuptools'	06:16
CarlFK	ok, um.. let me sudo -s and set the path	06:17
mithro	sudo apt-get python3-setuptools I think	06:17
CarlFK	too late.	06:18
CarlFK	DNA = 110000001100000011000000110000001100000011000000110000001 (0x181818181818181)	06:18
CarlFK	I just re-run until it works, right ?	06:18
mithro	CarlFK: Yeah - for the moment :(	06:19
mithro	CarlFK: actually you might need to force it into serial mode	06:19
CarlFK	mithro: how do I do that ?	06:20
mithro	CarlFK: hdmi2usb-mode-switch --mode=serial	06:20
mithro	Oh, wait you are on an Atlys....	06:20
CarlFK	AssertionError: serial mode only valid on the opsis.	06:20
CarlFK	lol	06:20
mithro	CarlFK: Just try again maybe?	06:20
mithro	CarlFK: basically, the different FX2 firmware we are using fight with each other about how the FX2 should be configured	06:21
CarlFK	mithro: iv tried over 5 times. normally it gets it by the third when I was doing whatever commands I was doing for lca	06:21
mithro	CarlFK: What were you doing for LCA?	06:22
mithro	CarlFK: At some point we need to fix the FX2 firmware to actually properly set up the FX2 so we don't get those failures.	06:22
CarlFK	mithro: er, not lca. what someone at DC figured out and I dropped it in the lca repo:	06:24
CarlFK	https://github.com/xfxf/video-scripts/blob/master/carl/opsis_setup.sh#L29 # repeat these 3 lines until it works	06:24
tpb	Title: video-scripts/opsis_setup.sh at master · xfxf/video-scripts · GitHub (at github.com)	06:24
mithro	CarlFK: sure	06:24
mithro	CarlFK: those are opsis commands....	06:25
CarlFK	ah right. I haven't done this to an atlys	06:25
mithro	CarlFK: okay	06:26
mithro	CarlFK: can you try power cycling the Atlys and running the command a few times?	06:26
CarlFK	power cycled, re-ran, worked the first time ... well, new error	06:27
CarlFK	DNA = 100110010110101100010110111000010010100001101000001001011 (0x132d62dc250d04b)	06:28
CarlFK	loaded file /home/juser/HDMI2USB-mode-switch/hdmi2usb/firmware/spartan6/atlys/bscan_spi_xc6slx45.bit to pld device 0 in 4s 414024us	06:28
CarlFK	Info : JTAG tap: xc6s.tap tap/device found: 0x34008093 (mfg: 0x049 (Xilinx), part: 0x4008, ver: 0x3)	06:28
CarlFK	Error: Unknown flash device (ID 0x00ffffff)	06:28
mithro	CarlFK: there is one peculiarity with the Atlys if it is still running the stock firmware from Digilent	06:28
CarlFK	http://paste.ubuntu.com/17842244/	06:28
tpb	Title: Ubuntu Pastebin (at paste.ubuntu.com)	06:28
mithro	CarlFK: This board still has the stock Digilent firmware on it, right?	06:29
CarlFK	mithro: correct	06:29
mithro	CarlFK: let me dig up the special thing you need to do the first time	06:30
mithro	Dammit, where did I put it....	06:32
mithro	Neuron1k: we are running a little bit, will ping when we are ready	06:34
Neuron1k	mithro: ok	06:35
mithro	CarlFK: http://paginas.fe.up.pt/~ee06211/doku.php?id=digilent_atlys#interfacing_with_the_onboard_numonyx_n25q128	06:36
tpb	Title: digilent_atlys [Generation of Reconfigurable Circuits from Machine Code] (at paginas.fe.up.pt)	06:36
mithro	CarlFK: I'm sure I wrote better instructions down somewhere....	06:37
mithro	Neuron1k: In your blog you mentioned hacking on a model	06:37
mithro	Neuron1k: can you publish that code somewhere	06:37
Neuron1k	mithro: It is added on my github	06:38
mithro	Neuron1k: can you link it here	06:38
Neuron1k	https://github.com/Neuron1k/litejpeg/tree/master/test/model	06:38
tpb	Title: litejpeg/test/model at master · Neuron1k/litejpeg · GitHub (at github.com)	06:38
CarlFK	mithro: this? " The workaround is to set the J11 jumper on the board to avoid power on flash boot, thus leaving the flash in legacy SPI mode."	06:39
mithro	CarlFK: Yes	06:39
mithro	CarlFK: I think you need to remove the jumper after the board has booted while the power is still connected	06:41
mithro	CarlFK: Please write this down somewhere in the Wiki	06:43
CarlFK	mithro: k	06:44
CarlFK	mithro: is j11 next to the usb-A jack, says "HOST" and in the picture here it doesn't look like there are header pins to jump?	06:45
CarlFK	https://hdmi2usb.tv/img/digilent-atlys.jpg	06:45
mithro	CarlFK: it should definitely be around that big USB HOST connector	06:45
mithro	CarlFK: JP9 seems to be on one side and JP11 on the other	06:46
CarlFK	mithro: so .. jumper off, power, jumper on, hdmi2usb-mode-switch --flash-gateware... ?	06:47
mithro	CarlFK: no	06:47
mithro	CarlFK: jumper on, power on, jumper off, hdmi2usb-mode-switch --flash-gateware	06:47
CarlFK	mithro: welll ... it didn't error	06:56
CarlFK	mithro: I was expecting the hdmi2usb colorbars ,	06:57
mithro	CarlFK: Does jp11 on your board have a header? Or are you just shorting it with a screwdriver?	06:59
CarlFK	mithro: header	07:00
mithro	Neuron1k: heyo	07:03
mithro	Neuron1k: sorry about running late, we are now getting setup	07:03
Neuron1k	mithro: ok, I will join the call	07:04
*** panther has joined #timvideos		07:04
mithro	CarlFK: I might be back in a bit, otherwise I'll respond on my phone in 30 minutes	07:05
CarlFK	mithro: k	07:05
shenki	Neuron1k: are you on?	07:08
Neuron1k	shenki: yes	07:08
shenki	Neuron1k: ok. you keep dropping and coming back	07:09
Neuron1k	shenki: I can't see you	07:10
*** sb0 has quit IRC		07:10
shenki	Neuron1k: okay. we can't see you either	07:11
shenki	Neuron1k: you've joined, but no video nor voice	07:11
shenki	Neuron1k: do you want to try inviting us to a new hangout?	07:11
Neuron1k	shenki: ok, I will drop and could you invite me instead?	07:11
shenki	ok	07:11
mithro	Neuron1k: have you tried restarting your browser?	07:12
*** panther is now known as sab_123		07:13
*** aravind_ has joined #timvideos		07:23
*** aravind_ has quit IRC		07:28
mithro	CarlFK: did ti work?	07:28
mithro	CarlFK: Have to run, will be back tomorrow	07:32
mithro	CarlFK: I'll push some updates to the udev rules which fix the permissions on unbind tomorrow	07:33
CarlFK	k	07:33
CarlFK	mithro: isn't there another step?	07:33
CarlFK	mobydikc: you still around? I'm heading to bed.	07:35
mithro	CarlFK: what do you mean?	07:35
CarlFK	mithro: don't i need to send hdmi2usb.hex ?	07:36
mithro	As I said, I've never been game enough to flash the ERPROM on an Atlys	07:37
mithro	Which is where the .hex file goes	07:40
CarlFK	mithro: what does this do: fxload -B libusb -D vid=0x16c0,pid=0x06ad -t fx2lp -I hdmi2usb.hex	07:40
mithro	Loads it into ram	07:41
mithro	That doesn't persist after power cycle	07:42
mithro	Btw where does your fxload come from?	07:43
CarlFK	k - so I either need to do that, or hdmi2usb-mode-switch --load-fx2-firmware	07:44
mithro	Yes	07:44
CarlFK	root@cnt1:~/HDMI2USB-mode-switch# apt-cache policy fxload	07:44
CarlFK	500 http://archive.ubuntu.com/ubuntu xenial/universe amd64 Packages	07:44
tpb	Title: Index of /ubuntu (at archive.ubuntu.com)	07:44
CarlFK	hmm, not sure it works	07:44
mithro	CarlFK: the mode switch tool will aways get the right vid:pid matter what mode the Atlys is in	07:45
CarlFK	hdmi2usb-mode-switch --load-fx2-firmware /home/juser/HDMI2USB-firmware-prebuilt/archive/master/v0.0.0-639-g956a8b3/atlys/hdmi2usb/hdmi2usb.hex	07:46
CarlFK	no error, but still not colorbars	07:46
mithro	CarlFK: I haven't tested --load-fx-firmware recently but as the system has to load fx2 firmware to do the other operations, it would be weird if broken.	07:46
mithro	CarlFK: you should get color bars after the gateware	07:47
mithro	Oh, you need to power cycle after a flash command	07:47
CarlFK	where is your script to bounce the usb so I can connect a termial	07:47
CarlFK	ah, right	07:47
mithro	You should get the colour bars from when you turn the board on now	07:48
CarlFK	yep!	07:48
CarlFK	flterm --port /dev/ttyVIZ0 --speed 115200	07:49
CarlFK	git describe: v0.0.0-639-g956a8b3-dirty	07:49
CarlFK	mithro: almost 3am, I an go to bed now.	07:50
mithro	CarlFK: great you flashed the right file then	07:53
mithro	You flash the .bin, load the .bit	07:53
mithro	Logged a bug to make the tool check that.	07:54
CarlFK	mithro: yep. I am up for trying to flash the eprom, but not now. to sleepy.	07:54
mithro	CarlFK: I should do it first	07:54
mithro	As trying to recover if we screw it up is to hard to do over the net	07:55
CarlFK	mithro: I would like to do it in the next day or two. that gives me enough time to un-brick or get another one	07:55
mithro	And requires physical access to the board	07:55
mithro	I'll be around tomorrow	07:55
mithro	Probably from around midday again	07:55
CarlFK	sounds good	07:56
mithro	Anyway, my phone is almost flat	07:56
CarlFK	I am going to re-image the machine I was running everything from, we can try doing a normal install and see how that goes	07:56
CarlFK	thanks for your time	07:56
CarlFK	see ya	07:56
*** sb0 has joined #timvideos		08:05
*** cr1901_modern has quit IRC		08:21
*** cr1901_modern has joined #timvideos		08:39
seaLne	my fork of the tofe-vga has the schematic i'd done in it, never felt confident to suggest merging it https://github.com/sealne/HDMI2USB-TOFE-VGA	09:42
tpb	Title: GitHub - sealne/HDMI2USB-TOFE-VGA: A VGA capture board for use with TOFE-8x compatible boards like the Numato Opsis. (at github.com)	09:42
*** sab_123 has quit IRC		10:05
*** ssk1328 has joined #timvideos		10:29
*** sb0 has quit IRC		11:09
*** Bertl is now known as Bertl_oO		11:45
*** shenki has quit IRC		12:49
*** shenki has joined #timvideos		13:06
*** ChanServ sets mode: +v shenki		13:06
*** ssk1328_ has joined #timvideos		13:49
*** mithro_ has joined #timvideos		13:50
*** ChanServ sets mode: +o mithro_		13:50
*** Joelw_ has joined #timvideos		13:51
*** Niharika has quit IRC		13:54
*** ssk1328 has quit IRC		13:54
*** Joelw has quit IRC		13:54
*** _florent_ has quit IRC		13:54
*** mobydikc has quit IRC		13:54
*** mithro has quit IRC		13:54
*** ssk1328_ is now known as ssk1328		13:54
*** mobydikc has joined #timvideos		13:54
*** Niharika has joined #timvideos		13:55
*** _florent_ has joined #timvideos		14:00
*** mithro_ is now known as mithro		14:04
CarlFK	mithro: ping	15:29
*** ssk1328 has quit IRC		15:34
CarlFK	Heading to lug meeting. back in a few hours. bye!	19:23
*** CarlFK has quit IRC		19:23
*** CarlFK has joined #timvideos		19:50
*** ChanServ sets mode: +v CarlFK		19:50
*** puck has quit IRC		20:16
*** puck has joined #timvideos		21:00
*** Bertl_oO is now known as Bertl_zZ		22:11
*** CarlFK has quit IRC		22:11
*** CarlFK has joined #timvideos		23:39
*** ChanServ sets mode: +v CarlFK		23:39
mobydikc	CarlFK: heyo. I ran dvsmon, started voctocore and voctogui. then went to the terminal and did the ingest blackmagic thingy	23:41
mobydikc	I got a no property "mode" in decklinkvideosrc0 or something like that	23:41
mobydikc	Also dvsmon has a gmi.sh and an ingest --video-source blackmagichdmi. It's looks kind of like they do the same thing	23:42
CarlFK	mobydikc: hmm, can you put that box on the net and do a git pull?	23:42
CarlFK	dvsmon and gmi.sh stuff ... I couldn't get it to use the gst-master libs	23:43
mobydikc	probably. I would have thought it'd be ready to go since it worked in portland tho	23:43
mobydikc	does the "ingest" command in dvsmon need to be running?	23:43
CarlFK	yeah, im a little fuzzy.. but may as well stay current with what I jsut did this morning	23:43
CarlFK	um..	23:44
CarlFK	voctocore is the only thing that needs to be running - it is the server that everything else connects to	23:44
CarlFK	ingest is: read from something, send to core	23:45
mobydikc	is there something on the camera that needs to be set?	23:45
mobydikc	it's got a mode button three lights, CAM, PB, PC, and cam is lit up	23:46
CarlFK	if you have an hdmi monitor you can plug the cam into, then you can see if it is sending something out the hdmi port	23:46
CarlFK	mobydikc: im out in about 10 min	23:58
CarlFK	then back in a few hours	23:59

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