[Mesa-dev] How can we help improve Mesa stable releases and distro adoption?

Timur Kristóf timur.kristof at gmail.com
Fri Mar 26 16:14:36 UTC 2021


Hi Everyone,

We've talked about these issues in the RADV team recently, and we
believe they concern everyone else too, so I'd like to propose a
discussion about stable releases.

1. How can we help make the release manager's job easier?

Dylan, could you share with us how this process works? I admit that I'm
not sure how much effort needs to go into creating stable releases such
as 20.3.X. How much of it is automated, what do you need to do
manually? How much testing needs to go into a release?

Most importantly, what could we (the dev team) do to help you?

2. How can we help distro maintainers ship up to date stable releases?

For example, latest Ubuntu and Pop! OS are stuck at mesa 20.2.x, but
Ubuntu LTS and Linux Mint have it even worse at mesa 20.0.x (according
to distrowatch). I think this is sad because these are considered very
popular and are often suggested to beginners and gamers.

In reality, people who run these distros won't be able to use the
latest HW, won't benefit from any of the work we put into mesa drivers
during the past 6+ months and will experience all the bugs which we
have solved a long time ago.

(I know there are 3rd party repositories which cater to this, but 
that's not the same as getting 1st party support from an OS vendor.)

What could we do to make the job of distro package maintainers easier
and to compel them to upgrade to better supported releases?

Maybe we could make fewer new mesa releases per year, but support each
release for longer? Or try to better line up the mesa releases with
distro releases? I'd really like to hear it from a distro maintainer's
point of view, what keeps you on a release as old as 20.0 or 20.2 when
you know that this seriously hurts your user experience?

Best regards,
Timur



More information about the mesa-dev mailing list