[PATCH] sungem update

[Benjamin Herrenschmidt]

Hi !

> Sorry for taking so long to review this :-/

No worries.

> This looks fine, except for one thing.  The WOL_* registers don't exist
> in the Sun variants of the GEM chip, so it's not safe to try and program
> them on all chips blindly like these changes do.  Perhaps some bit test
> like:
> 
> 	if (pdev->vendor != PCI_VENDOR_ID_SUN)
> 		gp->flags |= GEM_HAS_WOL;
> 
> Then protect all the WOL_* register programming with this new
> GEM_HAS_WOL flag.

Ok, I'll have a look.

Ben.