| Message ID | 1442608462-1825-1-git-send-email-pshelar@nicira.com |
|---|---|
| State | Accepted |
| Headers | show |
On Fri, Sep 18, 2015 at 1:34 PM, Pravin B Shelar <pshelar@nicira.com> wrote: > From: Chris J Arges <chris.j.arges@canonical.com> > > Upstream commit: > openvswitch: allocate nr_node_ids flow_stats instead of num_possible_nodes > > Some architectures like POWER can have a NUMA node_possible_map that > contains sparse entries. This causes memory corruption with openvswitch > since it allocates flow_cache with a multiple of num_possible_nodes() and > assumes the node variable returned by for_each_node will index into > flow->stats[node]. > > Use nr_node_ids to allocate a maximal sparse array instead of > num_possible_nodes(). > > The crash was noticed after 3af229f2 was applied as it changed the > node_possible_map to match node_online_map on boot. > Fixes: 3af229f2071f5b5cb31664be6109561fbe19c861 > > Signed-off-by: Chris J Arges <chris.j.arges@canonical.com> > Acked-by: Pravin B Shelar <pshelar@nicira.com> > Acked-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> > Signed-off-by: David S. Miller <davem@davemloft.net> > > Upstream: bac541e4631(""openvswitch: allocate nr_node_ids flow_stats > instead of num_possible_nodes") > > Signed-off-by: Pravin B Shelar <pshelar@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com> This one could probably go back to 2.3 and 2.4 as well independent of the other upstream changes.
On Fri, Sep 18, 2015 at 1:38 PM, Jesse Gross <jesse@nicira.com> wrote: > On Fri, Sep 18, 2015 at 1:34 PM, Pravin B Shelar <pshelar@nicira.com> wrote: >> From: Chris J Arges <chris.j.arges@canonical.com> >> >> Upstream commit: >> openvswitch: allocate nr_node_ids flow_stats instead of num_possible_nodes >> >> Some architectures like POWER can have a NUMA node_possible_map that >> contains sparse entries. This causes memory corruption with openvswitch >> since it allocates flow_cache with a multiple of num_possible_nodes() and >> assumes the node variable returned by for_each_node will index into >> flow->stats[node]. >> >> Use nr_node_ids to allocate a maximal sparse array instead of >> num_possible_nodes(). >> >> The crash was noticed after 3af229f2 was applied as it changed the >> node_possible_map to match node_online_map on boot. >> Fixes: 3af229f2071f5b5cb31664be6109561fbe19c861 >> >> Signed-off-by: Chris J Arges <chris.j.arges@canonical.com> >> Acked-by: Pravin B Shelar <pshelar@nicira.com> >> Acked-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> >> Signed-off-by: David S. Miller <davem@davemloft.net> >> >> Upstream: bac541e4631(""openvswitch: allocate nr_node_ids flow_stats >> instead of num_possible_nodes") >> >> Signed-off-by: Pravin B Shelar <pshelar@nicira.com> > > Acked-by: Jesse Gross <jesse@nicira.com> > > This one could probably go back to 2.3 and 2.4 as well independent of > the other upstream changes. Thanks for review. I pushed the patch to master, branch-2.3 and branch-2.4.
diff --git a/datapath/flow_table.c b/datapath/flow_table.c index 409f9dd..c76f2a1 100644 --- a/datapath/flow_table.c +++ b/datapath/flow_table.c @@ -988,7 +988,7 @@ int ovs_flow_init(void) BUILD_BUG_ON(sizeof(struct sw_flow_key) % sizeof(long)); flow_cache = kmem_cache_create("sw_flow", sizeof(struct sw_flow) - + (num_possible_nodes() + + (nr_node_ids * sizeof(struct flow_stats *)), 0, 0, NULL); if (flow_cache == NULL)