[oi-dev] Screen distorted with new i915 driver

Wed Sep 21 22:16:43 UTC 2016

Hallo  Alex,

no worries, so far I suspended i915/drm-kms versus agpgart stack work for the time until VirtualBox will be totally completed with all desktop- and server- features and really finally officially be incorporated into Hipster (when I'm ready I shall upload it to  https://github.com/OpenSXCE-org and then Alexander Pyhalov and Hipster-core team will take over).

The irony is, that you and Apostolos and Ben have been right all along: 

[oi-dev] Fwd: The compatibility PARADOX of the old versus the new i915 DRM(and now KMS submodule) kernel driver on newer (but not too new) versus older chipsets
https://openindiana.org/pipermail/oi-dev/2016-September/004804.html
Now, however, it sounds very bad that on your Laptop Oracle Solaris 11.3 shows exactly the same problem.
It sounds like something different then, although your error message in Xorg.0.log looks very damn related.

After I finally continue with i915 aka fix Illumos' outdated and since 2010 unchanged Sun-OpenSolaris agpgart stack I'll create a preliminary iso for you named three and also I have bought me a similar legacy Laptop for testing last week.
Then let's see if you can get it working with what I will by then (next week) have completed. But given that it doesn't work on 11.3 either, chances are perhaps dim, but we shall see.

Thanks for having taken the hassle to help the Hipster-community by testing and reporting.
Next time we will also look into syslog via ssh (first boot to single user and start sshd). This will be the point when we see if it was related to the same AGPIOC_RW_GTT (lack of in old agpgart stack). If yes, it makes me wonder even more why it also shows the same problems on S11.3 then.

Pls. some patience until next week for i915.
Once again tx for testing!

Beste Gruesse aus Ost-Berlin,
%martin

>Среда, 21 сентября 2016, 20:47 UTC от Alexander Jung <a.jung at scilset.de>:
>
>Hello Martin,
>
>
>sorry for the delay, but i was changing residence and there where
      no possibility to do your tasks.
>But no the situation has changed ...
>
>In BIOS i can only change display settings, i tested with graphics
      device Integrated Graphics (Intel) and Discrete Graphics (Nvidia)
      and with Nvidia Optimus with OS detection for optimus enabled and
      disabled.
>While i choose intel graphic there is no external monitor, only
      for nvidia, but that's another problem. 
>
>So with intel graphics or nvidia optimus the inteldriver don't
      work, the screen is always distorted, with and without custom
      configured xorg.conf, only vesa is working.
>Disabling DRI and hw-acceleration and change resolution also don't
      change anything.
>
>The same behaviour with solaris11.3   
>
>
>Best regards,
>Alex 
>
>
>
>
>PS: I do not use windows at all ;)
>
>
>
>
>
>
>
>
>On 09/ 2/16 06:30 PM, Мартин Бохниг via
      oi-dev wrote:
>>Hallo  Alex ,
>>
>>
>>do I need to ask for each snippet with an extra email, is it
        some quiz here?
>>
>>First I _need_ to know if Solaris11.3 works for you or has the
        same problem (yes, I must know that).
>>Then you didn't respond to any of my other questions, such as
        stolen-ram gtt setting in Bios or GPU-choice setting in Bios.
>>Also: Did you follow my instructions to try several stolen ram
        settings (best multiples of 32MB without remainder)?
>>
>>*Then*: Why is it so hard to connect some external screen to
        VGA1 and repeat _all_ mentioned tests (including on both Hipster
        _and_ Sol11.3)?
>>*Plus*: At least you now said that this Xorg.0.log was created
        "with xorg.conf". How useful.
>>But what about adding the customizations I proposed, such as
        disabling DRI in step A, disabling hw-acceleration in step B,
        and hardwiring the resolution to a low config like  1024x768 at 60 ?
>>
>>Sorry Sir, you are not using Windows 10 Home Edition   ;)
>>Am I asking for too much when I say that I need to know how it
        behaves on Sol11.3?
>>You can download it from Oracle after you register there (for
        free).
>>S11.3 uses the same DRM/KMS driver, only that it didn't need to
        backport it because they do have a more modern gfx_private and
        do have more modern kernel memory management interfaces.
>>For this reason I need these comparision results. Because then
        this would give some hints.
>>It is also mandatory that you repeat the test with an externally
        connected Monitor (VGA1).
>>And if you do decide to create xorg.conf, why then don't you
        benefit from its customizational powers?
>>Just to have some whatever xorg.conf is little helpful, if you
        don't follow the other configuration instructions.
>>
>>
>>#####p.s. Note to Hipster team: #####
>>The Mate-Compiz diff backport was more complicated than
        originally expected (gconf interfaces no longer in MATE).
>>After 2 days of external jobbing I can finally continue with it.
>>As announced on Tuesday I created a working MATE build env on a
        SDHC Photo/Video-16GB card connected to and booted from my old
        Lifebook via USB2 card reader. So I think tomorrow it should be
        at a stage where it functions.
>>
>>
>>
>>Best regards,
>>%martin bochnig
>>East-Berlin
>>
>>
>>
>>>Пятница, 2 сентября 2016, 8:25 UTC от
        Alexander Jung  <a.jung at scilset.de> :
>>>
>>>Hello Martin,
>>>
>>>starting X without xorg.conf results in the same
                  proplem.
>>>
>>>The Xorg.0.log is here ...
>>>
>>>http://vorschau.zebworx.com/OI/Xorg.0.log-without-xorg.conf
>>>
>>>
>>>Best regards, 
>>>Alex
>>
>>
>>_______________________________________________
oi-dev mailing list
>>oi-dev at openindiana.org
>>https://openindiana.org/mailman/listinfo/oi-dev
>
>-- 
>s c i l s e t
>Alexander Jung
>Solberg 5
>77883 Ottenhöfen
>
>TELEFON: (0 78
          42) 99 04 001   
>E-MAIL: a.jung at scilset.de
>INTERNET: www.scilset.de
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openindiana.org/pipermail/oi-dev/attachments/20160922/3c07d7d4/attachment-0005.html>