[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Better work-around: Re: SoundBlaster Live (emu10k1) on PWS433

On Fri, Jul 19, 2002 at 10:17:50PM +0200, Dannis 't Hart wrote:
> On Fri, 2002-07-19 at 18:15, Jay Estabrook wrote:
> > Which suggests that for a 2.4.19 kernel, you might try changing those
> > lines in arch/alpha/kernel/core_cia.c to:
> > 
> >         __direct_map_base = 0x40000000;
> >         __direct_map_size = 0x40000000;
> > 
> > and change the mask to:
> > 
> > #define EMU10K1_DMA_MASK 0x7fffffff
> > 
> Well, yes indeed, this sounds reasonable, doesn't it?
> I tried it, and it works!!!! It's actually all that's required, as far
> as I can see, for now.

Yes, and I feel much better about this solution than allowing the
scatter/gather DMA to be used.

I think that the scatter/gather DMA will actually work without problems
when used across the PCI-PCI bridge (ie to one of the 32-bit slots),
because the transfers get broken up by the HW so that the 8K boundary
crossing is never done. But I'd rather avoid it completely by making
direct-map usable by all your devices, even the broken ones... ;-}

OBTW, I think you should be able to put the Voodoo3 3000 in one of the
64-bit slots (better performance when not going through the bridge),
if you tell SRM:

	>>> set pci_device_override -1

But do the above *before* you try moving the card... ;-}

Good luck.


Jay A Estabrook                            HPTC - LINUX support
Hewlett-Packard Company - MRO1-2/K15       (508) 467-2080
200 Forest Street, Marlboro MA 01752       Jay.Estabrook@hp.com

To UNSUBSCRIBE, email to debian-alpha-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Reply to: