On Mon, 2012-06-25 at 19:14 +0200, Sven Joachim wrote:
On 2012-06-25 17:25 +0200, Adam Jackson wrote:
This fixes the extra_mode walk to be much more conservative. I still think the whole idea is bogus and that guessing about clone mode sizes is a userspace policy decision, but apparently xrandr --newmode / --addmode is unreasonably burdensome.
This should fix a number of reported regressions, please test.
Does not help in my case, unfortunately: instead of a bogus 1680x945 resolution I get a bogus 1400x1050 rather than the correct 1280x1024.
Going to try Takashi's patch instead.
Takashi's patch will promite 1280x1024 to the default - which is correct - but you'll still see a 1400x1050 in the mode list, because your monitor claims a maximum pixel clock of 140MHz and maximum hsync of 81kHz, and 1400x1050@60 fits in that.
Fixing that would probably require additional quirk work to add "preferred mode is physical pixel size". EDID 1.4 redefines the "first detailed mode is preferred" bit to mean that anyway, but we're not currently using that to filter the mode list.
- ajax