| *** fujita[m] has quit IRC | 00:11 | |
| *** fujita[m] has joined #cip | 00:12 | |
| *** dl9pf has quit IRC | 00:39 | |
| *** dl9pf has joined #cip | 00:40 | |
| *** dl9pf has joined #cip | 00:40 | |
| *** tpollard has quit IRC | 01:07 | |
| *** iwamatsu has quit IRC | 07:00 | |
| *** rajm has joined #cip | 07:35 | |
| *** toscalix_ has joined #cip | 08:15 | |
| *** tpollard has joined #cip | 09:35 | |
| patersonc | wagi: ping | 10:18 |
|---|---|---|
| *** tpollard has quit IRC | 10:24 | |
| *** tpollard has joined #cip | 10:29 | |
| wagi | patersonc: yes I am around | 10:29 |
| patersonc | :) | 10:31 |
| patersonc | Do you know the upstreaming status of the IPC227E board? | 10:31 |
| patersonc | Is it supported in mainline or CIP v4.4/v4.19 Kernels? | 10:31 |
| wagi | No, I don't know the exact status but I am pretty sure it is working out of the box with mainline. It's just a normal x86 box. The only thing I suppose wont work directly is the profinet stuff | 10:32 |
| wagi | JanK normaly knows best. I can ping him when I am the office. Still home office due my sickness | 10:33 |
| patersonc | Thank you! | 10:35 |
| patersonc | I hope you feel better soon. | 10:35 |
| wagi | It's getting better. Last week I went to early back (CIP meeting) which resulted in a second infection | 10:36 |
| patersonc | :( | 10:36 |
| wagi | So I decided to stay at home until I am really cured | 10:36 |
| wagi | It's nice working at home | 10:37 |
| patersonc | Indeed | 10:37 |
| patersonc | I'll email cip-members about IPC227E support | 10:38 |
| wagi | That's a good idea | 10:38 |
| wagi | BTW: https://github.com/igaw/test-definitions/tree/preempt-rt | 10:38 |
| patersonc | :) | 10:38 |
| wagi | I fixed up the exting tests to report latency misses | 10:38 |
| wagi | I would like to add some more tests in the future | 10:39 |
| patersonc | I'll try and take a look at some point today. Thank you for sharing. | 10:39 |
| wagi | And obviously jitterdebugger | 10:39 |
| wagi | Thanks, I suppose the shell/python code is far perfect | 10:40 |
| wagi | Also the last patch (jitterdebugger) needs some more thinking on the workload | 10:45 |
| wagi | I added the '-c/--command' option to jitterdebugger which starts a background load | 10:46 |
| wagi | one of the background load is 'compile stuff' | 10:46 |
| *** tpollard has quit IRC | 10:46 | |
| wagi | so the filesystem needs sources installed to a toolchain | 10:46 |
| wagi | currently the code just knows where the stuff is | 10:47 |
| patersonc | Okay | 10:47 |
| wagi | this needs some more tweaking | 10:47 |
| wagi | also I figured, it would make sense to restart any background tasks if it terminates before jitterdebugger stops | 10:47 |
| wagi | so we could avoid 'while true; do hackbench; done' | 10:47 |
| wagi | furthermore, I think it would make sense to write the stdout/stderr of the background tasks directly into file | 10:48 |
| wagi | avoid ">/dev/null" stuff | 10:48 |
| wagi | and last thing, allowing multiple -c would also be really nice | 10:48 |
| wagi | don't know if that is a feature fest or a feature creep thing :) | 10:49 |
| wagi | I would like to write something like | 10:49 |
| patersonc | :D | 10:50 |
| wagi | jitterdebugger -c hackbench -c 'make -C ~/src -j4' | 10:50 |
| wagi | this documents exactly what is happening | 10:50 |
| wagi | well, maybe the compile thing needs to be put into a shell script though | 10:51 |
| wagi | If you have any input I am glad to hear it | 10:52 |
| *** tpollard has joined #cip | 10:55 | |
| wagi | Next thing I add is a testplan which combines everything | 10:57 |
| wagi | Then we try to run it on the cip lava instance | 10:57 |
| wagi | There is just one thing I am not confused | 10:58 |
| wagi | For most test we need to hand in the max latency value | 10:59 |
| wagi | which is device and kernel specific | 10:59 |
| wagi | not sure if that works out with the testplan | 10:59 |
| wagi | Ha! I just saw overlays are possible! | 11:01 |
| patersonc | Tests can report a measurement back to LAVA, such as latency | 11:05 |
| patersonc | Test jobs are submitted per-board, so we'd know what board & s/w was used for the said latency result | 11:06 |
| patersonc | Here's an example: https://lava.ciplatform.org/scheduler/job/613 | 11:07 |
| patersonc | Results: https://lava.ciplatform.org/results/613/0_cyclictest | 11:07 |
| patersonc | Note that the above wasn't run on an -rt Kernel ;) | 11:08 |
| *** tpollard has quit IRC | 12:25 | |
| *** tpollard has joined #cip | 12:30 | |
| *** toscalix_ has quit IRC | 13:20 | |
| *** tpollard has quit IRC | 13:34 | |
| *** toscalix has joined #cip | 14:00 | |
| wagi | patersonc: do you know how to write a job which runs a test plan? | 14:08 |
| wagi | I must miss something. Running the test plan with test-runner works | 14:09 |
| wagi | but my job doesn't run the test plan at all | 14:09 |
| patersonc | Is your job much different to https://lava.ciplatform.org/scheduler/job/613/definition? | 14:10 |
| patersonc | (obviously you'll be linking to a different test definition) | 14:10 |
| patersonc | Here is a link to the actual test plan: https://github.com/igaw/test-definitions/blob/master/automated/linux/cyclictest/cyclictest.yaml | 14:12 |
| wagi | not really | 14:12 |
| wagi | I though that is a test case | 14:13 |
| wagi | I was referring to: https://github.com/igaw/test-definitions/blob/preempt-rt/plans/rt/rt-performance.yaml | 14:13 |
| patersonc | Ah, yes | 14:13 |
| wagi | I want to run that one | 14:13 |
| wagi | test-runner -p plans/rt/rt-performance.yaml -g root@c2d | 14:13 |
| wagi | works | 14:14 |
| patersonc | Okay I'll have a look | 14:14 |
| wagi | but i get just an empty run when doing it via the usual test block | 14:14 |
| wagi | (path, repository, from, branch, name, ...) | 14:14 |
| wagi | I see that the agl guys are using it as well | 14:14 |
| wagi | http://docs.automotivelinux.org/master/docs/devguides/en/dev/reference/CIAT/training/Hands_on_lab_documentation.pdf | 14:15 |
| wagi | there must be an error in my configuration | 14:15 |
| wagi | https://pastebin.com/raw/Tq2w9Bcg | 14:16 |
| wagi | this is my job | 14:16 |
| wagi | A normal test is executed when added to the test block | 14:17 |
| wagi | patersonc: if you don't know it, it's okay, don't waste your time | 14:20 |
| patersonc | It looks okay to me... | 14:20 |
| patersonc | You could try running the same on the CIP lab so we can see the logs etc | 14:21 |
| patersonc | You've have to change the device though | 14:21 |
| wagi | https://pastebin.com/raw/MRB9wzRA | 14:21 |
| wagi | I would need to upload the rootfs | 14:21 |
| wagi | I still have a slow uplink | 14:22 |
| wagi | Right now, they are attaching my house to fibre | 14:22 |
| wagi | I feel the vibrations from the heavy machines outside :D | 14:22 |
| patersonc | Happy days | 14:23 |
| wagi | so in the log, it looks like the test plan is executed as normal test case | 14:23 |
| wagi | :D | 14:23 |
| wagi | We were waiting for this for 3 years | 14:23 |
| patersonc | Your test plan isn't specifying the repo branch. | 14:26 |
| patersonc | And it looks like the master branch of your test-definitions doesn't have jitterdebugger | 14:27 |
| wagi | ahaahaaaahaaaaaaaaa | 14:28 |
| wagi | I think the test-defintiont repos is over defining stuff | 14:30 |
| patersonc | Rather then a test plan, you can just call multiple tests from the overall test job definition. | 14:31 |
| wagi | and also mixing hardware configuraiton into test definition | 14:31 |
| wagi | Yes that works but I would like to group things together | 14:31 |
| patersonc | Most of the meta data in the test case definitions isn't used | 14:32 |
| patersonc | e.g. devices | 14:32 |
| patersonc | Okay I understand. Grouping would be useful. I haven't tried a test plan though tbh. | 14:33 |
| wagi | the branch was not the problem :( | 14:33 |
| patersonc | I didn't think it would be, given the other test cases were in the master branch ;) | 14:34 |
| patersonc | You could drop mungaip an email. He may have a better idea than me | 14:34 |
| wagi | /usr/bin/git clone -b preempt-rt https://github.com/igaw/test-definitions.git /var/lib/lava/dispatcher/tmp/123/lava-overlay-d7vtg4jp/lava-123/0/tests/0_rt-performance --depth=1 | 14:34 |
| wagi | so this looks good so far | 14:36 |
| wagi | yeah, I dig a bit further maybe I see the problem | 14:36 |
| *** tpollard has joined #cip | 14:42 | |
| *** apteryx has joined #cip | 15:13 | |
| apteryx | hi! | 15:13 |
| apteryx | I'm reviewing the CIP project to know if it'd be suitable as a source of stable, CVE-patched base components (toolchain, kernel, etc.). | 15:13 |
| apteryx | Where can I learn if a CVE has been patched in the project currently? I don't see commit activities regarding addressing of CVEs. | 15:14 |
| patersonc | Hello apteryx | 15:15 |
| apteryx | patersonc: hi :-) | 15:16 |
| patersonc | bwh has created a CVE tracker that can check for patches CVEs: https://gitlab.com/cip-project/cip-kernel/cip-kernel-sec | 15:16 |
| patersonc | Using this tool bwh has backported a number of CVE fixes to the LTS v4.4 and CIP v4.4 Kernels | 15:17 |
| patersonc | wrt other components, most of them are supported by the Debian LTS project, which CIP contributes to. | 15:18 |
| patersonc | Perhaps send an email to cip-dev@lists.cip-project.org where people more in the know then me can provide more information. | 15:19 |
| apteryx | patersonc: I see! Thanks for the pointers :-) | 15:23 |
| patersonc | You're welcome. | 15:23 |
| patersonc | Let us know if you have any other queries! | 15:24 |
| wagi | patersonc: https://validation.linaro.org/static/docs/v2/results-intro.html#results-test-suite | 15:27 |
| wagi | it looks like suits are defined in the job | 15:27 |
| *** bwh has joined #cip | 15:28 | |
| toscalix | apteryx: we patch upstream first | 16:10 |
| *** toscalix has quit IRC | 17:14 | |
| *** tpollard has quit IRC | 20:34 | |
| *** tpollard has joined #cip | 21:29 | |
| *** rajm has quit IRC | 22:44 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!