Skip to contentSkip to conversation

iptables problem upgrading kernel from 3.18.8 to 3.19.1

DISCUSSION David R 4 participants 4 messages Mar 8, 2015 – Mar 16, 2015

Subsystems (3)
lore
Share & tools
lore

David R

author
#0 lore raw

I've just had an exception to my "uneventful kernel upgrade" monotony.

My boot scripts failed when setting up the firewall due to this :-

xt_recent: hitcount (1) is larger than packets to be remembered (1)
for table xxxx

This is a completely straightforward

    iptables -A yyyy -j REJECT -p tcp --reject-with tcp-reset -m recent

--set --name xxxx --rsource

Looking at the history for xt_recent.c it looks like this was introduced in abc86d0f99242b7f142b7cb8f90e30081dd3c256 but maybe corrected in cef9ed86ed62eeffcd017882278bbece32001f86 ?

Whatever, 3.19.1 is still affected, it can be worked around by setting ip_pkt_list_tot in the module parameters.

Cheers
David

#1 lore raw

David R <[email protected]> wrote:

[ CC Pablo & stable@ ]

Quoted 7 lines

I've just had an exception to my "uneventful kernel upgrade" monotony.

My boot scripts failed when setting up the firewall due to this :-

xt_recent: hitcount (1) is larger than packets to be remembered (1)
for table xxxx

Quoted 8 lines

This is a completely straightforward

    iptables -A yyyy -j REJECT -p tcp --reject-with tcp-reset -m recent

--set --name xxxx --rsource

Looking at the history for xt_recent.c it looks like this was introduced in abc86d0f99242b7f142b7cb8f90e30081dd3c256 but maybe corrected in cef9ed86ed62eeffcd017882278bbece32001f86 ?

Right. I would recommend to revert abc86d0f99242b7f142b7cb8f90e30081dd3c256 in 3.19.y series rather than applying cef9ed86ed62, though.

Sorry,
Florian

#2 lore raw

From: Florian Westphal <[email protected]>
Date: Sun, 8 Mar 2015 18:55:53 +0100

Quoted 16 lines

David R <[email protected]> wrote:

[ CC Pablo & stable@ ]

Quoted 12 lines

I've just had an exception to my "uneventful kernel upgrade" monotony.

My boot scripts failed when setting up the firewall due to this :-

xt_recent: hitcount (1) is larger than packets to be remembered (1)
for table xxxx

This is a completely straightforward

    iptables -A yyyy -j REJECT -p tcp --reject-with tcp-reset -m recent

--set --name xxxx --rsource

Quoted 6 lines
Quoted 3 lines

Looking at the history for xt_recent.c it looks like this was introduced in abc86d0f99242b7f142b7cb8f90e30081dd3c256 but maybe corrected in cef9ed86ed62eeffcd017882278bbece32001f86 ?

Right. I would recommend to revert abc86d0f99242b7f142b7cb8f90e30081dd3c256 in 3.19.y series rather than applying cef9ed86ed62, though.

Greg, please queue up a revert of abc86d0f99242b7f142b7cb8f90e30081dd3c256 for 3.19.x -stable, if you haven't done so already.

Thanks.

maintainer
#3 lore raw
Quoted 23 lines · David Miller

From: Florian Westphal <[email protected]>
Date: Sun, 8 Mar 2015 18:55:53 +0100

Quoted 20 lines

David R <[email protected]> wrote:

[ CC Pablo & stable@ ]

Quoted 15 lines

I've just had an exception to my "uneventful kernel upgrade" monotony.

My boot scripts failed when setting up the firewall due to this :-

xt_recent: hitcount (1) is larger than packets to be remembered (1)
for table xxxx

This is a completely straightforward

    iptables -A yyyy -j REJECT -p tcp --reject-with tcp-reset -m recent

--set --name xxxx --rsource

Looking at the history for xt_recent.c it looks like this was introduced in abc86d0f99242b7f142b7cb8f90e30081dd3c256 but maybe corrected in cef9ed86ed62eeffcd017882278bbece32001f86 ?

Quoted 5 lines
Quoted 2 lines

Right. I would recommend to revert abc86d0f99242b7f142b7cb8f90e30081dd3c256 in 3.19.y series rather than applying cef9ed86ed62, though.

Greg, please queue up a revert of abc86d0f99242b7f142b7cb8f90e30081dd3c256 for 3.19.x -stable, if you haven't done so already.

Now done, thanks.

greg k-h