Created attachment 26875 [details]
xorg conf file

I would argue if this is still a bug need to fix...

In the F2F meeting, it's suggested to add "basic support for non-DRI" as the third configuration to be supported in Q2.

please upload log files for 1280x800 and 640x480 respectively with modedebug option on.

thanks
Ma Ling

(In reply to comment #3)
> In the F2F meeting, it's suggested to add "basic support for non-DRI" as the
> third configuration to be supported in Q2.
> 

ok. even though it doesn't look like a mode setting related issue..could you please re-assign for efficiency?

Well, see if Jesse has idea on this one.

The log seems to be from a KMS enabled session, which doesn't appear to have this problem.  With UMS I get a crash... debugging now.

Hm, no the crash was due to some of my test code, this works for me with and without KMS on GM45.  May have been fixed by a recent commit...

(In reply to comment #8)
> Hm, no the crash was due to some of my test code, this works for me with and
> without KMS on GM45.  May have been fixed by a recent commit...
> 

Jesse,it is still broken on my side with below commits,so reopen it:
Kernel:       (for-linus)0e7ddf7eeeef5aea85412120539ab5369577faeb
Libdrm:         (master)790097c51330090b2b7b90429b9ab8ddf259fd8e
Mesa:           (master)cdbcb051d93836861bc4d4e1027156111ac249d2
Xserver:                (master)9480725af0275b72b24b9e1cb3b68d5b97288d01
Xf86_video_intel:     
(master)f0270bbb47baed78a0ff6189ae20d3ac322ec02b
Kernel:       (for-linus)0e7ddf7eeeef5aea85412120539ab5369577faeb

Created attachment 27144 [details]
xorg.0.log

the correct log from UMS is attached. So sorry for last one.

Created attachment 27478 [details] [review]
return screen offset when needed

In the non-DRI case, the screen pixmap won't have a pointer assigned after it moves, so we need to check for that case when we fetch the offset.  Can you confirm this works for you?

I think this patch would be better; can you test it?

diff --git a/src/i830_driver.c b/src/i830_driver.c
index c125a73..7db273d 100644
--- a/src/i830_driver.c
+++ b/src/i830_driver.c
@@ -926,6 +926,7 @@ i830_xf86crtc_resize (ScrnInfoPtr scrn, int width, int height)
        i830_memory *new_front, *old_front;
        Bool        tiled;
        ScreenPtr   screen = screenInfo.screens[scrn->scrnIndex];
+       pointer     data = NULL;

        scrn->displayWidth = i830_pad_drawable_width(width, i830->cpp);
        tiled = i830_tiled_width(i830, &scrn->displayWidth, i830->cpp);
@@ -945,9 +946,11 @@ i830_xf86crtc_resize (ScrnInfoPtr scrn, int width, int height)
        i830_set_pixmap_bo(screen->GetScreenPixmap(screen),
                           new_front->bo);
        scrn->fbOffset = i830->front_buffer->offset;
+       if (!new_front->bo)
+           data = i830->FbBase + scrn->fbOffset;
        screen->ModifyPixmapHeader(screen->GetScreenPixmap(screen),
                                   width, height, -1, -1, scrn->displayWidth * i830->cpp,
-                                  NULL);
+                                  data);
        xf86DrvMsg(scrn->scrnIndex, X_INFO, "New front buffer at 0x%lx\n",
                   i830->front_buffer->offset);
        i830_set_new_crtc_bo(scrn);

Created attachment 27481 [details] [review]
Set screen pixmap devPrivate.ptr to new pixmap address

Here's a patch which correctly sets the screen pixmap to point at the new memory.

Yeah that last patch works for me too; it's probably a better one to apply than mine too.

Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>

Fix pushed:

commit a66357832388ba9db21a4b3bf8311d9d1f1ab308
Author: Keith Packard <keithp@keithp.com>
Date:   Tue Jul 7 14:13:57 2009 -0700

    non-DRI FB resize failed to assign the screen pixmap devPrivate.ptr (22328)

*** Bug 22076 has been marked as a duplicate of this bug. ***

Yifei, please verify with the latest xf86-video-intel master.

It works fine now.