Date: Thu, 14 Mar 2002 22:43:27 +0100 (CET) From: Mark Kettenis <kettenis@chello.nl> To: FreeBSD-gnats-submit@FreeBSD.org Subject: kern/35900: Changing RealTek 8139 MAC address fails Message-ID: <200203142143.g2ELhRf00921@elgar.kettenis.dyndns.org>
next in thread | raw e-mail | index | archive | help
>Number: 35900
>Category: kern
>Synopsis: Changing RealTek 8139 MAC address fails
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: freebsd-bugs
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Thu Mar 14 13:50:00 PST 2002
>Closed-Date:
>Last-Modified:
>Originator: Mark Kettenis
>Release: FreeBSD 4.4-RELEASE i386
>Organization:
>Environment:
System: FreeBSD elgar.kettenis.dyndns.org 4.4-RELEASE FreeBSD 4.4-RELEASE #8: Mon Mar 11 02:20:26 CET 2002 kettenis@elgar.kettenis.dyndns.org:/usr/obj/usr/src/sys/ELGAR i386
>Description:
Changing the MAC address for my Sitecom PCI ethernet card,
which is based on the RealTek 8139 chipset, doesn't work.
Apparently the MAC address isn't actually changed in the
hardware. Since the MAC address used by the kernel does get
changed, outgoing packets are sent with the new MAC address.
Any reply-packets will therefore be dropped by the hardware,
and never reach the kernel's networking code unless the
interface is placed in promiscuous mode.
>How-To-Repeat:
Enter the command:
# ifconfig rl0 ether xx:xx:xx:xx:xx:xx
And try to ping another host that would be reached through the
rl0 interface. It won't work unless the interface is placed in
promiscuous mode.
>Fix:
Apparently one has to enter "Config register write enable"
mode to be able to write the MAC address into the appropriate
hardware registers; the chipset documentation doesn't mention
this, but the Linux driver does the same thing.
The attached patch fixes the problem. This patch is against
4.4-RELEASE, but should apply fine to 5.0-CURRENT just as
well. The patch also changes the write access to the ID
registers (whih contain the MAC address) from single-byte
access to 4-byte access (which is what the chipset
documentation demands). The way I implemented this is a bit
of a hack (since ac_enaddr is only 6 bytes long) but there are
several drivers that use similar code.
It would be great if this patch, or something similar, would
be applied and end up in -STABLE someday.
--- /usr/src/sys/pci/if_rl.c.orig Thu Jul 19 20:33:07 2001
+++ /usr/src/sys/pci/if_rl.c Thu Mar 14 21:53:25 2002
@@ -1414,7 +1414,7 @@
struct rl_softc *sc = xsc;
struct ifnet *ifp = &sc->arpcom.ac_if;
struct mii_data *mii;
- int s, i;
+ int s;
u_int32_t rxcfg = 0;
s = splimp();
@@ -1426,10 +1426,15 @@
*/
rl_stop(sc);
- /* Init our MAC address */
- for (i = 0; i < ETHER_ADDR_LEN; i++) {
- CSR_WRITE_1(sc, RL_IDR0 + i, sc->arpcom.ac_enaddr[i]);
- }
+ /*
+ * Init our MAC address. Even though the chipset
+ * documentation doesn't mention it, we need to enter "Config
+ * register write enable" mode to modify the ID registers.
+ */
+ CSR_WRITE_1(sc, RL_EECMD, RL_EEMODE_WRITECFG);
+ CSR_WRITE_4(sc, RL_IDR0, *(u_int32_t *)(&sc->arpcom.ac_enaddr[0]));
+ CSR_WRITE_4(sc, RL_IDR4, *(u_int32_t *)(&sc->arpcom.ac_enaddr[4]));
+ CSR_WRITE_1(sc, RL_EECMD, RL_EEMODE_OFF);
/* Init the RX buffer pointer register. */
CSR_WRITE_4(sc, RL_RXADDR, vtophys(sc->rl_cdata.rl_rx_buf));
>Release-Note:
>Audit-Trail:
>Unformatted:
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200203142143.g2ELhRf00921>