7.1-R to RELENG_7 upgrade breaks re nic

[Pyun YongHyeon]

On Wed, Feb 25, 2009 at 10:47:07PM -0500, Steve Wills wrote:
> On Feb 25, 2009, at 7:38 PM, Pyun YongHyeon wrote:
> >I need more information for your hardware revision.
> >Would you show me dmesg output and revision number of if_re.c?
> 
> I assume you only need the re0 related output. If you need the full  
> dmesg, let me know.
> 
> re0: <RealTek 8168/8168B/8168C/8168CP/8168D/8111B/8111C/8111CP PCIe  
> Gigabit Ethernet> port 0x7e00-0x7eff mem 0xfd3ff000-0xfd3fffff, 
> 0xfd3f8000-0xfd3fbfff irq 16 at device 0.0 on pci8
> re0: Chip rev. 0x28000000
> re0: MAC rev. 0x00100000
> re0: Ethernet address: 00:1f:d0:af:1a:4c
> re0: [FILTER]
> re0: link state changed to UP
> 
> $FreeBSD: src/sys/dev/re/if_re.c,v 1.95.2.41 2009/02/09 01:38:01  
> yongari Exp $
> 
> I get 3 link state DOWN/UP notices when DHCP client starts. It works  

That's normal(Technically this is not correct behavior but it's the
way how it was implemented in driver). 

> for exactly 60 seconds after boot, then stops.

I guess re(4) thinks it lost established link. How about unplug and
then replug UTP cable? Would you show me "devinfo -rv | grep phy"?

>                                                Patch from earlier "fun  
> with if_re" thread didn't help,

Your issue is completely different one.    

>                                 if_re.c from -CURRENT failed to build.  

You have to use if_re.c/if_rl.c and if_rlreg.h from CURRENT to
build it on stable.

> Reverting back to rev 1.95.2.36.2.2 fixes it.
> 

Your controller looks like RTL8168D PCIe controller. ATM I have no
idea why if_re.c 1.95.2.41 does not work. I'll let you know if I
find a clue.