mirror of
https://github.com/torvalds/linux.git
synced 2026-06-15 10:32:26 +02:00
53b56da83d
After converting to use rcu for conntrack hash, one CPU may update
the ct->status via ctnetlink, while another CPU may process the
packets and update the ct->status.
So the non-atomic operation "ct->status |= status;" via ctnetlink
becomes unsafe, and this may clear the IPS_DYING_BIT bit set by
another CPU unexpectedly. For example:
CPU0 CPU1
ctnetlink_change_status __nf_conntrack_find_get
old = ct->status nf_ct_gc_expired
- nf_ct_kill
- test_and_set_bit(IPS_DYING_BIT
new = old | status; -
ct->status = new; <-- oops, _DYING_ is cleared!
Now using a series of atomic bit operation to solve the above issue.
Also note, user shouldn't set IPS_TEMPLATE, IPS_SEQ_ADJUST directly,
so make these two bits be unchangable too.
If we set the IPS_TEMPLATE_BIT, ct will be freed by nf_ct_tmpl_free,
but actually it is alloced by nf_conntrack_alloc.
If we set the IPS_SEQ_ADJUST_BIT, this may cause the NULL pointer
deference, as the nfct_seqadj(ct) maybe NULL.
Last, add some comments to describe the logic change due to the
commit
|
||
|---|---|---|
| .. | ||
| asm-generic | ||
| drm | ||
| linux | ||
| misc | ||
| mtd | ||
| rdma | ||
| scsi | ||
| sound | ||
| video | ||
| xen | ||
| Kbuild | ||