Bug 14317 - [G35 HDMI PATCH] HDTV display is shifted to the left
Summary: [G35 HDMI PATCH] HDTV display is shifted to the left
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: unspecified
Hardware: x86 (IA32) Linux (All)
: medium normal
Assignee: Wang Zhenyu
QA Contact: Xorg Project Team
URL: http://en.wikipedia.org/wiki/EDID
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-01-31 05:26 UTC by Arjen
Modified: 2008-07-28 18:03 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Current xorg config file (3.47 KB, text/plain)
2008-03-26 09:26 UTC, Arjen
no flags Details
Last X.org log file (33.55 KB, text/plain)
2008-03-26 09:26 UTC, Arjen
no flags Details
enable sdvo hdmi output (10.45 KB, patch)
2008-06-04 19:18 UTC, Hong Liu
no flags Details | Splinter Review

Description Arjen 2008-01-31 05:26:46 UTC
I have a Asus P5E-VM-HDMI mainboard with Intel G35 chipset. I've a 
Toshiba  37Z3030 (fullhd) TV connected to it using the HDMI-output.

Although it more or less works, the image displayed on the TV is shifted 
about 35 pixels to the left and 15 pixels to the top (i.e. the upper 
left corner of the image is behind the upper left corner of the bezel of 
the screen) and on the right and bottom side I have blank pixels.

This happens with Ubuntu Linux 710 Gutsy (Xorg-core 1.3 and intel driver 
2.1.1), the current Hardy version (Intel 2.2.0+git20080107 and Xorg 
1.4.1~git20080118) and today's git-checkout of the Intel driver with 
Hardy's Xorg. And using 32bits or 64bits didn't matter, although 
currently I'm using 32bits.

I have currently no modelines or resolution configurations in my 
xorg.conf, nor did specifying any help. I couldn't really find earlier 
mail in the archive about this issue, except for some remarks that it 
was supposedly fixed in 2.2.0/1.4.0.

As a sidenote I can't seem to get the screen working with 1920x1080 at 
60Hz, X only detects 1920x1080 at 0Hz (and 720x576 and 640x480) and 
refuses to operate with any other modeline. The detected frequencies 
seem to indicate it operates at 50Hz.

In windows I was able to select 24, 25, 30, 50 and 60 Hz for 1920x1080 
and the image is correctly positioned on the display.

This is the output of xrandr --verbose:

Screen 0: minimum 320 x 200, current 1920 x 1080, maximum 1920 x 1920
TMDS-1 connected 1920x1080+0+0 (0x5c) normal (normal left inverted right 
x axis y axis) 708mm x 398mm
         Identifier: 0x5b
         Timestamp:  2042559
         Subpixel:   horizontal rgb
         Clones:
         CRTC:       0
         CRTCs:      0 1
         EDID_DATA:
                 00ffffffffffff005262050100000000
                 0011010380693b780a0dc9a057479827
                 12484c20000001010101010101010101
                 010101010101023a80d072382d40102c
                 4580c48e2100001e8c0ad09020403120
                 0c405500c48e21000018000000fc0054
                 4f53484942412d54560a2020000000fd
                 00173d0f440f000a202020202020013b
   1920x1080 (0x5c)  148.5MHz +HSync +VSync
         h: width  1920 start 2448 end 2492 total 2640 skew    0 clock 
  56.2KHz
         v: height 1080 start 1084 end 1089 total 1125           clock 
  50.0Hz
   720x576 (0x5d)   27.0MHz -HSync -VSync
         h: width   720 start  732 end  796 total  864 skew    0 clock 
  31.2KHz
         v: height  576 start  581 end  586 total  625           clock 
  50.0Hz
   640x480 (0x5e)   25.2MHz -HSync -VSync
         h: width   640 start  656 end  752 total  800 skew    0 clock 
  31.5KHz
         v: height  480 start  490 end  492 total  525           clock 
  60.0Hz
Comment 1 Archibald Baal 2008-02-20 09:19:31 UTC
Is xrandr pulling the 2nd EDID block?  I only see one in your --verbose output.  The next to last byte in the EDID block is "01", indicating that the display has another EDID block with more supported modes and detailed timings defined.  See the "CEA EDID Timing Extension Version 3 data format" section of for details.

This may not directly affect the shift to the left and up (the timings specified in the first EDID block are Consumer Electronics Association standards for 1080p 50Hz and should work properly), but probably indicates why Windows can detect extra resolutions and X does not.
Comment 2 Archibald Baal 2008-02-20 09:21:04 UTC
(In reply to comment #1)
> See the "CEA EDID Timing Extension Version 3 data format" section of for
> details.

Sorry, apparently URLs are unwelcome.  My line above should have said "the Wikipedia entry for EDID".
Comment 3 Arjen 2008-02-20 09:59:36 UTC
That output is all I get, and I don't see any option to indicate to xrandr that it should try to do so.
Comment 4 Olivier Arsac 2008-02-20 10:23:03 UTC
I have indeed the same bug on a different display (Sharp 42xd1e)
Image shifted to the left by about 30 pixels.

Using hardy (xserver-xorg 1:7.3+10ubuntu, kernel 2.6.24.7-generic)

xrandr --verbose
Screen 0: minimum 320 x 200, current 1920 x 1080, maximum 1920 x 1920
VGA disconnected (normal left inverted right x axis y axis)
    Identifier: 0x4b
    Timestamp:  18750
    Subpixel:   unknown
    Clones:       CRTCs:      0 1
TMDS-1 connected 1920x1080+0+0 (0x4d) normal (normal left inverted right x axis y axis) 1152mm x 648mm
    Identifier: 0x4c
    Timestamp:  18750
    Subpixel:   horizontal rgb
    Clones:       CRTC:       0
    CRTCs:      0 1
    EDID_DATA:
        00ffffffffffff004d10e80f01010101
        00100103807341782a1bbea25534b326
        144a5200000001010101010101010101
        010101010101023a80d072382d40102c
        458080884200001e023a801871382d40
        582c450080884200001e000000fc0053
        484152502048444d490a2020000000fd
        00313d0f4b11000a2020202020200159
  1920x1080 (0x4d)  148.5MHz +HSync +VSync
        h: width  1920 start 2448 end 2492 total 2640 skew    0 clock   56.2KHz
        v: height 1080 start 1084 end 1089 total 1125           clock   50.0Hz
  1920x1080 (0x4e)  148.5MHz +HSync +VSync
        h: width  1920 start 2008 end 2052 total 2200 skew    0 clock   67.5KHz
        v: height 1080 start 1084 end 1089 total 1125           clock   60.0Hz
Comment 5 Wang Zhenyu 2008-03-25 23:46:22 UTC
Could you attach xorg.conf in use and xorg log?
Comment 6 Arjen 2008-03-26 09:26:10 UTC
Created attachment 15475 [details]
Current xorg config file

Here is the requested xorg.conf, its mostly auto-detect.
Comment 7 Arjen 2008-03-26 09:26:49 UTC
Created attachment 15476 [details]
Last X.org log file

Here is the most recent log file for X
Comment 8 Hong Liu 2008-06-04 19:18:49 UTC
Created attachment 16921 [details] [review]
enable sdvo hdmi output

Please try the attached patch to see if SDVO-HDMI is working.
Comment 9 Arjen 2008-06-05 05:35:11 UTC
I've compiled a git-checkout of the Intel driver with your patch applied.

With that, the image seems to be on the right place of the TV, no black borders exist below or besides the image anymore.

It still doesn't detect all those other modes my TV supports (like 1280x720 and several other variants of 1920x1080), but that may be a different issue.
Comment 10 Hong Liu 2008-06-05 17:50:04 UTC
(In reply to comment #9)
> I've compiled a git-checkout of the Intel driver with your patch applied.
> 
> With that, the image seems to be on the right place of the TV, no black borders
> exist below or besides the image anymore.
> 
> It still doesn't detect all those other modes my TV supports (like 1280x720 and
> several other variants of 1920x1080), but that may be a different issue.
> 

Thanks for trying the patch, I will look at the modeline problem later.
Comment 11 Michael Fu 2008-07-27 19:56:50 UTC
zhenyu, would you please take over the patch in comment# 8? If it's accepted into master, we can mark this bug as fixed. modeline issue will be tracked in another bug. thanks.
Comment 12 Wang Zhenyu 2008-07-28 18:03:52 UTC
Hong's patch has already merged upstream and be in 2.4.0, so close this one.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.