Loading custom DSDT

[Rodd Clarkson]

On Tue, 2005-10-25 at 19:54 +0100, Manuel Moreno wrote:
> Dave Jones wrote:
> > On Tue, Oct 25, 2005 at 10:43:10AM -0400, John W. Linville wrote:
> >  
> >  > I would like to see a patch like this upstream.  I'm not sure how to
> >  > convince Len Brown to let it in.
> >  > 
> >  > I don't think putting it in Fedora is necessarily a bad idea, but it
> >  > is not a slam dunk either (based on support concerns as pointed-out
> >  > elsewhere).
> > 
> > The Intel folks have refused to merge it upstream as they'd rather
> > fix the interpretor to work around broken tables.  Some other vendors
> > are also very helpful in the "Whack BIOS vendor on the head" dept
> > when bad tables get reported to them.
> > 
> > If Fedora carried such a patch, we could forget all about ever
> > looking at fixing ACPI bugs that get reported, as I can guarantee
> > we'd get reports that conveniently 'forgot' to mention they've hacked
> > their DSDT in wierd and wonderful ways.
> > 
> > The ability to screw up AML is hurrendously easy, and the number of
> > wannabe AML hackers frankly, scares me witless.  A lot of folk seem
> > to think that things are as simple as ..
> > 
> > - disassemble DSDT
> > - fix up warnings from AML compiler
> > - put DSDT into initrd.
> > 
> > This is *wrong* on so many levels.  For one, even if it does fix
> > the problems the user was seeing, how does it help the next user
> > that hits the problem on the same hardware ?  We can't expect
> > every user to have to patch their DSDT.  The correct answer
> > is "Fix the BIOS", or where that isn't feasible "Work around
> > it in the interpretor" (Especially if its a widespread problem).
> > 
> > 		Dave
> > 
> 
> I _strongly_ disagree with that and so seems to be the opinion of 
> Ubuntu, Mandriva and Novel/SuSe developpers/users and I'm pretty sure 
> that they have fair good thinking heads too. Are we facing perhaps the 
> NIH RedHat syndrome or is it a desperate and stubborn USA-centric point 
> of view from RH far too afraid from USA barrister greed.

As a user I don't know what a DSDT table is and I don't care.

As a user (and someone who moved from Windows over 8 years ago) I do
worry that Linux might start taking the same "don't fix what's busted -
work around it" thinking that has made Windows (along with other things)
so brain dead.  If something is busted, then fix what's busted.  Don't
make something else work with the bustedness.  That just means we've got
two busted things to fix (and that we're going backward).

I admire the decision (which Dave supports) not to move too far from the
mainstream kernel.  The list of things that might have been added just
because someone thought they were a "Good Idea(TM)" are long and often
pointless.  As far as I'm aware, we've only just got to a point were
we've removed most of the 'busted' add-ins in the kernel from Redhat's
earlier days.  Let's not go back in that direction.


Rodd
-- 
"It's a fine line between denial and faith.
 It's much better on my side"