Created attachment 17367 [details]
reg_dumper_after_vt

Created attachment 17368 [details]
xorg.0.log

Created attachment 17369 [details]
xorg conf file

Haien, does this happen with the stable driver?

(In reply to comment #4)
> Haien, does this happen with the stable driver?
> 

Yes,there is the same issue happen with stable driver.

Haien, does this issue happen in every VT switch, or just the first time (bug#15187)?

(In reply to comment #6)
> Haien, does this issue happen in every VT switch, or just the first time
> (bug#15187)?
> 

yes ,it happened every switch(X->console).

I experience something similar (the same?) on my 945GM. Text console is displayed fine, but switching back to X only shows a black screen with the mouse pointer. The keyboard does not work anymore in that state.

I met the same problem with G45, which have one DVI and one HDMI. Here is a workaround that works for me.

diff --git a/src/i830_hdmi.c b/src/i830_hdmi.c
index b738463..7d31f87 100644
--- a/src/i830_hdmi.c
+++ b/src/i830_hdmi.c
@@ -218,7 +218,8 @@ i830_hdmi_init(ScrnInfoPtr pScrn, int output_reg)

     intel_output->dev_priv = dev_priv;
     intel_output->type = I830_OUTPUT_HDMI;
-    intel_output->pipe_mask = ((1 << 0) | (1 << 1));
+    intel_output->pipe_mask = (1 << 1);
+    //intel_output->pipe_mask = ((1 << 0) | (1 << 1));
     intel_output->clone_mask = (1 << I830_OUTPUT_HDMI);

     /* Set up the DDC bus. */

Interesting, Peng, could you paste your X log?

Created attachment 17816 [details]
Xorg.log without workaround

Xorg.log without workaround

Created attachment 17817 [details]
Xorg.log with workaround

Xorg.log with workaround

I once thought this maybe related to DSPARB. But it doesn't help when I program the register same as GM45.

Peng, I tested without any pipe kludge on one G45 here, hdmi attached to a sony bravia 1080p. VT switch works fine here. 

Strange. This problem happens on both of my G45 and Q45. Actually this is not specific for HDMI. The integrated DVI interface will also be covered by the hdmi code. 

Haien can reproduce this bug on the DVI port on our G45, with either DVI connection or DVI->VGA adapter.
He can't reproduce this bug on the HDMI port on our G45.

(In reply to comment #16)
> Haien can reproduce this bug on the DVI port on our G45, with either DVI
> connection or DVI->VGA adapter.
> He can't reproduce this bug on the HDMI port on our G45.

Same to me on the DVI port. I can also see the problem on HDMI.

BTW, I am testing HDMI on DELL 2408WFP.

  firstly,this bug still can reproduce on G45( has DVI && HDMI port) when using DVI->VGA, and we cannot see anything even startup information with DVI->DVI,but VT-switch works well with HDMI->HDMI.
 secondly,it works well on q45(only has a VGA port) with VGA->VGA.
 when we disconnect HDMI and connect DVI->VGA on G45, we can see "HDMI connected" in xrandr info,following is the output of xrandr:
xrandr -q:
Screen 0: minimum 320 x 200, current 1920 x 1200, maximum 1920 x 1920
VGA connected 1024x768+0+0 (normal left inverted right x axis y axis) 0mm x 0mm
   1024x768       60.0*
   800x600        60.3
   640x480        59.9
HDMI-1 disconnected (normal left inverted right x axis y axis)
HDMI-2 connected 1920x1200+0+0 (normal left inverted right x axis y axis) 519mm x 324mm
   1920x1200      60.0*+
   1600x1200      59.9
   1280x1024      75.0     59.9
   1152x864       74.8
   1024x768       75.1     60.0
   800x600        75.0     60.3
   640x480        75.0     60.0
   720x400        70.1

haien, Does my fix work for you ?

(In reply to comment #19)
> haien, Does my fix work for you ?
> 

I use the patch on your Comment #9 ,and it doesn't work on our G45.

(In reply to comment #18)

>  when we disconnect HDMI and connect DVI->VGA on G45, we can see "HDMI
> connected" in xrandr info,following is the output of xrandr:
> xrandr -q:
> Screen 0: minimum 320 x 200, current 1920 x 1200, maximum 1920 x 1920
> VGA connected 1024x768+0+0 (normal left inverted right x axis y axis) 0mm x 0mm
>    1024x768       60.0*
>    800x600        60.3
>    640x480        59.9
> HDMI-1 disconnected (normal left inverted right x axis y axis)
> HDMI-2 connected 1920x1200+0+0 (normal left inverted right x axis y axis) 519mm
> x 324mm
>    1920x1200      60.0*+
>    1600x1200      59.9
>    1280x1024      75.0     59.9
>    1152x864       74.8
>    1024x768       75.1     60.0
>    800x600        75.0     60.3
>    640x480        75.0     60.0
>    720x400        70.1
> 

haien, what monitor do you connect using DVI->VGA after you disconnect HDMI? I suspect this behaviour above might be a dup of our internal bug 345. 

(In reply to comment #15)
> Strange. This problem happens on both of my G45 and Q45. Actually this is not
> specific for HDMI. The integrated DVI interface will also be covered by the
> hdmi code. 
> 

Peng, does your Q45 have D-sub VGA port, or have integrated DVI ports only?

(In reply to comment #22)
> (In reply to comment #15)
> > Strange. This problem happens on both of my G45 and Q45. Actually this is not
> > specific for HDMI. The integrated DVI interface will also be covered by the
> > hdmi code. 
> > 
> Peng, does your Q45 have D-sub VGA port, or have integrated DVI ports only?

My Q45 only has two integrated DVI ports, no VGA port available.

(In reply to comment #23)
> (In reply to comment #22)
> > (In reply to comment #15)
> > > Strange. This problem happens on both of my G45 and Q45. Actually this is not
> > > specific for HDMI. The integrated DVI interface will also be covered by the
> > > hdmi code. 
> > > 
> > Peng, does your Q45 have D-sub VGA port, or have integrated DVI ports only?
> 
> My Q45 only has two integrated DVI ports, no VGA port available.
> 

I see. I think it confirms my suspect. I was just talking with zhenyu about this... thanks, Peng.

(In reply to comment #21)
> haien, what monitor do you connect using DVI->VGA after you disconnect HDMI? I
> suspect this behaviour above might be a dup of our internal bug 345. 

Yes, it's exactly #345. We wanted to track only one issue per bug report.

Could you test if this is still true with my patch on internal #345?

(In reply to comment #26)
> Could you test if this is still true with my patch on internal #345?
> 

with your patch ,we can get a black screen with message shows "cannot display this video mode" .

*** Bug 17506 has been marked as a duplicate of this bug. ***

*** Bug 17562 has been marked as a duplicate of this bug. ***

Hi, didn't see this one when looking through the list.  Anyhow, I have a similar problem with the Gigabyte G43 board.  I am using the VGA port (have had zero luck getting any output on DVI port).  I can use the vesa driver, but that is obviously unsatisfactory.  I will try the aforementioned workaround when I get a chance.

No luck with the workaround.  Though I suppose that is for HDMI only, anyway.

Created attachment 18887 [details]
unable to display console after starting X

Here's my log with ModeDebug enabled.  I started X, switched to VT1 (Monitor says: Out of Range), switched back to X (displays OK), then Ctrl-Alt Backspace (again monitor says: Out of Range).  Then I used the Magic SysRq hotkeys to reboot since I was stuck with no display.

You can see the LeaveVT,EnterVT,LeaveVT progression in the log.

Created attachment 19149 [details] [review]
DVI vt switch patch

Please test this one with G45/43 on DVI.

I tested this using xorg-server 1.5, mesa 7.2, libdrm from git and xf86-video-intel from git. It works! I am able to switch from X to text console and back.

I applied the patch to 2.4.2 and I can confirm that it seems to work fine. I'm using a Asus P5Q-EM motherboard with DVI.

Thanks for testing, I've pushed patch. Close.

verified.