On Mon, Jul 23, 2007 at 07:39:20PM +0200, Quentin Garnier wrote:
> On Mon, Jul 23, 2007 at 01:07:36PM -0400, diro@nixsyspaus.org wrote:
> > On Mon, Jul 23, 2007 at 05:40:00PM +0200, Quentin Garnier wrote:
> > > On Mon, Jul 23, 2007 at 12:42:00AM -0400, diro@nixsyspaus.org wrote:
> > > > http://mail-index.netbsd.org/netbsd-bugs/2007/07/21/0001.html
> > > > 
> > > > I looked through the sources for azalia.c, azalia.h, and azalia_codec.c and
> > > > reviewed some private correspondence from another NetBSD user. I am making time
> > > > in my schedule to work on this; however, it's difficult as i am running 3.1
> > > > instead of -current, so i don't think i'd be able to test any of the code other
> > > > than for syntax.
> > > > 
> > > > Also, i wasn't sure if it was an issue with the driver or something with the
> > > > motherboard. I would like to start working on these sorts of issues in NetBSD.
> > > > In a situation like this, what's the best place to start? What's most likely the
> > > > cause of the problem? Please advise.
> > > 
> > > It's a BIOS bug.  More precisely, the BIOS doesn't support 32-bits OSes,
> > > at least those that don't do PAE.
> > > 
> > > Check the motherboard's vendor site for upgrades or use NetBSD/amd64...
> > 
> > Hmm, it's strange that the BIOS would be 64 bit and the processor 32bit. I tried
> 
> Well, it might just be unfriendly to non-PAE systems.
> 
> > booting with amd64 and that was not possible. Do you think swapping the
> > processor with a 64bit processor would make this work? Is there another way to
> > modify the BIOS to support PAE? Strange that it doesn't, because it's a fairly
> > new board.
> 
> NetBSD is the one thing not supporting PAE here.
 

Oh, i thought NetBSD had PAE :0


> > The US representative of the company says that no firmware upgrades are
> > available for the motherboard. It's very difficult to find firmware for this
> > particular model, the Durabook S14D (D14RI).
> 
> You might have some setting in the BIOS regarding PCI memory spaces and
> where they should be mapped.


I'll check again, but the BIOS is so featureless. It's like they used the oldest
possible version of AMIBIOS available.


> Otherwise you'll have to ask nicely for someone to finish the work of
> merging amd64 and i386 pmaps so that i386 can have PAE support almost
> for free :-)
> 
> -- 
> Quentin Garnier - cube@cubidou.net - cube@NetBSD.org
> "You could have made it, spitting out benchmarks
> Owe it to yourself not to fail"
> Amplifico, Spitting Out Benchmarks, Hometakes Vol. 2, 2005.


I'd love if that was pulled into 3.1.1 or 4.0. The place i work is making
everyone transition to a VoIP system very soon and i need audio to work on my
laptop. The thought of running some other OS in a VM just to get audio is just 
horrifying to me. Please, save me from such a fate! :)