> Okay, first off; FireGL 1000 Pro. This is a Diamond card. Based on
> Permedia2 chipset. I'm having some serious kernel level problems with it.
> [pci.h]
> /* #define PCI_VENDOR_ID_TI 0x104c
> lame hack for the FireGL. it's only seeing the TI RAMDAC! */
> #define PCI_VENDOR_ID_DIAMOND 0x104c
> /* new entry - maybe someday it'll make production :) -prj */
> #define PCI_DEVICE_ID_3DLABS_PERMEDIA2 0x3d07
> [end]
> [pci.c]
> /* fresh hack line -prj */
> case PCI_VENDOR_ID_DIAMOND: return "Diamond Multimedia";
> DEVICE( 3DLABS, 3DLABS_PERMEDIA2,"PERMEDIA2"),
> [end]
[Which vendor do you mean here? 3Dlabs or Diamond Multimedia]
> I do something wrong there? I've spoken with Diamond techs on numerous
> occassions, and they have all confirmed that it's seeing the TI RAMDAC,
> and the card is powered by a Permedia2.
Is this *really* so? I guess Diamond or 3Dlabs really messed this one.
What devices do other Permedia2 boards report?
0x104c has been registered to Texas Instruments (that means that
TI paid money for it) and it would be very strange if linux claimed
that it belongs to Diamond Multimedia.
Old FireGL1000 reports two devices 3Dlabs Delta and 3Dlabs Permedia
correctly registered with 0x3d3d,0x0003 and 0x3d3d,0x0004.
So is TI RAMDAC the only pci-device FireGL1000 Pro reports?
I think linux should show exactly same pci-devices that board reported
and not make assumptions on which board they are on. That same TI RAMDAC
may well be on other boards too.
-maf