On 2023-11-22 13:21, Marcin Juszkiewicz wrote:
W dniu 22.11.2023 o 13:26, Leif Lindholm pisze:
On 2023-11-22 11:11, Sami Mujawar wrote:

[SAMI] The proposal above looks good to me. This may be slightly off topic, but can we also enable edk2-platform upstream CI as well, please? This would be helpful to catch issues much earlier. [/SAMI]

Yes, this is a problem we need to solve, but we don't have the time or resources to set up an upstream CI. What we've been thinking of
is to let maintainers set up their own CI infrastructure and then
have that perform any target-specific tasks and report back to
upstream CI. It's been a few months since I last discussed this with
Mike, but I think we were looking at https://docs.github.com/en/actions/hosting-your-own-runners/managing-self-hosted-runners/about-self-hosted-runners
as a possible tool.

This is probably not something we would like to tie into the edk2 mergify workflow (it would add too much delay), but localised to edk2-platforms.

Any help with implementing that would be most appreciated :)

I can write CI jobs which would run tests for QEMU platforms:

- virt/x86-64 (i440fx/q35)
- virt/aarch64

These live in the main edk2 repo though, so are not part of the problem that needs solving.

- sbsa-ref/aarch64

Sbsa-ref is something I am working on daily.

If Github Actions is all what's needed then it can be done using platform provided runners (no self-hosted needed).

We need the self-hosted for the physical platforms, which are the main part of the problem. But setting something like this up for the edk2-platforms platforms we can test with qemu could be a good first step.

/
    Leif



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#111605): https://edk2.groups.io/g/devel/message/111605
Mute This Topic: https://groups.io/mt/102746762/21656
Group Owner: devel+ow...@edk2.groups.io
Unsubscribe: 
https://edk2.groups.io/g/devel/leave/9847357/21656/1706620634/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-


Reply via email to