[OpenIndiana-discuss] Nobody for whom OI/Hipster Mate on i915/KMS is working fine wants to report?

Fri Sep 9 02:14:01 UTC 2016

Wow, that's actually bad, but nevertheless on the other side cool.
Because finally we have something!

Tnx for following my wish for more info.

It comes down to 
Sep  8 23:59:26 openindiana drm: [ID 588002 kern.warning] WARNING: [drm:drm_agp_rw_gtt:629] AGPIOC_RW_GTT failed 6

Guess where - in syslog!!!!!
And you are the first person for whom it is not working who actually - finally - sent me syslog.
Not sure why others found it that hard.
And I in turn instantly recall how I added AGPIOC_RW_GTT to the agpgart stack (need to look, where and how).
I needed to, because in Oracle's port it is referenced in a header:

https://java.net/projects/solaris-x11/sources/x-s12-clone/content/open-src/kernel/efb/src/agpgart.h?rev=1528
line 127: #define AGPIOC_RW_GTT       _IOW(AGPIOC_BASE, 15, agp_rw_gtt_t)

while in OS/Net's old agpgart and friends it was absent (from the top of my head).
Perhaps I made a mistake there!!!

You have to be aware, that agpgart is still mandatory for older Intel-GPU's. The term "agp" is misleading.
While Randy Fishel laughed about me when I pointed him to that, there are explicit remarks stating the very same in the drm/kms src.

Also the pciids Oracle assigned in Solaris 11.0 and 11.1 proved this.
Look into the December archives of oi-dev.

Another interesting fact appears to be, that in Xorg.0.log there actually isn't any sign of any abnormal termination, rather the somewhat odd:

[   208.865] (++) using VT number 7

(although firstly vconsole doesn't work with my backport [because of reasons related to the too outdated gfx_private so far in OS/Net, but TS is working on modernization] and secondly vconsole support probably isn't incorporated into the Hipster test iso [although I never tried, will do that now]).

So before I build an improved agpgart stack for you for extra tests that you can perform after installation to hdd, please could you please please also be the first person who makes a comparison test run with Oracle Solaris 11.1 or higher (best 11.3)?
And then let us know if it works betetr there?

I already have the suspicion that agpgart and friends will make the difference for you.

UPDATE regarding  [   208.865] (++) using VT number 7:

In  https://media-cdn.ubuntu-de.org/forum/attachments/00/45/4998857-Xorg.0.txt which I found here:

https://forum.ubuntuusers.de/topic/vesa-treiber-werden-benutzt-an-stelle-von-den-/  

the same line was found in Xorg.0.log on LinUX.
So let's ignore this for now.

Our problem is 99% related to agpgart.
Please the famous 3 child's wishes:

Pls. test if Sol11.1++ works for you on the same hardware with the 100% identical BIOS settings
Pls. install this Hipster DVD to hdd (for later creating test BE's and testing changed bins).
Pls. then follow my steps to test a modernized aggart stack which I will compile soon.

Tnx for your valuable testing work, 
%martin bochnig

>Пятница,  9 сентября 2016, 0:22 UTC от bentahyr at chez.com:
>
>So I downloaded the ISO from  http://buildzone.oi-build.r61.net/OI-install-tests-201608 , burnt and booted up.... to a black screen.
>( I can see grub, initial illumos boot up screen, then initial openindiana text console where it sets hostname and a few other things, then pitch back)
>
>Please find scanpci, dmesg and Xorg.0.log :
>https://share.gns.cri.nz/3UKLCK0PSZCC/scanpci.txt.html
>https://share.gns.cri.nz/0M0CQO77SEZR/dmesg.txt.html
>https://share.gns.cri.nz/HIA27PA2Y6J0/Xorg.0.log.html
>
>To add a few more interesting bits : 
>root at openindiana:/var/adm# isainfo -v
>64-bit amd64 applications
>        vmx xsave sse4.1 ssse3 cx16 sse3 sse2 sse fxsr mmx cmov amd_sysc
>        cx8 tsc fpu
>32-bit i386 applications
>        vmx xsave sse4.1 ssse3 ahf cx16 sse3 sse2 sse fxsr mmx cmov sep cx8
>        tsc fpu
>
>root at openindiana:/var/adm# ls -l /dev/dri
>total 2
>lrwxrwxrwx 1 root root 42 Sep  8 23:58 card0 -> ../../devices/pci at 0,0/display at 2:controlD64
>lrwxrwxrwx 1 root root 36 Sep  8 23:58 card1 -> ../../devices/pci at 0,0/display at 2:drm0
>
>Best regards.
>Ben
>
>----- Mail original -----
>De:  bentahyr at chez.com
>À: "Discussion list for OpenIndiana" < openindiana-discuss at openindiana.org >
>Envoyé: Jeudi 8 Septembre 2016 21:59:23
>Objet: Re: [OpenIndiana-discuss] Nobody for whom OI/Hipster Mate on i915/KMS is working fine wants to report?
>
>Hmm Ok, I'll do some tests tomorrow with ISO from buildzone.oi-build.r61.net.
>I'll gather requested info and tell you how it goes.
>
>Best regards.
>Ben
>
>----- Mail original -----
>De: "Alexander Pyhalov" < alp at rsu.ru >
>À: "Discussion list for OpenIndiana" < openindiana-discuss at openindiana.org >
>Envoyé: Jeudi 8 Septembre 2016 20:27:23
>Objet: Re: [OpenIndiana-discuss] Nobody for whom OI/Hipster Mate on i915/KMS is working fine wants to report?
>
>On 09/ 8/16 05:50 AM,  bentahyr at chez.com wrote:
>> Hi Martin,
>>
>> I got my hands on a Intel based laptop for testing.
>> Downloaded and burnt the ISO from  http://dlc-int.openindiana.org/hipster/20160816/
>>
>
>Hi.
>Wait, wait, wait... ISOs from 
>http://dlc-int.openindiana.org/hipster/20160816/ doesn't have KMS bits 
>integrated. Only OI_MATE* ISOs from 
>http://buildzone.oi-build.r61.net/OI-install-tests-201608/ have new 
>Intel DRM driver.
>-- 
>Best regards,
>Alexander Pyhalov,
>system administrator of Southern Federal University IT department
>
>_______________________________________________
>openindiana-discuss mailing list
>openindiana-discuss at openindiana.org
>https://openindiana.org/mailman/listinfo/openindiana-discuss
>
>_______________________________________________
>openindiana-discuss mailing list
>openindiana-discuss at openindiana.org
>https://openindiana.org/mailman/listinfo/openindiana-discuss
>
>_______________________________________________
>openindiana-discuss mailing list
>openindiana-discuss at openindiana.org
>https://openindiana.org/mailman/listinfo/openindiana-discuss