Saturday, 2018-01-20

*** tpb has joined #timvideos00:00
mithroI'm here now00:27
CarlFK[m]me too00:36
CarlFK[m]micolous: are poking at download-prebuilt-firmware.py00:37
CarlFK[m]it would be nice if it made more noise00:37
CarlFK[m]curl http://api.github.com/repos/timvideos/HDMI2USB-firmware-prebuilt/contents/archive/master/00:42
CarlFK[m]micolous: archive_url = "https://api.github.com/repos/{owner}/{repo}/contents/archive/{branch}/".format(**details)00:48
CarlFK[m]http*S* works much better00:48
micolous[m]ah great00:50
*** williammckee has quit IRC00:59
*** wcmckee has joined #timvideos01:00
CarlFK[m]we are escapting to find lunch01:00
*** futarisIRCcloud has quit IRC02:00
CarlFK[m]micolous: back from lunch - poking atlys again02:18
CarlFK[m]wandering off -- will be back in... 30 min?02:26
micolous[m]ok02:30
micolous[m]Well I got it flashed and loaded the FX2 firmware02:30
micolous[m]Image is kinda trippy, using the test pattern02:31
micolous[m]I'll send you a photo02:31
micolous[m]I suspect that the image buffer is layed out incorrectly for Atlys02:32
xobsmithro: 403 on or1k package02:52
*** rohitksingh-demo has joined #timvideos02:58
*** futarisIRCcloud has joined #timvideos03:01
mithroxobs: Hrm?03:04
xobscan't access the or1k .tar.bz2 files03:05
mithroOh03:05
mithro403 not 40403:05
mithroxobs: try again now?03:06
xobsmithro: Much better.  Thanks!03:06
xobsmithro: Why are there two versions of the or1k with the same timestamp?03:20
mithroxobs: No-stdc and newlib03:22
xobsmithro: Much better.  A lot less segfaultey03:25
felix_i'm at the door of the office now03:57
mithroshenki: https://github.com/mithro/qemu-litex/pull/1/files04:03
tpbTitle: Temp commit so Joel can continue doing stuff... by mithro · Pull Request #1 · mithro/qemu-litex · GitHub (at github.com)04:03
shenkimithro: you were missing some registers04:11
mithroshenki: I'm pretty sure they were added after I started that patch :-P04:11
*** futarisIRCcloud has quit IRC05:11
*** sb0 has quit IRC05:13
*** CarlFK has joined #timvideos05:14
*** ChanServ sets mode: +v CarlFK05:14
CarlFK[m]micolous: I've got a few min to think about Atlys - in our last episode.. I forget...05:19
micolous[m]Our last episode is that with the current FW, I get the trippy images I sent you05:19
CarlFK[m]can you see when it was built?05:20
CarlFK[m]use one of the 6 somewhat supported term programs, or your fav one...05:20
micolous[m]2018-01-1905:21
CarlFK[m]sweet05:21
CarlFK[m]I was expecting 2012 or so05:22
micolous[m]Did this solution even exist then? :P05:22
CarlFK[m]or so05:22
micolous[m]I seem to recall wrangling Twinpacts that year05:22
CarlFK[m]15 maybe?05:23
micolous[m]Anyways, that's not important ;)05:23
CarlFK[m]for a long time we didn't commit builds to non volleyball05:23
CarlFK[m]thanks - I think that gives _florent_  something to look at05:24
CarlFK[m]garbage display and console05:24
_florent_CarlFK[m]: what? fixing atlys?05:25
CarlFK[m]yes!05:25
CarlFK[m]hi _florent_05:25
CarlFK[m]https://github.com/timvideos/HDMI2USB-litex-firmware/issues/34705:25
tpbTitle: USB streaming is corrupted on Atlys · Issue #347 · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com)05:25
_florent_hi05:25
_florent_ok we'll see that together05:26
CarlFK[m]have you arrived in Sydney ?05:26
_florent_no, i'll be there tomorrow (in the afternoon)05:26
CarlFK[m]groovy - have a safe flight, get lots of rest, I'm going for coffee now05:27
_florent_thanks, see you tomorrow05:28
*** sb0 has joined #timvideos05:40
*** futarisIRCcloud has joined #timvideos05:46
*** tvCommitBot has joined #timvideos05:58
tvCommitBot[qemu-litex] mithro pushed 1 new commit to master: https://git.io/vNedH05:58
tvCommitBotqemu-litex/master dc90b35 Tim 'mithro' Ansell: liteeth: Fix for changes in LiteEth upstream.05:58
*** tvCommitBot has left #timvideos05:58
CarlFK[m]micolous: can we borrow your projector for the noc so we can project all the streams06:12
micolous[m]It's 1024x768, is that OK?06:12
CarlFKsure06:13
CarlFKit is mostly to make sure it is doing something, not if it is doing it well06:14
mithrorohitksingh-demo: I've merged the ident fix06:17
rohitksingh-demomithro: saw that, thanks! :)06:17
mithrorohitksingh-demo: you can delete your branch now :-P06:18
*** sasha_ has joined #timvideos06:25
rohitksingh-demosure!06:26
shenkimithro: grep -h "^#" ./build/arty_net_or1k.linux/software/include/generated/csr.h ./build/arty_net_or1k.linux/software/include/generated/mem.h | grep -v include > ~/dev/kernels/litex/include/dt-bindings/litex.h06:27
*** nbags[m] has joined #timvideos06:27
mithroshenki: No network...06:40
mithroshenki: Getting the following...06:40
mithro# [   23.400000] liteeth e0007000.mac eth0: LITEETH_READER_READY timed out06:40
mithro[   23.690000] liteeth e0007000.mac eth0: LITEETH_READER_READY timed out06:40
mithro[   23.720000] liteeth e0007000.mac eth0: LITEETH_READER_READY timed out06:40
mithroshenki: I do get the following....06:42
mithrohttps://www.irccloud.com/pastebin/tcGu6CLC/06:42
tpbTitle: Snippet | IRCCloud (at www.irccloud.com)06:42
mithroBut at least the hdmi2usb firmware and netbooting is still working under qemu06:44
mithroGoing to test real hardware06:44
*** ChanServ sets mode: +v micolous07:04
*** swalladge has joined #timvideos07:15
CarlFK[m]tumbleweed (IRC): what is the av server's hostname07:21
*** CarlFK has quit IRC07:21
CarlFK[m]av avserver... um...  server ?07:21
tumbleweedCarlFK[m]: gw07:21
CarlFK[m]gw07:21
shorne_mithro: I listended to the amp hour, good talk!07:33
*** CarlFK has joined #timvideos07:41
*** ChanServ sets mode: +v CarlFK07:41
nbags[m]Yeah me too, and that's what landed me here. Interested in the project and as a way into learning FPGA07:42
shorne_nbags[m]: we are currently working on getting litex better fpga board+linux support for OpenRISC07:44
shorne_you can check out the HDMI2USB-litex-firmware07:44
shorne_on github07:45
nbags[m]Cool, is this the workshop at LCA?07:49
shorne_well, I guess we have been working on this for the last few months working up to the workshop07:54
*** futarisIRCcloud has quit IRC07:56
mithronbags[m]: Yes08:11
mithronbags[m]: Are you coming to LCA?08:11
mithrohey shorne_08:19
mithrohow are things?08:19
shorne_mithro: things are ok, getting started on a gcc port of or1k08:27
shorne_call me crazy08:28
mithroshorne_: crazy08:28
mithroshenki: Working ping on hardware!08:29
shorne_mithro: anything I can do? docs, patch review, updates?08:31
mithroshorne_: Have you tried the Miniconf set up instructions recently?08:31
shorne_nope, on artix7 board?08:32
shorne_I can do that and give feedback08:32
mithroshorne_: Yeah08:33
mithroshorne_: http://j.mp/pre-fpga-lca201808:34
tpbTitle: FPGA Miniconf Set Up Instructions - Google Docs (at j.mp)08:34
mithroshorne_: It would also be good to see if you can build a rootfs for the or1k which works on our stuff -- at the moment the only person who can seems to be shorne_08:34
mithros/shorne_/shenki /08:34
shorne_cool, my hands are full right now with baby. ill try to get to it tonight. I have done 2 different rootfs's and they worked08:35
shorne_do you need insxns for that in the miniconf?08:36
shorne_mithro: or do you want to use a rootfs blob?08:36
mithroshorne_: We currently use a rootfs blob - but I would like some changes to the blob but shenki has run off08:37
shorne_mithro: buildroot right? what changes? when do you need them?08:38
mithroshorne_: I assume buildroot -- we can make do without them - but the sooner the better08:38
shorne_mithro: ok, are the changes required listed somewhere?08:40
shorne_or we just need something?08:40
mithroshorne_: Nope! Not yet :-P -- I only got to the stage about thinking of them08:40
shorne_ok, maybe ist them in the miniconf doc around the rootfs download part. Ill see it there08:41
shorne_mithro: "list" them08:42
mithroshorne_: Started putting them here -> https://docs.google.com/document/d/1dRWx6Ve1VlD11g6WaVvxSl2tqWE-vIBcX5UEiWff6bE/edit08:43
tpbTitle: or1k rootfs for Miniconf tuning - Google Docs (at docs.google.com)08:43
shorne_ok . cool08:43
shorne_I think i can do it08:44
mithroshorne_: Don't worry if you can't08:45
mithroshorne_: I'm sure there will be more small nice to haves as we go forward....08:45
shorne_its ok, ill get started and document the results there08:46
shorne_mithre: I probably wont use systemd :)08:47
shorne_but just will have an hour or so tonight and tmorrow08:47
nbags[m]mithro: yes but I'm volunteering so i can't participate in the miniconf09:13
nbags[m]Looks great though!09:20
*** sasha_ has quit IRC09:20
rohitksingh-demomithro: something is really bad with cmod_a7....it resets my usb port every time after loading gateware...and we don't know FT2232 connections since schematic is proprietary and also we have no idea from where is the 12MHz clock coming from (directly sourced from crystal? not likely! from the FT2232? maybe yes, but OpenOCD takes control of that chip while programming)09:22
mithrorohitksingh-demo: Could you test the other cmod_a7 and see if it behaves the same way?09:23
rohitksingh-demomithro: okay09:23
rohitksingh-demomithro: okay the other one is 35t...rebuilding the gateware09:24
mithrorohitksingh-demo: I doubt its a bad board.... -- but it is worth checking I guess?09:28
*** ewen has joined #timvideos09:32
ewenmithro: In case it helps, I tried rebuilding from latest git master of arty/linux, and booting.  qemu boot seemed to get stuck forever (spinning CPU in qemu) at point it should be tftp booting, apparently without any tftp packets over tap0.  Boot on actual hardware still worked.09:34
ewenActual hardware says:09:34
ewenInitializing random number generator... done.09:34
ewenStarting network: ip: socket: Function not implemented09:34
ewenip: socket: Function not implemented09:34
ewenFAIL09:34
ewenStarting dropbear sshd: FAIL09:34
mithroewen: Try "git submodules update" and try again?09:34
ewenmithro: I thought I'd done that, but will try again.09:34
rohitksingh-demomithro: same with second one...I suspect openocd might be doing something to the ftdi which forces it to behave weirdly? interesting part is that bitstream is being programmed correctly, and there is some garbage on flterm and then ftdi resets09:36
rohitksingh-demoopenocd's ftdi interface basically takes control of all 16  pins of the ftdi channels...09:36
mithrorohitksingh-demo: Possibly?09:36
mithroewen: What commit are you on?09:37
rohitksingh-demowho knows how the rest of the pins are connected09:37
rohitksingh-demomaybe one of the pin generates the 12MHz clock09:37
rohitksingh-demoI'm 99% sure09:37
rohitksingh-demoand when openocd takes over ftdi control, it messes up the clock09:38
ewenmithro: [email protected]:/src/fpga-miniconf/litex-buildenv$ git log | head -109:39
ewencommit 96d85866bf8c6fc27216c15986da2bd34df4f1f209:39
mithroewen: That looks right09:39
ewenYes, it seems to be latest I can see in GitHub web interface.  Have done "git submodule update" now and "make clean"; rebuilding and will see what happens.09:40
*** CarlFK has quit IRC09:45
felix_mithro, rohitksingh-demo: i'm at the front door of the office now and have a pizza for you09:48
mithroOh09:49
mithroWe ordered food09:49
mithroI'll come down and fetch you09:49
felix_meh09:49
felix_ok09:49
felix_i was asked to bring you food09:50
*** rohitksingh-demo has quit IRC09:51
mithrofelix_: Thanks anyway09:53
ewenmithro: Appears to be same result: works on actual hardware; qemu boot doesn't seem to generate packets on tap0, or make any progress (but consumes 1 core).09:54
mithroewen: oh... I think I know what it is09:54
ewenmithro: aha, sounds promisnig09:55
ewen(promising)09:55
* ewen tries rebooting entire laptop to clean up host kernel state (back in a bit)10:07
*** ewen has quit IRC10:07
*** ewen has joined #timvideos10:12
ewenmithro: No change after rebooting host kernel (ie, tap0 appears started correctly; tftpd appears started correctly; but qemu BIOS boot makes no progress after "Fetching from: UDP/6069" -- which should in theory either work or time out)10:13
ewenmithro: On a hunch, I rolled back to commit where I tested new BIOS TFTP booting (920d5122ed10111df16a3af70c913bc2b5729280), and that one seems to work (ie, boots to login: prompt in qemu)10:30
mithroewen: sorry, food turned up and we are eating10:33
ewenmithro: All good.  I'd guessed food had turned up :-)10:33
thaytano/10:33
thaytanhow's hacking?10:33
ewenmithro: I'll try a bisect between bad/good commits and see if I can narrow it down10:34
mithroewen: I'm pretty sure I know what's up10:34
ewenBisecting: 6 revisions left to test after this (roughly 3 steps)10:34
mithroewen: can you make sure you have the commit 96d810:35
ewencommit 96d85866bf8c6fc27216c15986da2bd34df4f1f210:36
ewenmithro:?10:37
ewenDate:   Sat Jan 20 16:56:36 2018 +110010:37
mithroewen: then make sure you qemu is dc90b35410:37
ewenmithro: It is not; seems much older10:38
ewenpwd; git log | head -110:38
ewen/src/fpga-miniconf/litex-buildenv/third_party/qemu-litex10:38
ewencommit d8ebfa9e0cc95cf7f7cb1a1014fb81ecc1c150df10:38
ewenDate:   Tue Oct 17 00:58:47 2017 +110010:38
mithroYeah - the Build-qemu doesn't pull changes10:38
ewenmithro: And it's not a submodule (just checked).10:39
ewenmithro: Sounds like this might break some miniconf users?10:39
mithroYeah10:40
mithroNeed to fix :-)10:40
ewenmithro: :-)   That "we've changed a lot, should probably test it all still works" was why I randomly decided to retest this evening....10:40
mithrothaytan: a bunch of us are at Google hacking on FPGAs right now10:40
mithroewen: thanks! It is really appreciated.10:41
thaytanmithro, I might make it tomorrow evening10:43
thaytanbut we drove down from Tamworth today, so I'm a bit tired tonight10:43
ewenmithro: I've added a note to the prep doc along side the qemu-boot-linux as a possible hint to people if this happens.10:45
ewenmithro: Not a doc change suggestion as hopefully fixed before anyone notices...10:45
*** CarlFK has joined #timvideos10:51
*** ChanServ sets mode: +v CarlFK10:51
ewenmithro: Just rebuilt after "cd third_party/qemu-litex && git pull", and master on litex-buildenv now boots to login: prompt in qemu.10:55
ewenmithro: So I think fix confirmed.10:55
ewenmithro: created issue as well so there's a reminder of interdependency not yet expressed mechanically: https://github.com/timvideos/litex-buildenv/issues/1411:06
tpbTitle: third_party/qemu-litex not automatically updated if changed · Issue #14 · timvideos/litex-buildenv · GitHub (at github.com)11:06
*** CarlFK has left #timvideos11:08
*** CarlFK has joined #timvideos11:09
*** ChanServ sets mode: +v CarlFK11:09
ewenmithro: FYI: configs/litex_defconfig (for Linux kernel) does not have CONFIG_NET (at all), which I think explains why in-Linux networking isn't happening.11:16
ewenmithro: It appears the rootfs loaded actually has a dropbear in it (sshd), which would probably work if networking were enabled, and there was a NIC driver working.11:16
ewenHave added a couple of comments to https://docs.google.com/document/d/1dRWx6Ve1VlD11g6WaVvxSl2tqWE-vIBcX5UEiWff6bE/edit that you mentioned earlier so this bit doesn't get lost11:19
tpbTitle: or1k rootfs for Miniconf tuning - Google Docs (at docs.google.com)11:19
thaytanwhat is your plan for tomorrow?11:32
*** sb0 has quit IRC11:35
*** pschulz01 has joined #timvideos11:40
*** ewen has quit IRC11:49
mithrothaytan: come here and hack again11:50
thaytanmithro, and probably move to UTS at some point?11:50
mithroYeah11:51
*** rohitksingh-demo has joined #timvideos11:54
rohitksingh-demomithro: is tftp boot working? mine is stuck at "Fetching from: UDP/6069"12:02
mithrorohitksingh-demo: In qemu or on hardware?12:03
rohitksingh-demoqemu12:03
rohitksingh-demoworking in hardware?12:03
rohitksingh-demolemme check12:03
mithrorohitksingh-demo: Pull the latest qemu (cd third_party/qemu; git pull)12:03
rohitksingh-demomithro: okay12:05
mithrorohitksingh-demo: I guess you missed were Ewen ran into exactly the same problem :-P12:07
rohitksingh-demomithro: ah okay12:09
mithrorohitksingh-demo: Can you give mithro/master a go -- the ./scripts/build-qemu.sh script should automatically pull the latest change now...12:19
rohitksingh-demomithro: sure! wait a few minutes for the results12:20
mithrorohitksingh-demo: Thanks!12:24
rohitksingh-demomithro: PackagesNotFoundError: The following packages are not available from current channels:12:24
rohitksingh-demo  - binutils-or1kx-elf=2.2812:24
mithroor1kx ?12:25
rohitksingh-demoyup12:25
rohitksingh-demooops12:25
mithro!= or1k12:25
rohitksingh-demothat bad habit I caught from mor1kx :p12:26
mithrohttps://github.com/timvideos/litex-buildenv/pull/15/files12:29
tpbTitle: Random collection of fixes. by mithro · Pull Request #15 · timvideos/litex-buildenv · GitHub (at github.com)12:29
*** pschulz01 has quit IRC12:29
*** pschulz01 has joined #timvideos12:38
rohitksingh-demomithro: on arty : https://paste.fedoraproject.org/paste/89DDwbwaM7lCWRvja~Z6jQ12:48
tpbTitle: Untitled - Modern Paste (at paste.fedoraproject.org)12:48
rohitksingh-demomithro: on qemu: https://paste.fedoraproject.org/paste/g1PxB0yr3nvjyjozNHFyaw12:48
tpbTitle: Untitled - Modern Paste (at paste.fedoraproject.org)12:48
mithrorohitksingh-demo: I'm missing the commands you ran to get into that state?12:49
rohitksingh-demojust followed the document commands12:49
rohitksingh-demomake gateware12:50
rohitksingh-demothen make image12:50
rohitksingh-demothen make firmware-connect12:50
rohitksingh-demofor qemu: ./scripts/build-qemu.sh12:50
mithrorohitksingh-demo: What is your network config look like?13:00
rohitksingh-demomithro: you mean /etc/network/interfaces?13:00
mithrorohitksingh-demo: no - I mean "ifconfig"13:01
*** sb0 has joined #timvideos13:02
rohitksingh-demomithro: oh...for some reason the ip address was a different one....now the arty works, thank you so much!13:03
rohitksingh-demomithro: qemu tftp still doesn't seem to be working13:14
*** pschulz01 has quit IRC13:19
*** rohitksingh-demo has quit IRC13:23
*** rohitksingh-demo has joined #timvideos13:25
*** rohitksingh-demo has quit IRC14:03
*** sb0 has quit IRC17:05
*** CarlFK has quit IRC17:41
*** CarlFK has joined #timvideos19:49
*** ChanServ sets mode: +v CarlFK19:49
*** micolous[m] has quit IRC22:08
*** deeprave has quit IRC22:08
*** Peetz0r has quit IRC22:08
*** mithro has quit IRC22:08
*** kierank has quit IRC22:08
*** jimmo has quit IRC22:08
*** Signotheque has quit IRC22:08
*** Peetz0r has joined #timvideos22:08
*** deeprave has joined #timvideos22:09
*** kierank has joined #timvideos22:09
*** jimmo has joined #timvideos22:09
*** mithro has joined #timvideos22:09
*** ChanServ sets mode: +o mithro22:09
*** Signotheque has joined #timvideos22:10
*** micolous[m] has joined #timvideos22:10
*** techman83 has quit IRC22:10
*** techman83 has joined #timvideos22:14
*** ChanServ sets mode: +v techman8322:14
*** shorne has joined #timvideos22:32
*** shorne_ has quit IRC22:35
*** futarisIRCcloud has joined #timvideos23:07
*** CarlFK has quit IRC23:25
mithroI'm just getting up now, will be in the office in like 3023:46

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