[v3.2.24] ct: endless loop when initializing netfilter/ct cache

Thomas Graf tgraf at suug.ch
Fri Nov 7 06:37:18 PST 2014


On 11/07/14 at 01:57pm, Holger Eitzenberger wrote:
> Hi Thomas,
> 
> > That makes sense. Insertion through pickup became exponetionally
> > more expensive.
> > 
> > The right thing to do here is to introduce a flag to disable the
> > dup check for individual caches and even cache types such as CT
> > which do not require the dup check at all.
> 
> Yes, the insert is O(n^2) since introducing dup check.  With
> thousands of conntracks it is currently unusable.
> 
> If you want I can cook an RFC path for that.

That sounds great. Thanks!



More information about the libnl mailing list