(please CC me when replying)
I just got a motherboard with the dual nic marvell chipset onboard.
Splendid board save for the driver issue I'm having :) I'm currently
using 2.6.19-gentoo-r4 which doesn't have any forcedeth patch applied, a
vanilla kernel seems to give the same results. This board is advertised
with the dual nics being capable of failover and teaming and other fancy
stuff. I've disabled this gadgetry in BIOS setup however.
Every now and then (and the period seems random) the NIC or driver
simply stops passing traffic [all the way] to userland or can't pass
packets from certain IPs. I see both problems. Sometimes the nic zones
out within a minute and sometimes it takes hours.
On eth1, the first nic, sits my cable modem. It periodically loses all
connectivity and tcpdump only shows the outgoing ARP for the gateway.
Restarting dhcpcd on the interface fixes it perfectly. ip link set eth1
down && then up does the same thing. Traffic is immediately seen with
no reconfiguration.
On eth2, the second nic, I have a desktop switch. The problem I see
there is that certain IPs can't be communicated with. I.e. eth2 is
10.0.0.1/24 and my printer is 10.0.0.15/24. It's been 10.0.0.15 for
years. If I try to ping it, ping never sees an ICMP answer however
tcpdump shows both the echo and echo_reply. Changing the IP to
10.0.0.12 and it is immediately pingable. Restarting the interface has
no effect. It doesn't matter what device I put 10.0.0.15 on, it's not
reachable.
No iptable rules on eth2 and only outbound SNAT to the interface's IP on
eth1. That is about the extent of firewalling. No fancy options in
/proc and no IP mgt daemons other than standard dhcpcd.
There is nothing in dmesg to indicate anything is amiss. There is
nothing in dhcp log, syslog, etc, to indicate anything has changed. For
the moment I have a 60 second loop to restart eth1 but it's a bit of a
bother to have everything stall frequently.
Any "me toos" out there, any suggestions?
-david
============================================================
10.0.0.0/24 dev eth2 proto kernel scope link src 10.0.0.1
69.167.96.0/21 dev eth1 proto kernel scope link src 69.167.97.200
127.0.0.0/8 dev lo scope link
default via 69.167.96.1 dev eth1
7: eth1: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:18:f3:82:55:01 brd ff:ff:ff:ff:ff:ff
inet 69.167.97.200/21 brd 69.167.103.255 scope global eth1
inet6 fe80::218:f3ff:fe82:5501/64 scope link
valid_lft forever preferred_lft forever
9: eth2: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:18:f3:82:5e:45 brd ff:ff:ff:ff:ff:ff
inet 10.0.0.1/24 brd 10.0.0.255 scope global eth2
inet6 fe80::218:f3ff:fe82:5e45/64 scope link
valid_lft forever preferred_lft forever
00:10.0 Bridge: nVidia Corporation MCP55 Ethernet (rev a2)
Subsystem: ASUSTeK Computer Inc. Unknown device cb84
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
Latency: 0 (250ns min, 5000ns max)
Interrupt: pin A routed to IRQ 21
Region 0: Memory at fe02a000 (32-bit, non-prefetchable) [size=4K]
Region 1: I/O ports at b000 [size=8]
Region 2: Memory at fe029000 (32-bit, non-prefetchable) [size=256]
Region 3: Memory at fe028000 (32-bit, non-prefetchable) [size=16]
Capabilities: [44] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA
PME(D0+,D1+,D2+,D3hot+,D3cold+)
Status: D0 PME-Enable+ DSel=0 DScale=0 PME-
Capabilities: [70] MSI-X: Enable- Mask- TabSize=8
Vector table: BAR=2 offset=00000000
PBA: BAR=3 offset=00000000
Capabilities: [50] Message Signalled Interrupts: Mask+ 64bit+
Queue=0/3 Enable-
Address: 0000000000000000 Data: 0000
Masking: 00000000 Pending: 00000000
Capabilities: [6c] HyperTransport: MSI Mapping
00:11.0 Bridge: nVidia Corporation MCP55 Ethernet (rev a2)
Subsystem: ASUSTeK Computer Inc. Unknown device cb84
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
Latency: 0 (250ns min, 5000ns max)
Interrupt: pin A routed to IRQ 20
Region 0: Memory at fe027000 (32-bit, non-prefetchable) [size=4K]
Region 1: I/O ports at ac00 [size=8]
Region 2: Memory at fe026000 (32-bit, non-prefetchable) [size=256]
Region 3: Memory at fe025000 (32-bit, non-prefetchable) [size=16]
Capabilities: [44] Power Management version 2
Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA
PME(D0+,D1+,D2+,D3hot+,D3cold+)
Status: D0 PME-Enable+ DSel=0 DScale=0 PME-
Capabilities: [70] MSI-X: Enable+ Mask- TabSize=8
Vector table: BAR=2 offset=00000000
PBA: BAR=3 offset=00000000
Capabilities: [50] Message Signalled Interrupts: Mask+ 64bit+
Queue=0/3 Enable-
Address: 0000000000000000 Data: 0000
Masking: 00000000 Pending: 00000000
Capabilities: [6c] HyperTransport: MSI Mapping
(AKA)
00:10.0 0680: 10de:0373 (rev a2)
Subsystem: 1043:cb84
Flags: bus master, 66MHz, fast devsel, latency 0, IRQ 21
Memory at fe02a000 (32-bit, non-prefetchable) [size=4K]
I/O ports at b000 [size=8]
Memory at fe029000 (32-bit, non-prefetchable) [size=256]
Memory at fe028000 (32-bit, non-prefetchable) [size=16]
Capabilities: [44] Power Management version 2
Capabilities: [70] MSI-X: Enable- Mask- TabSize=8
Capabilities: [50] Message Signalled Interrupts: Mask+ 64bit+
Queue=0/3 Enable-
Capabilities: [6c] HyperTransport: MSI Mapping
00:11.0 0680: 10de:0373 (rev a2)
Subsystem: 1043:cb84
Flags: bus master, 66MHz, fast devsel, latency 0, IRQ 20
Memory at fe027000 (32-bit, non-prefetchable) [size=4K]
I/O ports at ac00 [size=8]
Memory at fe026000 (32-bit, non-prefetchable) [size=256]
Memory at fe025000 (32-bit, non-prefetchable) [size=16]
Capabilities: [44] Power Management version 2
Capabilities: [70] MSI-X: Enable+ Mask- TabSize=8
Capabilities: [50] Message Signalled Interrupts: Mask+ 64bit+
Queue=0/3 Enable-
Capabilities: [6c] HyperTransport: MSI Mapping
eth1: forcedeth.c: subsystem: 01043:cb84 bound to 0000:00:10.0
ACPI: PCI Interrupt Link [AMC1] enabled at IRQ 20
ACPI: PCI Interrupt 0000:00:11.0[A] -> Link [AMC1] -> GSI 20 (level,
low) -> IRQ 20
PCI: Setting latency timer of device 0000:00:11.0 to 64
forcedeth: using HIGHDMA
eth2: forcedeth.c: subsystem: 01043:cb84 bound to 0000:00:11.0
ACPI: PCI Interrupt Link [AAZA] enabled at IRQ 23
ACPI: PCI Interrupt 0000:00:0e.1[B] -> Link [AAZA] -> GSI 23 (level,
low) -> IRQ 23
PCI: Setting latency timer of device 0000:00:0e.1 to 64
David Ford <david <at> blue-labs.org> writes:
> Any "me toos" out there, any suggestions?
Big "me too" here. I use this board to run "workstation and home samba services"
and it seems that under stress, the connection gets lost. Can't put my finger on
the when/what/why too.
I get the feeling though that when I use the ethernet-device locally _and_ it
gets used through samba (externally) the problem is more likely to occur.
If anyone gets this, please tell me. As David says: too nice a board to have
this problem :)
Thanx,
Jeroen Roodhart
Jeroen Roodhart wrote:
> > Any "me toos" out there, any suggestions?
>
> Big "me too" here. I use this board to run "workstation and home samba services"
> and it seems that under stress, the connection gets lost. Can't put my finger on
> the when/what/why too.
I've seen some issues with forcedeth too, but only with 2.6.20-rc5
and -rc6 as far as I can say. The box stopped receiving packets, but a
"ifconfig eth0 promisc" 'fixed' it.
It happened rarely and I haven't been really able to reproduce it on
demand so far.
Looking through my syslog I think
|Jan 28 22:16:44 melchior kernel: [28411.397053] Losing some ticks... checking if CPU frequency changed.
|Jan 28 22:18:12 melchior kernel: [28488.534330] warning: many lost ticks.
|Jan 28 22:18:12 melchior kernel: [28488.534333] Your time source seems to be instable or some driver is hogging int
|Jan 28 22:18:12 melchior kernel: [28488.534350] rip nv_nic_irq_other+0x2/0x160 [forcedeth]
|Jan 28 22:18:12 melchior kernel: [28488.534352] Falling back to HPET
sounds suspicios.
|Jan 25 18:17:20 melchior kernel: [ 2875.164597] Losing some ticks... checking if CPU frequency changed.
|Jan 25 18:18:57 melchior kernel: [ 2929.230774] warning: many lost ticks.
|Jan 25 18:18:57 melchior kernel: [ 2929.230777] Your time source seems to be instable or some driver is hogging interupts
|Jan 25 18:18:57 melchior kernel: [ 2929.230785] rip handle_IRQ_event+0x1a/0x53
|Jan 25 18:18:57 melchior kernel: [ 2929.230787] Falling back to HPET
This is on an ASUS M2N-SLI Deluxe.
Right now, it again only works with eth0 promisc, but interrupts I
can at least see interrupts for all three registered eth0 msi
interrupt handlers:
|279: 13239534 PCI-MSI-edge eth0
|280: 2515250 PCI-MSI-edge eth0
|281: 6268650 PCI-MSI-edge eth0
and "sleep 1" later:
|279: 13239635 PCI-MSI-edge eth0
|280: 2515259 PCI-MSI-edge eth0
|281: 6268661 PCI-MSI-edge eth0
--
Tobias PGP: http://9ac7e0bc.uguu.de
このメールは十割再利用されたビットで作られています。