4.6 DM/i915 test report on Bay Trail (Celeron J1800)\

[Pete Wright]

On 07/25/2016 23:13, Natasha Kerensikova wrote:
> Hello,
> 
> on Monday 25 July 2016 at 10:54, Jonathan Anderson wrote:
>> I've found on my Skylake machine that the modesetting driver is more 
>> stable and produces fewer artifacts than SNA. I got to this driver by 
>> deleting my /usr/local/etc/X11/xorg.conf.d/XX-intel-driver.conf 
>> entirely: without a configuration directive to use the Intel X driver, 
>> it just used the KMS stuff transparently.
> 
> Do you mean the "scfb" driver, instead of the intel one?
> 

scfb does indeed work - but without acceleration obviously.  it is
usable for shell and basic web browsing imho.

the modesetting xorg drivre will properly detect an intel GPU as long as
the i915kms kernel module is loaded - i've been running this config for
several days on my skylake system without issues.
> 
> It doesn't seem to be present on the CFT image, because when I removed
> all xorg conf, it showed exactly the same symptoms as my SNA testing, so
> I guess (at least on the CFT image) X defaults to intel driver with SNA
> acceleration.
> 

the CFT image is most likely out of date.  if you install that image on
a disk, pulling the latest commits from github then rebuilding the world
and kernel should get you proper support for modesetting.

cheers,
-pete

-- 
Pete Wright
pete at nomadlogic.org