Created
June 17, 2019 15:01
-
-
Save danrue/e6b2a2f96458d24276abae54a81dbbd9 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
PRESENT | |
- Tim Bird - Sony - Fuego project | |
- Major Hayden - CKI Project @ Red Hat | |
- Dave Pigott - Linaro - Test lab tech lead | |
- Michal Simek - Xilinx | |
- Chris Paterson - CIP Project / Renesas | |
- Kishore - Bosch India | |
- Charles Oliveira - Linaro - Squad | |
- Dan Rue - LKFT @ Linaro, kernelCI | |
- Mark Brown | |
- Remi Duraffort - LAVA @ Linaro | |
- Vishal | |
- Daniel Sangorrin - Toshiba | |
- Maria Hogberg - Linaro | |
- Sasha Levin- Microsoft, Stable kernel | |
- Daniel Diaz - LKFT @Linaro | |
- Matt Hart - Linaro, KernelCI, PDU | |
- Luca Di Stefano - Linaro | |
- Anders Roxell - Linaro | |
MINUTES | |
- Time isn’t that great for PST. Matty to reach out to KH to bring up this | |
point on the ML | |
- [matt] I asked Kevin to discuss on the ML | |
- [drue] i think it’s as good as it can be without splitting into two | |
meetings | |
- [tim] was hoping for Kevin Hilman to be on the call, who would prefer a | |
better time for west coast US | |
- [dave] we could alternate the meetings between early and late | |
- [tim] good minutes really help. +1 on alternating meetings | |
- [maria] how to send minutes to the list | |
- [tim] conversion processes slow things down | |
- [dan] it's useful to send to list as plain text | |
- [dan] willing to do conversion, allows document to stay mostly read-only | |
- [mark] do both, so people can easily see | |
- BigQuery Test Result Aggregation Demo - Tom Gall / Dan Rue | |
- [dan] Tom traveling, demo not available for call | |
- Attributes: BigQuery has some useful attributes | |
- Free to use | |
- data may be difficult to get out for visualization | |
- is mostly just SQL | |
- [maria] Can each project do it? | |
- [dan] yes | |
- [dan] how to use the data is an open question | |
- [tim] related to the thread on bug reporting on kernel summit list. Need to | |
do two things: | |
- Browse data | |
- Generate reports | |
- [tim] it’s a good experiment, even if it’s not useful, we may learn | |
valuable things | |
- [tim] biggest risk is we put the data there and nobody uses it | |
- PDU/USB/Relay universal control - Dave P | |
- Haven’t converted the doc to rst yet | |
- Actually started, but doing it in md if no objections | |
- [matt] Looks nice in github - approved :) | |
- [Dave P] Can someone please add this doc link to the meeting notification? | |
- [tim] good place on wiki for link is: https://elinux.org/Test_Standards | |
- Revisiting test name standards discussion from last meeting | |
- [tim] i would be happy to have other people interested in this, but I’m not | |
sure how to move forward | |
- AI - Tim to start a document on test names | |
- [tim] test definitions | |
- [tim] I plan to go back and maybe propose a set of fields for test | |
definitions as a starting point | |
- [drue] anyone interested in trying to add their data to bigquery? | |
- Major raised his hand and Tim volunteered Daniel :) | |
- [major] we (CKI) have been looking for a way to make our data easily | |
available to others with a standard format/structure | |
- [daniel] a proxy library would be helpful to convert projects’ json into | |
sql | |
- [dan] - will talk to Tom and propose something on the automated testing | |
list | |
- [mark] - people at ARM also maybe interested | |
- [tim] Bug reporting thread on kernel summit / syzbot | |
- https://lists.linuxfoundation.org/pipermail/ksummit-discuss/2019-May/006389.html | |
- (be sure to follow thread into June) | |
- Subject: Squashing bugs! | |
https://lists.linuxfoundation.org/pipermail/ksummit-discuss/2019-June/thread.html | |
- [drue] LKFT has poor reporting and would love to tie into something | |
- [mark] bug trackers full of noise or not updated regularly are a problem | |
- Conferences updates? | |
- [tim] ATS is coming along. Have two sponsors so far | |
- CFP and registration should go live in a week or two | |
- Plumbers and CKI hackfest | |
- [major] working on getting a place to host it still. Proposed agenda was | |
posted to the mailing list | |
- [tim] what is the format? | |
- [major] i think it will be pretty open. We don’t want to monopolize | |
the agenda. Not necessarily slides/presentations. More brainstorming | |
and hacking | |
- [drue] Agenda management for this meeting.. How to get items added before | |
hand? | |
- [maria] i requested items before hand but did not receive many. I could try | |
getting minutes and reminders out which might help. I’ll send a meeting | |
reminder a week ahead of the meeting. Any ideas/suggestions to make this | |
meeting more valuable please speak up. | |
- [Michal] from embedded perspective, on driver testing, would like to be able | |
to validate a driver before sending upstream | |
- [michal] Space to put together test cases, to verify a driver matches spec. | |
There are some tests available in LAVA for example, but it’s not clear what | |
needs to be run. | |
- [maria] is this related to test standards, so that you can tag a test for a | |
particular requirement (similar to naming) | |
- [michal] i just want to know what i have to run before a driver is | |
considered validated. If there’s a bug, there should be a test | |
- [tim] for each driver type/class/subsystem, there should be a kselftest. | |
There are subsystems that have tests that gate code. Best place for this is | |
kselftest | |
- [mark] i don’t know that kselftest is the best place for everything, for | |
example graphics and v4l, partly for historical reasons and partly because | |
of their dependencies | |
- Problem becomes how do you test something on hardware that is hard to | |
test without destroying hardware (power regulator) | |
- [tim] need to define the api between test rig and test code. If it requires | |
hardware, it’s not standardized (not industry standard for hw testing). So | |
tests are not portable between lab setups | |
- [tim] for example power regulators, if you want to do it you have to have | |
a spec or convention that you can pass to | |
- [mark] with my subsystem maintainer hat on - i don’t care if your driver | |
works, i care if it makes it difficult to maintain the subsystem | |
- [mark] i think your main point is if there is some requirement, you should | |
be able to find it | |
- [mark] this would be a good topic for the maintainer summit list. | |
Especially with regard to documentation about where to list test | |
requirements | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment