Bug 13195 - support EDID 1.4 format
Summary: support EDID 1.4 format
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Server/General (show other bugs)
Version: 7.2 (2007.02)
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Adam Jackson
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: xorg-7.4
  Show dependency treegraph
 
Reported: 2007-11-12 05:13 UTC by Francis Galiegue
Modified: 2008-02-24 17:35 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
Driver version 1.7.4, no 915resolution hack - failure (62.14 KB, text/plain)
2007-11-13 02:47 UTC, Francis Galiegue
no flags Details
Driver version 1.7.4, with 915resolution hack - SUCCESS (61.93 KB, text/plain)
2007-11-13 02:48 UTC, Francis Galiegue
no flags Details
Driver version 2.1.0, no 915resolution hack - failure (62.08 KB, text/plain)
2007-11-13 02:49 UTC, Francis Galiegue
no flags Details
Driver version 2.1.0, with 915resolution hack - failure (62.00 KB, text/x-log)
2007-11-13 02:50 UTC, Francis Galiegue
no flags Details
Driver version 2.1.1, no 915resolution hack - failure (62.02 KB, text/plain)
2007-11-13 02:50 UTC, Francis Galiegue
no flags Details
Driver version 2.1.1, with 915resolution hack - failure (62.02 KB, text/plain)
2007-11-13 02:51 UTC, Francis Galiegue
no flags Details
The xorg.conf file - UNCHANGED between drivers (3.02 KB, text/plain)
2007-11-13 02:51 UTC, Francis Galiegue
no flags Details
Output of 915resolution -l, after applying - overriden mode is 800x600, 8bpp (847 bytes, text/plain)
2007-11-13 02:54 UTC, Francis Galiegue
no flags Details
xorg.log, with preferredmode option - SUCCESS (62.63 KB, text/plain)
2007-11-20 04:58 UTC, Francis Galiegue
no flags Details

Description Francis Galiegue 2007-11-12 05:13:04 UTC
Setup:
* machine is a Dell OptiPlex 745;
* Gentoo 2007.0, native x86_64;
* xorg 7.2;
* Intel driver 2.1.0 (KO), 1.7.4 (OK);
* graphics chipset Intel 82Q963/965, integrated to motherboard;
* monitor is a Dell E178WFP.

The monitor can do 1440x900 at 60 Hz (the specsheet, available on the net, says so - it even gives the HorizSync and VertRefresh which I've added to xorg.conf just in case). I have made the modeline using gtf 1440 900 60.

I have tried:

* 1.7.4 driver, no 915resolution: KO;
* 1.7.4 driver, with 915resolution: OK;
* 2.1.0 driver, with or without 915resolution: KO.

Relevant sections from xorg.conf:

Section "Device"
        Identifier  "Card0"
        Driver      "i810"
        Option      "NoDDC" "0"
        Option      "DRI" "1"
        VendorName  "Intel Corporation"
        BoardName   "82Q963/Q965 Integrated Graphics Controller"
        BusID       "PCI:0:2:0"
EndSection

Section "Monitor"
        Identifier   "Monitor0"
        VendorName   "Dell"
        ModelName    "E178WFP"
        HorizSync 30-83
        VertRefresh 56-75
        Option       "DPMS" "1"
        Modeline "1440x900"  106.47  1440 1520 1672 1904  900 901 904 932  -HSyn
c +Vsync
EndSection

Section "Screen"
        Identifier "Screen0"
        Device     "Card0"
        Monitor    "Monitor0"
        DefaultDepth 24
        [...]
        SubSection "Display"
                Viewport   0 0
                Depth     24
                Modes "1440x900"
        EndSubSection
EndSection

Output from the X server (short!):

(==) Log file: "/var/log/Xorg.0.log", Time: Mon Nov 12 13:50:04 2007
(==) Using config file: "/etc/X11/xorg.conf"
(II) Module already built-in
(WW) I810: No matching Device section for instance (BusID PCI:0:2:1) found
(II) Module already built-in
(EE) I810(0): EDID Version 1.4 not yet supported
(II) Module already built-in
(WW) I810(0): Failed to set up write-combining range (0xc0000000,0x10000000)

Is that due to the lack of EDID support? Should I report more of the xorg.conf?

(OTOH, on an x86 Pentium M laptop, I didn't need to tell anything to the same version of this driver so that it could correctly drive my laptop screen with a 1280x768@57Hz resolution. It's a builtin Intel 965 AFAIR).
Comment 1 Hong Liu 2007-11-12 18:52:39 UTC
Please turn on option Modedebug and provide the xorg.conf and xorg.log.
Comment 2 Francis Galiegue 2007-11-13 02:47:57 UTC
Created attachment 12502 [details]
Driver version 1.7.4, no 915resolution hack - failure
Comment 3 Francis Galiegue 2007-11-13 02:48:44 UTC
Created attachment 12503 [details]
Driver version 1.7.4, with 915resolution hack - SUCCESS
Comment 4 Francis Galiegue 2007-11-13 02:49:39 UTC
Created attachment 12504 [details]
Driver version 2.1.0, no 915resolution hack - failure
Comment 5 Francis Galiegue 2007-11-13 02:50:08 UTC
Created attachment 12505 [details]
Driver version 2.1.0, with 915resolution hack - failure
Comment 6 Francis Galiegue 2007-11-13 02:50:33 UTC
Created attachment 12506 [details]
Driver version 2.1.1, no 915resolution hack - failure
Comment 7 Francis Galiegue 2007-11-13 02:51:10 UTC
Created attachment 12507 [details]
Driver version 2.1.1, with 915resolution hack - failure
Comment 8 Francis Galiegue 2007-11-13 02:51:56 UTC
Created attachment 12508 [details]
The xorg.conf file - UNCHANGED between drivers
Comment 9 Francis Galiegue 2007-11-13 02:54:30 UTC
Created attachment 12509 [details]
Output of 915resolution -l, after applying - overriden mode is 800x600, 8bpp
Comment 10 Francis Galiegue 2007-11-13 02:55:46 UTC
OK, here are all Xorg.0.log with different driver versions, and 915resolution hack on/off. I've also joined the output of 915resolution -l, WITH the hack applied. I've overriden mode 32, was 800x600, 8bpp, is now 1440x900, 32bpp.
Comment 11 Eric Anholt 2007-11-14 15:18:44 UTC
You're looking for "PreferredMode" "1440x900" in your Monitor section these days.  It was decided that we couldn't really support the nasty old way of specifying a preferred mode without hurting autodetection.

The real issue here is that we're not respecting the version 1.4 EDID from your monitor.  It's probably as much as bumping the maximum version allowed in interpret_edid.c, though I haven't reviewed the 1.3 -> 1.4 changes closely.
Comment 12 Francis Galiegue 2007-11-18 02:17:16 UTC
Hello,

France here so the transports are on strike, couldn't get back to my machine since my last report... I'm taking the opportunity to ask a question.

You say I should try PreferredMode in the Monitor section. Do I leave the modeline in the monitor section or do I zap it? Also, need I leave the Modes section in the Display subsection of the Screen subsection or can I remove it too?
Comment 13 Francis Galiegue 2007-11-20 04:58:32 UTC
Created attachment 12649 [details]
xorg.log, with preferredmode option - SUCCESS

OK, I tried PreferredMode.I added:

Option "PreferredMode" "1440x900"

in the Monitor section. I went "back" to version 2.1.0 of the driver, tried it out...

And it WORKS. Log attached. However, I still had to let the Modeline section describing what 1440x900 was. Lack of EDID biting again?
Comment 14 Michael Fu 2008-01-07 01:24:45 UTC
apprently, this is a bug request to support EDID 1.4 format.. changed the summary line from "Intel driver = 2.1.0 fails to comply to forced widescreen modeline (82Q963/965), 1.7.4 OK (with 915resolution...)" to " support EDID 1.4 format"
Comment 15 Hong Liu 2008-01-08 17:23:49 UTC
Adam Jackson is working on EDID 1.4 support.

Would you please try to pull the latest X server to have a try?

Thanks,
Hong
Comment 16 Hong Liu 2008-01-08 22:16:08 UTC
Reassign to Adam, since he is working on EDID 1.4 support.
Comment 17 Francis Galiegue 2008-01-09 05:50:41 UTC
OK, I'll try and build the latest X server from sources, although I don't know when:

* this is my work machine and I need it daily,
* I've never compiled X from sources - but then documentation exists for this, I'll manage to find it.

Don't expect an update before February though, I'm quite busy these days.

Thanks for the updates!
Comment 18 Adam Jackson 2008-01-10 15:18:11 UTC
This was a bug in xserver 1.2, resolved by commit fb8364bca30fe9268e807b0a9a3ebf875ee1fce2.  xserver 1.3 and later should warn about, but not reject, EDID 1.4 blocks.

xserver 1.5 will make additional use of the new information present in EDID 1.4.  Most of the interesting bits of EDID 1.4 are implemented now, but there is still some polish left to be done.

If your distribution has a package for xserver 1.3 or later, that should work fine.
Comment 19 Benjamin Close 2008-01-11 02:38:20 UTC
Bugzilla Upgrade Mass Bug Change

NEEDSINFO state was removed in Bugzilla 3.x, reopening any bugs previously listed as NEEDSINFO.

  - benjsc
    fd.o Wrangler


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.