Re: Nvidia binary drivers fail to install on Fedora 32

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sun, Mar 29, 2020 at 6:18 PM Kevin Kofler <kevin.kofler@xxxxxxxxx> wrote:
>
> Dominik 'Rathann' Mierzejewski wrote:
> > I don't maintain an kmods, but from what I observe in the package
> > commits, they break with each new Fedora kernel release and the
> > maintenance burden is even higher for older nVidia driver versions.
> >
> > You're welcome to help with automation, but I'm afraid hunting for
> > patches or writing them yourself will be a constant manual job in this
> > case.
>
> Well, if the kmod breaks, the akmod almost certainly breaks too, the same
> way, you just don't notice it (at least until the end user bug reports start
> coming in). That is another reason why the akmod approach is flawed.
>
> So, sure, there is work that cannot be automated, but that is exactly the
> part of the work that has to be done for akmods as well, and an automated
> kmod build process would automatically alert you as soon as the kernel gets
> released. (Ideally, as soon as it gets built in Fedora Koji, or at least as
> soon as it hits updates-testing, whereas most users will only notice the
> akmod breaking after it hits the stable updates.)
>

This goes back to the thing I've been talking about in other threads
for years: unnecessary grunt work to force people to manually do
rebuilds which could be triggered via automation.

While these days, we say the reason why Fedora doesn't do kmod
packages is because we'd rather have all the stuff in the upstream
kernel, I remember that much of the original reasoning for yanking
them out was that it was too troublesome for everyone trying to
maintain kmod packages to do rebuilds each time the kernel was bumped.
The user experience sucked because we didn't have (and still don't
have) a good mechanism for ensuring kmods were built and submitted
along with every kernel update.

Funnily enough, this same problem has manifested in a different, more
annoying and critical form: the gcc <-> annobin issue. Every single
gcc update has managed to break briefly whenever they forget to
rebuild annobin along with it.

Maybe one day we'll get smart dependency rebuilding that actually
submits to the build system. Koschei is just a few steps shy of being
able to do just that...




--
真実はいつも一つ!/ Always, there's only one truth!
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx




[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Users]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]

  Powered by Linux