>Number:         24043
>Category:       port-i386
>Synopsis:       sip driver not working with sis962 integrated ethernet
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    port-i386-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 10 22:33:02 UTC 2004
>Closed-Date:
>Last-Modified:
>Originator:     Roland McGrath
>Release:        current/200401030000
>Organization:
>Environment:
NetBSD dhcp-0.sf.frob.com 1.6ZG NetBSD 1.6ZG (GENERIC) #0: Sun Jan  4 19:24:46 UTC 2004  cyber@tgm.netbsd.org:/autobuild/HEAD/i386/OBJ/autobuild/HEAD/src/sys/arch/i386/compile/GENERIC i386
>Description:
See PR 23481 for some related context.  I had that same problem (bad MAC address detection and no PHY detection) with 1.6.2-RC4.  In current (if_sip.c rev 1.86), the address and media detection/autonegation work fine.  However, I can't actually get packets out (or maybe in too).
After failing to send packets, it reports "sip0: device timeout".

The probe messages are thus:

sip0 at pci0 dev 4 function 0: SiS 900 10/100 Ethernet, rev 0x91
sip0: interrupting at irq 5
sip0: Ethernet address 00:10:dc:98:ed:b7
ukphy0 at sip0 phy 1: Generic IEEE 802.3u media interface
ukphy0: OUI 0x000004, model 0x0020, rev. 1
ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto

FWIW, the specs for my motherboard (MSI 651M Combo-L) say what it has is
"SiS 962L integrated MAC + RealTek RTL8201BL PHY".  This hardware works fine under Linux 2.4.22.

>How-To-Repeat:
Bring up sip0.  Try to use it (e.g. dhclient).  No packets seen on the wire.  After a few seconds trying to send packets, "sip0: device timeout" seen.

I would be happy to do anything I can to help debug the problem,
but I don't really know where to start.  BTW, I really want to be running the stable branch.  I backported the if_sip.c changes from -current to 1.6.2-RC4 and that solved the problem of not detecting the PHY/media/address, but behaves in the same losing way for actual packets.
This report is for the -current kernel, though I hope to backport the fix to 1.6.2 after it's resolved in -current.
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: