Friday, 2017-06-23

*** tpb has joined #timvideos00:00
mithropaddatrapper: I will have time this afternoon to help you debug the descriptor thing00:42
mithropaddatrapper: We should also have our weekly sync up meeting00:42
*** sb0 has quit IRC02:40
*** rohitksingh_work has joined #timvideos03:44
Ishan_Bansalmithro : Hi04:18
*** rohitksingh_wor1 has joined #timvideos04:45
*** rohitksingh_work has quit IRC04:47
mithroIshan_Bansal: Hi!04:56
mithroIshan_Bansal: so, we should get you a pull request set up04:57
mithroIshan_Bansal: I've got time now if you want me to help walk you through it05:01
Ishan_Bansalmithro : I got a lecture now , are u available after 2 hours or so05:29
mithroIshan_Bansal: I plan to be05:29
Ishan_Bansalmithro : I will see you than05:30
mithroIshan_Bansal: great!05:32
mithroIshan_Bansal: your lecture on anything interesting? :-P05:33
paddatrappermithro: ok cool06:55
Ishan_Bansalmithro: I am back.07:03
Ishan_Bansalmithro : ya, the lectures are on computer vision stuff.07:03
mithroAh cool07:03
mithroI'm going to run away for about 30 minutes - do you want to see if you can get started without me?07:04
Ishan_Bansalmithro: ok, no problem07:04
mithroNeed to grab some food while it's still available07:05
paddatrapperfood is good :)07:06
paddatrappermithro: got it working! Now to remove all the extra stuff in the main function...07:26
*** rohitksingh_work has joined #timvideos07:43
*** rohitksingh_wor1 has quit IRC07:45
paddatrappermithro: the issue was I need to run the RENUMERATE() macro to cause the FX2 to disconnect and reconnect when the new firmware is loaded07:51
mithroBack now!08:28
mithropaddatrapper: yay!08:28
paddatrapperSo now to fix the descriptors...08:28
mithroIshan_Bansal: did you get anywhere?08:29
mithropaddatrapper: wireshark is really helpful for that08:31
mithropaddatrapper: Have you learnt how to look at the ASM the compiler is producing yet?08:31
paddatrappermithro: yeah, I' mkinda counting on it08:31
paddatrappermithro: vim descriptors.asm08:31
mithropaddatrapper: Yeah08:32
*** rohitksingh_work has quit IRC08:33
*** rohitksingh_work has joined #timvideos08:45
mithropaddatrapper: Anything I can help with?08:55
paddatrappermithro: not currently, wireshark is being very helpful08:55
Ishan_Bansalmithro: Are you there for about an hour or so.08:56
mithroIshan_Bansal: I can be08:56
mithroIshan_Bansal: Shall we get the pull request going?08:57
Ishan_Bansalmithro : I have to submit some documents to the college as I am going to home on 25 th , so busy with that but as soon as I get time I will made that.08:59
mithroIshan_Bansal: Okay09:00
Ishan_Bansalmithro : So I need to fork the repository as I did with the Nueron1k/litejpeg and than after adding my changes have to made the pull request for the timvideos. Is that it is supposed to be.09:02
mithroIshan_Bansal: yes09:02
mithroIshan_Bansal: a pull request maps to a branch in git09:03
Ishan_Bansalmithro : okay.09:04
mithroIshan_Bansal: So the idea is that you develop on a branch - committing frequently, once you are happy with the code you create another branch with the cleaned up version and send that as a pull request09:04
Ishan_Bansalgot it.09:06
mithroIshan_Bansal: Once you send a pull request, I'll then review it by putting comments and such09:13
mithroIshan_Bansal: I think we should aim to get one pull request done per week09:13
paddatrappermithro: do you know if there is a particular packet that lists the entire descriptor? At the moment I'm having to find sections by chance09:22
tpbTitle: USB in a NutShell - Chapter 5 - USB Descriptors (at
tpbTitle: USB in a NutShell - Chapter 6 - USB Requests (at
mithroThey might be helpful?09:24
paddatrapperthanks that second one is helpful. The first is a repeat of the spec09:24
mithropaddatrapper: I think it's a little more concise09:25
mithropaddatrapper: You can also use libusb to manually do the descriptor request stuff if you want09:27
tpbTitle: libusb: USB descriptors (at
paddatrappermithro: that's an idea. At the moment I'm trying to reconcile the device requests you linked with what I'm seeing in wireshark09:28
mithroUSB in a Nutshell is a pretty good reference09:28
mithropaddatrapper: BTW Have you tried the bulkloop demo with C descriptors?09:28
paddatrappere.g. get status: actual response is 0x02a0 while is should only have a length of 2 bytes, with the last two bits being used...09:29
paddatrappermithro: No I haven't09:29
paddatrapperah ok, never mind get status is correct09:30
paddatrappermithro: Seems the host is never requesting for descriptors - get descriptors request never sent09:32
mithropaddatrapper: That normally means there is an error earlier09:32
paddatrapperdmesg gives the error: usb 2-2: device descriptor read/8, error -11009:33
mithropaddatrapper: Go look up error -110?09:33
mithroUSB error -110 means power exceeded, the host could not provide enough electric power for the pendrive to operate. Because it has not enough power also the USB stick cannot provide the device descriptor to the host, as a result it cannot be identified and so on. Maybe your motherboard it's overloaded with devices that consume all the available pwoer.09:34
mithropaddatrapper: Could mean you have the current requirement in the descriptor stuff wrong09:35
mithropaddatrapper: or you have a ground loop / short in how you have connected your boards together09:35
mithropaddatrapper: Another option is to check how it differs from a known working request09:36
paddatrappermithro: I've set max power to 250 which is 500mA09:36
mithropaddatrapper: I would set it to something like 100mA or lower09:37
paddatrapperI only have one board connected to usb, nothing else to interfear09:37
paddatrapperok will try that09:37
paddatrappereven setting it to 50mA didn't help09:42
mithropaddatrapper: Yeah - it's a good idea to have something working to compare too09:44
paddatrappermithro: I've been comparing the descriptors to the ones from the webcam09:45
tpbTitle: HDMI2USB-fx2-firmware/reference_descriptors at audio · paddatrapper/HDMI2USB-fx2-firmware · GitHub (at
mithropaddatrapper: I mean working get descriptors09:50
paddatrapperAh ok09:50
mithroIE See if works09:53
tpbTitle: fx2lib/examples/bulkloop at linux-descriptors · mithro/fx2lib · GitHub (at
paddatrappermithro: yup it works10:02
mithropaddatrapper: Okay, start modifying the descriptors of that :-P10:02
paddatrappergoing through the descriptors10:02
mithropaddatrapper: I'm just about head out - do you have a quick moment to discuss goals for this week?12:14
*** rohitksingh_wor1 has joined #timvideos12:16
*** rohitksingh_work has quit IRC12:16
*** rohitksingh_wor1 has quit IRC12:39
*** sb0 has joined #timvideos14:03
paddatrappermithro: sorry didn't realise my IRC connection was gone. Ping me when you're back around15:22
mithroI'm just about to head home to bed15:22
paddatrapperok. I'll be around from about 9 tomorrow if that works better for you?15:24
mithropaddatrapper: So for this week I think we should target getting the descriptors working enough that the linux audio subsystem detects the device with a stretch goal of trying to get some data packets going?15:24
paddatrappermithro: Sounds reasonable15:24
paddatrapperEspecially working from the bulkloop stuff15:24
mithroI'll try and look at all the pull requests tomorrow15:25
*** CarlFK has quit IRC22:08
*** CarlFK has joined #timvideos22:10
*** ChanServ sets mode: +v CarlFK22:10
*** hyades has joined #timvideos22:24

Generated by 2.13.1 by Marius Gedminas - find it at!