Re: Using PXE+GRUB to boot Linux on a Thinkpad T20

[Patrick J. LoPresti]

Donald Becker <address@hidden> writes:

> On 29 Jul 2002, Patrick J. LoPresti wrote:
> >
> > How can I determine the subsystem?
> 
> lspci -s 0:3 -v -n

Subsystem: 10b7:6356

> > Interesting.  Is pci-scan slated for inclusion in the stock kernel?
> 
> No.  The pci-scan interface was defined in 1999 to combine PCI
> detection, PCI power managemet and hot-swap PCI/CardBus in one
> cross-version compatible interface.  The interface was reject in
> favor of a initially-simpler PCI detection function that didn't
> address all of the issues.

Um, does this mean that Donald Becker is no longer contributing to the
stock kernel?  If so, that sounds like an enormous loss for Linux...

I am not qualified to evaluate the technical merits of the interfaces.
But I do know that your driver is working for me while the one in the
stock kernel is not.

As a sysadmin, I also know that local changes are very expensive in
the long run; I am a big believer in sticking with vendor-supplied
tools.  So I consider it important to see this fixed in the stock
kernel, or at least in Red Hat's.

Do you have any suggestions for how I, as a simple end-user, can help
make that happen?  (Would there be any point in lobbying the kernel
maintainers?)

> > Well, it is not a -5 interface.  It's a "3c1556B mini-PCI".  No "-5";
> 
> What's the driver detection message?

It says:

  3c59x.c: v0.99Xd 7/26/2002 Donald Becker, address@hidden
  eth0: Overriding PCI latency timer (CFLT) setting of 0, new value is 32.
  eth0: 3Com 3c1556B mini-PCI at 0x1400, 00:00:86:43:cd:46, IRQ 11

...then it goes on about mapping the CardBus functions.

Incidentally, is the proper name for this chipset "3c556B" or
"3c1556B"?  If the latter, somebody should probably tell these folks:

  http://pciids.sourceforge.net/iii/?i=10b7


 - Pat