LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* Possible ipv6 regression
@ 2011-01-16 20:24 Hans de Bruin
  2011-01-18 19:52 ` [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6 Hans de Bruin
  0 siblings, 1 reply; 6+ messages in thread
From: Hans de Bruin @ 2011-01-16 20:24 UTC (permalink / raw)
  To: LKML

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.

-- 
Hans

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6
  2011-01-16 20:24 Possible ipv6 regression Hans de Bruin
@ 2011-01-18 19:52 ` Hans de Bruin
  2011-01-18 20:06   ` Jesse Gross
  0 siblings, 1 reply; 6+ messages in thread
From: Hans de Bruin @ 2011-01-18 19:52 UTC (permalink / raw)
  To: LKML, jesse

On 01/16/2011 09:24 PM, Hans de Bruin wrote:
> 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 <jesse@nicira.com>
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 <jesse@nicira.com>
     Signed-off-by: David S. Miller <davem@davemloft.net>


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)

-- 
Hans


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6
  2011-01-18 19:52 ` [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6 Hans de Bruin
@ 2011-01-18 20:06   ` Jesse Gross
  2011-01-18 21:42     ` Hans de Bruin
  0 siblings, 1 reply; 6+ messages in thread
From: Jesse Gross @ 2011-01-18 20:06 UTC (permalink / raw)
  To: Hans de Bruin; +Cc: LKML, netdev

On Tue, Jan 18, 2011 at 11:52 AM, Hans de Bruin <jmdebruin@xmsnet.nl> wrote:
> On 01/16/2011 09:24 PM, Hans de Bruin wrote:
>>
>> 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 <jesse@nicira.com>
> 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 <jesse@nicira.com>
>    Signed-off-by: David S. Miller <davem@davemloft.net>
>
>
> 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/

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6
  2011-01-18 20:06   ` Jesse Gross
@ 2011-01-18 21:42     ` Hans de Bruin
  2011-01-18 22:03       ` Eric Dumazet
  0 siblings, 1 reply; 6+ messages in thread
From: Hans de Bruin @ 2011-01-18 21:42 UTC (permalink / raw)
  To: Jesse Gross; +Cc: LKML, netdev

[-- Attachment #1: Type: text/plain, Size: 3491 bytes --]

On 01/18/2011 09:06 PM, Jesse Gross wrote:
> On Tue, Jan 18, 2011 at 11:52 AM, Hans de Bruin<jmdebruin@xmsnet.nl>  wrote:
>> On 01/16/2011 09:24 PM, Hans de Bruin wrote:
>>>
>>> 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<jesse@nicira.com>
>> 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<jesse@nicira.com>
>>     Signed-off-by: David S. Miller<davem@davemloft.net>
>>
>>
>> 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

-- 
Hans


[-- Attachment #2: config.gz --]
[-- Type: application/gzip, Size: 14483 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6
  2011-01-18 21:42     ` Hans de Bruin
@ 2011-01-18 22:03       ` Eric Dumazet
  2011-01-19 13:17         ` Hans de Bruin
  0 siblings, 1 reply; 6+ messages in thread
From: Eric Dumazet @ 2011-01-18 22:03 UTC (permalink / raw)
  To: Hans de Bruin; +Cc: Jesse Gross, LKML, netdev

Le mardi 18 janvier 2011 à 22:42 +0100, Hans de Bruin a écrit :
> On 01/18/2011 09:06 PM, Jesse Gross wrote:
> > On Tue, Jan 18, 2011 at 11:52 AM, Hans de Bruin<jmdebruin@xmsnet.nl>  wrote:
> >> On 01/16/2011 09:24 PM, Hans de Bruin wrote:
> >>>
> >>> 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<jesse@nicira.com>
> >> 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<jesse@nicira.com>
> >>     Signed-off-by: David S. Miller<davem@davemloft.net>
> >>
> >>
> >> 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



^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6
  2011-01-18 22:03       ` Eric Dumazet
@ 2011-01-19 13:17         ` Hans de Bruin
  0 siblings, 0 replies; 6+ messages in thread
From: Hans de Bruin @ 2011-01-19 13:17 UTC (permalink / raw)
  To: LKML

On 01/18/2011 11:03 PM, Eric Dumazet wrote:
> Le mardi 18 janvier 2011 à 22:42 +0100, Hans de Bruin a écrit :
>> On 01/18/2011 09:06 PM, Jesse Gross wrote:
>>> On Tue, Jan 18, 2011 at 11:52 AM, Hans de Bruin<jmdebruin@xmsnet.nl>   wrote:
>>>> On 01/16/2011 09:24 PM, Hans de Bruin wrote:
>>>>>
>>>>> 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.
>>>>>
...

Just received a fix from one of the netdev's

-- 
Hans

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2011-01-19 13:16 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-01-16 20:24 Possible ipv6 regression Hans de Bruin
2011-01-18 19:52 ` [regression] 2.6.37+ commit 0363466866d9.... breaks tcp ipv6 Hans de Bruin
2011-01-18 20:06   ` Jesse Gross
2011-01-18 21:42     ` Hans de Bruin
2011-01-18 22:03       ` Eric Dumazet
2011-01-19 13:17         ` Hans de Bruin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).