Erik Andersen (andersen@inconnect.com) wrote
> Um. How about we include the patch which helps avoid the need to
> specify mem="XXX" for most people, and the few machines that do not
> work (i.e. have broken BIOSs) get to specify mem="XXX" just like
> before. The need of many is greater than the need of few, as long as
> few are still accomodated. I don't see how leaving out the patch
> helps anything.
> Broken BIOSs will still (as before) require that people specify memory
> size. This patch just reduces the population that needs to add the
> additional args to the kernel. The guy who got 1meg instead of 64megs
> detected had to add kernel command line args before the patch also,
> right?
Yes, and no, I'd say, being the guy who got 1meg ,-}}
Well, I agree with you, the "need of many is greater than the need of
few" ! And no: I didn't have to set memory size with kernel
parameter...
Sure, I'll have to use "mem=XXX", so will others... I don't have a
problem with that, but I believe it might be confusing for others, since
it's hard to acually spot the problem when BOOTING, since the kernel
goes on initializing - but with 1 MB of mem, this is a rather problem.
Alan also was of the oppinion to have it dropped on 2.0.xx
Regards
raY
Keep up the great work !