cvs commit: src/sys/dev/sound/pci emu10k1.c
scottl at freebsd.org
Wed Jun 16 02:41:23 GMT 2004
On Tue, 15 Jun 2004, David O'Brien wrote:
> On Mon, Jun 14, 2004 at 01:10:12AM +0200, Marius Strobl wrote:
> > On Sun, Jun 13, 2004 at 10:12:02PM +0000, David E. O'Brien wrote:
> > > obrien 2004-06-13 22:12:02 UTC
> > >
> > > FreeBSD src repository
> > >
> > > Modified files:
> > > sys/dev/sound/pci emu10k1.c
> > > Log:
> > > Add PCI identifier for Dell modified SBLive! card
> > >
> > > Submitted by: Joseph Dunn <joseph at magnesium.net>
> > >
> > Wasn't there consensus that simply adding its PCI ID isn't enough to
> > get this card working?
> No, not with the current emu10k1 driver. I think my commit gets us
> closer to working with the Dell emu10k1 hardware. I really don't think
> some dell-emu10k1 driver will be written vs. adding extra bits to this
> emu10k1 driver if needed. So getting the proper driver to at least probe
> and try to work with the hardware are steps closer than not even trying
> to work with the Dell emu10k1 hardware.
> -- David (obrien at FreeBSD.org)
No, please don't add ids to drivers without confirmation that the
hardware will work with the driver. You are muddying the 'support'
issue, and this gets you no more than what 'pciconf' can give you.
By your logic, we should just suck the entire pci id database into
the kernel, and you know that thatis silly.
For someone who is usually so concerned about FreeBSD standards and
what goes into the tree, this is very unlike you. Please take out
this change until there is confirmation that it will work. If it
is te same Dell SbLive card that I've heard about, then the emu10k1
driver won't be able to do a thing with it, period. It may even, as
Peter hinted, cause a lockup on attach.
More information about the cvs-src