Re: [patch 2.5] tg3.c: pci_{save,restore}_extended_state

[Jeff Garzik <jgarzik@pobox.com>]

On Fri, Jan 24, 2003 at 03:24:53PM -0500, Wiedemeier, Jeff wrote:
> On Fri, Jan 24, 2003 at 03:05:38PM -0500, Jeff Garzik wrote:
> > That implies we should be disabling PCI_MSI_FLAGS_ENABLE when we first
> > initialize each board, if hardware reset does not automatically do that
> > for us...
> 
> A true hardware reset does reset this bit. It should only be disabled
> arbitrarily if the intent is to *never* use MSI. 

Yes, that's the intent :)


> If the intent is to just not use MSI on tg3 devices, I can use the pci
> quirks to make sure that MSI gets turned off for tg3 devices.

hmmm, maybe I am missing something?

AFAICS, this is a per-driver decision, and needs to be done at the
driver level, in the tg3 driver source.

A quirk would need to be carried to every other PCI-capable
architecture, and every PCI-capable arch would need to be changed if
tg3 suddenly started using MSI again.

	Jeff