Skip to contentSkip to conversation

2.6.19 and up to 2.6.20-rc2 Ethernet problems x86_64

DISCUSSION Sid Boyce 3 participants 9 messages Dec 28, 2006 – Jan 4, 2007 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore

Sid Boyce

author
#0 lore raw

I first saw the problem on the 64x2 box after upgrading to 2.6.19. The network appeared OK with ifconfig and route -n, but I had no network access. Pinging any other box, the box was responding, but no response was received by the 64x2. I checked cables and the switch which were OK, so I figured I had a faulty receiver on the on-board ethernet. I plugged in a PCI eepro100, reconfigured the network and it's been fine ever since.

I upgraded to 2.6.20-rc2 (also 2.6.19 and later) on the x86_64 laptop and that also experienced the same problem - different cards in both boxes. Downgraded to 2.6.19-rc5 and network is up.

================== 64x2 box ======================
64: None 00.0: 10701 Ethernet
[Created at net.115]
UDI: /org/freedesktop/Hal/devices/net_00_01_6c_e6_ab_7a Unique ID: usDW.ndpeucax6V1
Parent ID: rBUF.siZp+507g89
SysFS ID: /class/net/eth0
SysFS Device Link: /devices/pci0000:00/0000:00:14.0 Hardware Class: network interface
Model: "Ethernet network interface"
Driver: "forcedeth"
Device File: eth0
HW Address: 00:01:6c:e6:ab:7a
Link detected: yes
Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #39 (Ethernet controller)
*********** PROBLEM **************
--------------------------------------------------------------------
eepro100 working (64x2)
=======================
65: None 01.0: 10701 Ethernet
[Created at net.115]
UDI: /org/freedesktop/Hal/devices/net_00_d0_b7_13_e5_ea Unique ID: L2Ua.ndpeucax6V1
Parent ID: JNkJ.HVgIlgOrmpC
SysFS ID: /class/net/eth1
SysFS Device Link: /devices/pci0000:00/0000:00:10.0/0000:04:08.0 Hardware Class: network interface
Model: "Ethernet network interface"
Driver: "eepro100"
Device File: eth1
HW Address: 00:d0:b7:13:e5:ea
Link detected: yes
Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #46 (Ethernet controller)
=======================================================================

Acer 1501LCe x86_64 laptop.
============================
49: udi = '/org/freedesktop/Hal/devices/pci_14e4_169c'
info.bus = 'pci'
info.linux.driver = 'tg3'
info.parent = '/org/freedesktop/Hal/devices/computer' info.product = 'NetXtreme BCM5788 Gigabit Ethernet' info.udi = '/org/freedesktop/Hal/devices/pci_14e4_169c' info.vendor = 'Broadcom Corporation' linux.hotplug_type = 1 (0x1)
linux.subsystem = 'pci'
linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:0c.0' linux.sysfs_path_device = '/sys/devices/pci0000:00/0000:00:0c.0' pci.device_class = 2 (0x2)
pci.device_protocol = 0 (0x0)
pci.device_subclass = 0 (0x0)
pci.linux.sysfs_path = '/sys/devices/pci0000:00/0000:00:0c.0' pci.product = 'NetXtreme BCM5788 Gigabit Ethernet' pci.product_id = 5788 (0x169c)
pci.subsys_product = 'Unknown (0x0046)' pci.subsys_product_id = 70 (0x46)
pci.subsys_vendor = 'Acer Incorporated [ALI]' pci.subsys_vendor_id = 4133 (0x1025)
pci.vendor = 'Broadcom Corporation'
pci.vendor_id = 5348 (0x14e4)

============== tcpdump on barrabas when pinged from Boycie ===== 01:45:32.089245 arp who-has Boycie.site tell barrabas.site 01:45:32.089323 arp reply Boycie.site is-at 00:0a:e4:4e:a1:42 (oui Unknown) 01:45:32.172445 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 41, length 64
01:45:32.172480 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 41, length 64
01:45:33.189534 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 42, length 64
01:45:33.189569 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 42, length 64
01:45:34.205635 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 43, length 64
01:45:34.205667 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 43, length 64
01:45:35.221734 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 44, length 64
01:45:35.221767 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 44, length 64
01:45:36.237835 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 45, length 64
01:45:36.237868 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 45, length 64
01:45:37.254956 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 46, length 64
01:45:37.254991 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 46, length 64
01:45:38.272044 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 47, length 64
01:45:38.272077 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 47, length 64
01:45:39.288136 IP Boycie.site > barrabas.site: ICMP echo request, id 8463, seq 48, length 64
01:45:39.288170 IP barrabas.site > Boycie.site: ICMP echo reply, id 8463, seq 48, length 64
170 packets captured
340 packets received by filter
0 packets dropped by kernel

Boycie sees just a few packets occasionally when pinged from barrabas, but says its dropped no packets.

Also strange, I can scp a file from Boycie. barrabas:/usr/src/linux-2.6.20-rc1-git5 # scp Boycie:/IFCONFIG / Password:
IFCONFIG 100% 907
0.9KB/s 00:00
barrabas:/usr/src/linux-2.6.20-rc1-git5 # o /IFCONFIG
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42
inet addr:192.168.10.5 Bcast:255.255.255.255 Mask:255.255.255.0 inet6 addr: fe80::20a:e4ff:fe4e:a142/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:577 errors:0 dropped:0 overruns:0 frame:0 TX packets:569 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000
RX bytes:65202 (63.6 Kb) TX bytes:49404 (48.2 Kb) Interrupt:20

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:26 errors:0 dropped:0 overruns:0 frame:0 TX packets:26 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0
RX bytes:1866 (1.8 Kb) TX bytes:1866 (1.8 Kb)

barrabas:/usr/src/linux-2.6.20-rc1-git5 # ssh Boycie ifconfig Password:
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42
inet addr:192.168.10.5 Bcast:255.255.255.255 Mask:255.255.255.0 inet6 addr: fe80::20a:e4ff:fe4e:a142/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:728 errors:0 dropped:0 overruns:0 frame:0 TX packets:710 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000
RX bytes:83960 (81.9 Kb) TX bytes:65125 (63.5 Kb) Interrupt:20

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:26 errors:0 dropped:0 overruns:0 frame:0 TX packets:26 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0
RX bytes:1866 (1.8 Kb) TX bytes:1866 (1.8 Kb)
Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#1 lore raw
Quoted 5 lines · Sid Boyce

I first saw the problem on the 64x2 box after upgrading to 2.6.19. The network appeared OK with ifconfig and route -n, but I had no network access. Pinging any other box, the box was responding, but no response

...

Quoted 4 lines

barrabas:/usr/src/linux-2.6.20-rc1-git5 # ssh Boycie ifconfig Password:
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42

          inet addr:192.168.10.5  Bcast:255.255.255.255  Mask:255.255.255.0

This Bcast isn't probably what you need.

Regards,
Jarek P.

#3 lore raw
In reply to a message not in the archive
Quoted 13 lines · Sid Boyce
Quoted 12 lines · Jarek Poplawski
Quoted 4 lines · Sid Boyce

I first saw the problem on the 64x2 box after upgrading to 2.6.19. The network appeared OK with ifconfig and route -n, but I had no network access. Pinging any other box, the box was responding, but no response

...

Quoted 5 lines

barrabas:/usr/src/linux-2.6.20-rc1-git5 # ssh Boycie ifconfig Password:
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42

         inet addr:192.168.10.5  Bcast:255.255.255.255  
         Mask:255.255.255.0
Quoted 8 lines
Quoted 6 lines

This Bcast isn't probably what you need.

Regards,
Jarek P.

Corrected on the one box where it was not correct, problem is still there.

There are many things to suspect yet:
- firewall,
- switch,
- routing,
- ifconfig,
- other misonfigured box,
- connecting
and so on.

I think you should try with some linux networking group at first and if you really think it's driver then [email protected] (instead of linux-kernel@).

If you could send full ifconfig, route -n (or ip route if you use additional tables) and tcpdump (all packets) from both boxes while pinging each other and a few words how it is connected (other cards, other active boxes in the network?) maybe something more could be found.

Cheers,
Jarek P.

PS: Sorry for late responding.

author
#4 lore raw
Quoted 39 lines · Jarek Poplawski
Quoted 11 lines · Sid Boyce
Quoted 10 lines · Jarek Poplawski
Quoted 3 lines · Sid Boyce

I first saw the problem on the 64x2 box after upgrading to 2.6.19. The network appeared OK with ifconfig and route -n, but I had no network access. Pinging any other box, the box was responding, but no response

...

Quoted 5 lines

barrabas:/usr/src/linux-2.6.20-rc1-git5 # ssh Boycie ifconfig Password:
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42

         inet addr:192.168.10.5  Bcast:255.255.255.255  
         Mask:255.255.255.0
Quoted 7 lines
Quoted 6 lines

This Bcast isn't probably what you need.

Regards,
Jarek P.

Corrected on the one box where it was not correct, problem is still there.

There are many things to suspect yet:
- firewall,
- switch,
- routing,
- ifconfig,
- other misonfigured box,
- connecting
and so on.

I think you should try with some linux networking group at first and if you really think it's driver then [email protected] (instead of linux-kernel@).

If you could send full ifconfig, route -n (or ip route if you use additional tables) and tcpdump (all packets) from both boxes while pinging each other and a few words how it is connected (other cards, other active boxes in the network?) maybe something more could be found.

Quoted 6 lines

Cheers,
Jarek P.

PS: Sorry for late responding.

I have a problem with posting to linux-kernel and netdev, my mail got
returned as SPAM, now it just gets dropped. postmaster says the filter
is seeing a sub-string of something that is filtered. I shall try under my other email address.

Everything is fine with a eepro100 on the 64x2 box that gave the same problem with a nVidia Corporation MCP51 Ethernet Controller (rev a1) using the forcedeth module. On the x86_64 laptop the problem is with a Broadcom NetXtreme BCM5788 using the tg3 module. Switching back to a 2.6.18.2 kernel, there is no problem.
With all configurations of cards on both, route -n is the same on all kernels and instantly reports back. With >=2.6.19 on the laptop, netstat -r takes a very long time before returning the information ~30 seconds, instantly on 2.6.18.2.
Boycie:~ # netstat -r
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt
Iface
192.168.10.0 * 255.255.255.0 U 0 0 0
eth0

loopback        *               255.0.0.0       U         0 0          0 lo
default         Smoothie.site   0.0.0.0         UG        0 0          0

eth0
Boycie:~ # route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use
Iface

192.168.10.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0
127.0.0.0       0.0.0.0         255.0.0.0       U     0      0        0 lo
0.0.0.0         192.168.10.102  0.0.0.0         UG    0      0        0 eth0
Signature · 4 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
author
#8 lore raw
In reply to Jarek Poplawski
Quoted 39 lines · Jarek Poplawski
Quoted 11 lines · Sid Boyce
Quoted 10 lines · Jarek Poplawski
Quoted 3 lines · Sid Boyce

I first saw the problem on the 64x2 box after upgrading to 2.6.19. The network appeared OK with ifconfig and route -n, but I had no network access. Pinging any other box, the box was responding, but no response

...

Quoted 5 lines

barrabas:/usr/src/linux-2.6.20-rc1-git5 # ssh Boycie ifconfig Password:
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42

         inet addr:192.168.10.5  Bcast:255.255.255.255  
         Mask:255.255.255.0
Quoted 7 lines
Quoted 6 lines

This Bcast isn't probably what you need.

Regards,
Jarek P.

Corrected on the one box where it was not correct, problem is still there.

There are many things to suspect yet:
- firewall,
- switch,
- routing,
- ifconfig,
- other misonfigured box,
- connecting
and so on.

I think you should try with some linux networking group at first and if you really think it's driver then [email protected] (instead of linux-kernel@).

If you could send full ifconfig, route -n (or ip route if you use additional tables) and tcpdump (all packets) from both boxes while pinging each other and a few words how it is connected (other cards, other active boxes in the network?) maybe something more could be found.

Quoted 6 lines

Cheers,
Jarek P.

PS: Sorry for late responding.

I have a problem with posting to linux-kernel and netdev, my mail got returned as SPAM, now it just gets dropped. postmaster says the filter is seeing a sub-string of something that is filtered.

Everything is fine with a eepro100 on the 64x2 box that gave the same problem with a nVidia Corporation MCP51 Ethernet Controller (rev a1) using the forcedeth module. On the x86_64 laptop the problem is with a Broadcom NetXtreme BCM5788 using the tg3 module. Switching back to a 2.6.18.2 kernel, there is no problem.
With all configurations of cards on both, route -n is the same on all kernels and instantly reports back. With >=2.6.19 on the laptop, netstat -r takes a very long time before returning the information ~30 seconds, instantly on 2.6.18.2.
Boycie:~ # netstat -r
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt
Iface
192.168.10.0 * 255.255.255.0 U 0 0 0
eth0

loopback        *               255.0.0.0       U         0 0          0 lo
default         Smoothie.site   0.0.0.0         UG        0 0          0

eth0
Boycie:~ # route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use
Iface

192.168.10.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0
127.0.0.0       0.0.0.0         255.0.0.0       U     0      0        0 lo
0.0.0.0         192.168.10.102  0.0.0.0         UG    0      0        0 eth0

Regards
Sid.

Signature · 4 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support
Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
author
#9 lore raw
In reply to Jarek Poplawski
Quoted 39 lines · Jarek Poplawski
Quoted 11 lines · Sid Boyce
Quoted 10 lines · Jarek Poplawski
Quoted 3 lines · Sid Boyce

I first saw the problem on the 64x2 box after upgrading to 2.6.19. The network appeared OK with ifconfig and route -n, but I had no network access. Pinging any other box, the box was responding, but no response

...

Quoted 5 lines

barrabas:/usr/src/linux-2.6.20-rc1-git5 # ssh Boycie ifconfig Password:
eth0 Link encap:Ethernet HWaddr 00:0A:E4:4E:A1:42

         inet addr:192.168.10.5  Bcast:255.255.255.255  
         Mask:255.255.255.0
Quoted 7 lines
Quoted 6 lines

This Bcast isn't probably what you need.

Regards,
Jarek P.

Corrected on the one box where it was not correct, problem is still there.

There are many things to suspect yet:
- firewall,
- switch,
- routing,
- ifconfig,
- other misonfigured box,
- connecting
and so on.

I think you should try with some linux networking group at first and if you really think it's driver then [email protected] (instead of linux-kernel@).

If you could send full ifconfig, route -n (or ip route if you use additional tables) and tcpdump (all packets) from both boxes while pinging each other and a few words how it is connected (other cards, other active boxes in the network?) maybe something more could be found.

Quoted 6 lines

Cheers,
Jarek P.

PS: Sorry for late responding.

I have a problem with posting to linux-kernel and netdev, my mail got returned as SPAM, now it just gets dropped. postmaster says the filter is seeing a sub-string of something that is filtered. I shall try posting under another email address.

Everything is fine with a eepro100 on the 64x2 box that gave the same problem with a nVidia Corporation MCP51 Ethernet Controller (rev a1) using the forcedeth module. On the x86_64 laptop the problem is with a Broadcom NetXtreme BCM5788 using the tg3 module. Switching back to a 2.6.18.2 kernel, there is no problem. With all configurations of cards on both, route -n is the same on all kernels and instantly reports back. With >=2.6.19 on the laptop, netstat -r takes a very long time before returning the information ~30 seconds, instantly on 2.6.18.2.
Boycie:~ # netstat -r
Kernel IP routing table

Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
192.168.10.0    *               255.255.255.0   U         0 0          0 eth0
loopback        *               255.0.0.0       U         0 0          0 lo
default         Smoothie.site   0.0.0.0         UG        0 0          0 eth0
Boycie:~ # route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
192.168.10.0    0.0.0.0         255.255.255.0   U     0      0        0 eth0
127.0.0.0       0.0.0.0         255.0.0.0       U     0      0        0 lo
0.0.0.0         192.168.10.102  0.0.0.0         UG    0      0        0 eth0
Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#5 lore raw
In reply to Sid Boyce
Quoted 1 line · Sid Boyce

Jarek Poplawski wrote:

...

Quoted 5 lines
Quoted 5 lines

If you could send full ifconfig, route -n (or ip route if you use additional tables) and tcpdump (all packets) from both boxes while pinging each other and a few words how it is connected (other cards, other active boxes in the network?) maybe something more could be found.

...

Quoted 9 lines

Everything is fine with a eepro100 on the 64x2 box that gave the same problem with a nVidia Corporation MCP51 Ethernet Controller (rev a1) using the forcedeth module. On the x86_64 laptop the problem is with a Broadcom NetXtreme BCM5788 using the tg3 module. Switching back to a 2.6.18.2 kernel, there is no problem.
With all configurations of cards on both, route -n is the same on all kernels and instantly reports back. With >=2.6.19 on the laptop, netstat -r takes a very long time before returning the information ~30 seconds, instantly on 2.6.18.2.

This could be a problem with DNS. Could you do all tests (including pinging) with -n option?

I've read your other message on netdev and see you have firewall working and addresses from various networks in logs. I think it would be much easier to exclude possible network config errors and try to isolate pinging problems by connecting (with switch or even crossed cable if possible) only 2 boxes with firewalls and other net devices disabled and try to repeat this pinging with tcpdumps.

Jarek P.

#6 lore raw
Quoted 30 lines · Jarek Poplawski
Quoted 1 line · Sid Boyce

Jarek Poplawski wrote:

...

Quoted 5 lines
Quoted 5 lines

If you could send full ifconfig, route -n (or ip route if you use additional tables) and tcpdump (all packets) from both boxes while pinging each other and a few words how it is connected (other cards, other active boxes in the network?) maybe something more could be found.

...

Quoted 9 lines

Everything is fine with a eepro100 on the 64x2 box that gave the same problem with a nVidia Corporation MCP51 Ethernet Controller (rev a1) using the forcedeth module. On the x86_64 laptop the problem is with a Broadcom NetXtreme BCM5788 using the tg3 module. Switching back to a 2.6.18.2 kernel, there is no problem.
With all configurations of cards on both, route -n is the same on all kernels and instantly reports back. With >=2.6.19 on the laptop, netstat -r takes a very long time before returning the information ~30 seconds, instantly on 2.6.18.2.

This could be a problem with DNS. Could you do all tests (including pinging) with -n option?

I've read your other message on netdev and see you
have firewall working and addresses from various networks in logs. I think it would be much easier to exclude possible network config errors and try
to isolate pinging problems by connecting (with
switch or even crossed cable if possible) only 2
boxes with firewalls and other net devices disabled
and try to repeat this pinging with tcpdumps.

Quoted 3 lines

Jarek P.

It seems >=2.6.19 and the SuSEfirewall are incompatible.

Turning off the firewall on the laptop cured the problem using 2.6.20-rc2 and 2.6.20-rc3. I thought it was off. On the 64x2 box, the firewall is turned off, so I shall have to reconfigure the network to use the Gigabit ethernet instead of the eepro100 in case I turned it off after seeing the bug on that box, but didn't stop the firewall. There is still the problem where the ethernet doesn't get configured with acpi=off which I shall post to the acpi devel list, but this is not a showstopper for me. ifconfig eth0 192.168.10.5 returns SIOCSIFFLAGS: Function not implemented.
Thanks and Regards
Sid.

Signature · 3 lines
Sid Boyce ... Hamradio License G3VBV, Licensed Private Pilot
Emeritus IBM/Amdahl Mainframes and Sun/Fujitsu Servers Tech Support Specialist, Cricket Coach
Microsoft Windows Free Zone - Linux used for all Computing Tasks
#7 lore raw

On Wed, Jan 03, 2007 at 04:53:58PM +0000, Sid Boyce wrote:
...

Quoted 1 line

It seems >=2.6.19 and the SuSEfirewall are incompatible.

Actually, many programs could be incomapatible with newer kernel versions, so sometimes upgrades or at least recompilations are needed.

...

Quoted 3 lines

There is still the problem where the ethernet doesn't get configured with acpi=off which I shall post to the acpi devel list, but this is not a showstopper for me. ifconfig eth0 192.168.10.5 returns SIOCSIFFLAGS:

I don't know acpi enough but as far as I know some mainboards can't work properly with acpi off, so it's probably not a bug.

Cheers,
Jarek P.