Realtek RTL8169 on FreeBSD 5.4: no carrier.

Thu Aug 11 19:16:35 GMT 2005

>>>I recently installed FreeBSD 5.4 on an ABIT AA-8 DuraMax and all went
>>>well.  All hardware detected properly and everything was running great,
>>>until I got to configuring my network.  ifconfig shows my onboard
>>>gigabit LAN as "status: no carrier"
>>>
>>>I can successfully ping localhost and the IP that was assigned to re0
>>>(192.168.1.31).
>>>
>>>when I plug an ethernet cable from my FreeBSD box to my router, I get
>>>"status: no carrier."  Oddly, when I plug an ethernet cable from my
>>>FreeBSD box to my laptop's LAN port, I get "status: active".  The lights
>>>on the ethernet jack indicate the same.
>>>
>>>Additionally, if I manually set the media with the following command:
>>># ifconfig re0 media 10baseT/UTP mediaopt full-duplex
>>>The status magically switches to "active" and I can use my ethernet!
>>>
>>>I know that there are known problems with RealTek chipsets, but it is
>>>listed in the 5.4 supported hardware list.
>>>
>>>Bottom line is that the onboard LAN is detected, installed, and working
>>>properly, but it seems as if the driver can't properly detect when a
>>>cable is plugged into the jack.
>>>
>>>I was hoping someone could help.
>>>
>>>uname -a:
>>>--------------------------------------
>>>FreeBSD db.domain.com 5.4-RELEASE FreeBSD 5.4-RELEASE #0: Sun May 8
>>>10:21:06 UTC 2005
>>>root at harlow.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC i386
>>>
>>>ifconfig:
>>>--------------------------------------
>>>re0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
>>>options=18<VLAN_MTU,VLAN_HWTAAGGING>
>>>inet 192.168.1.31 netmast 0xffffff00 broadcast 192.168.1.255
>>>inet6 ...
>>>ether 00:50:8d:eb:e5:be
>>>media: Ethernet autoselect (none)
>>>status: no carrier
>>>lo0: flags=8049<UP,LOOPBACK,RUNNING,MULTICAST> mtu 16384
>>>inet 127.0.0.1 netmask 0xff000000
>>>inet6 ::1 prefixlen 128
>>>inet6 fe80::1%lo0 prefixlen 64 scopeid 0x4
>>>
>>>
>>>Relevant dmesg:
>>>--------------------------------------
>>>re0: <RealTek 8169S Single-chip Gigabit Ethernet> port 0xee00-0xeeff mem
>>>0xfbfff000-0xfbfff0ff irc 16 at device 1.0 on pci1
>>>miibus0: <MII bus> on re0
>>>rgephy0: <RTL8169S/8110S media interface> on miibus0
>>>rgephy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseTX,
>>>1000baseTX-FDX, auto
>>>re0: Ethernet address: 00:50:8d:eb:e5:be
>>>
>>>
>>>pciconf -lv:
>>>--------------------------------------
>>>re0 at pci1:1:0: class=0x020000 card=0x1039147b chip=0x816910ec rev=0x10
>>>hdr=0x00
>>>vendor = 'Realtek Semiconductor'
>>>device = 'RTL8169 Gigabit Ethernet Adapter
>>>class = network
>>>subclass = ethernet
>>>
>>>pciconf -r pci1:1:0 0:0xff
>>>--------------------------------------
>>>816910ec 02b00007 02000010 00002008
>>>0000ee01 fbfff000 00000000 00000000
>>>00000000 00000000 00000000 1039147b
>>>00000000 000000dc 00000000 40200110
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 f7c20001
>>>00000000 00000000 00000000 00000000
>>>00000000 00000000 00000000 00000000

>> Regrettably, i always encountered this problem.  I spoke about that since
>> the middle of 2004, and didn't really receive feedback on this.  I try a
>> lot of things but none worked better than the other.
>>
>> To not forget about it, i filled a bug report on this particular problem,
>> see PR kern/80005 for more details.
>>
>> The last thing i want to give another try is to upgrade to RELENG_6,
>> since i currently follow the RELENG_5 branch.  But i am not *very*
>> confident about that...
>>
>> Sorry not to have better answer to give you.

> what irq information for re0 is listed when you do "vmstat -i"?

I get this one:
 $ vmstat -i
 interrupt                          total       rate
 irq1: atkbd0                          10          0
 irq8: rtc                       11094577        127
 irq11: acpi0                       13495          0
 irq13: npx0                            1          0
 irq14: ata0                       197330          2
 irq15: ata1                           85          0
 irq17: fwohci0                         1          0
 irq19: re0                         10079          0
 irq20: ohci0                           1          0
 irq21: ohci1                           2          0
 irq22: ohci2                         225          0
 irq0: clk                        8667646         99
 Total                           19983452        230

Is there any problem here?

> theres known issues for gigabit devices like this. I know some people
> have fixed the "no carrier" issue (with broadcom gigabit) by using the
> windows ndis driver. perhaps you can give that a try?

Yes, at the beginning of this year.  Without much success.

> before using the ndis driver give this patched mod a try:
>  http://people.freebsd.org/~wpaul/re/

I tried these the past, but don't remember this solved anything.

> another common issue with gigabit devices is when the interface is set
> up to "autoselect".  Does the the diode on the ethernet card and the
> diode on the switch/hub/whatever blink at the same time and then turn
> off (or something along the line)?

Since it is a notebook, i didn't get any diode on the ethernet adapter.
Just get one on the network switch.

When booting and before loading the kernel i get a green light (there is
a link and is ok).  During the load of the kernel, the diode gets orange
(there is a link but something is wrong).  When starting the DHCP client,
the light symply disappeared.  After tens of second (~one to two minutes),
the light reappeared and become green.

> the fix for this is to manually setup the mediatype to a fixed speed and
> duplex. That has been a fix for many people.

Also tried this using '/etc/start_if.re0'.  Same problem here.

Thanks,
-- 
-jpeg.