Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 10 Nov 2019 16:43:12 +0100
From:      Tomasz CEDRO <tomek@cedro.info>
To:        "Steve O'Hara-Smith" <steve@sohara.org>
Cc:        Polytropon <freebsd@edvax.de>, "James E. Pace" <james@pacehouse.com>,  FreeBSD Questions Mailing List <freebsd-questions@freebsd.org>
Subject:   Re: 12.1 on Thinkpad - problem making graphics work
Message-ID:  <CAFYkXj=tpwVhuNPbcjAdS9pYT3Z91cJHT_ZPCgTrGvdVTz%2BJaQ@mail.gmail.com>
In-Reply-To: <20191110144830.096ce11b0668dd0721359cfa@sohara.org>
References:  <CAN1es4Jqy17VZeJkEFrUtX872929eLt78Bmrc20jPcernhk88A@mail.gmail.com> <CAN1es4Lab0uvZRPb51PaSjDDOH%2B9ZQTQedi1xvKO%2BCbnsbr20g@mail.gmail.com> <20191108194652.50c4f8e7c87ec76b9abc6e19@sohara.org> <CAFYkXjkWLwfZaBwR6pXRp18ZWV76B2%2BBwUeuyKVTgv6Trzy_qQ@mail.gmail.com> <20191108200005.21a9cdac18587cc36bd7cb01@sohara.org> <CAFYkXj=zofebWOtNhdzc8PhHC5xbWSQGf2pmQNywq1hjSDFVUg@mail.gmail.com> <20191109154827.42b0b2e1.freebsd@edvax.de> <20191109183356.76307bdc4ab6f0a1f68c0acd@sohara.org> <CAFYkXjnu2GDD%2BcY900nXQMRJ4jwzk0vdob5ESxjg24FhC6BHOA@mail.gmail.com> <20191110144830.096ce11b0668dd0721359cfa@sohara.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Steve O'Hara-Smith napisa=C5=82:

> On Sun, 10 Nov 2019 13:22:52 +0100
> Tomasz CEDRO wrote:
>
> > I hope FreeBSD will stay that way and remain its high quality standards=
.
> > Breaking kernel API/ABI with each release is definitely NOT the way to
> go.
>
>         This was a late (post release) discovered bug not intentional
> breakage.
>

Sure. Bugs happen. Workaround was quickly found. However, if you look at
GitHub discussion [1] a solution is proposed that will consider API
changes, different versions variants for different releases, modules and
packages infrastructure complication. This sounds the Linux way not the BSD
way. This seems to be the root cause of the problem..

The problem would never happen in the first place when API was designed in
a way it does not change with next versions and releases. No further
complications to the core and infrastructure is necessary. No
"bleeding-edge" accepted into kernel.

While some "DRM vs DRM2" issues may result like removing drm.ko from kernel
at all (it has no use anyway). Why the upstream does not land here into
drm.ko?

Another story is about loading external modules from ports directly into
kernel.. and updating them with PKG. If the module was reviewed under
strict core merge rules and probably tested with CI before things like this
would not happen. Kernel modules restricted for update with freebsd-update
could also increase security and stability.

https://github.com/FreeBSDDesktop/kms-drm/issues/183

I really LOVE FreeBSD. I just do not want it to become a mess like the rest
of the world, just because others do :-)

Best regards,
Tomek

--
CeDeROM, SQ7MHZ, http://www.tomek.cedro.info

>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAFYkXj=tpwVhuNPbcjAdS9pYT3Z91cJHT_ZPCgTrGvdVTz%2BJaQ>