Bug#241256: possible discover bug ? (Was Re: Bug#241256: Debian-Installer bug report)

[Sven Luther <sven.luther@wanadoo.fr>]

On Sun, Apr 04, 2004 at 01:19:55AM -0500, David Nusinow wrote:
> On Thu, Apr 01, 2004 at 11:30:47AM +0200, Sven Luther wrote:
> > On Thu, Apr 01, 2004 at 11:34:05AM +0900, Takushi Fujiwara wrote:
> > > ps.
> > > Installer detect iMac G3(Early 2001) Build in Ethernet but module is not
> > > installed.
> > 
> > The :
> > 
> >     < >   GMAC (G4/iBook ethernet) support (OBSOLETE, use Sun GEM) 
> >  
> > driver has been replaced by the : 
> > 
> >     <*>   Sun GEM & Apple GMAC support 
> > 
> > which should do as fine for you, and is indeed builtin. This should work
> > without problem for you, or else i would like to know about it.
> > 
> > The more probable reason for what i seem to understand your problem is,
> > is that discover/discover2 doesn't know about this change, and is not
> > able to load the sungem driver for your hardware, this would be a
> > bugreport against discover if this was the case, but to be certain, we
> > need more info.
> 
> Right. To fix this, or even check on it, I'll need the output of lspci
> and lspci -n. Only a couple of things in discover's current database use
> the sungem driver, so if there are more now, I'd like to get them in.
> Even if the driver is compiled in to the kernel right now, I'd still
> like the entry in the database for completeness sake as well as a
> fallback.

What about using the kernel source as inspiration ? 

static struct pci_device_id gem_pci_tbl[] = {
        { PCI_VENDOR_ID_SUN, PCI_DEVICE_ID_SUN_GEM,
          PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL },

        /* These models only differ from the original GEM in
         * that their tx/rx fifos are of a different size and
         * they only support 10/100 speeds. -DaveM
         * 
         * Apple's GMAC does support gigabit on machines with
         * the BCM54xx PHYs. -BenH
         */
        { PCI_VENDOR_ID_SUN, PCI_DEVICE_ID_SUN_RIO_GEM,
          PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL },
        { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_UNI_N_GMAC,
          PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL },
        { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_UNI_N_GMACP,
          PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL },
        { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_UNI_N_GMAC2,
          PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL },
        { PCI_VENDOR_ID_APPLE, PCI_DEVICE_ID_APPLE_K2_GMAC,
          PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0UL },
        {0, }
};

#define PCI_VENDOR_ID_SUN               0x108e
#define PCI_DEVICE_ID_SUN_GEM           0x2bad
#define PCI_DEVICE_ID_SUN_RIO_GEM       0x1101
#define PCI_VENDOR_ID_APPLE             0x106b
#define PCI_DEVICE_ID_APPLE_UNI_N_GMAC  0x0021
#define PCI_DEVICE_ID_APPLE_UNI_N_GMACP 0x0024
#define PCI_DEVICE_ID_APPLE_UNI_N_GMAC2 0x0032
#define PCI_DEVICE_ID_APPLE_K2_GMAC     0x004c

Taken from the sungem driver of the 2.4.25-powerpc kernel.

Friendly,

Sven Luther