Thursday, 2016-07-21

*** tpb has joined #timvideos00:00
*** sb0 has joined #timvideos01:18
*** nueces has joined #timvideos02:03
*** sab_123 has joined #timvideos03:58
*** nueces has quit IRC04:09
*** rohitksingh_work has joined #timvideos04:15
mithrosab_123: Formatting requirements are not the only thing you should be checking on the pull request05:22
mithrosab_123: But they are required05:23
*** ssk1328 has joined #timvideos05:27
ssk1328mithro: The pull request looks okay05:28
sab_123mithro, ok05:28
ssk1328mithro: I merged my float16 conversion branch csc-update with this branch05:28
mithrosab_123: What do you think of  splitting the fx2lib examples into their own repositories? IE Maybe like each having fx2lib-firmware-ftdi  fx2lib-firmware-cdc fx2lib-firmware-audio ....06:35
mithrossk1328: would that make things easier / harder for you?06:36
sab_123mithro, Sure. Sounds good.06:36
ssk1328mithro: No problem06:37
mithrosab_123: I'd like you to think about it and have some reasoning why you think it is good / bad06:37
mithrossk1328: opps that was suppose to be "sab_123: would that make things easier / harder for you?"06:37
mithrosab_123: I didn't mean via email - chatting via IRC would be fine06:48
sab_123mithro, ah okay06:49
sab_123mithro, do you want to discuss it now?06:55
mithrosab_123: The indent tool is making the code formatted a certain way, right?08:16
mithrosab_123: so if there is a diff, then the code has not been formatted properly08:16
sab_123mithro, We are not sure of that. Many times I ran it through indent but you still found some issues08:17
sab_123like wrapping, new lines etc08:17
mithrosab_123: s/indent/an indent tool/08:17
mithrosab_123: that is what I meant by "evaluate and figure out which tools are the right choice"08:18
sab_123mithro, okay08:18
mithrosab_123: if the tool is letting through code which isn't formatted correctly, then that is bad08:18
mithrosab_123: Checking formatting is a hard thing to do, so some tools might miss things08:19
sab_123Well its not letting code without formatting. Its just what we expect is different08:19
mithrosab_123: Its why generally in python world we use *both* pep8 and pylint08:19
mithrosab_123: they catch different set of problems08:20
sab_123mithro, okay08:20
sab_123mithro, I just have 1 question08:21
tpbTitle: Implemented UART in accordance with the API definition in softuart.h to allow the example code to access printf() functionality by RacingTornado · Pull Request #18 · mithro/fx2lib · GitHub (at
sab_123mithro, why do you say that?08:21
mithrosab_123: because just from inspecting the code I can see a whole bunch of bugs / issues08:22
sab_123mithro, uart_api.h needs to merged still. So I havent edited the header file08:23
sab_123mithro, only soft_uart.h and uart_main.c are the 2 main files08:23
mithrosab_123: I don't understand what you mean08:24
sab_123mithro, That pull request ideally is only 2 files08:24
sab_123I dont see anything in uart_main.c apart from a printf and configuration of UART08:25
sab_123mithro, so the only file left where there could possibly be bugs is the soft_uart.h08:25
mithrosab_123: You can create a branch which only has the files, then create a separate branch where you do a merge of that plus the dependencies08:27
sab_123mithro, okay08:27
mithrosab_123: Do you want me to review #17 again?08:30
sab_123mithro, not right now08:30
sab_123I really havent changed much08:30
mithroThere is a comment on that "I have changed things to the best of my knowledge."08:30
tpbTitle: Created a UART API definition by RacingTornado · Pull Request #17 · mithro/fx2lib · GitHub (at
sab_123Ah thats 1708:31
sab_123Well yes, thats just the API08:31
mithrosab_123: and you have performed all the checks in your list08:32
sab_123well I was working on the linting tools, but from what I understand this was already okay08:33
tpbTitle: Created a UART API definition by RacingTornado · Pull Request #17 · mithro/fx2lib · GitHub (at
mithrosab_123: As mentioned multiple times, you should check every pull request for all the things in your check list08:35
sab_123I am in the process of doing that line by line08:35
sab_123using the python scripts08:35
sab_123and write to an excel file and upload it08:35
mithrosab_123: You should also be manually checking08:36
mithrosab_123: until you are 100% confident that automation is working, even then you should do a manual check08:36
sab_123mithro, ok08:36
mithrosab_123: It is like 100 lines, it should be pretty easy to check manually too08:37
mithroI'm going to dinner08:37
mithrobe back in an hour08:37
sab_123mithro, okay08:37
*** Bertl_zZ is now known as Bertl09:23
mithroxfxf: ping?10:30
*** ssk1328 has quit IRC10:45
*** sab_123 has quit IRC11:31
*** sab_123 has joined #timvideos11:31
*** danielki has joined #timvideos12:01
*** panther_ has joined #timvideos12:07
*** sab_123 has quit IRC12:07
*** panther_ is now known as sab_12312:07
*** rohitksingh_work has quit IRC13:18
*** ssk1328 has joined #timvideos13:22
ssk1328mithro: I am building gateware by inserting my modules in video pipeline, after make load-gateware the USB device automatically disconnects and disappears from lsusb. ANy reason why this is happening?13:25
mithrossk1328: The FPGA resets the FX2 chip after it boots so it can load the FX2 firmware13:26
ssk1328mithro: So shoudl I wait?13:27
mithroIt should come up as the UVC camera+serial port after a couple of seconds13:27
mithroAs you don't care about the USB functionality at the moment, it might be worth just using the mixer target and not resetting the FX213:28
ssk1328mithro: You mean change something in opsis_hdmi2usb file13:29
ssk1328mithro: In the target folder13:29
mithrossk1328: No, "export TARGET=videomixer"13:30
ssk1328mithro: Ok13:30
mithrossk1328: The videomixer target also builds gateware much faster then the hdmi2usb target13:30
mithrossk1328: we don't regularly test/use the mixer target, so test that it works without your changes first13:31
ssk1328mithro: Okay13:32
mithrossk1328: As the FX2 will no longer get firmware from the FPGA you'll have to use modeswitch to get it into serial/jtag modes13:32
ssk1328mithro: From my understanding, there is no videomixer target in make file and also no such file in target folder, so I guess you are reffering to using 'video' instead of 'videomixer'13:34
mithrossk1328: oh, yeah it's called VideoMixerSoC in the file13:35
ssk1328But I set TARGET=video?13:35
*** danielki has quit IRC13:36
ssk1328mithro: But anyways I need to rebuilt the bit file, so I am running make gateware again. Lets see how soon it ends13:36
mithrossk1328: yes13:36
ssk1328mithro: I have inserted both float16 conversions and floatmult in the video pipeline and it seemed to be working correctly13:37
mithrossk1328: What is it doing?13:39
ssk1328right now connected hdmi_in0 to hdmi_out0, which looked good13:40
mithrossk1328: But what is it multiplying?13:40
ssk1328mithro: So I hardcoded one of the input of multiplier to 0.5, to see if this works13:41
ssk1328mithro: So the piprline at the input is like this RBG->RGB16f->RGB16fMult->RGB->YCbCr->YCbCr4:2:213:42
mithrossk1328: so it's making things darker or lighter?13:44
ssk1328mithro: It looked darker13:44
ssk1328mithro: But it didn't look perfect, so I thought I should try with hardcoded 1 as input, and since then I am stuck13:45
mithrossk1328: Well, if I understand correctly you are multiplying by 0.5 ?13:45
ssk1328mithro: Yeah, i understand logically it should be darker13:45
mithrossk1328: If you use an value of 1 it should be perfect, right?13:46
ssk1328mithro: Yeah, that was what I was hoping to test, but I was stuck in error which I mentioned earlier13:47
ssk1328mithro: Hence couldn't test it13:47
mithroWhat error?13:47
ssk1328mithro: The USB device disappearing when I do make load-gateware with my generated gateware13:48
ssk1328mithro: And then I can't do anything without power cycling the board13:49
ssk1328mithro: Which then erases the uploaded gateware13:49
ssk1328mithro: And this loop goes on!13:49
mithrossk1328: It worked with 0.5 and not with 1.0?13:49
ssk1328mithro: Yeah, but I encountered the same problem the first time I tried to upload the "0.5" gateware file, then I rebuilt the gateware again (read waiting for 20 mins) and then when I uploaded it worked13:51
ssk1328mithro: I know the error is not related to hardcoded constant13:51
*** rohitksingh has joined #timvideos13:54
*** danielki has joined #timvideos13:54
mithrossk1328: Does the device disappear from the USB forever?13:55
ssk1328mithro: Yeah, I waited for like 5-10 seconds13:56
ssk1328mithro: But anyways the videomixer thing you suggested seemed to work13:56
mithrossk1328: You can debug it by looking at the FX2 reset and I2C lines13:58
ssk1328mithro: What exactly do I look at?13:58
ssk1328mithro: Some pins on the board?13:58
*** rohitksingh1 has joined #timvideos14:00
mithrossk1328: yes14:01
*** rohitksingh has quit IRC14:01
ssk1328mithro: So I need to check for voltages at those pins, I don't have access to an multimeter right now though!14:04
mithrossk1328: no, you'll need a logic analyzer14:04
ssk1328mithro: Ohh, I might take time for me to get hold of that, so maybe later14:05
*** rohitksingh has joined #timvideos14:33
*** rohitksingh1 has quit IRC14:35
*** rohitksingh1 has joined #timvideos14:39
*** rohitksingh has quit IRC14:39
*** rohitksingh1 has quit IRC14:50
*** rohitksingh has joined #timvideos15:11
*** Bertl is now known as Bertl_oO15:18
*** sab_123 has quit IRC16:32
*** rohitksingh has quit IRC17:19
*** ssk1328 has quit IRC17:35
*** [d__d] has quit IRC18:33
*** [d__d] has joined #timvideos18:36
*** [d__d] has quit IRC18:47
*** [d__d] has joined #timvideos18:52
*** danielki has quit IRC19:32
*** danielki has joined #timvideos20:40
*** danielki has quit IRC21:31
*** danielki has joined #timvideos21:50
*** CarlFK has quit IRC22:36
*** Bertl_oO is now known as Bertl_zZ23:26
*** paradisaeidae has joined #timvideos23:42
*** paradisaeidae has quit IRC23:45
*** paradisaeidae_ has joined #timvideos23:46

Generated by 2.13.1 by Marius Gedminas - find it at!