Hello; I'm lost as to where to look in the kernel, as to why I'm
receiving a black screen with startx
DRI seems to be fine, I'm noticing a sema_init but am not sure if this
is the cause of the blackness.
any direction as to where I might look would be helpful.
regards;

-- 
Justin P. Mattock
--

[ message continues ]

Kernel version?

A git-bisect should get you down the the offending commit within an
hour or three.  http://www.kernel.org/doc/local/git-quick.html has some
instructions.

But if you're on 2.6.26-rcX, be sure to use the very latest - a pile of
bad DRI code got reverted recently.  This morning, I think.
--

[ message continues ]

On Wed, May 21, 2008 at 5:34 AM, Andrew Morton

Thanks for the help, From what I can remember I noticed this around
2.6.26-rc1. It could be an issue with DRI, but then again it could be
something else. With running some more tests I'm noticing  a conflict
with libdri.so, not recognizing the module or the module not
recognizing libdri.so.  As for the latest git I'll have to pull, and
see if it makes  a difference.  If not then I'll have to sit and keep
working this problem until I can figure what is happening  "even if my
eye ball's fall out of my sockets", and I walk around feeling like
I've had brain surgery".
regards;
-- 
Justin P. Mattock
--

[ message continues ]

Which X video driver are you using?


cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

--

[ message continues ]

Hello; thanks for the feedback, from previous posts, you know what
module I'm talking about, So with that in mind I'm going to go
to the next question: From error messages "there isn't"
the only leeway or light at the end of the tunnel I see is with
libdri.so i.g. If I boot the system up with out messing with anything
I will receive
a black screen after issueing startx, (from looking at what I see, the
screen will flash one time then black, it should flash or blink two
times
before giving me my lovely picture(wallpaper) then if I remove
libdri.so everything starts up as usual so with that in mind since I
received
a post about DRI being all F*&^ked up then maybee that is the culprit,
I think All I really need right now is info on the xserver and the
kernel.
What mechanisms are used or files that are responsible for the two too coexist.
regards;

-- 
Justin P. Mattock
--

[ message continues ]

First, you really need to try the latest kernel from the git
repository (or latest nightly snapshot). The DRI 'mechanisms' you're
asking about were broken in 2.6.26-rc1. They got fixed very, very
recently (past 24 hours). So if you haven't tried the latest kernel
from -git yet, then, respectfully, you're not going to get very far.

Please try the latest kernel first, and then let us know what happens
-- given your description there's a good chance the problem is already
fixed. Also, please report back with the output of `git describe` for
the version you try, so that everyone can see what exact version you
were having trouble with.
--

[ message continues ]

you use ati or nvidia binary driver?



-- 
Thanks,
Oliver
--

[ message continues ]

Alright: I'm using 2.6.26-rc3-00122-g748bf99, I pulled if I can
remember Monday night and I did notice drmP.h and the rest of the drm
files,
but, will try and pull again to see if it makes a difference. As for
the question about what video card nvidia or ati, I'm stuck with ati,
went into apple yesterday to see if that would be easy to change out,
but unfortunately I would have to get a new laptop.
To bad apple didn't use an intel graphics card, that linux already has
a module for.
As for error messages looking at /var/log/Xorg.0.log it seems the
Xserver can't finish completing  DRIScreenInit
the /dev/dri/video0 is created, but then just stops after that, no
conversation with the module and dri after /dev/dri/video0 is created.
At this point, I'll try and pull and see if it makes a difference, and
if not then I'll have to just keep trying.
That's why right now I just need any kind of info about files or
mechanisms that  are used in this situation so I can at least try and
connect the dots with this issue.

regards;
-- 
Justin P. Mattock
--

[ message continues ]

O.K. I've got my 8-balls back, from what I could tell there was three
areas that needed attention to:
1)  add nopage back into the kernel
2) The DRI needed to be reverted back to it's original state, or
modified due to some issue.
3) find_task_by_pid( pid) on the module side needed to be changed to
find_task_by_vpid( pid) or else you'll have an issue with init_pid_ns

I did have the latest DRI updates yesterday, but should have been
using nopage to build the module instead of fault.
so three weeks ago if the DRI wasn't adjusted, and if I would of found
find_task_by_(v)pid(pid) then I would of been O.K.(but am unsure at
the moment, still need to do some more tests) Now after up and running
I can confirm that suspend successfully wakes up, So I can go ahead
and look into appletouch surviving suspend, and the mighty mouse
scroll function surviving suspend as well(then play Myst).
Overall this was a good educational exercise for me since I suck at
reading and writing.
Again thank you guy's for helping me out with this,
regards;

-- 
Justin P. Mattock
--

[ message continues ]

Well, I for one am still confused. Is there an issue with the latest
kernel.org kernel and Xorg opensource drivers that need to be
addressed? In other words, are other people going to hit the problem
you had? If so, a patch, or more technical description of what's going
on would benefit others that would otherwise have to do everything you
just did.

Or are you using the closed-source nvidia drivers? (No judgement from
me if you are, but it would really help clear up what the heck is
going on.)

Thanks,

Ray
--

[ message continues ]

Yes, the whole problem was that he was trying to use an incompatible
kernel with the closed source ATI driver.

Lee
--

[ message continues ]

Just so I can have suspend working properly, now the next question is
seeing if radeohd can wakeup from suspend
That way I don't have to be confined to using one module
But for the time being I need a nice cold beer from this headache..  : - )
regards;

-- 
Justin P. Mattock
--

[ message continues ]