> Le Mardi 25 Novembre 2003 10:11, vous avez écrit : > > Didn't you report you applied my rivafb patch to your 2.6.0-test tree > > recently? Doess it still not work properly? The only thing I cannot get > > working is the backlight, and the cursor on the initial console is weird. > I confirm: > My kernel is catched from bk benh'stree the 21 november. It does not
Please verify the rivafb patch is in there by grep set_ivad_contrast drivers/video/riva/fbdev.c (while in the 2.6.0 source dir). What rsync mirror did you use? Maybe it's not been updated lately. Ben applied my patch on Nov. 16 FWIW. > work with my GeForce4 MX420 + Apple Studio Display 15 > The 2.4.23-pre7-ben works fine (the display of charactes in console seems > slower than in offb). If above test shows a couple of lines containing set_ivad_contrast, the patch has been applied. Please check there's a OF property corresponding to your the panel in your device tree (something with EDID in the property name). I've only had a 15'' flat panel iMac for testing, your display might be different. > > looking into the code, it seems to be possible to give a flatpanel parameter > at boot. No? I tested with video=rivafb,flatpanel rivafb=flatpanel etc... > I cannot find something that makes it work. video=rivafb:flatpanel I suspect. You can also try forceCRTC=0 or forceCRTC=1 If that still fails, 1) try to figure out what the 2.4 kernel thinks the display is hooked up to, and what the display parameters are as read from the EDID property, and 2) try to hard code that for 2.6. Also, have a look at the EDID parsing code in 2.6 radeonfb. Michael