wm driver on SuperMicro ... quirks?

[Havard Eidnes <he%uninett.no@localhost>]

Hi,

I run a number of NetBSD hosts which get their network
connectivity via the use of the "wm" driver.

I am observing a few ... "quirks" which I thought I would ask if
anyone else is seeing -- so far not reported as a PR.

Both of the hosts mentioned in 1) and 2) below run NetBSD/amd64
10.0.

1) One host with an i82576 (wm) + i82566 (igphy) is connected to
   an old Juniper ex4200 switch.

   At startup, it completes auto-negotiation at 1G FDX.  However,
   after a while (while I'm not looking), the host and switch
   apparently agree to settle on 100Mbit/s FDX instead.  This
   does "wonders" when the nightly backup run commences, it
   manages to saturate the capacity.

   If I manually do "ifconfig wm0 down; ifconfig wm0 up", it once
   again completes auto-negotiation at 1G FDX.  However, that is
   not a permanent fix, as sure enough, it later on decides on
   its own to settle on 100Mbit/s FDX instead.

2) Another host with the same combination runs wm0 and wm1 in an
   LACP bundle, connected to the same switch as above.  The
   problem seen in this combination is that occasionally the host
   decides that wm0 is no longer operational, and the switch
   "agrees".  The host may sit there for an hour with only one
   member in the LACP bundle, and apprently at random decides to
   bring the interface back up again.  Manually forcing an
   interface reset with "ifconfig wm0 down; ifconfig wm0 up"
   gives immediate remediation of this particular instance, "of
   course".

I've tried replacing the copper patch cable with a new one, to no
avail.

Hmm...  Zooming out, both these hosts are using SuperMicro
motherboards (both appear to use H8DGU boards) and AMD processors
(6212 and 6274 respectively)

One of them is also reporting the occasional

wm0: device timeout (txfree 4095 txsfree 63 txnext 82)

but it's not clear that event is connected to either of the above
two types of events.

Anyone have any comments or are seeing anything similar?

Parts of dmesg:

ppb0: PCI Express capability version 2 <Root Port of PCI-E Root Complex> x4 @ 5.
0GT/s
ppb0: link is x4 @ 2.5GT/s
pci1 at ppb0 bus 2
pci1: i/o space, memory space enabled, rd/line, wr/inv ok
wm0 at pci1 dev 0 function 0, 64-bit DMA: 82576 1000BaseT Ethernet (rev. 0x01)
wm0: interrupting at ioapic1 pin 20
wm0: PCI-Express bus
wm0: 16384 words (16 address bits) SPI EEPROM, version 1.43, Image Unique ID e6060000
wm0: RX packet buffer size: 64KB
wm0: Ethernet address 00:25:90:86:b2:a0
wm0: Copper
wm0: 0x374040<SPI,PCIE,NEWQUEUE,ASF_FIRM,ARC_SUBSYS,MANAGE,WOL>
igphy0 at wm0 phy 1: i82566 10/100/1000 media interface, rev. 1
igphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
wm1 at pci1 dev 0 function 1, 64-bit DMA: 82576 1000BaseT Ethernet (rev. 0x01)
wm1: interrupting at ioapic1 pin 21
wm1: PCI-Express bus
wm1: 16384 words (16 address bits) SPI EEPROM, version 1.43, Image Unique ID e6060000
wm1: RX packet buffer size: 64KB
wm1: Ethernet address 00:25:90:86:b2:a1
wm1: Copper
wm1: 0x274040<SPI,PCIE,NEWQUEUE,ASF_FIRM,ARC_SUBSYS,WOL>
igphy1 at wm1 phy 1: i82566 10/100/1000 media interface, rev. 1
igphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
ahcisata0 at pci0 dev 17 function 0: ATI Technologies SB700-SB900 SATA Controller (AHCI mode) (rev. 0x00)

and

ppb1 at pci0 dev 4 function 0: ATI Technologies RD890 PCI Express Bridge GPP Port D (rev. 0x00)
ppb1: PCI Express capability version 2 <Root Port of PCI-E Root Complex> x4 @ 5.0GT/s
ppb1: link is x4 @ 2.5GT/s
pci2 at ppb1 bus 7
pci2: i/o space, memory space enabled, rd/line, wr/inv ok
wm0 at pci2 dev 0 function 0, 64-bit DMA: 82576 1000BaseT Ethernet (rev. 0x01)
wm0: interrupting at ioapic0 pin 16
wm0: PCI-Express bus
wm0: 16384 words (16 address bits) SPI EEPROM, version 1.43, Image Unique ID e6060000
wm0: RX packet buffer size: 64KB
wm0: Ethernet address 00:30:48:f5:15:80
wm0: Copper
wm0: 0x374040<SPI,PCIE,NEWQUEUE,ASF_FIRM,ARC_SUBSYS,MANAGE,WOL>
igphy0 at wm0 phy 1: i82566 10/100/1000 media interface, rev. 1
igphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
wm1 at pci2 dev 0 function 1, 64-bit DMA: 82576 1000BaseT Ethernet (rev. 0x01)
wm1: interrupting at ioapic0 pin 17
wm1: PCI-Express bus
wm1: 16384 words (16 address bits) SPI EEPROM, version 1.43, Image Unique ID e6060000
wm1: RX packet buffer size: 64KB
wm1: Ethernet address 00:30:48:f5:15:81
wm1: Copper
wm1: 0x274040<SPI,PCIE,NEWQUEUE,ASF_FIRM,ARC_SUBSYS,WOL>
igphy1 at wm1 phy 1: i82566 10/100/1000 media interface, rev. 1
igphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto

Regards,

- Havard