Tuesday, 2017-12-12

*** tpb has joined #timvideos00:00
*** mithro has joined #timvideos00:01
*** ChanServ sets mode: +o mithro00:01
*** ivodd has joined #timvideos00:01
*** tumbleweed has joined #timvideos00:01
*** tumbleweed has joined #timvideos00:01
*** Signotheque has joined #timvideos00:01
*** jimmo has quit IRC00:22
*** jimmo has joined #timvideos00:22
*** springermac has quit IRC02:22
*** springermac has joined #timvideos02:22
*** mikal_ has joined #timvideos03:41
*** mikal has quit IRC03:45
*** rohitksingh_work has joined #timvideos03:52
mithroshenki: So you had success last night?04:35
shenkimithro: somewhat. packets get sent, and i know that they are being recieved, but they don't make it all the way through the rx path04:37
shenkiand quite often we hit an unaligned access in the network code rx path04:37
mithroshenki: You mean they are not coming back?04:37
shenkithey come back. i get the irq04:37
shenkiand i read the data04:37
shenkibut they don't appear in userspace (ie, ping doesn't get a response, even tho the response is coming in on the wire and causing an interrupt)04:38
mithroshenki: That sounds like something inside Linux not working?05:04
mithroshenki: Okay, so you need help from Linux kernel people rather than me now :-P05:07
shenkimithro: you could write an unaligned memory handler for us05:09
mithroshenki: Why is it generating unaligned memory reads?05:09
shenkiit's reading data from the network buffer05:11
mithroshenki: But shouldn't gcc be doing the work to fixing that up?05:12
shenkiit's a buffer that comes in off the wire, containing the ethernet frame05:14
shenkii'm doing what should be the right things to make the network stack align the data, but it is not working05:14
mithroshenki: I assumed the unaligned access was reading from the Ethernet SRAM region?05:16
shenkino, it's from ram05:17
shenkithe packet has been copied into memory by this stage05:17
mithrocan you attach gdb and see what is going on?05:18
shenkii know what's going on05:20
shenkiit's reading the ip header from an unaligned address within the skb05:20
mithroshenki: Well, I guess you have it under control...05:22
shenkimithro: you don't want to help write the unaligned memory handler?05:23
mithroshenki: why would it be needed?05:23
shenkiso the kenrel can read unaligned memory05:24
shenkiit's something most kenrel ports have05:24
shenkikernel, dammit05:24
CarlFK[m]mithro (IRC): make conda - I guess I did that. what do I do to 'activate'  or something add to the path05:40
mithroCarlFK[m]: I'm pretty sure it tells you at the end?05:40
CarlFK[m]mithro (IRC): mabye, but that is gone.  now I get errors http://paste.ubuntu.com/26168077/05:41
tpbTitle: Ubuntu Pastebin (at paste.ubuntu.com)05:41
CarlFK[m]found it ansible logs         "source /home/juser/HDMI2USB-litex-firmware/scripts/enter-env.sh"05:45
CarlFK[m]WARNING:root:unbind-helper not found, will have to run as root!05:48
CarlFK[m](H2U P=opsis) [email protected]:~$ sudo hdmi2usb-mode-switch05:48
CarlFK[m]sudo: hdmi2usb-mode-switch: command not found05:48
CarlFK[m]I think I have head that one before.05:48
CarlFK[m]https://hdmi2usb.tv/firmware/ TODO: Can someone please accept my PR :p06:00
shenkimithro: something that would be really useful is exposing the reset line (the same signal that is hooked up to the RESET button) as a csr register06:01
mithroshenki: That would be pretty easy to do...06:01
shenkicool :)06:03
mithroshenki: Do you want me to walk you through it?06:06
shenki# ping
shenkiPING ( 56(84) bytes of data.06:44
shenki64 bytes from icmp_seq=1 ttl=64 time=9.06 ms06:45
shenki64 bytes from icmp_seq=2 ttl=64 time=2.21 ms06:45
shenki🚢 ship it06:45
mithroshenki: Opps somehow I got distracted with doing work...06:48
cr1901_modern_with_ doing work, or _from_ doing work :P?06:49
mithroshenki: Just create a GPIOOut register kind of like this -> https://github.com/timvideos/HDMI2USB-litex-firmware/blob/master/gateware/led.py#L1106:49
tpbTitle: HDMI2USB-litex-firmware/led.py at master · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com)06:49
* cr1901_modern highly relates to one of the two06:49
mithroInstead of pads give it a Signal()06:50
mithroshenki: Then connect the Signal() to the reset signal inside the CRG module06:51
mithroshenki: Somewhere around here -> https://github.com/timvideos/HDMI2USB-litex-firmware/blob/master/targets/arty/base.py#L77-L8406:52
tpbTitle: HDMI2USB-litex-firmware/base.py at master · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com)06:52
*** rohitksingh_work has quit IRC06:54
*** rohitksingh_work has joined #timvideos06:56
mithroI have to run06:57
*** fumblehool has joined #timvideos07:09
mithrocr1901_modern: you still around at all?07:57
mithroAhh well, I'll try and catch you tomorrow08:01
mithroReally going to bed now....08:01
felix_hm, i wonder how much time the volunteering at lca will take and if that will collide with being at the fpga miniconf. and which type of ticket i should order for the miniconf; the flight to australia already cost me quite some money, but i don't want to take a sponsored ticket from someone who is also interested but really can't afford that otherwise09:21
xfxffelix_: it's a ticket and accomodation for the week, as a thankyou for volunteering on the core AV crew for the week.  i sent the email you're probably responding to but i don't think we've met, hi :)09:56
xfxfmithro: how many of your crew are going to be involved in the FPGA miniconf? because that'll indeed conflict with AV volunteer duties, so i may be mis-reading your spreadsheet09:57
xfxf(one or two people in that situation doesn't bother me too much at all, but it occurs to me it might be almost everyone you've invited, which might present some setup challenges over the weekend)09:59
felix_hi xfxf10:00
xfxf(although.. for the FPGA hackers, i mostly care that we have a working firmware/gateware with the functionality we need, and we have somebody around to help us troublesheet if we have issues.  which hopefully we won't!)10:00
xfxfmithro: also, the current plan for the keynote venue is to use HDMI2USB's as vision switchers, because UTS don't have any and originally insisted on using a Windows laptop in the top area and a giant remote clicker for Powerpoint.  so we should probably get on the same page there to ensure it'll be reliable...10:01
felix_i'm not too familiar with the inner workings of the hdmi2usb and i don't know much about video recording stuff. i'll check the boxes for av and venue helper, is this right?10:06
xfxfi'd imagine mithro already has plans for you :)10:06
xfxfi'll wait until he's awake + responds10:07
felix_ok, then i'll wait until he's awake again. don't really know how urgent the "please register for that" was10:08
xfxfa volunteer ticket isn't too much of an issue if you're helping at the hackfest and then after the FPGA miniconf; it's more just as long as half the intended AV crew doesn't disappear for the FPGA miniconf10:08
xfxfa few days is fine - i said urgent mostly so people did it and didn't forget about it :)10:08
xfxfand i'm trying to confirm numbers before I book accomodation10:09
xfxfi'm guessing only 4 or 5 of the 14 people i have on my list will be going to the FPGA miniconf so this is likely totally fine10:10
*** sb0 has quit IRC10:34
*** fumblehool has quit IRC11:08
*** sb0 has joined #timvideos11:11
*** sb0_ has joined #timvideos12:00
*** rohitksingh_work has quit IRC12:31
*** shorne_ is now known as shorne13:19
mithrofelix_: I'm not expecting you to pay for anything extra13:26
mithrofelix_: I already have FPGA Miniconf hardware for you and Rohit13:26
cr1901_modernmithro: I am around now13:27
mithrofelix_: I would like you to participate in the A/V  volunteer roster so that you get a feel for what environment PhotonSDI will be used in and the type of stresses it might be put under :-)13:27
cr1901_modern(shouldn't you still be in bed?)13:27
mithrocr1901_modern: Yes, but I woke up and couldn't get back to sleep so came into work instead13:27
cr1901_modernmithro: That's dedication :)! Or you getting your hours out of the way early13:28
*** tsglove has quit IRC13:33
*** tsglove has joined #timvideos13:35
*** tsglove has joined #timvideos13:38
*** tsglove has joined #timvideos13:40
*** tsglove has quit IRC13:42
*** kierank has joined #timvideos13:51
felix_mithro: ok, sounds like a plan. i'd definitely like to try to get a better understanding of the usage of the things i'm developing13:53
felix_when trying to register for the fpga miniconf, i get this error: "14:09
felix_    There are no products available from category: FPGA Miniconf14:09
mithrofelix_: You don't need to register for the FPGA Miniconf15:11
mithro(You explicitly -- everyone else does)15:12
felix_i got the volunteer ticket and the registration of the fpga miniconf is on that page. so what should i select there? i also wonder if i should get a ticket for the penguin dinner and which hotel i should choose or if my accommodation booking will happen outside of that ordering page15:15
mithrofelix_: Your hotel booking will probably happen outside that page15:16
CarlFK[m]felix_: did you select one of the 6 or so fpga hardware choices?  like usb a or c?15:24
CarlFK[m]good morning all15:24
felix_i selected the sponsored with usb a option15:25
felix_or should i have selected no board there and i'll get the board directly from mithro?15:26
mithrofelix_: Don't sign up for anything at the FPGA miniconf15:26
felix_ok, i selected no selection instead15:29
CarlFK[m]mithro (IRC): accept my PR so I can make it better15:39
CarlFK[m]petty sure it is out of date with filenames15:39
CarlFK[m]mithro (IRC): is there a spreadsheet or something with build/results?15:44
CarlFK[m]this is what I think needs testing:  http://paste.ubuntu.com/26168128/ (http://paste.ubuntu.com/26168128/) and I would like to log the results somewhere15:45
tpbTitle: Ubuntu Pastebin (at paste.ubuntu.com)15:45
mithroCarlFK[m]: yes15:48
mithroCarlFK[m]: No we don't have any spreadsheet or anything15:48
mithroCarlFK[m]: I'll try and look at your PR in a minute15:48
CarlFK[m]take you time, i'm making breakfast15:49
felix_hmm, i thought that there was some minimal migen/misoc or litex example with only a blinking led or some characters sent via a serial port, but i don't find that any more. most examples i found already used the ddr memory, but i don't want to spend too much time on getting that working right now for the ac70115:55
cr1901_modernmithro: As a matter of policy, does every commit to HDMI2USB need to compile? B/c I don't think I can promise that during the rebase.16:08
cr1901_modern(Makefile conflict)16:10
mithrofelix_: Getting the DDR working should be pretty trivial. You can just comment out the ddr16:11
cr1901_modernmithro: Disregard16:13
cr1901_modernmithro: Just FYI, it's not possible to test whether changes to the Makefile during a rebase break until _after_ the rebase is done thanks to >>16:21
cr1901_modernversion_data.sh bombing if "ref HEAD is not a symbolic ref"16:22
cr1901_modern(which it won't be during a rebase conflict)16:22
FelixVicr1901_modern: ping - any further thoughts on the flash interface?17:34
cr1901_modernFelixVi: No and I'm really busy the next few days. I'll talk to you privately later tonight about what we should do.17:38
FelixVicr1901_modern: No worries, we can also talk about it next week17:40
*** samsagaz has quit IRC18:30
*** MaZderMind has quit IRC20:09
*** MaZderMind has joined #timvideos20:09
*** bunnie_ has quit IRC20:37
*** mcarden has quit IRC20:37
*** mcarden has joined #timvideos20:38
*** jea has quit IRC20:38
*** Guest4752 has quit IRC20:42
CarlFK[m]MaZderMind (IRC): Do you have notes on what you installed from where to flash your Opsis?20:44
*** bunnie has joined #timvideos20:44
CarlFK[m]I have both a stretch and xenail box, and get various errors from modeswitch20:44
MaZderMindCarlFK[m]: i blindly ran the getting-started script and successfully used the modeswitch tool. I've put my board aside until january, because of ccc congress preparations20:45
CarlFK[m]what getting-started script?20:46
CarlFK[m]does "make conda" sound familiar?20:46
*** jea has joined #timvideos20:46
*** Peetz0r has quit IRC20:47
*** Peetz0r has joined #timvideos20:47
*** Guest4752 has joined #timvideos20:50
CarlFK[m]mithro (IRC): docs :p20:58
*** Guest4752 is now known as puck21:36
*** CarlFK has quit IRC23:29

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