On Tue, Jan 18, 2005 at 06:43:50PM +0100, Thiemo Seufer wrote: > Horms wrote: > > On Mon, Jan 17, 2005 at 08:30:24PM +0100, Georg Wittenburg wrote: > > > Package: kernel-source-2.6.10 > > > Version: 2.6.10-3 > > > Severity: normal > > > > > > > > > When the intelfb module is compiled into the kernel all consoles remain > > > blank until the X server starts up. Using it as a module is not an > > > option for non-CRT screens (see > > > http://www.xfree86.org/~dawes/intelfb.html, which I'm assuming not to be > > > outdated on this issue). This makes the driver unusable for laptops. > > > > hlh has already mentioned that we need to walk away from > > the modular fb code, for reasuns such as this. I am not > > sure what unplesant side effects that is going to > > have in relation to mkinitrd and the debian installer. > > But perhaps 2.6.10 would be a good place to start the > > rollback. > > Does this mean all fb drivers would need to be built in?
I spoke to dilinger about this on IRC, and my intepretation of hch's (not hlh, he is a different guy) remarks seem to have been a bit off. I now think that he was only talking about vesafb. Which along with intelfb would make two drivers. What may well be a good solution, which was kind of what I was thinking of, and what dilinger also thought might work, is to put loading the fb drivers into initrd. That may help a lot of problems. But as for vesafb, there seems to be a consensus that it needs to be non-modular. And the same seems to be true of intelfb, so they seem destined to be built into the kernel. I guess the real trick is finding which ones need to be in the kernel, and which can live as modules. Hopefuly there aren't to many of the former. -- Horms -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
