Saturday, 2015-10-03

*** tpb has joined #timvideos00:00
mithroseaLne: did you sign up for the champion program?01:41
*** thaytan_ has joined #timvideos04:14
seaLnemithro: yes07:00
mithroseaLne: I have a fix for the Makefile issue07:03
seaLnewhich issue?07:06
mithromithro: seaLne: 'make firmware' first then 'make gateware'07:07
seaLnei'm just up so was going to work on the readme after some coffee07:07
seaLnewhat is the "firmware"? the name isn't too obvious?07:10
mithroseaLne: will explain in a sec07:10
seaLnefor the lm32?07:10
mithroseaLne: let me finish replying to the people on hackaday and upload this patch :)07:11
seaLnenp :)07:11
mithroseaLne: HackADay replying done I think07:23
mithro - :-)07:24
tpbTitle: Numato Opsis: FPGA-based open video platform | Hackaday (at
seaLnegood explanations, you also seem to have got off quite lightly for hackaday comments :)07:33
mithroseaLne: It is the weekend, the professional trolls are sleeping :)07:35
seaLnehmm weird doing make gateware without firmware first is now not giving an error (after make clean)07:55
seaLneah clean doesn't seem to actually clean build/misoc/software/include/generated/08:02
*** CarlFK has quit IRC08:19
seaLnemithro: what do you think of these changes?
tpbTitle: HDMI2USB-misoc-firmware/scripts at master · sealne/HDMI2USB-misoc-firmware · GitHub (at
seaLnesplit it up a bit and added the make lm32-firmware08:28
seaLnegoing to also add a test to about /opt/Xilinx/14.7/ISE_DS/settings64.sh08:29
*** CarlFK has joined #timvideos08:44
*** ChanServ sets mode: +v CarlFK08:44
*** miselin has quit IRC08:52
tpbTitle: expand bootstrap README and add test for /opt/Xilinx/14.7/ISE_DS/ by sealne · Pull Request #85 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroseaLne: see :-P08:57
tpbTitle: Travis should check that `make clean` actually cleans up all the files which are created. · Issue #83 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
seaLnewell that was my first attempt at using github, only used normal git servers before09:00
mithroseaLne: Pull request generally looks good!09:07
mithroseaLne: I added two small comments09:07
seaLnejust seen them09:07
mithroseaLne: if you make the fixes I requested and then force push to the same branch you sent the pull request from, then it will update the pull request09:07
tumbleweedoh, before I forget, some more feedback from pyconza:09:21
mithroseaLne: Do you know how to use "git rebase --interactive" to rewrite history and fix your older commit messages?09:21
tumbleweedwe had a few machines that had HDMI ports that didn't seem to be useable. I think they may have been Optimus (connected to the nvidia GPU, not the intel one in use)09:22
tumbleweedthey'd see the atlys, but not be able to output anything09:22
tumbleweedCarlFK: did you see any?09:22
seaLnemithro: let me have a look09:24
mithroseaLne: You almost got it right09:31
mithroseaLne: at the end, instead of doing a merge you should do a force push effectively wiping out the old commits09:32
* seaLne is now a bit lost09:34
seaLnelet me read some more09:35
mithroseaLne: Have you used gitk before?09:35
seaLnealso my git usage has been fairly basic09:35
mithroseaLne: I recommend installing it and starting it up - it's a great way to visualize what you just did09:35
seaLneso i should be able to remove the extras with git rebase i think? but i'm a bit stuck with "Cannot rebase: You have unstaged changes" and a bit scared of messing it up more09:44
seaLneit does feel like it would be a lot simpler to start again but that would be cheating09:45
mithroseaLne: okay09:45
mithroseaLne: first thing is - do a "git status" and show me what you have there09:46
seaLnemodified:   scripts/README.md09:46
mithroseaLne: you can use "git diff" to see what changes you have09:47
seaLneyeah, which seems to show it compared to before i started09:47
mithroseaLne: does it say anything about being in the middle of a rebase?09:48
seaLneno, previously i tried git rebase --abort and git reset HEAD which probably messed things up more09:49
mithroseaLne: okay09:50
mithroseaLne: which branch are you on - still master?09:50
seaLnesorry for all the newbie problems09:50
mithroseaLne: great! So, now you just need master to point at the result of the rebase you did earlier09:50
mithroseaLne: to do that09:51
mithroseaLne: do a "git reset --hard 2dc1a10afceb1a5cca83306c07c67f466fc33c69"09:51
seaLnethat looks better now09:52
mithroseaLne: Now, when you tried to push to your repo it probably said something like "can't push, need to merge"09:52
mithroseaLne: and you did a "git merge origin/master" which created that merge commit that you can see at the end of your commit09:53
mithroseaLne: this one ->
seaLnemy last push didn't seem to complain. i had just done "git push"09:53
seaLne(i haven't done anything yet since the reset)09:54
mithroseaLne: If you try a "git push origin master" now I think it will complain?09:56
seaLne! [rejected]        master -> master (non-fast-forward)09:57
*** travis-ci has joined #timvideos09:58
travis-ci[mithro/HDMI2USB-misoc-firmware/travis-makefile-fix#1] (d48baa1): The build failed. (
*** travis-ci has left #timvideos09:58
*** travis-ci has joined #timvideos09:59
travis-ci[timvideos/HDMI2USB-misoc-firmware/travis-makefile-fix#180] (946ee71): The build passed. (
*** travis-ci has left #timvideos09:59
mithroseaLne: So what git is complaining about is that if you push your master into the remote repository it is going to cause a bunch of commits to "disappear"10:00
mithroseaLne: which is normally bad10:01
seaLnewhich is this case we want?10:01
mithroseaLne: but in this case, you have deliberately use git rebase to "rewrite history"10:01
mithroseaLne: with git, a commit is dependent on it's contents - so when you changed things the commits become different10:02
mithroseaLne: Git doesn't know that your new rewritten commits are related to those old commits, so it doesn't think what you are about to do is a good idea :)10:03
mithroseaLne: but we can tell git that we know better (and potentially shoot ourselves in the foot if you aren't careful)10:03
mithroseaLne: "git push origin master --force"10:04
mithroseaLne: then git will look at you strangely and go "oh well, I'll do as you ask" making the old commits disappear and the new commits effectively take their place10:05
seaLnethanks, looks fixed now10:05
mithroseaLne: Now if we look at
tpbTitle: Expand bootstrap README and add test for /opt/Xilinx/14.7/ISE_DS/ by sealne · Pull Request #85 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroseaLne: we see only your 3 commits10:06
mithroseaLne: and its like you did everything perfectly the first time :-P10:06
mithroseaLne: Time travel with git ;)10:06
mithroseaLne: you have to be careful with this type of workflow, as you can potentially delete your work10:08
mithroseaLne: git keeps a thing called the "reflog" which can be used to rescue stuff in a git repo10:11
seaLneso the problem was that i didn't properly finish the rebasing editing messages and fixup before pushing?10:11
mithroseaLne: the problem is when you finished doing the rebase, you merged the commits from your old branch10:12
seaLnewhat should i have done after git rebase -i HEAD~1010:14
mithroseaLne: "git push origin master --force" probably10:16
seaLnewithout force suceded though10:16
seaLnei had just done git push10:16
mithroseaLne: between the "git rebase -i HEAD~10" and "git push" you ran a "git merge origin/master" or similar command10:17
seaLnenot entirely sure, thanks for the mini tutorial10:18
mithroseaLne: I'm sure you ran some type of "git merge" command between your rebase and push because that is really the only way you get that merge commit you pushed10:19
mithroseaLne: It takes a while to get use to doing rebase stuff10:25
mithroseaLne: I think this pull request is probably good to merge then10:27
mithroseaLne: want to take a look at my pull request and see what you think?10:33
seaLneis gateware a misoc name for it? its a bit unclear to me when i first looked what it was10:37
mithroseaLne: well, things get a bit confusing when you have a softcore on your FPGA10:38
mithroseaLne: so we call what is loaded onto the FPGA the "gateware"10:38
mithroseaLne: because it is basically what configured the "gates" of the FPGA (Field Programmable *Gate* Array)10:39
seaLneah, never made that connection10:39
seaLnethe stuff in #84 though looks good10:39
mithroseaLne: Just need to wait for Travis to report I haven't broken anything, then I'll merge that10:41
mithroseaLne: Got time to take a wack at a bug?10:42
seaLnemaybe, need to sort some stuff in the garden10:44
mithroseaLne: If you want to continue hacking on infrastructure it would be good to look at
tpbTitle: Travis should check that `make clean` actually cleans up all the files which are created. · Issue #83 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
seaLneso there is more needed than your make clean fix?10:46
seaLneah sorry thats a test10:46
mithroseaLne: If you want to get your hands dirty on actually hacking on some code which runs on the FPGA / softcore - or
tpbTitle: Support resetting an input without resetting the output · Issue #80 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroseaLne: yeah - a test which I'm pretty sure we'll fail at the moment :P10:47
seaLnei'll have a poke in a few hours10:47
mithroseaLne: Do you know C?10:50
*** travis-ci has joined #timvideos10:51
travis-ci[mithro/HDMI2USB-misoc-firmware/travis-makefile-fix#2] (cce3d42): The build was fixed. (
*** travis-ci has left #timvideos10:51
seaLnei can read it a bit, i'm not really a programer10:52
seaLnedone some basic c/c++ stuff before10:52
mithroseaLne: okay - the C code we use is pretty basic but occasionally does something tricky10:55
mithroseaLne: plenty of things which can be done to improve our infrastructure if you prefer that10:57
mithroseaLne: :-)10:58
tpbTitle: Issues · timvideos/HDMI2USB-misoc-firmware · GitHub (at
mithroseaLne: Thanks for your help so far! You are now 1 of 6 contributors to the firmware! :)11:10
mithrowell, I'm going to get some dinner11:22
*** travis-ci has joined #timvideos11:24
travis-ci[timvideos/HDMI2USB-misoc-firmware/travis-makefile-fix#187] (cce3d42): The build passed. (
*** travis-ci has left #timvideos11:24
*** travis-ci has joined #timvideos11:41
travis-ci[mithro/HDMI2USB-misoc-firmware/travis-makefile-fix#4] (da32723): The build was fixed. (
*** travis-ci has left #timvideos11:41
*** travis-ci has joined #timvideos12:15
travis-ci[mithro/HDMI2USB-misoc-firmware/travis-makefile-fix#3] (83e70bb): The build was fixed. (
*** travis-ci has left #timvideos12:15
*** travis-ci has joined #timvideos12:32
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#186] (a457961): The build passed. (
*** travis-ci has left #timvideos12:32
mithro_florent_: Those "videos" that you linked are pretty cool12:51
mithro_florent_: The looks pretty cool too!12:52
_florent_yes, asciinema is really fine12:52
mithroEveryone should checkout -> https://asciinema.org12:53
tpbTitle: asciinema - Record and share your terminal sessions, the right way (at
mithro_florent_: so how does the second video actually work?12:53
mithro_florent_: the Ethernet sim one?12:53
_florent_the soc is running in simulation with lwip on the lm32 and I'm pinging it12:54
*** travis-ci has joined #timvideos12:54
travis-ci[timvideos/HDMI2USB-misoc-firmware/travis-makefile-fix#189] (83e70bb): The build passed. (
*** travis-ci has left #timvideos12:54
_florent_tcp/ip software stack12:54
mithro_florent_: the lm32 is running in qemu, how is the ethernet hardware being simulated?12:55
_florent_in the video the soc is: lm32 + sdram + ethernet + uart simulated with verilator12:56
mithro_florent_: oh!12:56
_florent_lm32 is not running with qemu12:56
_florent_it's simulated12:56
_florent_but since verilator is very fast, it is able to simulate the soc at ~1Mhz12:56
_florent_wich is already fast for a simulation12:57
_florent_and in fact usable to develop software12:57
mithro_florent_: yeah - that should be fast enough to do any of the stuff I was suggesting12:57
seaLnewhich videos are you refering to?13:07
tpbTitle: Get lm32 firmware running under qemu to enable testing without hardware · Issue #86 · timvideos/HDMI2USB-misoc-firmware · GitHub (at
tpbTitle: asciicast:17175 - asciinema (at
tpbTitle: asciicast:18224 - asciinema (at
seaLneah, thanks13:07
*** rohitksingh has joined #timvideos14:41
*** travis-ci has joined #timvideos14:54
travis-ci[timvideos/HDMI2USB-misoc-firmware/travis-makefile-fix#191] (da32723): The build has errored. (
*** travis-ci has left #timvideos14:54
*** rohitksingh has quit IRC14:58
*** rohitksingh has joined #timvideos15:12
*** 7GHABBZAL has joined #timvideos15:23
7GHABBZAL[timvideos/HDMI2USB-misoc-firmware/travis-makefile-fix#191] (da32723): The build passed. (
*** 7GHABBZAL has left #timvideos15:23
*** CarlFK has quit IRC16:15
*** travis-ci has joined #timvideos16:34
travis-ci[timvideos/HDMI2USB-misoc-firmware/master#193] (6da250e): The build passed. (
*** travis-ci has left #timvideos16:34
*** rohitksingh has quit IRC17:28
seaLnecan i cause travis to fail a build with return 1?17:32
*** rohitksingh has joined #timvideos17:42
seaLnehmm ok that isn't quite right17:46
seaLnemissing pipe17:53
seaLnemithro: make clean check seems to be pretty much working now i think18:50
*** rohitksingh has quit IRC20:44
*** CarlFK has joined #timvideos22:03
*** ChanServ sets mode: +v CarlFK22:03

Generated by 2.13.1 by Marius Gedminas - find it at!