question about RT_TABLE_MAX

question about RT_TABLE_MAX

[Dan Carpenter]

RT_TABLE_MAX is 0xFFFFFFFF.  It's always compared against an unsigned
int so the checks against it don't test anything.

net/decnet/dn_table.c
   828  struct dn_fib_table *dn_fib_get_table(u32 n, int create)
   829  {
   830          struct dn_fib_table *t;
   831          struct hlist_node *node;
   832          unsigned int h;
   833  
   834          if (n < RT_TABLE_MIN)
   835                  return NULL;
   836  
   837          if (n > RT_TABLE_MAX)
                    ^^^^^^^^^^^^^^^^
Never true.
   838                  return NULL;

net/decnet/dn_table.c
   874  struct dn_fib_table *dn_fib_empty_table(void)
   875  {
   876          u32 id;
   877  
   878          for(id = RT_TABLE_MIN; id <= RT_TABLE_MAX; id++)
                                       ^^^^^^^^^^^^^^^^^^
Always true.
   879                  if (dn_fib_get_table(id, 0) == NULL)
   880                          return dn_fib_get_table(id, 1);
   881          return NULL;
   882  }

net/ipv4/fib_rules.c
   122  static struct fib_table *fib_empty_table(struct net *net)
   123  {
   124          u32 id;
   125  
   126          for (id = 1; id <= RT_TABLE_MAX; id++)
                             ^^^^^^^^^^^^^^^^^^
Always true.
   127                  if (fib_get_table(net, id) == NULL)
   128                          return fib_new_table(net, id);
   129          return NULL;
   130  }

Maybe it should be id < RT_TABLE_MAX?

regards,
dan carpenter

Re: question about RT_TABLE_MAX

[David Miller]

From: Dan Carpenter <dan.carpenter@oracle.com>
Date: Mon, 12 Nov 2012 17:51:35 +0300

> RT_TABLE_MAX is 0xFFFFFFFF.  It's always compared against an unsigned
> int so the checks against it don't test anything.

It used to be 256, that's why the spurious tests are there, they
just need to be cleaned up.

This change happened in commit:

commit b801f54917b7c6e8540f877ee562cd0725e62ebd
Author: Patrick McHardy <kaber@trash.net>
Date:   Thu Aug 10 23:12:34 2006 -0700

    [NET]: Increate RT_TABLE_MAX to 2^32

    Signed-off-by: Patrick McHardy <kaber@trash.net>
    Signed-off-by: David S. Miller <davem@davemloft.net>