Date: Fri, 22 Aug 2003 20:25:00 +0200 From: Marius Strobl <marius@alchemy.franken.de> To: Pete Bentley <pete@sorted.org> Cc: freebsd-sparc64@freebsd.org Subject: Re: Netra X1 Message-ID: <20030822202500.A41470@newtrinity.zeist.de> In-Reply-To: <BB9130B8-D48A-11D7-A4E7-000A959F6E56@sorted.org>; from pete@sorted.org on Fri, Aug 22, 2003 at 11:23:54AM %2B0100 References: <20030822013101.A24406@newtrinity.zeist.de> <BB9130B8-D48A-11D7-A4E7-000A959F6E56@sorted.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Aug 22, 2003 at 11:23:54AM +0100, Pete Bentley wrote: > On Friday, August 22, 2003, at 12:31 AM, Marius Strobl wrote: > > Just curious, could you please try the attached patch and check if the > > MAC addresses get set and then set local-mac-address in the > > OpenFirmware > > and check again if the dc interfaces get different MAC addresses? > > The patch should work (I'm just not sure about the different MAC > > addresses > > part), > > I'll have a play with it - it looks helpful but not quite right. > > For example if dc_read_eeprom() ever gets a non-zero mac address it'll > skip over the OF code. Also if local-mac-address? is true it looks like Which will be the case if someone puts a PCI card with a Davicom DM9102 into a SUN box. > it'll leave a zeroed MAC address on the interface, but for > local-mac-address? I couldn't test this but this part is actually code from Jake Burkholder so I'm confident that it'll work. > being false, it'll program the interface with the hosts's MAC address. > > That last case is handy for me and possibly others running X1's... > > > however Maxime doesn't want to have this in the device driver but > > at a higher level. > > I'd agree with that - otherwise doesn't every ethernet driver on sparc64 > need extra code to decide whether to respect the local-mac-address > setting? > For all drivers except hme(4) and gem(4) I wouldn't check it all. AFAIK local-mac-address is only an issue with hme(4) and gem(4) add-in cards, which ship with a ROM containing their own fcode including the MAC address. The onboard hme(4) and gem(4) NICs don't have seperate fcode, at least the boards with a single interface. I don't know what SUN did with the boards that have multiple ethernet interfaces onboard, like the Netra T1 200 with two gem(4) or Netra X1 with two dc(4). Therefor I asked you to test the patch, I guess others would also be interested to know if it's necessary to check local-mac-address for the onboard dc(4) interfaces on the Netra X1. Marius
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030822202500.A41470>