elisa-tech / workgroups Goto Github PK

• what we have to offer, reasons to engage, areas to engage, perspectives to bring to play. :: need people to have vision here, and willing to review. Domain content.
• Clear instructions on how to engage with workgroups of interest.
• Standard LF code of conduct, respect for individuals applies

Currently, the callgraph-tool accepts various input formats for creating the callgraph database from the compilation logs. Modify the tool to accept exclusively format defined with http://clang.llvm.org/docs/JSONCompilationDatabase.html.
For the existing formats that use raw buildlog create a tool script that will convert it to JSON compilation database.
Note: LL_CLANG format accepts build dir and is not affected by this change.

Paul Albertella, Kate, Michael, Nicole to look at #OpenAPS, and see if we can engage that community.

Original wording:

Some analysis of set of potential use cases, possibly a framework.

Discussed on the first workshop:

• find a leader/main moderator/interested on providing recommendations on best practices.
• Kate to find community to assess impact of ideas in real project, before taking to Linux.

Currently the callgraph is only relying on function name as identificator, which works the most time but results in the callgraph database errors occassionaly. If there are functions with the same name only the first one detected ends up in the database. As the processing is done in a multithreaded fashion the results will differ from run to run.

The parser is implemented with hand-crafted patterns that might break as clang IR format is evolving. Rely on the Clang/LLVM codebase to parse out the IR relevant data. Rewrite current IR parser C++ back-end and supporting build files.

Min, @bulwahn and the local coordinator to take care

Looking for input from Kuka, BMW, Toyota: Are there profiles?

• Kuka: Stefan S. will look at examples here to provide
• BMW: Lukas will look for
• Toyota: Ando will look for.

Call a meeting when material available, and then look and seeing what is possible.

Oscar will volunteer to look at tools and libraries and comment on them.

Hi,
My name is Jeremy. I'm interresting by your project but I don't know how to start. I'm architect software developer and I can help to create some tools. I use Linux everyday in my work (Ubuntu, Debian, Raspbian, etc...) and i would like to participate. I know tools as Kubernetes, Docker, Gitlab with cluster deployments inside linux environments (embedded devices or cloud), some services as SQL or NoSQL Databases (postgreSQL, MongoDb, InfluxDb, Redis, DynamoDb, ...), Real time TCP protocols as HTTP, Nats, WebSockets, MQTT and programming languages as JavaScript, Ruby and C++. Often I like to test new technologies as blockchain, IOT system or ML tools.

Can I help ?

Thank you

Gold deck to use to explain the problem and need for participants:

• Pain problems

• Use Cases for ELISA - medical equipment, industry automation, autonomous systems (vehicle, factories, robots).

• Why use linux - new technologies available quicker, no licensing fees, total cost over lifecycle.

• Wider community to draw on for security issues. ELISA is build a wider community focused on safety issues.

• Kate to take first pass - next Tuesday.

• Nicole, Nicholas, Olaf to review Wed/Thurs.

• Paul to arrange presentation by the STPA expert at CodeThink: Shaun presented STPA in a webinar.
• If further followed up, create connection between John Thomas from MIT and the ELISA group

From the meeting minutes:

Open STPA - Sean to give presentation, Paul to arrange. What like, what hate.
Robin is interested in this. Handbook has been sent to list. Sean talking through it. Learning experience. Use case that is being done in the open then, can we also support it.
AI: Paul to take lead on this, and they write a “hello paul” email. Spin something up with ELISA.
Ask John Thomas from MIT - webinars & training. Interested in applying further. Paul to connect Kate with them.

Possibibly Very Difficult: The callgraph database for Linux source tree is built as a post processing step of the kbuild process. Establish the indexer build process as a Kbuild target inside of the Linux OS community.

In order to improve the testing and verify the callgraph database consistency over various versions of tool/LLVM create a canonical source project with all the relevant edge cases that need to be detected to form the proper call graph.

First task is to break down next steps here.

• Robin to look into ARM hosting in September
• Min to determine a suitable date for the participants
  After the workshop:
• Finalize the meeting minutes and summary on the second workshop
• Ask for feedback from participants
• Write summary of feedback

ELISA needs to investigate state of the art update policies of security critical systems and DevOps approaches towards increased software quality to understand which concepts on reliability and stability are used there.

Action items from first workshop:

• Lukas to provide pointers to analysis and tools available to consider.
  • show first attempts of root-cause and safety impact analysis, also share current development
• Phillipp to structure content and put on github to improved access

On the first workshop, we discussed:

• Organise Weekly Sync, Thursday 15:00 - 16:00 CEST: intended as open hangout for participants
• Set up of a Flock channel (asked Min on 2019-06-26 11:20 CEST for support)

Create information material to help ambassadors (reference material online, including slides to walk through, whitepapers for offline reading)

• Nicole: Outreach to Bitkom Forum: Nicole to provide overview at next working group meeting.
• Nicole: Provide feedback to Kate on Gold Deck

Collect from the case studies or from a common understanding how certain tools are used in the kernel development process, i.e., kernel documentation.