Re: 2.6.23-rc2-mm2

[Andrew Morton <akpm@linux-foundation.org>]

On Fri, 10 Aug 2007 11:33:41 -0600
ebiederm@xmission.com (Eric W. Biederman) wrote:

> Gabriel C <nix.or.die@googlemail.com> writes:
> 
> > Michal Piotrowski wrote:
> >> Andrew Morton pisze:
> >>>
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.23-rc2/2.6.23-rc2-mm2/
> >>>
> >> 
> >> Yet another sysctl table check failed
> >> 
> >> [   88.949055] Netfilter messages via NETLINK v0.30.
> >> [   89.485399] nf_conntrack version 0.5.0 (16384 buckets, 65536 max)
> >> [ 89.491892] sysctl table check failed:
> > /net/netfilter/nf_conntrack_generic_timeout .3.19.13 Missing strategy
> >> [   89.558178] xt_state: Unknown symbol nf_conntrack_untracked
> >> [   89.563942] xt_state: Unknown symbol nf_ct_l3proto_module_put
> >> [   89.569870] xt_state: Unknown symbol nf_ct_l3proto_try_module_get
> >> [   90.852319] NET: Registered protocol family 17
> >>
> >
> > And maybe this ?
> 
> Could be.
> 
> sysctl table check failure looks legitimate.
> At a quick skim I can't tell for certain if failure to register the
> sysctl table will keep the module from loading but it might.
> 
> It looks like another canidate on the pile of kill the broken
> binary sysctl.
> 

There seems to be rather a lot of damage here.

I assume that the sysctl changes are what caused the netfilter oopses.

- nf_conntrack_init() calls nf_conntrack_expect_init() which fails due to
  sysctl problems.  

- nf_conntrack_init() bales out without calling nf_conntrack_helper_init()

  So nf_ct_helper_hsize never gets initialised.

- Later, netfilter client code calls helper_hash(), which gets a
  divide-by-zero due to nf_ct_helper_hsize==0.


yeah, that's a netfilter bug, but we're trying to get kernels tested here. 
If I'm feeling energetic I'll drop the sysctl changes and do rc2-mm3. 
Probably I won't feel energetic, but we'll need a lot of fixes here before
I can release the sysctl changes in another -mm, please.