Re: [oi-dev] [OpenIndiana-discuss] !!! Debug bins for developers: Working (but on my hw instable) Intel-KMS test-bins now uploaded for testing

[Мартин Бохниг]

 Hi Aurelian,

BIG THANKS for your quick response and instant testing.
It was my mistake.
Somehow I picked one of the not so good versions (another version already 
uploaded).

With the new one however it should actually work (also on your hw).
If in doubt try to manually re-start (or check with svcs that the service is 
both enabled and online)

/lib/svc/method/ogl-select start

This ensures that your system has the correct symlinks to (in this case) mesa- 
libGL*.* in place.


Regards,
%martin



>Среда,  6 января 2016, 10:21 +01:00 от Aurélien Larcher 
>:
>
>Dear Martin,
>thank you for your dedicated work !
>
>Here is some feedback on a ThinkCentre M91p with Intel HD Graphics 2000 to
>cheer you up:
>
>1) Loaded your driver binary:
>
>alarcher@pikku:~$ modinfo| grep i915
> 89 f7cca000  97b28 107   1  i915 (I915 DRM driver)
>
>node name:  display
>Vendor: Intel Corporation
>Device: 2nd Generation Core Processor Family
>Integrated Graphics Controller
>Sub-Vendor: Lenovo
>binding name:   pci8086,102
>devfs path: /pci@0,0/display@2
>bus addr:   2
>pci path:   0,2,0
>compatible name:
>(pci8086,102.17aa.3078.9)(pci8086,102.17aa.3078)(pci17aa,3078)(pci8086,102.9)(pci8086,102)(pciclass,03)(pciclass,0300)
>driver name:i915
>instance:   0
>driver state:   Attached
>
>2) Running now with comfortable resolution:
>
>alarcher@pikku:~$ xrandr
>Screen 0: minimum 320 x 200, current 1920 x 1080, maximum 8192 x 8192
>VGA1 connected 1920x1080+0+0 (normal left inverted right x axis y axis)
>510mm x 287mm
>
>So the driver does allow me to fire a Gnome session, neat :)
>
>Now some issues:
>
>alarcher@pikku:~$ glxinfo
>name of display: :0.0
>intel_do_flush_locked failed: Invalid argument
>
>Applications cannot cope with video rendering and there are glitches at the
>bottom of the screen, it seems that the last four (?) pixel rows are
>repeated while some preceding ~ (n - 8) are garbled.
>Some logs attached may help you to pinpoint the cause, let me know if you
>need more information.
>In any case, this is an encouraging step :)
>Best regards,
>
>Aurelien
>
>
>On Wed, Jan 6, 2016 at 9:19 AM, Мартин Бохниг < opens...@mail.ru > wrote:
>
>>
>>
>>  http://opensxce.org/intelkms_working_testbins/
>>  http://opensxce.org/intelkms_working_testbins/INSTALL.txt
>>  http://openindiana.org/pipermail/oi-dev/2016-January/004043.html
>> < http://openindiana.org/pipermail/oi-dev/2016-January/004043.html >
>>
>>
>> Better versions later today.
>> I really don't know where to start and where to stop. It is such a mess.
>> Actually this release was planned to happen in a sophisticated manner.
>> But time is running and running. You community members are probably tired
>> of only hearing stories about delays, deferrals and instabilities etc  .
>>
>> Therefore BANG! You have some first bins.
>> All the code and everything follows later today / next night / tomorrow,
>> asap.
>> Literally  ...
>>
>> --
>> Мартин Бохниг
>>
>> ___
>> oi-dev mailing list
>>  oi-dev@openindiana.org
>>  http://openindiana.org/mailman/listinfo/oi-dev
>>
>
>
>
>-- 
>---
>Praise the Caffeine embeddings
>___
>openindiana-discuss mailing list
>openindiana-disc...@openindiana.org
>http://openindiana.org/mailman/listinfo/openindiana-discuss

___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] R.I.P: Ian Murdock got murdered by USA-"Police"

[Nikola M]

On 01/ 6/16 07:49 AM, Мартин Бохниг wrote:
*Ian Murdock ,* a 
programmer and founder of DEBorah-IAN and I-nd-IA-a whose achievements 
are too many to list, "committed suicide" after being beaten 
and raped by the Police.


Here his absolutely shocking tweets  :

https://archive.is/cPE6h



There is also topic on openindiana-discuss list about this. It is 
obvious to me that normal people do Not comit "suicide" After police get 
their hands on them.


This is the reminder of need for tighten connections between people and 
organizations, so that no one is left behind and is protected by 
community of people who care and have channels of independently 
expressing and direct real lines of help.


___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

[oi-dev] Ouch: better versions re-uploaded __/__ Re: !!! Debug bins for developers: Working (but on my hw instable) Intel-KMS test-bins now uploaded for testing

[Мартин Бохниг]

 Ouch, I said: I grabbed them blindly from douzens of snapshots.
A quick test showed that these were very bad samples.

As said: The newer they are, the more stuff is changed and experimental.
(later you will see that in src). But it all had absolutely nothing to do with 
the instabilities on my Gigabyte mainboard, as I now finally found out that the 
Oracle 11.1 LiveDVD without any changed bits runs into the same conditions on 
my mainboard (with even another CPU and mem now).



Here an earlier version: with custom tracing but without debugging output.
Built with gcc4.8.

However, to give you a replacement for the lack of debugging output with the 
now newly uploaded bins here my local /var/adm/messages which grew to 39MB in 
only some few hours of runtime:   
http://svr4.opensxce.org/intelkms_working_testbins/messages


http://svr4.opensxce.org/intelkms_working_testbins/drm_gem_kms and  
http://svr4.opensxce.org/intelkms_working_testbins/i915   now re-uploaded 
better versions (should bring you into compiz on most Intel gear, including the 
latest internal HD gpu's.

_How-ever_: This is only a preliminary upload for you to test a bit.
The real release was not planned for today.

In fact it is Ian Murdocks unexpected death which caused me to write something 
(I learned of it just a few hours ago on a Russian freedom blog site!)

http://arpadhaizy.livejournal.com/3447260.html

His last tweets, which got vaporized by USA-Twitter (but which were backed up 
before) :

https://archive.is/cPE6h


R.I.P Ian Murdock!
I’m really wondering how I could think badly about him in the past.
It is very brave what he wrote, said and did.




>Среда,  6 января 2016, 11:19 +03:00 от Мартин Бохниг :
>
>
>
>http://opensxce.org/intelkms_working_testbins/
>http://opensxce.org/intelkms_working_testbins/INSTALL.txt
>http://openindiana.org/pipermail/oi-dev/2016-January/004043.html
>
>
>Better versions later today.
>I really don't know where to start and where to stop. It is such a mess.
>Actually this release was planned to happen in a sophisticated manner.
>But time is running and running. You community members are probably tired of 
>only hearing stories about delays, deferrals and instabilities etc  .
>
>Therefore BANG! You have some first bins.
>All the code and everything follows later today / next night / tomorrow, asap.
>Literally  ...
>
>-- 
>Мартин Бохниг
>___
>oi-dev mailing list
>oi-dev@openindiana.org
>http://openindiana.org/mailman/listinfo/oi-dev

___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] [OpenIndiana-discuss] !!! Debug bins for developers: Working (but on my hw instable) Intel-KMS test-bins now uploaded for testing

[Aurélien Larcher]

Dear Martin,

>
> BIG THANKS for your quick response and instant testing.
> It was my mistake.
> Somehow I picked one of the not so good versions (another version already
> uploaded).
>
> With the new one however it should actually work (also on your hw).
>

In the mean time I had rebooted with the new versions and I can confirm
that it works without glitches. ;)


> If in doubt try to manually re-start (or check with svcs that the service
> is both enabled and online)
>
> /lib/svc/method/ogl-select start
>
> This ensures that your system has the correct symlinks to (in this case)
> mesa- libGL*.* in place.
>

Thanks for the advice, it seems things are in place.

I can see repeated warnings in the logs, which I paste here in case they
are of interest:

Jan  6 10:42:21 pikku drm_gem_kms: [ID 986297 kern.warning] WARNING:
[drm:drm_gem_object_alloc_internal_normal:535] >dma_hdl is NEGATIVE:
>dma_hdl:ff02e3ecb3e8 dev->devinfo:ff02c932f2b0
dma_attr:c0093720 cb:0 >dma_hdl:ff02e3ecb3e8
obj->dma_hdl:ff02e510e100
Jan  6 10:42:21 pikku drm_gem_kms: [ID 280825 kern.warning] WARNING:
[drm:drm_gem_object_init:643] drm/src/drm_gem.c 526: dev:ff02cc7e8d40
obj:ff02e6215ac0 size:16384 gen:60
Jan  6 10:42:21 pikku drm_gem_kms: [ID 881333 kern.warning] WARNING:
[drm:drm_gem_object_internal:594] drm/src/drm_gem.c 478:
dev:ff02cc7e8d40 obj:ff02e6215ac0 size:16384

Best regards

Aurelien
___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] R.I.P: Ian Murdock got murdered by USA-"Police"

[Мартин Бохниг]

I am absolutely in SHOCK  :

I did not notice anything.
We have luck that no thermonuclear war got started while I was fiddeling with 
KMS (I hope that this is so far still correct?).

I regret I ever said a bad word about Ian in 2007.
I had no clue what a good heart he has (very much unlike Scott McNeally that 
idiot, who once tweeted in 2012: "More millions of tax payer's money are wasted 
for native americans/Indians").


Although, the long answer is, that it was not correct in the first place, let's 
see what USA-'s nice partner Saudi Arabia was doing earlier last year in Yemen:
Confirmed - Israeli - Saudi Arabia Tactical Nuclear Strike on Yemen (Neutron 
Bomb)

https://www.youtube.com/results?search_query=yemen+nuke


Dead children in Yemen (Western "media" never "reports") :
https://www.google.de/search?q=yemen+dead+children=firefox-a=org.mozilla:en-US:official=lnms=isch=X=0ahUKEwiMlfDzkpXKAhWmg3IKHduUAvcQ_AUIBygB=1237=780




>Среда,  6 января 2016, 12:58 +01:00 от Nikola M :
>
>On 01/ 6/16 07:49 AM, Мартин Бохниг wrote:
>>Ian
  Murdock , a programmer and founder of DEBorah-IAN
  and I-nd-IA-a whose achievements are too many to list,
  "committed suicide" after being beaten and raped by the
  Police.
>>Here
  his absolutely shocking tweets :
>>
>>https://archive.is/cPE6h
>
>There is also topic on openindiana-discuss list about this. It is
obvious to me that normal people do Not comit "suicide" After police
get their hands on them.
>
>This is the reminder of need for tighten connections between people
and organizations, so that no one is left behind and is protected by
community of people who care and have channels of independently
expressing and direct real lines of help.
>
>___
>oi-dev mailing list
>oi-dev@openindiana.org
>http://openindiana.org/mailman/listinfo/oi-dev

___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] [OpenIndiana-discuss] !!! Debug bins for developers: Working (but on my hw instable) Intel-KMS test-bins now uploaded for testing

[Aurélien Larcher]

Hi Martin,


> Could you please try to keep it enabled by default (unless you are wring
> your Ph.D. thesis in another window)?
>

I will continue running the driver with Compiz enabled for testing.
A bit more than one hour, working on the machine without any major hiccup.
The only thing was Firefox disappearing when enabling Compiz.

Best regards

Aurelien

-- 
---
Praise the Caffeine embeddings
___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] [OpenIndiana-discuss] !!! Debug bins for developers: Working (but on my hw instable) Intel-KMS test-bins now uploaded for testing

[Мартин Бохниг]

 Ah, great.
I'm releaved to hear that it also works on other hardware out of the box :)

As said: Such skilled instant testers like you are essential, Thank You again!!
Now - these bins are still complete crap compared to what it would be if I 
cleaned out all the experimental settings.

I wonder that i still works at all like that.
However: What you see as "warnings" is perfectly normal.
Your /var/adm/messages will grow up quickly.
These tracings are only a help to myself (don't belong there normally).
If something crashed one tails backwards to find out something, that's all.
In theory(!) DRM functions like an perfect clockwork, until at some point (on 
my mainboard) it stops doing so  :(

Now the only real question is, is this related to my mainboard ('s BIOS) or not.
If not then it is a general bug in Sun/Oracle's port (not at all related to my 
backport), because 11.1 has the same problems as I foolishly enough found out 
only yesterday (I only tested it regularily back then before Dec24th, before 
X11 cam up at all with the backport).
But I made the mistake that after December 24th I never _thoroughly_ tested a 
normal user session on 11.1 (due to a lack of hdd's I installed Hipster over 
the old test install).

Could you please try to keep it enabled by default (unless you are wring your 
Ph.D. thesis in another window)?

I only have one amd computer (unmodified 2005' Ultra20 with Opteron 1.8GHz) and 
this Intel based box also in a U20 enclosure 
(System Configuration: Gigabyte Technology Co., Ltd. H61M-S2V-B3
BIOS Configuration: Award Software International, Inc. F8 03/27/2012 with now 
again Celeron G530 aka Sandy-Celeron 2.4GHz Socket 1155).

I have no other Intel hardware with integrated GPU.

And the i7 2700K is listed as emergency sale here:   
http://www.ebay.de/itm/Intel-Core-i7-2700K-3-5-GHz-8-Vcore-Prozessor-CPU-100-OK-not-3770K-2600k-30GHz-/172052290217?hash=item280f1d9aa9:g:YagAAOSwa-dWjIMy



The bigmouthed "person" who wanted to give me a job in Brussels lost patience 
after KMS was not finished in 2 days (well, he is a capitalist running a ZFS 
admin biz).

Ahh, forget that.
I'm happy enough if you others on your non-Gigabyte hw won't run into these 
instabilities after some minutes, hours or even days of running a DRM gnome 
session with compiz and doing normal work, browsing etc.

Then all the last 2 weeks were only wasted due to my cheap H61 based Gigabyte 
board! 
_IF_ that was the case: Hooray.
Unfortunately I'm afraid it is rather a bug in DRM itself.


If you google for "gpu fault drm" or the like - even on non Oracle OS's such as 
on LinUX 4.x kernels - DRM can sometimes cause a lot of problems in certain 
revisions.
If that's would causes the mentioned problems, then we will have a hard time 
attempting to fix them.

I cannot expect you and everybody's test results (also over a longer testing 
session, with compiz effects enabled and FF running etc.)



Best regards,
%martin


___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] R.I.P: Ian Murdock got murdered by USA-"Police"

[Nikola M]

On 01/ 6/16 01:07 PM, Мартин Бохниг wrote:


I am absolutely in SHOCK  :



Please Martin, sorry to hear the news, but let's have list stay 
on-topic. Thank you a lot!



___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] R.I.P: Ian Murdock got murdered by USA-"Police"

[Мартин Бохниг]

>Среда,  6 января 2016, 13:58 +01:00 от Joerg Schilling 
>:
>
>Nikola M < minik...@gmail.com > wrote:
>
>> On 01/ 6/16 01:07 PM, ?? ?? wrote:
>> >
>> > I am absolutely in SHOCK  :
>> >
>>
>> Please Martin, sorry to hear the news, but let's have list stay 
>> on-topic. Thank you a lot!
>
>And BTW: Ian committed suicide during an episode of his bipolar desease, see 
>the typical entries in Ians tweets and the blog from Bruce Perens.
>
>Jörg

You were there?
How can you verify who wrote these lines?

___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

Re: [oi-dev] Again "new" (in fact older) KMS test bins uploaded

[Aurélien Larcher]

Hello Martin,
no worries, this is anyway a great step forward.
I rebooted and the glitch is gone, things have been stable for now.
I will make sure I get a kernel dump next time that happens.
Please get some rest ;)

Aurelien

On Wed, Jan 6, 2016 at 3:50 PM, Мартин Бохниг  wrote:

> Hello,
>
> I just wanted to upload a better version anyway (status from 20151230).
> Then I saw your unfortunate message.
>
> Before having the opportunity to review your crash (still not yet in bed,
> but must stand up in 3 hrs!) here first different less messed with bins
> based on cleaner src:
>
>
> http://svr4.opensxce.org/intelkms_working_testbins/
>
> ../ INSTALL.txt 
>   
>   06-Jan-2016 08:042815driver_aliases 
>    
>   06-Jan-2016 08:03 36Kdrm_gem_kms 
>   
>   06-Jan-2016 14:38633Ki915 
>  
>   06-Jan-2016 14:38  1Mmessages 
>  
>   06-Jan-2016 09:43 39M
>
>
> Tnx for testing and reporting!
> Will come back to this specific dump past next midnight CET.
>
> Best regards, %martin
>
>
>
> Среда, 6 января 2016, 15:19 +01:00 от Aurélien Larcher <
> aurelien.larc...@gmail.com>:
>
> Hello,
> after 5 hours or so, I did manage to crash it:
>
> Jan 6 15:03:08 pikku savecore: [ID 570001 auth.error] reboot after panic:
> BAD TRAP: type=e (#pf Page fault) rp=ff000f733ac0 addr=568 occurred in
> module "unix" due to a NULL pointer dereference
>
> Attached the messages log before the crash, unfortunately it got truncated
> (or beheaded).
> After rebooting the glitch at the bottom of the screen is now visible.
> Best regards
>
> Aurelien
>
> On Wed, Jan 6, 2016 at 11:41 AM, Aurélien Larcher <
> aurelien.larc...@gmail.com
> > wrote:
>
> > Hi Martin,
> >
> >
> >> Could you please try to keep it enabled by default (unless you are wring
> >> your Ph.D. thesis in another window)?
> >>
> >
> > I will continue running the driver with Compiz enabled for testing.
> > A bit more than one hour, working on the machine without any major
> hiccup.
> > The only thing was Firefox disappearing when enabling Compiz.
> >
> > Best regards
> >
> > Aurelien
> >
> > --
> > ---
> > Praise the Caffeine embeddings
> >
>
>
>
> --
> ---
> Praise the Caffeine embeddings
> ___
> openindiana-discuss mailing list
> openindiana-disc...@openindiana.org
> 
> http://openindiana.org/mailman/listinfo/openindiana-discuss
>
>
>
> ___
> oi-dev mailing list
> oi-dev@openindiana.org
> http://openindiana.org/mailman/listinfo/oi-dev
>



-- 
---
Praise the Caffeine embeddings
___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

[oi-dev] Again "new" (in fact older) KMS test bins uploaded

[Мартин Бохниг]

 Hello,

I just wanted to upload a better version anyway (status from 20151230).
Then I saw your unfortunate message.

Before having the opportunity to review your crash (still not yet in bed, but 
must stand up in 3 hrs!) here first different less messed with bins based on 
cleaner src:


http://svr4.opensxce.org/intelkms_working_testbins/
../
INSTALL.txt06-Jan-2016 08:042815
driver_aliases 06-Jan-2016 08:03 36K
drm_gem_kms06-Jan-2016 14:38633K
i915   06-Jan-2016 14:38  1M
messages   06-Jan-2016 09:43 39M

Tnx for testing and reporting!
Will come back to this specific dump past next midnight CET.

Best regards, %martin



>Среда,  6 января 2016, 15:19 +01:00 от Aurélien Larcher 
>:
>
>Hello,
>after 5 hours or so, I did manage to crash it:
>
>Jan  6 15:03:08 pikku savecore: [ID 570001 auth.error] reboot after panic:
>BAD TRAP: type=e (#pf Page fault) rp=ff000f733ac0 addr=568 occurred in
>module "unix" due to a NULL pointer dereference
>
>Attached the messages log before the crash, unfortunately it got truncated
>(or beheaded).
>After rebooting the glitch at the bottom of the screen is now visible.
>Best regards
>
>Aurelien
>
>On Wed, Jan 6, 2016 at 11:41 AM, Aurélien Larcher <
>aurelien.larc...@gmail.com > wrote:
>
>> Hi Martin,
>>
>>
>>> Could you please try to keep it enabled by default (unless you are wring
>>> your Ph.D. thesis in another window)?
>>>
>>
>> I will continue running the driver with Compiz enabled for testing.
>> A bit more than one hour, working on the machine without any major hiccup.
>> The only thing was Firefox disappearing when enabling Compiz.
>>
>> Best regards
>>
>> Aurelien
>>
>> --
>> ---
>> Praise the Caffeine embeddings
>>
>
>
>
>-- 
>---
>Praise the Caffeine embeddings
>___
>openindiana-discuss mailing list
>openindiana-disc...@openindiana.org
>http://openindiana.org/mailman/listinfo/openindiana-discuss

___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev

[oi-dev] !!! Debug bins for developers: Working (but on my hw instable) Intel-KMS test-bins now uploaded for testing

[Мартин Бохниг]

http://opensxce.org/intelkms_working_testbins/
http://opensxce.org/intelkms_working_testbins/INSTALL.txt
http://openindiana.org/pipermail/oi-dev/2016-January/004043.html


Better versions later today.
I really don't know where to start and where to stop. It is such a mess.
Actually this release was planned to happen in a sophisticated manner.
But time is running and running. You community members are probably tired of 
only hearing stories about delays, deferrals and instabilities etc  .

Therefore BANG! You have some first bins.
All the code and everything follows later today / next night / tomorrow, asap.
Literally  ...

-- 
Мартин Бохниг___
oi-dev mailing list
oi-dev@openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev