>> Or fbutils. > >fblevel uses the PMU_IOC_SET_BACKLIGHT ioctl; unless the ix86 guys >(and others) implement that one I don't see the point in adding fblevel to >fbutils ?
We need some new fbdev ioctls to handle backlight and other things like video mirroring, etc... This is not simple. For example, backlight can be tricky since it is not always controller by the video chip itself (some powerbooks control it via the PMU, some via the ATI chip backlight control registers for example). We could move the current ioctl to the fbdev easily, but our current mecanism works fine in single-headed configurations only. Similar issues happen with things like mirroring & multihead. Some chips can do mirroring only (I have code for mach64 to do that), so in this case, a single ioctl is enough to enable/disable it. But most chips now are real multihead. They can do either mirroring using different timings than the LCD (useful as most LCDs are 60Hz, which is really low for a CRT), or do multihead. It's not simple to figure out a clean ioctl API to be able to use all those features, especially since mirroring & dual head are fundamentally different at the fbdev level while they concern usually the same chipset features. Enabling/Disabling mirroring can mean disabling/enabling the second head, etc... Ben.
