vge0 not autonegotiating to 1000baseTX full duplex in 7.1

[Pyun YongHyeon]

On Sun, Mar 08, 2009 at 11:56:41PM -0700, David Ehrmann wrote:
> Pyun YongHyeon wrote:
> >On Sun, Mar 08, 2009 at 05:10:58PM -0700, David Ehrmann wrote:
> >  
> >>Pyun YongHyeon wrote:
> >>    
> >>>On Sat, Mar 07, 2009 at 09:22:02PM -0800, David Ehrmann wrote:
> >>> 
> >>>      
> >>>>It's been reported before, but I haven't seen anything new.  vge 
> >>>>devices 
> >>>>        
> >>>Because I don't have access to the hardware it looks like hard to
> >>>fix.
> >>>
> >>> 
> >>>      
> >>>>won't autonegotiate to gigabit speeds, and if I set the media to 
> >>>>1000baseTX, ifconfig reports "no carrier."  This was with two different 
> >>>>interfaces on the other end, one a switch, the other another computer 
> >>>>(but not a vge one).
> >>>>
> >>>>Any ideas?
> >>>>   
> >>>>        
> >>>Would you show me the output of dmesg?(Only vge(4) related one)
> >>>Also show me the output of "devinfo -rv | grep phy".
> >>> 
> >>>      
> >>dmesg:
> >>
> >>vge0: <VIA Networking Gigabit Ethernet> port 0xe800-0xe8ff mem 
> >>0xfeaffc00-0xfeaf
> >>fcff irq 28 at device 0.0 on pci3
> >>vge0: Reserved 0x100 bytes for rid 0x14 type 3 at 0xfeaffc00
> >>miibus0: <MII bus> on vge0
> >>ip1000phy0: <IC Plus IP1001 10/100/1000 media interface> PHY 22 on miibus0
> >>ip1000phy0:  10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseTX, 
> >>1000bas
> >>eTX-FDX, auto
> >>vge0: WARNING: using obsoleted if_watchdog interface
> >>vge0: bpf attached
> >>vge0: Ethernet address: 00:40:63:xx:xx:xx
> >>ioapic1: routing intpin 4 (PCI IRQ 28) to vector 49
> >>vge0: [MPSAFE]
> >>vge0: [ITHREAD]
> >>
> >>
> >>devinfo -rv | grep phy
> >>               ip1000phy0 pnpinfo oui=0x90c3 model=0x19 rev=0x0 at 
> >>               phyno=22
> >>           ukphy0 pnpinfo oui=0x4063 model=0x32 rev=0xa at phyno=1
> >>
> >>    
> >
> >Would you try attached patch? Due to lack of hardware access I
> >don't know whether it helps or not(Just compilation tested).
> >  
> I have no idea how you did it,

Hmm... it was just a guess and this time I was lucky. :-)

>                                but the patch seems to have worked.  

Glad to hear that.

> ifconfig reports 1000baseTX, and I can nc zeros at more than 50 MB/s.
> 
> Is this patch "stable?"  Should it break any existing functionality?  

As you might guess, only you can say "how stable" for this patch. 

> What was the problem that it fixed?  Will it be in the next 7.x 

I guess newer IC Plus PHYs require correct next page bit so I
just read the autonegotiation advertisement register to cache the
next page capability. All other changes are to use
mii_phy_add_media() KPI which will remove duplicated codes in phy
drivers.

> release?  8.0?
> 

Yes, I'll commit the patch soon and make sure to MFC to 7-stable.

> Here's the link to the bug id for it:
> http://www.freebsd.org/cgi/query-pr.cgi?pr=130846
> 

Ok, I'll handle the PR.

> Thank you!

No problem. Thanks for testing!