[PATCH nf] netfilter: ebtables: Fix argument order to ADD_COUNTER

[PATCH nf] netfilter: ebtables: Fix argument order to ADD_COUNTER

[Todd Seidelmann]

The ordering of arguments to the x_tables ADD_COUNTER macro
appears to be wrong in ebtables (cf. ip_tables.c, ip6_tables.c,
and arp_tables.c).

This causes data corruption in the ebtables userspace tools
because they get incorrect packet & byte counts from the kernel.
---
  net/bridge/netfilter/ebtables.c | 8 ++++----
  1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/net/bridge/netfilter/ebtables.c 
b/net/bridge/netfilter/ebtables.c
index c8177a8..4096d8a 100644
--- a/net/bridge/netfilter/ebtables.c
+++ b/net/bridge/netfilter/ebtables.c
@@ -221,7 +221,7 @@ unsigned int ebt_do_table(struct sk_buff *skb,
              return NF_DROP;
          }

-        ADD_COUNTER(*(counter_base + i), 1, skb->len);
+        ADD_COUNTER(*(counter_base + i), skb->len, 1);

          /* these should only watch: not modify, nor tell us
           * what to do with the packet
@@ -959,8 +959,8 @@ static void get_counters(const struct ebt_counter 
*oldcounters,
              continue;
          counter_base = COUNTER_BASE(oldcounters, nentries, cpu);
          for (i = 0; i < nentries; i++)
-            ADD_COUNTER(counters[i], counter_base[i].pcnt,
-                    counter_base[i].bcnt);
+            ADD_COUNTER(counters[i], counter_base[i].bcnt,
+                    counter_base[i].pcnt);
      }
  }

@@ -1280,7 +1280,7 @@ static int do_update_counters(struct net *net, 
const char *name,

      /* we add to the counters of the first cpu */
      for (i = 0; i < num_counters; i++)
-        ADD_COUNTER(t->private->counters[i], tmp[i].pcnt, tmp[i].bcnt);
+        ADD_COUNTER(t->private->counters[i], tmp[i].bcnt, tmp[i].pcnt);

      write_unlock_bh(&t->lock);
      ret = 0;
--
1.8.3.1

Re: [PATCH nf] netfilter: ebtables: Fix argument order to ADD_COUNTER

[Florian Westphal]

Todd Seidelmann <tseidelmann@linode.com> wrote:
> The ordering of arguments to the x_tables ADD_COUNTER macro
> appears to be wrong in ebtables (cf. ip_tables.c, ip6_tables.c,
> and arp_tables.c).
> 
> This causes data corruption in the ebtables userspace tools
> because they get incorrect packet & byte counts from the kernel.

Please resend with a 'Signed-off-by' tag and the
'Fixes:' tag that I provided when I reviewd this patch.

It will help stable maintainers to figure out which trees
need this fix applied.

[PATCH nf] netfilter: ebtables: Fix argument order to ADD_COUNTER

[Todd Seidelmann]

The ordering of arguments to the x_tables ADD_COUNTER macro
appears to be wrong in ebtables (cf. ip_tables.c, ip6_tables.c,
and arp_tables.c).

This causes data corruption in the ebtables userspace tools
because they get incorrect packet & byte counts from the kernel.

Fixes: d72133e628803 ("netfilter: ebtables: use ADD_COUNTER macro")
Signed-off-by: Todd Seidelmann<tseidelmann@linode.com>
---
My last submission that included Florian Westphal's requests
seems to have been lost in the ether, so I'm resending.
Apologies for the spam.

  net/bridge/netfilter/ebtables.c | 8 ++++----
  1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c
index c8177a8..4096d8a 100644
--- a/net/bridge/netfilter/ebtables.c
+++ b/net/bridge/netfilter/ebtables.c
@@ -221,7 +221,7 @@ unsigned int ebt_do_table(struct sk_buff *skb,
              return NF_DROP;
          }

-        ADD_COUNTER(*(counter_base + i), 1, skb->len);
+        ADD_COUNTER(*(counter_base + i), skb->len, 1);

          /* these should only watch: not modify, nor tell us
           * what to do with the packet
@@ -959,8 +959,8 @@ static void get_counters(const struct ebt_counter *oldcounters,
              continue;
          counter_base = COUNTER_BASE(oldcounters, nentries, cpu);
          for (i = 0; i < nentries; i++)
-            ADD_COUNTER(counters[i], counter_base[i].pcnt,
-                    counter_base[i].bcnt);
+            ADD_COUNTER(counters[i], counter_base[i].bcnt,
+                    counter_base[i].pcnt);
      }
  }

@@ -1280,7 +1280,7 @@ static int do_update_counters(struct net *net, const char *name,

      /* we add to the counters of the first cpu */
      for (i = 0; i < num_counters; i++)
-        ADD_COUNTER(t->private->counters[i], tmp[i].pcnt, tmp[i].bcnt);
+        ADD_COUNTER(t->private->counters[i], tmp[i].bcnt, tmp[i].pcnt);

      write_unlock_bh(&t->lock);
      ret = 0;
--
1.8.3.1

Re: [PATCH nf] netfilter: ebtables: Fix argument order to ADD_COUNTER

[Florian Westphal]

Todd Seidelmann <tseidelmann@linode.com> wrote:
> The ordering of arguments to the x_tables ADD_COUNTER macro
> appears to be wrong in ebtables (cf. ip_tables.c, ip6_tables.c,
> and arp_tables.c).
> 
> This causes data corruption in the ebtables userspace tools
> because they get incorrect packet & byte counts from the kernel.
> 
> Fixes: d72133e628803 ("netfilter: ebtables: use ADD_COUNTER macro")
> Signed-off-by: Todd Seidelmann<tseidelmann@linode.com>

Acked-by: Florian Westphal <fw@strlen.de>

Re: [PATCH nf] netfilter: ebtables: Fix argument order to ADD_COUNTER

[Pablo Neira Ayuso]

On Wed, Aug 14, 2019 at 10:54:16AM -0400, Todd Seidelmann wrote:
> The ordering of arguments to the x_tables ADD_COUNTER macro
> appears to be wrong in ebtables (cf. ip_tables.c, ip6_tables.c,
> and arp_tables.c).
> 
> This causes data corruption in the ebtables userspace tools
> because they get incorrect packet & byte counts from the kernel.

Applied, thanks.