| *** rajm has joined #cip | 06:42 | |
| *** toscalix has joined #cip | 07:21 | |
| *** Yamamodo has joined #cip | 08:36 | |
| *** tpollard has quit IRC | 08:38 | |
| *** tpollard has joined #cip | 08:39 | |
| *** pave1 has joined #cip | 08:54 | |
| *** sangorrin has joined #cip | 08:58 | |
| *** hiraku has joined #cip | 08:58 | |
| sangorrin | wow new interface in freenode | 08:58 |
|---|---|---|
| gavinlai | szlin is not in office today. I will chair today's meeting. | 09:00 |
| gavinlai | #startmeeting CIP IRC weekly meeting | 09:00 |
| brlogger | Meeting started Thu Jun 27 09:00:38 2019 UTC and is due to finish in 60 minutes. The chair is gavinlai. Information about MeetBot at http://wiki.debian.org/MeetBot. | 09:00 |
| brlogger | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 09:00 |
| brlogger | The meeting name has been set to 'cip_irc_weekly_meeting' | 09:00 |
| *** brlogger changes topic to " (Meeting topic: CIP IRC weekly meeting)" | 09:00 | |
| gavinlai | #topic rollcall | 09:00 |
| *** brlogger changes topic to "rollcall (Meeting topic: CIP IRC weekly meeting)" | 09:00 | |
| gavinlai | please say hi if you're here | 09:00 |
| pave1 | hi | 09:00 |
| Yamamodo | hi | 09:00 |
| *** kazu has joined #cip | 09:01 | |
| fujita[m] | hi | 09:01 |
| patersonc | hi | 09:01 |
| hiraku | hi | 09:01 |
| kazu | hi | 09:01 |
| sangorrin | hi | 09:01 |
| gavinlai | #topic AI review | 09:02 |
| *** brlogger changes topic to "AI review (Meeting topic: CIP IRC weekly meeting)" | 09:02 | |
| gavinlai | 1. Provide the script for CIP kernel config collection - bwh | 09:02 |
| *** vidda has joined #cip | 09:02 | |
| pave1 | I don't see bwh on the rollcall. | 09:02 |
| gavinlai | #link https://lists.cip-project.org/pipermail/cip-dev/2019-June/002506.html | 09:02 |
| gavinlai | I guess bwh is not around. I will keep this action item | 09:02 |
| gavinlai | #action Provide the script for CIP kernel config collection - bwh | 09:03 |
| gavinlai | 2. List real time kernel questions to ask Daniel Wagner - szlin | 09:03 |
| gavinlai | szlin is collecting the questions. He plans to send the questions next week. | 09:03 |
| gavinlai | #action List real time kernel questions to ask Daniel Wagner - szlin | 09:03 |
| gavinlai | 3. Try updating CIP RT kernel to 4.19.50 - Pavel | 09:03 |
| gavinlai | pave1: (wave | 09:04 |
| pave1 | So... I have something that builds for me. 4.19 turned out to be relatively easy. | 09:04 |
| pave1 | I posted pointers to the list. | 09:04 |
| pave1 | What I really need at the moment is someone to step up and run the tests for me. | 09:04 |
| pave1 | I'm currently configuring test machine, because testing on development machines is unpleasant. | 09:05 |
| pave1 | But that will result in "does it boot" tests, not "is it realtime" tests. | 09:05 |
| patersonc | If you have a Kernel that is built for RT it should be easy enough for us to test with Daniel's jittertest | 09:06 |
| patersonc | Using the LAVA lab | 09:06 |
| sangorrin | yeah or rt-tests | 09:07 |
| pave1 | Ok, that would be great. Can we coordinate it after the meeting? | 09:07 |
| patersonc | Sure | 09:07 |
| pave1 | Thanks! | 09:07 |
| gavinlai | :) | 09:07 |
| sangorrin | https://github.com/Linaro/test-definitions/tree/master/automated/linux/cyclictest | 09:08 |
| sangorrin | pave1: you will probably want to use Chris's CI | 09:08 |
| sangorrin | for the kernel build | 09:08 |
| patersonc | :) | 09:08 |
| pave1 | I hope someone can configure that for me ;-). | 09:09 |
| patersonc | Sure | 09:09 |
| gavinlai | move to next topic? | 09:10 |
| pave1 | Yes. | 09:10 |
| gavinlai | #topic Kernel maintenance updates | 09:10 |
| *** brlogger changes topic to "Kernel maintenance updates (Meeting topic: CIP IRC weekly meeting)" | 09:10 | |
| pave1 | So... I reviewed 4.19.54 and 4.19.56 (regular size) and then tiny releases of 4.19.52 and 4.19.55 | 09:11 |
| pave1 | I played with the realtime, as discussed before. | 09:11 |
| pave1 | I guess I should mention... I'll be traveling next week. | 09:11 |
| pave1 | ...that's all from me. | 09:11 |
| gavinlai | pave1: thanks | 09:11 |
| gavinlai | any comments? | 09:12 |
| gavinlai | #topic Kernel testing | 09:12 |
| *** brlogger changes topic to "Kernel testing (Meeting topic: CIP IRC weekly meeting)" | 09:12 | |
| sangorrin | Unrelated to Pavel's comments but as you may have seen I have been sending some patches for cip-kernel-sec. | 09:12 |
| gavinlai | #undo | 09:13 |
| brlogger | Removing item from minutes: #topic Kernel testing | 09:13 |
| pave1 | I guess I should relay... | 09:13 |
| sangorrin | These scripts are very useful to track down CVE IDs | 09:13 |
| patersonc | Good work | 09:13 |
| sangorrin | I added functionality for tracking product kernels | 09:13 |
| Yamamodo | sangorrin: thanks your help for cip-kernel-sec, we expect this tool very well:) | 09:14 |
| gavinlai | sangorrin: the scripts are really helpful | 09:14 |
| pave1 | Nobuhiro says: I reviewd v4.4.163, and I released v4.4.182-cip34 and v4.19.52-cip4. | 09:14 |
| pave1 | And I am working | 09:14 |
| pave1 | I am testing to release a CIP kernel based on v4.19.56 tomorrow. | 09:14 |
| sangorrin | great to know thanks | 09:14 |
| pave1 | (End of Nobuhiro's message) | 09:14 |
| gavinlai | next topic ? | 09:15 |
| gavinlai | #topic Kernel testing | 09:16 |
| *** brlogger changes topic to "Kernel testing (Meeting topic: CIP IRC weekly meeting)" | 09:16 | |
| patersonc | Hello | 09:16 |
| patersonc | #info lab-cip-mentor should be online today with a qemu machine. Mentor will then add the SIMATIC IPC227E and Cyclone V DE0-Nano-SoC boards. | 09:16 |
| patersonc | #info We now have direct access to the CIP LAVA master (thank you Ryan). This should make maintenance a bit easier. | 09:16 |
| patersonc | #info I've created a linux-cip-ci repository that creates build and test Docker containers that can be used to build the linux-cip Kernel using GitLab CI/CD. | 09:16 |
| patersonc | Example usage of this can be found in the chris/linux-4.19.y-cip-ci-test branch in the linux-cip GitLab repo. | 09:17 |
| patersonc | #link https://gitlab.com/cip-playground/linux-cip-ci | 09:17 |
| patersonc | There's a lot functionality to add, but I'll get there :) | 09:17 |
| patersonc | Discussion topics: | 09:17 |
| patersonc | 1) | 09:17 |
| patersonc | How would the Kernel maintainers like me to submit the gitlab-ci.yml files to the CIP Kernel branches? Mailing List or merge request in GitLab? | 09:17 |
| patersonc | Should they be applied to the kernel.org repo? Or just the GitLab mirror? | 09:17 |
| pave1 | Umm. | 09:18 |
| pave1 | I don't think we want them in kernel repositories. | 09:18 |
| pave1 | Those are normally not part of kernel sources, right? | 09:18 |
| patersonc | There aren't any in mainline, no | 09:19 |
| patersonc | Another option is I create a separate repo for the CI files | 09:19 |
| pave1 | We keep other stuff needed for build in cip-kernel-config repository. | 09:19 |
| pave1 | Would it be ok to put those there, too? | 09:19 |
| pave1 | git@gitlab.com:cip-project/cip-kernel/cip-kernel-config.git | 09:19 |
| patersonc | I can, but there is a special project type in GitLab that is just for CI/CD builds | 09:20 |
| sangorrin | The problem is that cross-triggers are paid | 09:20 |
| patersonc | I'll have an experiment | 09:21 |
| sangorrin | if you use two repositories (kernel and tests), then you can not trigger the tests when you push new patches to the kernel unless you have paid support (if i remember correctly) | 09:21 |
| pave1 | Ok, I don't really know how gitlab makes money. | 09:21 |
| sangorrin | btw, patersonc: the script has changed a lot!! | 09:21 |
| patersonc | sangorrin: Okay, I'll have a look | 09:22 |
| pave1 | So that means both separate repository and putting it in cip-kernel-config have this "give us money" disadvantage? | 09:22 |
| sangorrin | I think so, but patersonc says he will have a second look | 09:23 |
| patersonc | sangorrin: You might be right: https://docs.gitlab.com/ee/ci/ci_cd_for_external_repos/ | 09:23 |
| gavinlai | patersonc: any other discussion topics? | 09:23 |
| patersonc | Is CIP using a free account then? | 09:23 |
| pave1 | Aha, ok. In such case putting it into gitlab kernel sources would make some sense. | 09:23 |
| sangorrin | https://docs.gitlab.com/ee/ci/multi_project_pipelines.html | 09:23 |
| sangorrin | patersonc: I think so | 09:24 |
| patersonc | gavinlai: yes | 09:24 |
| patersonc | At the moment the backups from the LAVA master are only stored locally. | 09:24 |
| patersonc | Is it okay to create a private repository on CIP's GitLab pages to store the backups? | 09:25 |
| patersonc | Downside is that all of the CIP 'owners' will get access to keys etc. Does anyone have a problem with this? | 09:25 |
| pave1 | Or maybe... someone is running mirroring from kernel.org to gitlab. Maybe same script can trigger the CI? | 09:25 |
| patersonc | pave1: let's discuss after the meeting | 09:26 |
| pave1 | patersonc: ok | 09:26 |
| patersonc | Any objections to storing LAVA backups in CIP's GitLab account? | 09:27 |
| sangorrin | do you mean logs? | 09:27 |
| sangorrin | or the source code? | 09:27 |
| pave1 | Dunno. Storing keys in gitlab does not sound right. | 09:28 |
| patersonc | sangorrin: Everything | 09:28 |
| patersonc | Source is already in GitLab, but private details are just stored locally | 09:28 |
| Yamamodo | patersonc: do you mean you want to upload private detail to gilab for LAVA ? | 09:29 |
| pave1 | That sounds like a right design to me :-). I don't think 3rd parties should be trusted with private data. | 09:29 |
| patersonc | Yes. We'd make the repo private. | 09:29 |
| patersonc | Open to other ideas? | 09:30 |
| sangorrin | private repositories need a paid account though | 09:30 |
| patersonc | Not on Gitlab | 09:30 |
| sangorrin | oh | 09:30 |
| patersonc | It's github you have to pay | 09:30 |
| pave1 | If it is just a backup, encryption could help? | 09:31 |
| patersonc | True | 09:31 |
| patersonc | #info lab-cip-mentor is now up and running: https://lava.ciplatform.org/scheduler/worker/lab-cip-mentor | 09:32 |
| patersonc | (thank you mungaip ) | 09:33 |
| sangorrin | that was fast | 09:33 |
| sangorrin | good work | 09:33 |
| patersonc | I'll re-think where to store backups. | 09:33 |
| patersonc | gavinlai: let's move on | 09:33 |
| gavinlai | :) | 09:33 |
| gavinlai | #topic CIP Core | 09:34 |
| *** brlogger changes topic to "CIP Core (Meeting topic: CIP IRC weekly meeting)" | 09:34 | |
| patersonc | #action Work out a solution for LAVA master backups - patersonc | 09:34 |
| gavinlai | patersonc: thanks :) | 09:34 |
| kazu | I'm still updating the "Package decision process" based on the action items in the last meeting | 09:34 |
| kazu | would like to share its URL by the next Fri or Mon... | 09:34 |
| kazu | That's all from me, | 09:35 |
| kazu | daniel: do you have more? | 09:35 |
| sangorrin | iwg20m support in ISAR is about to occur | 09:35 |
| sangorrin | Jan is upgrading isar-cip-core to the latest ISAR | 09:35 |
| sangorrin | because that ISAR contains my bmap patch | 09:35 |
| sangorrin | once that upgrade is done, I want to deploy the isar-cip-core builds to AWS and trigger LAVA jobs | 09:36 |
| sangorrin | including RT kernels | 09:36 |
| gavinlai | next topic? | 09:37 |
| gavinlai | #topic Software update | 09:37 |
| *** brlogger changes topic to "Software update (Meeting topic: CIP IRC weekly meeting)" | 09:37 | |
| sangorrin | #info We are working on the PoC demo for OSSJ2019 | 09:37 |
| sangorrin | #info Raw image updates with Hawkbit on the server side working#info TODO: finish u-boot script, use librsync, rollback example | 09:38 |
| sangorrin | shit | 09:38 |
| sangorrin | #info Raw image updates with Hawkbit on the server side working | 09:38 |
| sangorrin | #info TODO: finish u-boot script, use librsync, rollback example | 09:38 |
| sangorrin | we have two more weeks '=( | 09:38 |
| patersonc | :) | 09:38 |
| gavinlai | sangorrin: thanks | 09:39 |
| gavinlai | #topic AOB | 09:39 |
| *** brlogger changes topic to "AOB (Meeting topic: CIP IRC weekly meeting)" | 09:39 | |
| gavinlai | #endmeeting | 09:41 |
| brlogger | Meeting ended Thu Jun 27 09:41:17 2019 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 09:41 |
| brlogger | Minutes: https://irclogs.baserock.org/meetings/cip/2019/06/cip.2019-06-27-09.00.html | 09:41 |
| brlogger | Minutes (text): https://irclogs.baserock.org/meetings/cip/2019/06/cip.2019-06-27-09.00.txt | 09:41 |
| brlogger | Log: https://irclogs.baserock.org/meetings/cip/2019/06/cip.2019-06-27-09.00.log.html | 09:41 |
| *** brlogger changes topic to "Civil Infrastructure Platform Project. Find the logs at https://irclogs.baserock.org/cip/" | 09:41 | |
| gavinlai | Thanks all! | 09:41 |
| sangorrin | #thanks | 09:41 |
| Yamamodo | thanks | 09:41 |
| pave1 | Thank you! | 09:41 |
| kazu | Thanks! | 09:41 |
| patersonc | Thanks | 09:41 |
| hiraku | Thanks | 09:42 |
| patersonc | Continuing the gitlab-ci.yml discussion.... | 09:42 |
| patersonc | I've just created a CI/CD only project. | 09:42 |
| patersonc | But what this actually does is create another mirror of the Kernel, then expect you to add the .gitlab-ci.yml files | 09:42 |
| patersonc | We may as well just add them to the mirror we already have :) | 09:43 |
| *** kazu has quit IRC | 09:43 | |
| patersonc | I thought that it would have just created a repo that only contains .gitlab-ci.yml files that get triggered by the repo it's concerned with, but I was wrong | 09:44 |
| pave1 | Well, if it works separate, I'd keep it separate like that. | 09:44 |
| pave1 | Overhead is on gitlab side, and I guess it can share the repository data just fine, so... | 09:44 |
| patersonc | It just means having two almost identical mirrors of linux-cip | 09:44 |
| patersonc | It also means you won't get the nice green ticks on commits in the repo that everyone uses | 09:45 |
| sangorrin | I think the gitlab mirror | 09:45 |
| sangorrin | is something where we can experiment | 09:45 |
| patersonc | Green tick e.g.: https://gitlab.com/cip-project/cip-kernel/linux-cip/commits/chris/linux-4.19.y-cip-ci-test | 09:46 |
| sangorrin | people should clone the kernel from kernel.org anyway | 09:46 |
| patersonc | sangorrin: Yep | 09:46 |
| sangorrin | btw multi-project triggers would be $19/month (silver) | 09:46 |
| sangorrin | https://about.gitlab.com/pricing/ | 09:46 |
| pave1 | Ok, so I use the kernel.org repo, so I don't care that much. I'd expect linux-cip to be same on kernel.org and gitlab... | 09:47 |
| patersonc | sangorrin: I'm not 100% sure if we need that though. We can set up triggers that tell different CI builds to start based on the outcome of others. | 09:47 |
| pave1 | But I used your link and don't see green ticks anyway. | 09:47 |
| patersonc | pave1: Try https://gitlab.com/cip-project/cip-kernel/linux-cip/pipelines/68279776 | 09:48 |
| pave1 | patersonc: We do syncing between kernel.org and gitlab "manually", right? It should be possible to fire any builds at that moment. | 09:48 |
| patersonc | pave1: Nope. GitLab mirrors automatically once an hour I think | 09:48 |
| pave1 | Aha, ok. | 09:48 |
| patersonc | Or once every 15 minutes | 09:49 |
| sangorrin | To mirror means we can't have a yaml file there, doesn't it? | 09:49 |
| patersonc | That I'm not sure on until we give it a go | 09:49 |
| sangorrin | I mean, wouldn't it get overwritten every time? | 09:49 |
| pave1 | Aha, I logged in, and now I seen the green ticks. | 09:49 |
| patersonc | The mirror hasn't seen any issues with having separate branches | 09:50 |
| patersonc | pave1: Would you be willing to do a quick test? | 09:50 |
| sangorrin | ok, please check then | 09:50 |
| patersonc | pave1: Or I can | 09:50 |
| patersonc | We should cherry-pick 74ace00414d13123e9a5d76284605750318cfd6a to linux-4.19.y-cip | 09:51 |
| patersonc | And then trigger the mirror to run an update | 09:51 |
| sangorrin | Do we want a test per commit? | 09:51 |
| sangorrin | build + LAVA test | 09:52 |
| pave1 | patersonc: You can. | 09:52 |
| pave1 | sangorrin: Don't do that. | 09:52 |
| patersonc | pave1: Thanks. | 09:52 |
| patersonc | sangorrin: I think it'll end up being a test per update from the mirror | 09:52 |
| patersonc | (if there are new commits) | 09:52 |
| patersonc | So if 20 patches are merged at the same time, only 1 build/test job is run | 09:53 |
| pave1 | Greg likes to cherry-pick mainline commit, even when it is known buggy, and then merge a fix on top. | 09:53 |
| pave1 | If you do tests per commit, you'll get some failures. | 09:54 |
| sangorrin | I see, then probably we only need to test every release | 09:54 |
| sangorrin | when a new stable tag is released | 09:54 |
| sangorrin | or a new cip tag is released | 09:54 |
| pave1 | Every update from mirror seems right. | 09:54 |
| patersonc | I'll just try | 09:54 |
| pave1 | I push a change from mailing list, it is nice for it to be tested... before it is tagged etc. | 09:55 |
| sangorrin | the thing is that when we start doing LTP and other tests, each test job will require more and more time | 09:55 |
| patersonc | pave1: This could be done on a new branch? | 09:55 |
| sangorrin | so 1 test per hour might be too quick | 09:55 |
| pave1 | Normally the pushes are not that common. | 09:56 |
| patersonc | See https://gitlab.com/cip-project/cip-kernel/linux-cip/commits/linux-4.19.y-cip | 09:56 |
| patersonc | "This project is mirrored from https://git.kernel.org/pub/scm/linux/kernel/git/cip/linux-cip.git. Updated 18 minutes ago. This branch has diverged from upstream." | 09:56 |
| patersonc | So maybe it is an issue | 09:56 |
| pave1 | sangorrin: If there are already tests running, just wait till the lab is available again, then test whatever is current. | 09:58 |
| sangorrin | well, it has diverged because you added the yaml file, but does it affect the mirroring? | 09:58 |
| patersonc | pave1: You're right. It's the pushes to the cip branches that matter. And I assume you guys normally update to the latest LTS in one hit? | 09:58 |
| patersonc | Not sure we'll find out unless the branch changes on kernel.org | 09:59 |
| patersonc | Pavel, could you add a test commit? Can we force push later to remove? I assume no one would mind/notice? | 09:59 |
| sangorrin | Sorry, I have to leave my work now. | 09:59 |
| sangorrin | I hope you have a good discussion | 09:59 |
| sangorrin | bye | 09:59 |
| patersonc | sangorrin: Okay :) Safe journey home | 10:00 |
| pave1 | patersonc: yes, we do. So if tests are run per push, it should be fine. | 10:00 |
| *** sangorrin has quit IRC | 10:00 | |
| patersonc | pave1: Would you be able to push a quick test commit to the kernel.org repo? | 10:01 |
| pave1 | patersonc: I'm thinking if I'm comfortable doing that. | 10:02 |
| patersonc | on the linux-4.19.y-cip branch? | 10:02 |
| patersonc | pave1: Okay :) | 10:02 |
| patersonc | We could set up a completely new branch if you'd prefer? | 10:02 |
| pave1 | What about this one? | 10:02 |
| patersonc | Rather than pollute our 'release' branch? | 10:02 |
| pave1 | linux-4.19.y-cip-rt-rebase | 10:02 |
| patersonc | Can do. I'll need to add my commit to the gitlab repo first. | 10:03 |
| pave1 | Ok. | 10:03 |
| patersonc | Okay, done: https://gitlab.com/cip-project/cip-kernel/linux-cip/commits/linux-4.19.y-cip-rt-rebase | 10:07 |
| patersonc | Ready for you to push a change to kernel.org on that branch | 10:07 |
| patersonc | Then we'll sync the mirror | 10:07 |
| pave1 | Okay, let me push | 10:07 |
| pave1 | ... | 10:11 |
| pave1 | ...this is going slower than I expected. | 10:13 |
| patersonc | ha | 10:13 |
| pave1 | And I got rejected, because this is not fast forward. Ok. | 10:15 |
| patersonc | Maybe it safer to use a new non-protected branch? | 10:17 |
| pave1 | Maybe I should have created branch for testing after all, but that would not have helped with your test. | 10:21 |
| pave1 | Done. | 10:21 |
| *** vidda has quit IRC | 10:22 | |
| patersonc | What's it called? | 10:22 |
| pave1 | No, no branch for now. | 10:23 |
| pave1 | I have pushed the test changes to linux-4.19.y-cip-rt-rebase | 10:23 |
| pave1 | You can trigger the mirror now. | 10:23 |
| patersonc | Ah okay | 10:23 |
| patersonc | Update in progress | 10:25 |
| patersonc | Looks like it doesn't work | 10:26 |
| pave1 | You'll need to better describe what doesn't work. | 10:26 |
| pave1 | :-). | 10:26 |
| patersonc | Sorry! | 10:26 |
| patersonc | Your new commits don't get added to the gitlab mirror | 10:27 |
| patersonc | https://git.kernel.org/pub/scm/linux/kernel/git/cip/linux-cip.git/log/?h=linux-4.19.y-cip-rt-rebase | 10:27 |
| patersonc | I.e. my patch adding the .gitlab-ci.yml breaks the mirroring | 10:27 |
| patersonc | So it's not an option | 10:27 |
| pave1 | Ok, do you know how to unbreak the mirroring? | 10:28 |
| patersonc | I'll remove my file | 10:28 |
| pave1 | And we can still use second repository you set up for testing, right? I liked it better, anyway. | 10:28 |
| pave1 | You'll probably need to force-update to the older head. | 10:29 |
| patersonc | Yep | 10:29 |
| pave1 | Anything other to try? | 10:30 |
| patersonc | The only other options are to have a new GitLab repo, or to put the .gitlab-ci files into the kernel.org repo | 10:31 |
| patersonc | If maintainers want to CI test their code they will have to push branches to the new gitlab testing mirror. I guess this is okay? | 10:32 |
| pave1 | Ok, I'm not sure if I want to decide that. | 10:32 |
| pave1 | Here and now. | 10:32 |
| patersonc | That's okay | 10:33 |
| patersonc | I'll experiment and ask the ML | 10:33 |
| patersonc | Thank you for your help | 10:33 |
| pave1 | Neither option is obviously bad. And we may want to add some files to our repositories, anyway. | 10:33 |
| pave1 | Now... can I get you to test linux-4.19.y-cip-rt-rebase branch we played with? :-) | 10:33 |
| patersonc | True | 10:34 |
| patersonc | The other option is that we manually mirror/merge to the gitlab repo, but that's not ideal. | 10:34 |
| patersonc | :D | 10:34 |
| pave1 | Yes, it would be nice to automate it if we can. Other problem I see is that the .gitlab-ci file will be "all over" our repositories, so changing it will not be easy. | 10:37 |
| pave1 | Will it need to be updated often? | 10:38 |
| patersonc | Hopefully not. Only when we want to start testing new configurations | 10:38 |
| patersonc | https://gitlab.com/cip-project/cip-kernel/linux-cip/commits/linux-4.19.y-cip-rt-rebase is fixed now | 10:39 |
| pave1 | One more to go.. | 10:43 |
| patersonc | https://docs.gitlab.com/ee/ci/ci_cd_for_external_repos/ | 10:44 |
| patersonc | I've had a go with the above, but I get the same issues. As soon as the .gitlab-ci file is added the mirroring breaks | 10:45 |
| pave1 | I see https://gitlab.com/cip-project/cip-kernel/linux-cip/commits/linux-4.19.y-cip is now fixed, too? | 10:51 |
| pave1 | When you havetime, I'd like to talk about rt testing :-). | 10:52 |
| patersonc | Sure | 10:53 |
| patersonc | What tests do you want to run? | 10:53 |
| pave1 | cyclictest would be a good start. | 10:54 |
| pave1 | linux-4.19.y-cip-rt-rebase branch contains current code. | 10:55 |
| patersonc | Do you have a config file I can use? Or have you already built the Kernel? | 10:55 |
| pave1 | I don't have a kernel build for the boards in the lab. | 10:57 |
| pave1 | defconfig from cip-kernel-config repository would be fine. | 10:57 |
| pave1 | Plus CONFIG_PREEMPT_RT_FULL=y needs to be enabled to really enable the realtime. | 10:58 |
| patersonc | Aren't there more config changes than that for building RT though? | 10:58 |
| pave1 | You can just hold enter for the rest :-). | 10:58 |
| pave1 | But I see... we really should have suitable configs stored somewhere. | 10:59 |
| patersonc | Then I can grab them in the CI testing | 11:00 |
| pave1 | I don't understand the comment, but whatever config you get, it should be fine. | 11:01 |
| patersonc | I mean, it's easier for me to just use a config from a repo, then manually have to configure the Kernel it on the fly | 11:03 |
| pave1 | Can you do it manually, or do you want me to produce something? | 11:06 |
| patersonc | Maybe this time you could build the Kernel for me for the Renesas boards. By the next release we'll have CI going... | 11:06 |
| pave1 | Ok, let me try :-). | 11:07 |
| pave1 | It will take few minutes, maybe half an hour. | 11:09 |
| patersonc | No worries. I'm still trying to work out what to do with these .gitlab-ci.yml files :) | 11:09 |
| pave1 | I guess I can use cip-kernel-config/4.19/arm64/renesas_defconfig as a base? | 11:10 |
| patersonc | Yep | 11:11 |
| patersonc | Or just the arm64 defconfig | 11:11 |
| pave1 | I worked with renesas_defconfig before, let my try that one. | 11:14 |
| pave1 | What kind of images do you need for testing? | 11:15 |
| patersonc | We usually use Image for arm64 | 11:15 |
| pave1 | Ok, I can do that. And what dtb? | 11:16 |
| patersonc | r8a774c0-ek874.dtb | 11:16 |
| pave1 | Ok :-). | 11:16 |
| patersonc | The more I look into it, the more I think adding our .gitlab-ci.yml files to the kernel.org repo is the best option | 11:19 |
| patersonc | I'll create some patches and submit them to the ML for review there. | 11:22 |
| pave1 | Yep, that's probably best way. It is possible that placing .yml on kernel.org is easiest. | 11:24 |
| patersonc | Yea | 11:24 |
| patersonc | And is easily ignored by people who don't care about it :) | 11:24 |
| patersonc | pave1: What RFS needs to be used for the RT tests? | 11:33 |
| pave1 | No idea :-(. | 11:34 |
| pave1 | Ok, could you attempt cyclictest on images from http://atrey.karlin.mff.cuni.cz/~pavel/outgoing/cip/ ? | 11:34 |
| pave1 | It should fail, but I'd like to see failing test, first. | 11:35 |
| patersonc | Okay | 11:45 |
| patersonc | https://lava.ciplatform.org/scheduler/job/1533 | 11:46 |
| patersonc | It's just waiting for another test to run | 11:46 |
| patersonc | Btw, this is the result of running cyclictest on v4.19 non-RT Kernel, without any RT config options, on an arm32 board: https://lava.ciplatform.org/results/1530/0_cyclictest | 11:50 |
| *** Yamamodo has quit IRC | 11:50 | |
| patersonc | pave1: Your job is just starting | 11:54 |
| patersonc | Doh, it timed out. But it was working though. | 12:04 |
| *** Yamamodo has joined #cip | 12:13 | |
| patersonc | New job with larger timeouts: https://aws.amazon.com/ec2/instance-types/ | 12:14 |
| patersonc | ooops | 12:15 |
| patersonc | https://lava.ciplatform.org/scheduler/job/1535 | 12:15 |
| pave1 | Ok, so it is 60usec average, 1000usec maximum. Not really a realtime ... as expected. | 12:15 |
| pave1 | Let me update the binaries with realtime ones. | 12:15 |
| pave1 | Can you kill/restart it so it uses new binaries? | 12:16 |
| patersonc | Have you overwritten them in your server? | 12:17 |
| pave1 | Yes. | 12:17 |
| patersonc | Okay: https://lava.ciplatform.org/scheduler/job/1537 | 12:17 |
| pave1 | Thanks a lot! | 12:18 |
| patersonc | Not looking as good? | 12:25 |
| patersonc | pave1: Btw, I think you have a login to LAVA if you want to try yourself? Just select "resubmit" and amend the definition as required. | 12:27 |
| patersonc | Latest results: https://lava.ciplatform.org/results/1537/0_cyclictest | 12:27 |
| pave1 | Not looking good at all :-(. | 12:31 |
| pave1 | Thanks, I should be to re-run it myself. | 12:32 |
| *** Yamamodo has quit IRC | 12:34 | |
| patersonc | Let me know if you get stuck | 12:37 |
| *** Yamamodo has joined #cip | 12:41 | |
| *** Yamamodo has quit IRC | 13:26 | |
| patersonc | pave1: How's it going? | 13:29 |
| *** Yamamodo has joined #cip | 13:36 | |
| pave1 | Not good. | 13:39 |
| pave1 | My kernel is not realtime, and when I try "official" rt 4.19.13-cip1-rt1, it does not boot. https://lava.ciplatform.org/scheduler/job/1541 | 13:40 |
| patersonc | Strange | 13:41 |
| patersonc | I assume that the RT kernel isn't changing the debug serial output port for some reason? | 13:42 |
| pave1 | I'll find out :-). | 13:44 |
| pave1 | Question: is it possible to abort running job? | 13:44 |
| patersonc | Yes | 13:44 |
| patersonc | There is a cancel button at the top right | 13:44 |
| pave1 | I don't see it. I have "Actions", but it does not contain cancel. | 13:45 |
| pave1 | Aha, it finished. Let me retry. | 13:45 |
| patersonc | Should be a yellow button next to "resubmit" | 13:45 |
| pave1 | I see it now. Seems like it disappears after job finishes. | 13:50 |
| patersonc | Yep | 13:51 |
| pave1 | Hmm. Does 4.19.13-cip1 have enough support to boot on that board? | 13:54 |
| patersonc | Yep | 13:55 |
| patersonc | Oh shit | 13:56 |
| patersonc | No | 13:56 |
| patersonc | Not cip1 | 13:56 |
| patersonc | Sorry, I didn't realise the version you're using | 13:56 |
| pave1 | No problem, it is an old one. | 13:56 |
| pave1 | That solves the mystery. | 13:56 |
| patersonc | :) | 13:57 |
| *** Yamamodo has quit IRC | 13:59 | |
| *** Yamamodo has joined #cip | 14:00 | |
| pave1 | Is there another board we could use easily and which would work in 4.19.13-cip1 ? | 14:01 |
| patersonc | the iwg20m | 14:02 |
| patersonc | It's an arm board rather than arm64 | 14:02 |
| patersonc | https://gitlab.com/cip-project/cip-kernel/cip-kernel-config/blob/master/4.19/arm/renesas_shmobile_defconfig | 14:03 |
| patersonc | Example job with cyclictest: https://lava.ciplatform.org/scheduler/job/1530 | 14:04 |
| patersonc | I have to head home now pave1 | 14:04 |
| pave1 | Yes, thanks for all the help! | 14:04 |
| patersonc | Will be back bright and early tomorrow! Good luck :) | 14:04 |
| patersonc | Email me if you need anything, I'll check later tonight | 14:04 |
| pave1 | I don't expect to do much more today. Thanks! | 14:05 |
| patersonc | There are two iwg20m boards in the lab btw, so testing should be quicker ;) | 14:05 |
| pave1 | I'm not sure if I can manage two boards at once. I'll see. | 14:05 |
| patersonc | :) | 14:11 |
| patersonc | ttfn | 14:11 |
| pave1 | see you! | 14:11 |
| *** Yamamodo has quit IRC | 14:22 | |
| *** Yamamodo has joined #cip | 14:30 | |
| *** pave1 has quit IRC | 15:07 | |
| *** Yamamodo has quit IRC | 15:15 | |
| *** Yamamodo has joined #cip | 15:24 | |
| *** Yamamodo has quit IRC | 15:48 | |
| *** Yamamodo has joined #cip | 15:52 | |
| *** Yamamodo has quit IRC | 16:15 | |
| *** Yamamodo has joined #cip | 16:20 | |
| *** toscalix has quit IRC | 16:24 | |
| *** Yamamodo has quit IRC | 17:09 | |
| *** Yamamodo has joined #cip | 17:17 | |
| *** OutBackDingo has quit IRC | 17:51 | |
| *** OutBackDingo has joined #cip | 17:54 | |
| *** Yamamodo has quit IRC | 18:03 | |
| *** Yamamodo has joined #cip | 18:08 | |
| *** Yamamodo has quit IRC | 18:31 | |
| *** Yamamodo has joined #cip | 18:40 | |
| *** Yamamodo has quit IRC | 19:04 | |
| *** Yamamodo has joined #cip | 19:10 | |
| *** Yamamodo has quit IRC | 19:32 | |
| *** Yamamodo has joined #cip | 19:38 | |
| *** Yamamodo has quit IRC | 20:00 | |
| *** hiraku has quit IRC | 20:04 | |
| *** Yamamodo has joined #cip | 20:11 | |
| *** pave1 has joined #cip | 20:29 | |
| *** Yamamodo has quit IRC | 20:35 | |
| *** Yamamodo has joined #cip | 20:39 | |
| *** Yamamodo has quit IRC | 21:01 | |
| *** Yamamodo has joined #cip | 21:03 | |
| *** Yamamodo has quit IRC | 21:27 | |
| *** Yamamodo has joined #cip | 21:35 | |
| *** Yamamodo has quit IRC | 21:58 | |
| *** Yamamodo has joined #cip | 22:03 | |
| *** Yamamodo has quit IRC | 22:25 | |
| *** pave1 has quit IRC | 22:28 | |
| *** Yamamodo has joined #cip | 22:31 | |
| *** rajm has quit IRC | 22:48 | |
| *** Yamamodo has quit IRC | 22:53 | |
| *** Yamamodo has joined #cip | 22:59 | |
| *** Yamamodo has quit IRC | 23:21 | |
| *** Yamamodo has joined #cip | 23:26 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!