Re: [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6

Le mardi 18 janvier 2011 à 22:42 +0100, Hans de Bruin a écrit :

On 01/18/2011 09:06 PM, Jesse Gross wrote:

quoted

On Tue, Jan 18, 2011 at 11:52 AM, Hans de Bruin[off-list ref]  wrote:

quoted

On 01/16/2011 09:24 PM, Hans de Bruin wrote:

quoted

After last nights compile i lost the possibility to connect to ssh and
http over ipv6. The connection stops at syn_sent. connections to my
machine end in syn_recv. ping6 still works.

The bisect ended in:

0363466866d901fbc658f4e63dd61e7cc93dd0af is the first bad commit
commit 0363466866d901fbc658f4e63dd61e7cc93dd0af
Author: Jesse Gross[off-list ref]
Date:   Sun Jan 9 06:23:35 2011 +0000

    net offloading: Convert checksums to use centrally computed features.

    In order to compute the features for other offloads (primarily
    scatter/gather), we need to first check the ability of the NIC to
    offload the checksum for the packet.  Since we have already computed
    this, we can directly use the result instead of figuring it out
    again.

    Signed-off-by: Jesse Gross[off-list ref]
    Signed-off-by: David S. Miller[off-list ref]


ssh ::1  still works. And since dns still works I guess udp is not affected.
My nic is a:

09:00.0 Ethernet controller: Broadcom Corporation NetXtreme BCM5752 Gigabit
Ethernet PCI Express (rev 02)

Are you using vlans?  If so, can you please test this patch?
http://patchwork.ozlabs.org/patch/79264/

No I am not using vlans. The option is even not set in the config file. 
Except for the disk less bit is a straightforward setup:

   bash-4.1# ip addr show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
     link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
     inet 127.0.0.1/8 scope host lo
     inet6 ::1/128 scope host
        valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP 
qlen 1000
     link/ether 00:1c:23:2d:73:87 brd ff:ff:ff:ff:ff:ff
     inet 10.10.0.6/16 brd 10.10.255.255 scope global eth0
     inet6 2001:610:76e:0:21c:23ff:fe2d:7387/64 scope global dynamic
        valid_lft 86399sec preferred_lft 14399sec
     inet6 fe80::21c:23ff:fe2d:7387/64 scope link
        valid_lft forever preferred_lft forever
3: sit0: <NOARP> mtu 1480 qdisc noop state DOWN
     link/sit 0.0.0.0 brd 0.0.0.0
bash-4.1# ip route show
10.10.0.0/16 dev eth0  proto kernel  scope link  src 10.10.0.6
default via 10.10.0.1 dev eth0
bash-4.1# ip -6 route show
2001:610:76e::/64 dev eth0  proto kernel  metric 256  expires 86404sec
fe80::/64 dev eth0  proto kernel  metric 256
ff00::/8 dev eth0  metric 256
default via fe80::230:18ff:feae:75d8 dev eth0  proto kernel  metric 1024 
  expires 29sec hoplimit 64
bash-4.1# iptables -L
Chain INPUT (policy ACCEPT)
target     prot opt source               destination

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination
bash-4.1# ip6tables -L
Chain INPUT (policy ACCEPT)
target     prot opt source               destination

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination
bash-4.1# brctl show
bridge name     bridge id               STP enabled     interfaces
bash-4.1# cat /proc/cmdline
root=/dev/nfs nfsroot=10.10.0.2:/nfs/root/psion/,v3,tcp ro ip=::: ::dhcp 
BOOT_IMAGE=nightlybuild-psion

You could try "tcpdump -i eth0 ip6 -v"

I guess you receive frames with bad checksums

You can ask other system to not offload tx checksums

ethtool -K eth0 tx off

Please give result of (on both machines)
lspci -v