[i945] i945 don't recognize 1440x900 on external monitor in jaunty

[lugburz]

Binary package hint: xserver-xorg-video-intel

After the upgrade to ubuntu 9.04 it's not more possible to set the external monitor resolution to 1440x900, the native resolution of my external LCD panel, while the maximum possible is 1360x768

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 945GM/PM/GMS, 943/940GML and 945GT Express Memory Controller Hub [8086:27a0] (rev 03)
 Subsystem: Sony Corporation Device [104d:81e6]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
 Subsystem: Sony Corporation Device [104d:81e6]

[Geir Ove Myhr (gomyhr)]

Thank you for taking the time to report this bug. Could you please attach the file /var/log/Xorg.0.log, and the output of `lspci -vvnn` here. The best way to do this is to run `apport-collect 376722` from the command line on your computer, since that will include a lot of other useful information as well.

[lugburz (lugburz)]

Unfortunately I made an error setting the permission and I have to attach the data by hand.

[lugburz (lugburz)]

and a recente Xorg0.log

[Geir Ove Myhr (gomyhr)]

Thank you for the information. It looks like we need some more detailed information.
1. Which computer model do you have? What monitors do you have, and how are they connected to the computer (I don't see any trace of an external monitor in your Xorg.0.log)?
2. Could you upload the output of `xrandr --verbose` as well?
3. Could you add the following line to the "Device" section of xorg.conf:
  Option "ModeDebug" "true"
and upload the Xorg.0.log you get next time you log in?
4. Does the behaviour change if you comment out the line
  Option "AccelMethod" "uxa"
(or simply change "uxa" to "exa")?

[lugburz (lugburz)]

I attached the correct log file. The model is a sony vaio laptop, while the monitor is an LCD LG TV 19LS4D. Changing the uxa accelerator method to exa doesn't change anything.

[Geir Ove Myhr (gomyhr)]

Thank you for provided the more detailed information. The problem seems to be that the computer doesn't get the EDID information from the monitor. This is a known problem for a number of Sony Vaio laptops. Unfortunately, Sony ships laptops with monitors that have no ability to tell the computer what their preferred resolution is. There is some more information about this and what lines to add to xorg.conf to work around this at bug 288807. I'm going to mark this bug as a duplicate of that one. If your laptop model is not mentioned on that bug report, please leave a comment there.

[lugburz (lugburz)]

I'll see the original bug but the monitor that isn't correctly configured is the external 19 LCD monitor, the laptop panel is instead correctly configured. The other point is: why didn't this problem exist on the previous release?

[Geir Ove Myhr (gomyhr)]

I can't see any EDID information in the output of `xrandr --verbose` or Xorg.0.log. What are the correct resolution of your internal laptop monitor (1200x800?) Would it be possible for you to connect your external monitor (1440x900) via DVI and get the Xorg.0.log for that? Since this was working in an earlier version of ubuntu, do you have the possibility to start from an Intrepid LiveCD and attach `xrandr --verbose` output and Xorg.0.log (preferably with the ModeDebug option set)? Are you using your two monitors in mirrored mode or as an extended desktop?

Newer versions of xorg (or the driver) rely more heavily on the EDID information provided by the monitor that before. This is a standard that virtually all monitors have supported since around 2001. Unfortunately, many Vaio laptop panels don't support this, and these are the only monitors I know about that don't. Occationally, the EDID information in some models are slightly wrong, but in those cases it is possible to make xorg work around it. Without any EDID information, it is not possible to figure out what is connected at all.

[lugburz (lugburz)]

In the Xorg.0.log file attached to seem there are the information you requested:

After the line:
(II) intel(0): EDID for output VGA
...
(II) intel(0): Not using default mode "1440x900" (hsync out of range)
...
(II) intel(0): Printing probed modes for output VGA
(II) intel(0): Modeline "1360x768"x59.8 84.75 1360 1432 1568 1776 768 771 781 798 -hsync +vsync (47.7 kHz)
(II) intel(0): Modeline "1152x864"x60.0 81.62 1152 1216 1336 1520 864 865 868 895 -hsync +vsync (53.7 kHz)
(II) intel(0): Modeline "1024x768"x60.0 65.00 1024 1048 1184 1344 768 771 777 806 -hsync -vsync (48.4 kHz)
(II) intel(0): Modeline "800x600"x60.3 40.00 800 840 968 1056 600 601 605 628 +hsync +vsync (37.9 kHz)
(II) intel(0): Modeline "640x480"x59.9 25.18 640 656 752 800 480 490 492 525 -hsync -vsync (31.5 kHz)
(II) intel(0): EDID for output LVDS

When I use the external monitor I switch off the internal LCD panel. Using the intrepid live-cd I cannot configure the external monitor because the panel size is limited within 1200x800, in general on 8.10 I had many more modeline to choose, some of them not really working. In jaunty all the modelines seem to work, in some chases with artifacts in the borders.

[Geir Ove Myhr (gomyhr)]

Xorg.0.log always has those kind of lines after "EDID for output VGA" when it doesn't get any EDID information. Take a look at http://launchpadlibrarian.net/26967510/Xorg.0.vga.log for how it looks when the EDID information is acquired properly (in that case the there is only a VGA connected, it still prints "EDID for output HDMI-1" with no EDID information afterwards). Note especially that it dumps the EDID in hex, so that it possible to verify that it was interpreted right. The EDID in hex should also be written out by `xrandr --verbose` (the mentioned log is from bug 378617, take look at the xrandr --verbose output there when the monitor is connected via VGA).

You may check if you can get any EDID information using the tools get-edid (from the read-edid package, available for amd64 at https://launchpad.net/~intuitivenipple/+archive/ppa/ (see bug 242043), or ddcprobe (from the xresprobe package).

We also would like to know excactly what kind of Sony Vaio you have.

[lugburz (lugburz)]

I'm sorry for the long time for previous message.

The laptop model is Sony vgn-sz1m, now about 3 years old.

To attempt a work-around I added explicitly the modeline for 1440x900, using those parameters:

Modeline "1440x900" 103.97 1440 1496 1624 1856 900 900 902 933

the results is not very good: the desktop seems to be few percents wider than the panel.

The randr --verbos output now appears as follow:

Screen 0: minimum 320 x 200, current 1280 x 800, maximum 1680 x 1200
VGA connected (normal left inverted right x axis y axis)
 Identifier: 0x3b
 Timestamp: 22520
 Subpixel: unknown
 Clones:
 CRTCs: 0 1
 Transform: 1.000000 0.000000 0.000000
             0.000000 1.000000 0.000000
             0.000000 0.000000 1.000000
            filter:
  1440x900 (0xf8) 104.0MHz
        h: width 1440 start 1496 end 1624 total 1856 skew 0 clock 56.0KHz
        v: height 900 start 900 end 902 total 933 clock 60.0Hz
  1360x768 (0xf9) 84.8MHz -HSync +VSync
        h: width 1360 start 1432 end 1568 total 1776 skew 0 clock 47.7KHz
        v: height 768 start 771 end 781 total 798 clock 59.8Hz
  1152x864 (0xfa) 81.6MHz -HSync +VSync
        h: width 1152 start 1216 end 1336 total 1520 skew 0 clock 53.7KHz
        v: height 864 start 865 end 868 total 895 clock 60.0Hz
  1024x768 (0x42) 65.0MHz -HSync -VSync
        h: width 1024 start 1048 end 1184 total 1344 skew 0 clock 48.4KHz
        v: height 768 start 771 end 777 total 806 clock 60.0Hz
  800x600 (0x47) 40.0MHz +HSync +VSync
        h: width 800 start 840 end 968 total 1056 skew 0 clock 37.9KHz
        v: height 600 start 601 end 605 total 628 clock 60.3Hz
  640x480 (0x4c) 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 59.9Hz
LVDS connected 1280x800+0+0 (0x3e) normal (normal left inverted right x axis y axis) 0mm x 0mm
 Identifier: 0x3c
 Timestamp: 22520
 Subpixel: horizontal rgb
 Clones:
 CRTC: 1
 CRTCs: 1
 Panning: 0x0+0+0
 Tracking: 0x0+0+0
 Border: 0/0/0/0
 Transform: 1.000000 0.000000 0.000000
             0.000000 1.000000 0.000000
             0.000000 0.000000 1.000000
            filter:
 PANEL_FITTING: full_aspect
  supported: center full_aspect full
 BACKLIGHT_CONTROL: kernel
  supported: native legacy combination kernel
 BACKLIGHT: 7 (0x00000007) range: (0,7)
  1280x800 (0x3e) 83.5MHz *current +preferred
        h: width 1280 start 1352 end 1480 total 1680 skew 0 clock 49.7KHz
        v: height 800 start 803 end 809 total 831 clock 59.8Hz
  1024x768 (0x3f) 94.5MHz +HSync +VSync
        h: width 1024 start 1072 end 1168 total 1376 skew 0 clock 68.7KHz
        v: height 768 start 769 end 772 total 808 clock 85.0Hz
  1024x768 (0x40) 78.8MHz +HSync +VSync
        h: width 1024 start 1040 end 1136 total 1312 skew 0 clock 60.0KHz
        v: height 768 start 769 end 772 total 800 clock 75.0Hz
  1024x768...

[lugburz (lugburz)]

Dear Gear,
  You can maintain your idea about the source of the problem. But installing the old 2.4 version of the intel driver the problem disappeared.

The xrandr --verbose output shows the 1440x900 modeline, and the EDID informations. The only attention I need is to have a virtual size larger than 1440x900 in both dimensions.

[Geir Ove Myhr (gomyhr)]

lugburz, I don't have any intention of maintaining my idea for its own sake, and you seem to have demonstrated well that the 2.4 is able to get the EDID information from the external monitor. Could you attach Xorg.0.log with ModeDebug and the xorg.conf that you use for 2.4? Do I understand you right in that you have to add a Display subsection to the Screen section to set a bigger virtual for it to work with 2.4 and that this does not work with 2.6.3?

Would you mind testing Karmic alpha-2 available from http://cdimage.ubuntu.com/releases/9.10/ or a recent daily-live image from http://cdimage.ubuntu.com/daily-live/ ? If the problem is still there, could you attach the same information for that case (Xorg.0.log with ModeDebug and output of `xrandr --verbose`)? With this information we should be able to send the bug upstream. Would you be able and willing to follow up questions from upstream and test fixes they may provide?

[lugburz (lugburz)]

Hi Geir, yes, using the 2.4 driver without explicitly setting the virtual area to 1440x900 or larger I obtain only resolution within the area of the main display (1280x800), this happened also using 8.10.

I just tried karmic allpha-2 but having the same problem I'm experiencing now: resolution limited to 1360x768. The xrand --verbose output and Xorg.0.log in debugging mode are attached. the files have the suffix -karmic.

The xrand and Xorg.0.log files for the version 2.4 have the suffix xorg-2.4.

I hope I have attached all the files you asked.

[Geir Ove Myhr (gomyhr)]

I think we can push this bug upstream with the current information. Could you take a look at this proposed bug report for upstream and see if it looks all right? (the architecture is missing, do you have 32-bit or 64-bit?)

Title: [i945GM] EDID information of monitor on VGA not detected

Forwarding a bug from a ubuntu user:
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/376722

Bug description:
With a LCD LG TV 19LS4D connected to the VGA port on a Sony VAIO vgn-sz1m, the EDID information is not detected. The EDID information is detected with the 2.4 version of the driver, but not with 2.6.3 and above (also tested with 2.7.99). This results in the native resolution of the external monitor (1440x900) not being available. The maximum is 1360x768.

Since the internal LVDS has a native resolution of 1200x800, he needs to add
 SubSection "Display"
  Virtual 1680 1200
 EndSubSection
to the "Screen" section of xorg.conf in order to get the higher resolution even with the 2.4 driver.

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 945GM/PM/GMS, 943/940GML and 945GT Express Memory Controller Hub [8086:27a0] (rev 03)
 Subsystem: Sony Corporation Device [104d:81e6]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
 Subsystem: Sony Corporation Device [104d:81e6]

System environment:
-- chipset: 945GM
-- system architecture: XX-bit
-- xf86-video-intel: 2.6.3 and 2.7.99
-- xserver: 1.6.1.901 (1.6.2 RC 1)
-- kernel: Linux ubuntu 2.6.30-8-generic #9-Ubuntu SMP
-- Linux distribution: Ubuntu
-- Machine or mobo model: Sony VAIO vgn-szlm
-- Display connector: VGA

Attachments:
Xorg log in karmic
xrandr log in karmic
Xorg log in jaunty using 2.4
Xrandr log using 2.4 in jaunty
xorg.conf

[Geir Ove Myhr (gomyhr)]

Forwarding a bug from a ubuntu user:
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/376722

Bug description:
With a LCD LG TV 19LS4D connected to the VGA port on a Sony VAIO vgn-sz1m, the EDID information is not detected. The EDID information is detected with the 2.4 version of the driver, but not with 2.6.3 and above (also tested with 2.7.99). This results in the native resolution of the external monitor (1440x900) not being available. The maximum is 1360x768.

Since the internal LVDS has a native resolution of 1200x800, he needs to add
 SubSection "Display"
  Virtual 1680 1200
 EndSubSection
to the "Screen" section of xorg.conf in order to get the higher resolution even with the 2.4 driver.

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 945GM/PM/GMS, 943/940GML and 945GT Express Memory Controller Hub [8086:27a0] (rev 03)
 Subsystem: Sony Corporation Device [104d:81e6]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
 Subsystem: Sony Corporation Device [104d:81e6]

System environment:
-- chipset: 945GM
-- system architecture: 32-bit
-- xf86-video-intel: 2.6.3 and 2.7.99
-- xserver: 1.6.1.901 (1.6.2 RC 1)
-- kernel: Linux ubuntu 2.6.30-8-generic #9-Ubuntu SMP
-- Linux distribution: Ubuntu
-- Machine or mobo model: Sony VAIO vgn-szlm
-- Display connector: VGA

Files:
xorg.conf:
http://launchpadlibrarian.net/26749309/xorg.conf
Xorg-log from ubuntu 9.04 with driver version 2.4 installed:
http://launchpadlibrarian.net/27960658/Xorg.0.log-xorg-2.4
Output of `xrandr --verbose` with driver version 2.4:
http://launchpadlibrarian.net/27960705/xrand.log-xorg-2.4
Xorg-log from ubuntu 9.10 which has driver version 2.7.99:
http://launchpadlibrarian.net/27960645/Xorg.0.log-karmic
Output of `xrandr --verbose` with driver version 2.7.99:
http://launchpadlibrarian.net/27960755/xrandr.log-karmic

[Geir Ove Myhr (gomyhr)]

I should add that while the 2.4 logs says that the driver version is 2.4.1, there are quite a number of patches from 2.4.x making it closer to 2.4.3 in practice:
https://launchpad.net/ubuntu/+source/xserver-xorg-video-intel

[lugburz (lugburz)]

Yes, I'm using a 32 bit architecture.

[Geir Ove Myhr (gomyhr)]

I have reported the bug upstream. Please subscribe to the upstream bug report so that you may give the developers any additional information they might need. http://bugs.freedesktop.org/show_bug.cgi?id=22466

[Ling-ma (ling-ma)]

hi Geir
The root cause is your monitor is Digital monitor and not real analog display VGA.
II) intel(0): EDID Version: 1.3
II) intel(0): Digital Display Input

In our latest driver we guarantee monitor connected with VGA must be analog output by if (!edid_mon || DIGITAL(edid_mon->features.input_type) sentence.
Your monitor is pluged by DVI-I VGA convert as bug #21632?
Thanks
Ma Ling

[Geir Ove Myhr (gomyhr)]

Upstream is requesting information about how the VGA is connected. Could you please subscribe to the upstream bug report and reply there?

[Ling-ma (ling-ma)]

http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=200345011383&ih=010&category=11071&ssPageName=WDVW&rd=1

16:10 Aspect Ratio
3000:1 Contrast Ratio
160/160 Degree Viewing Angle
Digital Comb Filter
1440x900 Panel Resolution (Pixels)
5 (ms) Screen Response Rate
HD Ready
Freeview
1 x HDMI Connection
1 x Component Video Connection
1 x RS232 Connection
1 x Standard Scart Connection
1 x RGB Scart Connection
1 x PC D-Sub Connection

From above message, I assume you connect VGA cable to PC D-Sub connection, so if yes, I close the issue, please contact vendor and ask them why D-Sub port will provide digital EDID not analog EDID. Pls open it if you need.

Thanks
Ma Ling

[Geir Ove Myhr (gomyhr)]

Uptream has closed the bug report with NOTOURBUG status. Assuming that the monitor is connected via a regular VGA cable (D-SUB), the monitor actually gives an invalid EDID information, saying that it expects a digital signal and not an analog one. Upstream requested the information about the monitor cable, but didn't get it but found a page where it looked like the monitor has a VGA port. There are cases where this would be a valid EDID (for instance if using the analog part of a DVI-I interface).

The digital bit is the first bit in byte nr. 0x14 in the EDID information. In the EDID block from the intel-2.4 xrandr output this is the 9th and 10th digit on the second line and is equal to 0x80 which in binary is 1000 0000. If the 9th hexadecimal digit on the second line of an EDID block is 8-f the display is expecting a digital signal and 0-7 indicates an analog signal. If you look at other bug reports, all other VGA displays report analog signal (usually the 9th hex-digit is 6). For LVDS displays (laptop displays) it is the same as in your case.