post-352 drivers

Luca Boccassi luca.boccassi at gmail.com
Sat Feb 6 11:20:33 UTC 2016


On Sat, 2016-02-06 at 02:12 +0100, Andreas Beckmann wrote:
> On 2016-02-04 10:23, Luca Boccassi wrote:
> >> Actually realised it would take just a few minutes, so I did a quick
> >> rebuild. Seems to work fine on both my machines for games, videos,
> >> opencl and cuda.
> >>
> >> But reading the original bug reports it looks like it worked on some
> >> setups and not on others, so can't say if the underlying issue has been
> >> actually fixed or I'm just one of the lucky ones.
> > 
> > Note that I only tested building with ld.gold with 358, not 352.
> 
> Anyway, lets test with this patch disabled (but still included) on 352
> with the next upload.
> 
> I also backported all the new buildsystem stuff eliminating our
> conftest.h to 352 - give it a try :-)
> We can also easily backport this to 340, but not to 304 (at least it
> would take more work for 304), so we will stick with conftest.h for 304
> (supported until end of 2017)

Ran usual tests with latest trunk, everything was fine on both desktop
and optimus laptop, including opencl and cuda.

The only small issue, that I've already fixed and committed, was that
the XEN on ARM workaround patch had to be extended to fix conftest.sh
too. I'm not entirely sure why it was working OK on 358 though, it's a
bit strange.

Besides that it's all green as far as I can see. Including 358 by the
way, looks ready for experimental to me, I've been using it for a couple
of days.

> Re #813541: What do you think about adding a "nvidia-legacy-check"
> package that becomes a pre-depends of most packages, and move the
> preinst check for legacy hardware there. That should avoid having
> partial -current installation after the user chose "abort installation"
> to manually install a legacy driver.

That sounds like a good idea! Should solve the problem nicely.

Kind regards,
Luca Boccassi
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://lists.alioth.debian.org/pipermail/pkg-nvidia-devel/attachments/20160206/25e50929/attachment.sig>


More information about the pkg-nvidia-devel mailing list