Bug#765726: VGA misdetection and freeze
beck at ibh.de
Tue Oct 21 12:17:55 UTC 2014
On Mon, Oct 20, 2014 at 08:29:33PM +0200, Andreas Beckmann wrote:
> On 2014-10-20 20:11, Beck, Andre wrote:
> > Given that metamode, the result is an offset nVidia Logo (which you would
> > have seen hadn't you disabled it ;) - and the login window on the wrong
> > part of that virtual screen.
> Good catch!
Searching further, we find others were hit by this before and already
brought it up in the nV forums, e.g.
> anyway, there is a 343.22 driver (soon) in experimental
Given that version will push a lot of cards (including my NVS 5100M) to
legacy state (last supported by 340.XX) as of
this isn't really going to help, at least for some of those hit by the
issue. We'll see what nVidia comes up with for 340.XX, given a large
amount of hardware will be stuck on that version for years to come...
> something else you could try is to edit conftest.h in the dkms source
> directory and disable the line defining NV_DRM_AVAILABLE
> as that got ctivated in 340.46
I'll give this a spin as soon as time permits.
> and please file bug reports with nvidia
I'll consider that should the current activity regarding this issue lead
nowhere within the next weeks. Given that several high end notebooks with
a more enterprisey customer focus are on the report lists so far, I don't
expect this to be necessary, though.
BTW, would a more generic bug title be in order? Way more models seem to
be affected, and the monitor presence misdetection seems to be the core
issue (I'm even currently assuming my flickering and final freeze may be
due to more such misdetection events when running, causing the driver
to repeatedly try to adapt to supposedly changed monitor setups, with
the final near-freeze a result of those code paths usually not being
tested against such frequent changes).
Thanks & HTH,
Cool .signatures are so 90s...
-> Andre Beck +++ ABP-RIPE +++ IBH IT-Service GmbH, Dresden <-
More information about the pkg-nvidia-devel