| Summary: | [EDID] quirk needed for Sony KDF-60XS955 HDTV | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | xorg | Reporter: | Jeremy Stanley <fungi> | ||||||||||||||
| Component: | Driver/intel | Assignee: | Wang Zhenyu <zhenyu.z.wang> | ||||||||||||||
| Status: | RESOLVED FIXED | QA Contact: | Xorg Project Team <xorg-team> | ||||||||||||||
| Severity: | enhancement | ||||||||||||||||
| Priority: | low | Keywords: | NEEDINFO | ||||||||||||||
| Version: | 7.3 (2007.09) | ||||||||||||||||
| Hardware: | x86 (IA32) | ||||||||||||||||
| OS: | Linux (All) | ||||||||||||||||
| Whiteboard: | |||||||||||||||||
| i915 platform: | i915 features: | ||||||||||||||||
| Attachments: |
|
||||||||||||||||
Created attachment 18120 [details]
output of get-edid utility
Created attachment 18121 [details]
errors reported by get-edid utility
Created attachment 18122 [details]
Xorg log with ModeDebug on for intel video
Created attachment 18123 [details]
xorg.conf corresponding to log
Created attachment 18124 [details]
kernel ring buffer/dmesg output
Created attachment 18125 [details]
output from xrandr --verbose
Your log shows intel version is 2.3.2, please try 2.4.0 or git master. You EDID looks really odd, e.g 1920x540 mode. Try "NoDDC" option, and set a Modeline in Monitor section, and using PreferredMode option to assign it. See intel man page for more details, e.g Section "Monitor" Identifier "HDMI-1" Modeline "1280x720" xxx xxx... Option "PreferredMode" "1280x720" EndSection Section "Device" ... Option "NoDDC" "true" Option "monitor-HDMI-1" "HDMI-1" EndSection On Wed, Aug 06, 2008 at 10:32:49PM -0700, bugzilla-daemon@freedesktop.org wrote: [...] > Keywords| |NEEDINFO Info coming. I was indeed testing 2.3.2 from Debian rather than the 2.4.0 I thought I was using. I'm running into some compilation issues, but I'll reserve that for a post to the list. On a positive note, I must have fat-fingered the NoDDC option in my xorg.conf when I was trying it previously, and then (presumably mis)read a list post suggesting that NoDDC and IgnoreEDID were deprecated in favor of in-driver quirks. After re-trying, I *am* able to get the (2.2.0 and 2.3.2) driver to attempt 1280x720 resolution, but it's fairly messy-looking so I'm holding off marking this resolved until I am able to test 2.4.0 in the next few days. Okay, after testing with Intel video 2.4.0 and properly enabling the NoDDC option in my config, I was able to sync at 720p to my HDTV over the HDMI interface just fine. Turned out I also needed to compile against a newer DRM than was available in Debian/unstable or else I was getting some garbling and inconsistent starts (occasional blank screen instead of the crosshatch). Unless you need any additional information from me on it, this bug can be filed in the PEBCAK pile and closed. |
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.
I'm only effectively able to display 640x480 to my HDTV because I suspect the driver needs an EDID quirk for it. I'm connecting from a G35 onboard Asus P5E-VM HDMI to the HDMI input on my Sony KDF-60XS955 HDTV using a bog-standard HDMI cable. Running Debian/unstable, generic 2.6.25-2-686 Linux kernel, current xorg 7.3 debs (x11-common 1:7.3+14), and intel display driver compiled from 2.4 sources for SVDO HDMI support. With an ATI Radeon and a DVI->HDMI cable I was successfully using: Section "Device" Identifier "ATI Radeon 9000/RV250" Driver "ati" Option "IgnoreEDID" "on" EndSection Section "Monitor" Identifier "Sony KDF-60XS955 HDTV" HorizSync 15-46 VertRefresh 59-61 DisplaySize 320 180 Mode "1280x720" DotClock 74.25 HTimings 1280 1320 1376 1650 VTimings 720 722 728 750 EndMode Option "PreferredMode" "1280x720" EndSection Key here is that I have to override EDID checking because this HDTV only claims support for 480p and 1080i for some reason, but its native resolution is actually 720p. Trying IgnoreEDID (or NoEDID for the DDC module) with the intel driver didn't keep it from believing EDID according to the log, and I see in the list archives that adding a quirk in the driver is the preferred solution anyway. Granted, the timings I used are probably not optimal as I had to recenter the image through my television's service menus. I'm happy to patch, recompile and test whatever fixes you can recommend. If I find time to better familiarize myself with the source, I'll see if I can come up with any patches on my own as well. I'll attach output from get-edid, Xorg.0.log, xorg.conf, dmesg, and xrandr --verbose in following updates