Hi,
I have built today 2.6.34-rc5-git6 Linux-kernel on a 32-bit Debian/sid system.
Unfortunately, my machine hangs while startup at sshd-start (from
memory + written down):
[...]
EIP: is at inet6_csk_bind_conflict + 06xe/0xb7 [ipv6]
Pid: 2150, comm: sshd Not tainted $(uname -r) #1 30AC/HP Compaq nc6400
(RH72EA#ABD)
[...]
I suspect one of the latest IPv6 commits as culprit, especially:
commit f4f914b58019f0e50d521bbbadfaee260d766f95
net: ipv6 bind to device issue
...or...
commit f2228f785a9d97307aa8ba709088cfda6c3df73f
ipv6: allow to send packet after receiving ICMPv6 Too Big message with
MTU field less than IPV6_MIN_MTU
Could not test yet by reverting.
Kind Regards,
- Sedat -
P.S.: Attached kernel dot-config.
[ Please CC - I am not subscribed to LKML ]
[QUOTE]
On 25.04.2010 16:26, Jongman Heo wrote:
> I also hit this bug today.
I also hit similar bug, maybe it is the same.
> Doing git bisect, first bad commit was
>
> commit fda48a0d7a8412cedacda46a9c0bf8ef9cd13559
> tcp: bind() fix when many ports are bound
>
> Reverting above commit fixes the problem.
How to do it? Would you please publish a reverting patch?
Regards,
Piotr Hosowicz
[/QUOTE]
Hi,
did forget to mention 2.6.34-rc5-git5 was OK.
Revering this commit did not help:
commit f4f914b58019f0e50d521bbbadfaee260d766f95
net: ipv6 bind to device issue
After looking into net-2.6 GIT repository, "tcp: bind() fix when many
ports are bound" could cause indeed the problems here, too.
Building now....
Regards,
- Sedat -
P.S.: Attached 0001-Revert-tcp-bind-fix-when-many-ports-are-bound.patch
On 25.04.2010 17:11, Sedat Dilek wrote:
> [ Please CC - I am not subscribed to LKML ]
>
> [QUOTE]
>
> On 25.04.2010 16:26, Jongman Heo wrote:
>
>> I also hit this bug today.
>
> I also hit similar bug, maybe it is the same.
>
>> Doing git bisect, first bad commit was
>>
>> commit fda48a0d7a8412cedacda46a9c0bf8ef9cd13559
>> tcp: bind() fix when many ports are bound
>>
>> Reverting above commit fixes the problem.
>
> How to do it? Would you please publish a reverting patch?
>
> Regards,
>
> Piotr Hosowicz
>
> [/QUOTE]
>
> Hi,
>
> did forget to mention 2.6.34-rc5-git5 was OK.
>
> Revering this commit did not help:
>
> commit f4f914b58019f0e50d521bbbadfaee260d766f95
> net: ipv6 bind to device issue
>
> After looking into net-2.6 GIT repository, "tcp: bind() fix when many
> ports are bound" could cause indeed the problems here, too.
> Building now....
>
> Regards,
> - Sedat -
>
> P.S.: Attached 0001-Revert-tcp-bind-fix-when-many-ports-are-bound.patch
Thanks a lot. Applied and building now.
Regards,
Piotr Hosowicz
--
Z cyklu "Uroki demokracji", czyli pytania i odpowiedzi w teledurniejach:
- W kt?rym kraju znajduje si? Mount Everest?
- Hm, to nie Szkocja, prawda?
NP: Mark Knopfler - Cleaning My Gun
NB: 2.6.34-rc5-git5
[ CCing netdev ML ]
Confirmed: The revert-patch [1] fixes the problem here.
See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
Feel free to add a... Tested-by: Sedat Dilek <[email protected]>
- Sedat -
[1] https://patchwork.kernel.org/patch/94959/
[2] https://bugzilla.kernel.org/show_bug.cgi?id=15847
2010/4/25 Piotr Hosowicz <[email protected]>:
> On 25.04.2010 17:11, Sedat Dilek wrote:
>>
>> [ Please CC - I am not subscribed to LKML ]
>>
>> [QUOTE]
>>
>> On 25.04.2010 16:26, Jongman Heo wrote:
>>
>>> I also hit this bug today.
>>
>> I also hit similar bug, maybe it is the same.
>>
>>> Doing git bisect, first bad commit was
>>>
>>> commit fda48a0d7a8412cedacda46a9c0bf8ef9cd13559
>>> tcp: bind() fix when many ports are bound
>>>
>>> Reverting above commit fixes the problem.
>>
>> How to do it? Would you please publish a reverting patch?
>>
>> Regards,
>>
>> Piotr Hosowicz
>>
>> [/QUOTE]
>>
>> Hi,
>>
>> did forget to mention 2.6.34-rc5-git5 was OK.
>>
>> Revering this commit did not help:
>>
>> commit f4f914b58019f0e50d521bbbadfaee260d766f95
>> net: ipv6 bind to device issue
>>
>> After looking into net-2.6 GIT repository, "tcp: bind() fix when many
>> ports are bound" could cause indeed the problems here, too.
>> Building now....
>>
>> Regards,
>> - Sedat -
>>
>> P.S.: Attached 0001-Revert-tcp-bind-fix-when-many-ports-are-bound.patch
>
> Thanks a lot. Applied and building now.
>
> Regards,
>
> Piotr Hosowicz
>
> --
> Z cyklu "Uroki demokracji", czyli pytania i odpowiedzi w teledurniejach:
> - W którym kraju znajduje się Mount Everest?
> - Hm, to nie Szkocja, prawda?
> NP: Mark Knopfler - Cleaning My Gun
> NB: 2.6.34-rc5-git5
>
On 25.04.2010 18:31, Sedat Dilek wrote:
> [ CCing netdev ML ]
>
> Confirmed: The revert-patch [1] fixes the problem here.
I confirm, I've built a git6 kernel and it works fine.
> See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
>
> Feel free to add a... Tested-by: Sedat Dilek<[email protected]>
I added created and tested phrase in my archive. ;-) Thank you a lot. I
hope there will be no this error in git7.
Regards,
Piotr Hosowicz
> - Sedat -
>
> [1] https://patchwork.kernel.org/patch/94959/
> [2] https://bugzilla.kernel.org/show_bug.cgi?id=15847
>
> 2010/4/25 Piotr Hosowicz<[email protected]>:
>> On 25.04.2010 17:11, Sedat Dilek wrote:
>>>
>>> [ Please CC - I am not subscribed to LKML ]
>>>
>>> [QUOTE]
>>>
>>> On 25.04.2010 16:26, Jongman Heo wrote:
>>>
>>>> I also hit this bug today.
>>>
>>> I also hit similar bug, maybe it is the same.
>>>
>>>> Doing git bisect, first bad commit was
>>>>
>>>> commit fda48a0d7a8412cedacda46a9c0bf8ef9cd13559
>>>> tcp: bind() fix when many ports are bound
>>>>
>>>> Reverting above commit fixes the problem.
>>>
>>> How to do it? Would you please publish a reverting patch?
>>>
>>> Regards,
>>>
>>> Piotr Hosowicz
>>>
>>> [/QUOTE]
>>>
>>> Hi,
>>>
>>> did forget to mention 2.6.34-rc5-git5 was OK.
>>>
>>> Revering this commit did not help:
>>>
>>> commit f4f914b58019f0e50d521bbbadfaee260d766f95
>>> net: ipv6 bind to device issue
>>>
>>> After looking into net-2.6 GIT repository, "tcp: bind() fix when many
>>> ports are bound" could cause indeed the problems here, too.
>>> Building now....
>>>
>>> Regards,
>>> - Sedat -
>>>
>>> P.S.: Attached 0001-Revert-tcp-bind-fix-when-many-ports-are-bound.patch
>>
>> Thanks a lot. Applied and building now.
>>
>> Regards,
>>
>> Piotr Hosowicz
>>
>> --
>> Z cyklu "Uroki demokracji", czyli pytania i odpowiedzi w teledurniejach:
>> - W kt?rym kraju znajduje si? Mount Everest?
>> - Hm, to nie Szkocja, prawda?
>> NP: Mark Knopfler - Cleaning My Gun
>> NB: 2.6.34-rc5-git5
>>
>
--
Grupa marze? w eliminacjach M? :
Zimbabwe, Alaska, Grenlandia, Antarktyda i Zair.
NP: Chickenfoot - Oh Yeah
NB: 2.6.34-rc5-git6
Le dimanche 25 avril 2010 à 18:36 +0200, Piotr Hosowicz a écrit :
> On 25.04.2010 18:31, Sedat Dilek wrote:
> > [ CCing netdev ML ]
> >
> > Confirmed: The revert-patch [1] fixes the problem here.
>
> I confirm, I've built a git6 kernel and it works fine.
>
> > See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
> >
> > Feel free to add a... Tested-by: Sedat Dilek<[email protected]>
>
> I added created and tested phrase in my archive. ;-) Thank you a lot. I
> hope there will be no this error in git7.
>
Did you test the proposed fix ?
[PATCH] ipv6: Fix inet6_csk_bind_conflict()
Commit fda48a0d7a84 (tcp: bind() fix when many ports are bound)
introduced a bug on IPV6 part.
We should not call ipv6_addr_any(inet6_rcv_saddr(sk2)) but
ipv6_addr_any(inet6_rcv_saddr(sk)) because sk2 can be IPV4, while sk is
IPV6.
Reported-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Eric Dumazet <[email protected]>
---
diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
index b4b7d40..3a4d92b 100644
--- a/net/ipv6/inet6_connection_sock.c
+++ b/net/ipv6/inet6_connection_sock.c
@@ -48,7 +48,7 @@ int inet6_csk_bind_conflict(const struct sock *sk,
ipv6_rcv_saddr_equal(sk, sk2))
break;
else if (sk->sk_reuse && sk2->sk_reuse &&
- !ipv6_addr_any(inet6_rcv_saddr(sk2)) &&
+ !ipv6_addr_any(inet6_rcv_saddr(sk)) &&
ipv6_rcv_saddr_equal(sk, sk2))
break;
}
Rebuild ipv6 kernel-modules by:
$ make M=net/ipv6
...and copied net/ipv6/*.ko and net/ipv6/netfilter/*.ko files manually
to the right place.
Applied your patch (seen on netdev ML) already and booted into new kernel.
Works, thanks.
Tested-by: Sedat Dilek <[email protected]>
- Sedat -
On Sun, Apr 25, 2010 at 6:39 PM, Eric Dumazet <[email protected]> wrote:
> Le dimanche 25 avril 2010 à 18:36 +0200, Piotr Hosowicz a écrit :
>> On 25.04.2010 18:31, Sedat Dilek wrote:
>> > [ CCing netdev ML ]
>> >
>> > Confirmed: The revert-patch [1] fixes the problem here.
>>
>> I confirm, I've built a git6 kernel and it works fine.
>>
>> > See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
>> >
>> > Feel free to add a... Tested-by: Sedat Dilek<[email protected]>
>>
>> I added created and tested phrase in my archive. ;-) Thank you a lot. I
>> hope there will be no this error in git7.
>>
>
> Did you test the proposed fix ?
>
>
> [PATCH] ipv6: Fix inet6_csk_bind_conflict()
>
> Commit fda48a0d7a84 (tcp: bind() fix when many ports are bound)
> introduced a bug on IPV6 part.
> We should not call ipv6_addr_any(inet6_rcv_saddr(sk2)) but
> ipv6_addr_any(inet6_rcv_saddr(sk)) because sk2 can be IPV4, while sk is
> IPV6.
>
> Reported-by: Michael S. Tsirkin <[email protected]>
> Signed-off-by: Eric Dumazet <[email protected]>
> ---
> diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
> index b4b7d40..3a4d92b 100644
> --- a/net/ipv6/inet6_connection_sock.c
> +++ b/net/ipv6/inet6_connection_sock.c
> @@ -48,7 +48,7 @@ int inet6_csk_bind_conflict(const struct sock *sk,
> ipv6_rcv_saddr_equal(sk, sk2))
> break;
> else if (sk->sk_reuse && sk2->sk_reuse &&
> - !ipv6_addr_any(inet6_rcv_saddr(sk2)) &&
> + !ipv6_addr_any(inet6_rcv_saddr(sk)) &&
> ipv6_rcv_saddr_equal(sk, sk2))
> break;
> }
>
>
>
On 25.04.2010 18:55, Sedat Dilek wrote:
> Rebuild ipv6 kernel-modules by:
>
> $ make M=net/ipv6
>
> ...and copied net/ipv6/*.ko and net/ipv6/netfilter/*.ko files manually
> to the right place.
>
> Applied your patch (seen on netdev ML) already and booted into new kernel.
> Works, thanks.
>
> Tested-by: Sedat Dilek<[email protected]>
>
> - Sedat -
>
> On Sun, Apr 25, 2010 at 6:39 PM, Eric Dumazet<[email protected]> wrote:
>> Le dimanche 25 avril 2010 à 18:36 +0200, Piotr Hosowicz a écrit :
>>> On 25.04.2010 18:31, Sedat Dilek wrote:
>>>> [ CCing netdev ML ]
>>>>
>>>> Confirmed: The revert-patch [1] fixes the problem here.
>>>
>>> I confirm, I've built a git6 kernel and it works fine.
>>>
>>>> See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
>>>>
>>>> Feel free to add a... Tested-by: Sedat Dilek<[email protected]>
>>>
>>> I added created and tested phrase in my archive. ;-) Thank you a lot. I
>>> hope there will be no this error in git7.
>>>
>>
>> Did you test the proposed fix ?
Eric, but where to get the proposed fix as a patch? I looked at kernel's
bugzilla and there is no such thing. As for now I applied Sedat's
reverse patch and now I am booted fine in git6 kernel.
Regards,
Piotr Hosowicz
>> [PATCH] ipv6: Fix inet6_csk_bind_conflict()
>>
>> Commit fda48a0d7a84 (tcp: bind() fix when many ports are bound)
>> introduced a bug on IPV6 part.
>> We should not call ipv6_addr_any(inet6_rcv_saddr(sk2)) but
>> ipv6_addr_any(inet6_rcv_saddr(sk)) because sk2 can be IPV4, while sk is
>> IPV6.
>>
>> Reported-by: Michael S. Tsirkin<[email protected]>
>> Signed-off-by: Eric Dumazet<[email protected]>
>> ---
>> diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
>> index b4b7d40..3a4d92b 100644
>> --- a/net/ipv6/inet6_connection_sock.c
>> +++ b/net/ipv6/inet6_connection_sock.c
>> @@ -48,7 +48,7 @@ int inet6_csk_bind_conflict(const struct sock *sk,
>> ipv6_rcv_saddr_equal(sk, sk2))
>> break;
>> else if (sk->sk_reuse&& sk2->sk_reuse&&
>> - !ipv6_addr_any(inet6_rcv_saddr(sk2))&&
>> + !ipv6_addr_any(inet6_rcv_saddr(sk))&&
>> ipv6_rcv_saddr_equal(sk, sk2))
>> break;
>> }
>>
>>
>>
--
"Bardzo fajna firma, tylko czasami nie płaci. Do jej głównych
zalet należą: ekspres ciśnieniowy (ale tylko na jednym z pię-
ter) oraz fajne widoki z salki konferencyjnej. No i jakieś
pół roku temu odeszło 90% pracowników działu IT, ale może nie
mieli racji ;)"
NP: Chickenfoot - Soap On A Rope
NB: 2.6.34-rc5-git6
Le dimanche 25 avril 2010 à 19:01 +0200, Piotr Hosowicz a écrit :
> On 25.04.2010 18:55, Sedat Dilek wrote:
> > Rebuild ipv6 kernel-modules by:
> >
> > $ make M=net/ipv6
> >
> > ...and copied net/ipv6/*.ko and net/ipv6/netfilter/*.ko files manually
> > to the right place.
> >
> > Applied your patch (seen on netdev ML) already and booted into new kernel.
> > Works, thanks.
> >
> > Tested-by: Sedat Dilek<[email protected]>
> >
> > - Sedat -
> >
> > On Sun, Apr 25, 2010 at 6:39 PM, Eric Dumazet<[email protected]> wrote:
> >> Le dimanche 25 avril 2010 à 18:36 +0200, Piotr Hosowicz a écrit :
> >>> On 25.04.2010 18:31, Sedat Dilek wrote:
> >>>> [ CCing netdev ML ]
> >>>>
> >>>> Confirmed: The revert-patch [1] fixes the problem here.
> >>>
> >>> I confirm, I've built a git6 kernel and it works fine.
> >>>
> >>>> See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
> >>>>
> >>>> Feel free to add a... Tested-by: Sedat Dilek<[email protected]>
> >>>
> >>> I added created and tested phrase in my archive. ;-) Thank you a lot. I
> >>> hope there will be no this error in git7.
> >>>
> >>
> >> Did you test the proposed fix ?
>
> Eric, but where to get the proposed fix as a patch? I looked at kernel's
> bugzilla and there is no such thing. As for now I applied Sedat's
> reverse patch and now I am booted fine in git6 kernel.
>
I wont spend my Sunday time at filling bugzilla entries.
I sent a patch on netdev.
I finaly could test it (I have a very slow machine while traveling, and
compiling the thing took a *lot* of time)
Now back to family affairs ;)
On Sun, Apr 25, 2010 at 7:01 PM, Piotr Hosowicz <[email protected]> wrote:
> On 25.04.2010 18:55, Sedat Dilek wrote:
>>
[...]
>> On Sun, Apr 25, 2010 at 6:39 PM, Eric Dumazet<[email protected]>
>> wrote:
>>>
>>> Le dimanche 25 avril 2010 à 18:36 +0200, Piotr Hosowicz a écrit :
>>>>
>>>> On 25.04.2010 18:31, Sedat Dilek wrote:
>>>>>
>>>>> [ CCing netdev ML ]
>>>>>
>>>>> Confirmed: The revert-patch [1] fixes the problem here.
>>>>
>>>> I confirm, I've built a git6 kernel and it works fine.
>>>>
>>>>> See also "Bug 15847 - crash in inet6_csk_bind_conflict" [2].
>>>>>
>>>>> Feel free to add a... Tested-by: Sedat Dilek<[email protected]>
>>>>
>>>> I added created and tested phrase in my archive. ;-) Thank you a lot. I
>>>> hope there will be no this error in git7.
>>>>
>>>
>>> Did you test the proposed fix ?
>
> Eric, but where to get the proposed fix as a patch? I looked at kernel's
> bugzilla and there is no such thing. As for now I applied Sedat's reverse
> patch and now I am booted fine in git6 kernel.
>
> Regards,
>
> Piotr Hosowicz
>
For easy catching proposed patched to LKML look at <patchwork.kernel.org> [1].
There you find other mailing-lists and further patches.
Cut-N-Paste excerpts of Eric's email should also work :-).
- Sedat -
[1] https://patchwork.kernel.org/project/LKML/list/
[2] https://patchwork.kernel.org/patch/94961/