Thinkpad X40 and xorg-6.7.0-3

[Giuseppe Castagna]

Dear List,

   I have installed FC2 on my ThinkPad X40. Most of the features work
fine out-of-the-box apart from video, which has some annoying problems.
The machine uses an Intel 82852/855GM integrated graphic device. I have
tried standard FC2, changed color depths, upgraded to custom kernels and
Arjan's builds as well as to the xorg version in fedora update-testing
(6.7.0-3): the problems persist.
   The first noticeable problem is that whenever I close the laptop lid
and reopen it the upper part of the screen (about half of the size of
the gnome panel) gets corrupted since the image is shifted down of the
corresponding space. This is not annoying since a simple ctrl-alt-F1
ctrl-alt-f7 sequence restores everything.
   What is *really* annoying is that this corruption appears when I
switch to the external monitor as well (and the f1-f7 trick no longer
works since the image returns back on the LCD screen). This makes the
laptop not usable for presentations, which is one of the main uses of it
(I have to boot in WinXP to give a presentation).
   Also (but this is a minor problem), when I try to use xv with mplayer
on the external monitor, the mplayer window is blue.
   Does any of you have any suggestion on how I can overcome at least
the external monitor corruption problem? I attach my xorg.conf file.

  Thanks in advance

---Beppe---

[Peter Boy]

Am Mo, den 07.06.2004 schrieb Giuseppe Castagna um 09:46:
>     I have installed FC2 on my ThinkPad X40. Most of the features
> work
>  fine out-of-the-box [...] 
Did you try to add apci=off to the kernel's boot parameters (in
/boot/grub/grub.conf)? Most Thinkpads are currently incompatible with
the Linux ACPI implementation (but I don't know about the X40). APM in
those cases used to work fine (e.g. your special function keys should
work again).


Peter

[Jason L Tibbitts III]

> >>>> "PB" == Peter Boy
> <pboy(a)barkhof.uni-bremen.de&gt; writes: 
PB> Did you try to add apci=off to the kernel's boot parameters (in
PB> /boot/grub/grub.conf)?

I just brought up and X40 and am seeing the same issues.  ACPI suspend
is totally hosed.  ACPI sleep (close the lid) works except for the
graphics corruption.

APM suspend nearly works; I can suspend and resume without the X
server running, but as soon as I try to start X after a resume the
graphics system is completely wedged; the following lines are repeated
endlessly:

Jun  4 10:58:17 rashi kernel: mtrr: base(0xe0020000) is not aligned on a size(0x300000)
boundary
Jun  4 10:58:24 rashi kernel: [drm:i830_wait_ring] *ERROR* space: 130652 wanted 131064
Jun  4 10:58:24 rashi kernel: [drm:i830_wait_ring] *ERROR* lockup
Jun  4 10:58:26 rashi kdm[2861]: X server for display :0 terminated unexpectedly

If I suspend in graphics mode the machine locks hard.

BTW, this is all using the 2.6.6-414 kernel.  I see there's a newer
one out so I'll try it as well.  Otherwise it's off to bugzilla.

 - J<

[Giuseppe Castagna]

On Mon, 2004-06-07 at 16:13, Jason L Tibbitts III wrote:
>  >>>>> "PB" == Peter Boy
> <pboy(a)barkhof.uni-bremen.de&gt; writes:
>  APM suspend nearly works; I can suspend and resume without the X
>  server running, but as soon as I try to start X after a resume the
>  graphics system is completely wedged; the following lines are repeated
>  endlessly:
>  
>  Jun  4 10:58:17 rashi kernel: mtrr: base(0xe0020000) is not aligned on a size(0x300000)
> boundary
>  Jun  4 10:58:24 rashi kernel: [drm:i830_wait_ring] *ERROR* space: 130652 wanted 131064
>  Jun  4 10:58:24 rashi kernel: [drm:i830_wait_ring] *ERROR* lockup
>  Jun  4 10:58:26 rashi kdm[2861]: X server for display :0 terminated unexpectedly
>  
>  If I suspend in graphics mode the machine locks hard.
>   

If I close the lid I receive this error

ACPI breakpoint: Executed AML Breakpoint opcode

I think that the problem is the i810 driver. I thought it was a memory
allocation problem so I tried to reserve the memory in xorg.conf with
the option VideoRam 32768 (and also tried up to 128Mb ... I have 1Gb
ram) but it did not work. 

I posted the same message on the linux on Thinkpad mailing list and I
received this answer

On Mon, 2004-06-07 at 12:48, Stephan Frank wrote: 
>  I think you are out of luck here. Since there is not sufficient 
>  information available about the necessary register values there is 
>  currently no good support for this chipset in the xfree drivers apart 
>  from the stuff the standard BIOS calls provide. This also affects people 
>  which have high resolution displays (1400x1050 and more) since the BIOS 
>  is not know about such high resolutions. People are currently fiddling 
>  with BIOS/driver routine disassembling to work out how stuff is done 
>  under windows (see the xfree-devel list). For the moment I see two 
>  option for you: a) try out the binary only Intel Embedded Graphics 
>  drivers [1] (which according to the website provide Dual-display 
>  capabilities; however, I personally did not test since this driver 
>  currently does not support the 1400x1050 screen resolution I looked 
>  for), or b) use a commercial X-Server like Accelerated X [2]. 
I had already tried (a) but abandoned since the Inter driver works only
on 2.4.* kernels. I will try (b) since they provide demo drivers. In the
meanwhile I discovered this, 
http://sourceforge.net/projects/i855crt
which may be worth trying (but I need a monitor or a videobeamer to test
my external monitor exit and I do not have one here at home)

Thanks for your help.

---Beppe---

[Michal Jaegermann]

On Mon, Jun 07, 2004 at 09:13:42AM -0500, Jason L Tibbitts III wrote:
>  
>  APM suspend nearly works; I can suspend and resume without the X
>  server running, but as soon as I try to start X after a resume the
>  graphics system is completely wedged; 
A number of years ago I had a similar problem with an APM suspend
and a Toshiba laptop.  I simply added 'chvt 1' to a script executing
on a suspend event.  True, the laptop was waking up to a text
console but hitting 'Alt-Ctrl-F7' after that was not a big deal
and X worked afterwards with a restored session.

You can try if this kind of a ruse will work with your laptop.

What is passing for APM in a laptop my wife is using right now is
not worth to bother in any sense. :-)

   Michal

[Jason L Tibbitts III]

> >>>> "MJ" == Michal Jaegermann
> <michal(a)harddata.com&gt; writes: 
[chvt trick]

MJ> You can try if this kind of a ruse will work with your laptop.

Unfortunately it's not nearly that simple.  APM fares worse than ACPI;
just closing the lid kills the machine.  I've filed bug 125458 with
all the details:

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=125458

Note that this has significant bearing on fedora-test: kernel
2.6.6-1.422 made things measurably worse than 2.6.5-1.358.  Running
358, I can actually do an ACPI S3 sleep and resume from it, but the
graphics are completely dead.  422 just hangs hard at suspend time.

I think I tried all twelve possible combinations of kernel
(2.6.5-1.358, 2.6.6-1.422), ACPI/APM, and running X/running X but
switched to VT1/not running X.

 - J<

[Jason L Tibbitts III]

A quick update: I found that running the i586 2.6.5-1.358 kernel
(instead of the i686 one) under APM works for me to allow a proper
suspend/resume.  However, the 2.6.6-1.422 kernel seems to be less
stable running this way.  ACPI is still busted under either kernel,
with 422 being more busted.

This doesn't have any bearing on Giuseppe's problem; the screen
corruption is still there when switching to the external device and
nothing will get rid of it since switching back to a text console
turns off the external display.  I thought that perhaps xrandr might
help, but it switches off the external display as well.

 - J<

[Giuseppe Castagna]

On Mon, 2004-06-07 at 09:54, Peter Boy wrote:
>  Am Mo, den 07.06.2004 schrieb Giuseppe Castagna um 09:46:
>  >    I have installed FC2 on my ThinkPad X40. Most of the features work
>  > fine out-of-the-box [...]
>  
>  Did you try to add apci=off to the kernel's boot parameters (in
>  /boot/grub/grub.conf)?  
Yes .... crash :-)

I do not think it is a matter of ACPI, that is not just the lid part.
Indeed this would not explain why happens on the external monitor as
well. 

---Beppe---