[oi-dev] Test bins with full debug output - blindly chosen in a hurry _/_ Re[2]: MAJOR BREAKTHROUGH: On Solaris 11.0 and even Illumos it works now!!! __/__ Re[2]: [OpenIndiana-discuss] Sun/Oracle China's DRM//KMS Sol11.2 port backported to function on old-style gfxp_private from pre-2010 era but still immediatedly PANICS

Мартин Бохниг opensxce at mail.ru
Wed Jan 6 07:52:22 UTC 2016


Another quick info in advance:
Don't worry about the old drm code in Illumos which continues to be required for legacy ums Radeon drm: By simply renaming the newer drm to drm_gem_kms (as it resides in common it must be a hw-independent name, therefore I didn't chose drm_i915) this conflict was solved last week (/kernel/drv/amd64/i915now looks for /kernel/misc/amd64/drm_gem_kms which will also be used for Ati-Radeon in the future).

Now both can peacefully co-exists until we have radeon-kms some day (if we live long enough on this planet).


In 5 minutes you find test-bins (taken without even knowing 100% which bad vs. good version they are, will test them quickly) 
on http://opensxce.org/intelkms_working_testbins/

I cannot attach them due to size restrictions on the ml.

Test/install Instruction: 

0.) cp "i915" to /kernel/drv/amd64/i915
1.) cp drm_gem_kms to /kernel/misc/amd64/drm_gem_kms
2.) Make sure that /etc/driver_aliases contains exactly the following lines for i915:


i915 "pci8086,42"
i915 "pci8086,46"
i915 "pci8086,102"
i915 "pci8086,106"
i915 "pci8086,10a"
i915 "pci8086,112"
i915 "pci8086,116"
i915 "pci8086,122"
i915 "pci8086,126"
i915 "pci8086,152"
i915 "pci8086,156"
i915 "pci8086,15a"
i915 "pci8086,162"
i915 "pci8086,166"
i915 "pci8086,16a"
i915 "pci8086,402"
i915 "pci8086,406"
i915 "pci8086,40a"
i915 "pci8086,40b"
i915 "pci8086,40e"
i915 "pci8086,412"
i915 "pci8086,416"
i915 "pci8086,41a"
i915 "pci8086,41b"
i915 "pci8086,41e"
i915 "pci8086,422"
i915 "pci8086,426"
i915 "pci8086,42a"
i915 "pci8086,42b"
i915 "pci8086,42e"
i915 "pci8086,a02"
i915 "pci8086,a06"
i915 "pci8086,a0a"
i915 "pci8086,a0b"
i915 "pci8086,a0e"
i915 "pci8086,a12"
i915 "pci8086,a16"
i915 "pci8086,a1a"
i915 "pci8086,a1b"
i915 "pci8086,a1e"
i915 "pci8086,a22"
i915 "pci8086,a26"
i915 "pci8086,a2a"
i915 "pci8086,a2b"
i915 "pci8086,a2e"
i915 "pci8086,c02"
i915 "pci8086,c06"
i915 "pci8086,c0a"
i915 "pci8086,c0b"
i915 "pci8086,c0e"
i915 "pci8086,c12"
i915 "pci8086,c16"
i915 "pci8086,c1a"
i915 "pci8086,c1b"
i915 "pci8086,c1e"
i915 "pci8086,c22"
i915 "pci8086,c26"
i915 "pci8086,c2a"
i915 "pci8086,c2b"
i915 "pci8086,c2e"
i915 "pci8086,d02"
i915 "pci8086,d06"
i915 "pci8086,d0a"
i915 "pci8086,d0b"
i915 "pci8086,d0e"
i915 "pci8086,d12"
i915 "pci8086,d16"
i915 "pci8086,d1a"
i915 "pci8086,d1b"
i915 "pci8086,d1e"
i915 "pci8086,d22"
i915 "pci8086,d26"
i915 "pci8086,d2a"
i915 "pci8086,d2b"
i915 "pci8086,d2e"
i915 "pci8086,2562"
i915 "pci8086,2572"
i915 "pci8086,2582"
i915 "pci8086,2592"
i915 "pci8086,2772"
i915 "pci8086,27a2"
i915 "pci8086,27ae"
i915 "pci8086,2972"
i915 "pci8086,2982"
i915 "pci8086,2992"
i915 "pci8086,29a2"
i915 "pci8086,29b2"
i915 "pci8086,29c2"
i915 "pci8086,29d2"
i915 "pci8086,2a02"
i915 "pci8086,2a12"
i915 "pci8086,2a42"
i915 "pci8086,2e02.8086.2e02"
i915 "pci8086,2e12"
i915 "pci8086,2e22"
i915 "pci8086,2e32"
i915 "pci8086,2e42"
i915 "pci8086,3582"
i915 "pci8086,a001"
i915 "pci8086,a011"


Or simply rename your own /etc/driver_aliases and cp over mine which I will also place into  http://opensxce.org/intelkms_working_testbins/
in 5 minutes.

Then make a reconf reboot and you should end up in a Intel-KMS-DRM-GEM Gnome session!
Don't forget to also remove or rename /etc/X11/xorg.conf (or replace "vesa" with "intel" in the device section)

Then you can enable compiz via the right mouse button and gnome settings menues.
For this test  downloaded Alexander's March 30th 2015 Hipster  LiveDVD.

It's all a bit not as cleanly writted as I actually wanted, but on the other hand you finally can test something yourself.
code all that today on github (will write another email then).

Don't ask me about agp-gart: Randy was right with that, from Sol11.2 on!
But it appears that Sun China went also into the wrong direction for 11.0 and 11.1 (as clearly illustarted by Sol11.0 and 11.1's agp header files and pciids for the agp/gart/target stuff.

In fact you can completely remove agpgart and _also_ xsvc completely from the system, will still function.
All mappings through i195 which depends on drm_gem_kms which depends on gfx_private which depends on certain structures directly in unix.



Wow, simply starting to write helped.
Because for the long story there will never be enough time.

Be aware that the bins I upload first produce the most "strange patterns" somewhere, because those are the ones that contain most fiddelings.

Also note: If syslog is enabled these bins will soon riddle or even fill your disk due to all the verbode output written to /var/adm/messages.
Of course in some more hours you can have bins that are not only cleaned up, but also without DEBUG info and without any custom-tracings.




best regards for the moment,
till later today,
%m


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://openindiana.org/pipermail/oi-dev/attachments/20160106/3eb00f0b/attachment-0002.html>


More information about the oi-dev mailing list