[haiku-development] Re: Proposed Change to Common Accelerant for EDID

Axel Dörfler wrote:

Gerald Zajac <zajacg@xxxxxxxxxxxxx> wrote:

Anyway, attached is a diff file with my proposed changes.
As a temporary solution, I think it's okay -- but it's a problem that should be fixed in a generic way. 


Hi Axel;
Since you feel that my fix is okay as a temporary solution, would you please commit the changes since I do not have commit access. That diff file also made a couple changes to the base mode list, but those were done only to make those modes more compliant with the VESA spec's. Thank you.
BTW, have you tried 1152x864 @60, 70 or 72Hz with any of the Intel video chips? I'm wondering if you encountered results similar to what I obtained with the ATI and S3 chips. AFAIK, the intel_extreme, ati, and s3 accelerants are the only ones that use the common accelerant for building a mode list.
The Screen Preferences merely adjusts the pixel clock when it can not find a matching refresh rate for a specified resolution. Is this the correct way of doing it, or should it also be adjusting some of the other parameters in the mode info? 

It should also adjust the other parameters.
The propose_mode() function in the accelerant should only make sure that the graphics card and monitor can show the mode - the Screen preferences should therefore ask for good modes. It should use the GTF to compute the modes that are not in the card's mode list.
I agree. The Screen Preferences should use GTF to compute the modes that are not in the chip's mode list. Also, the Screen Preferences should use a refresh rate tolerance of probably 0.5% or 0.6% percent when deciding if a matching mode is found. I've noticed that it adjusts the pixel clock of modes where the refresh rate differs by only a fraction of a percent. It should be noted that the refresh rate calculations are not precise to an exact integer value. 

--Gerald

Other related posts: