Skip to contentSkip to conversation

net/mlx5e: Handle FIB events to update tunnel endpoint device

bug report Dan Carpenter 3 participants 7 messages Feb 10, 2021

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

1640 }
1641
1642 return encap;
1643 }

regards,
dan carpenter

#1 lore raw
Quoted 30 lines · Dan Carpenter

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

Quoted 5 lines
  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

Hi Dan,

Thanks for the bug report!

This looks a bit strange to me because none of the other users of this API handle positive error code in any special way (including reference netdevsim implementation). Maybe API itself should be fixed? Jiri, Ido, what do you think?

Regards,
Vlad

Quoted 8 lines

1640 }
1641
1642 return encap;
1643 }

regards,
dan carpenter

#2 lore raw
Quoted 37 lines · Vlad Buslov
Quoted 30 lines · Dan Carpenter

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

Quoted 5 lines
  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

Quoted 8 lines

Hi Dan,

Thanks for the bug report!

This looks a bit strange to me because none of the other users of this API handle positive error code in any special way (including reference netdevsim implementation). Maybe API itself should be fixed? Jiri, Ido, what do you think?

The other functions that call register_fib_notifier() return an int, but mlx5e_tc_tun_init() returns a pointer. I think that's why it was
flagged: "error: passing non negative 1 to ERR_PTR".

fib6_walk_continue() cannot return a positive value when called from register_fib_notifier(), but ignoring it means that you will keep getting the same static analysis error. What about:

net/ipv6/ip6_fib.c +3 -2
index f43e27555725..ef9d022e693f 100644
@@ -499,7 +499,7 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb,
 
                hlist_for_each_entry_rcu(tb, head, tb6_hlist) {
                        err = fib6_table_dump(net, tb, w);
-                       if (err < 0)
+                       if (err)
                                goto out;
                }
        }
@@ -507,7 +507,8 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb,
 out:
        kfree(w);
 
-       return err;
+       /* The tree traversal function should never return a positive value. */
+       return err > 0 ? -EINVAL : err;
 }
 
 static int fib6_dump_node(struct fib6_walker *w)
Quoted 13 lines

Regards,
Vlad

Quoted 8 lines

1640 }
1641
1642 return encap;
1643 }

regards,
dan carpenter

#3 lore raw
Quoted 42 lines · Ido Schimmel
Quoted 41 lines · Vlad Buslov
Quoted 30 lines · Dan Carpenter

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

Quoted 5 lines
  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

Hi Dan,

Thanks for the bug report!

Quoted 8 lines
Quoted 4 lines

This looks a bit strange to me because none of the other users of this API handle positive error code in any special way (including reference netdevsim implementation). Maybe API itself should be fixed? Jiri, Ido, what do you think?

The other functions that call register_fib_notifier() return an int, but mlx5e_tc_tun_init() returns a pointer. I think that's why it was
flagged: "error: passing non negative 1 to ERR_PTR".

Actually, some of them do. I mentioned netdevsim specifically because nsim_fib_create() returns pointer and also casts return value of register_fib_notifier() with ERR_PTR.

Quoted 20 lines

fib6_walk_continue() cannot return a positive value when called from register_fib_notifier(), but ignoring it means that you will keep getting the same static analysis error. What about:

net/ipv6/ip6_fib.c +1 -1
index f43e27555725..ef9d022e693f 100644
@@ -499,7 +499,7 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb,
 
                hlist_for_each_entry_rcu(tb, head, tb6_hlist) {
                        err = fib6_table_dump(net, tb, w);
-                       if (err < 0)
+                       if (err)
                                goto out;
                }
        }

net/ipv6/ip6_fib.c int fib6_tables_dump(struct net *net, struct notifier_block *nb, · line 514

Quoted 9 lines
net/ipv6/ip6_fib.c +2 -1
@@ -507,7 +507,8 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb,
 out:
        kfree(w);
 
-       return err;
+       /* The tree traversal function should never return a positive value. */
+       return err > 0 ? -EINVAL : err;
 }
 
 static int fib6_dump_node(struct fib6_walker *w)

Makes sense. You want me to send the fix or you will do it?

Quoted 13 lines
Quoted 13 lines

Regards,
Vlad

Quoted 8 lines

1640 }
1641
1642 return encap;
1643 }

regards,
dan carpenter

#4 lore raw
Quoted 84 lines · Vlad Buslov
Quoted 42 lines · Ido Schimmel
Quoted 41 lines · Vlad Buslov
Quoted 30 lines · Dan Carpenter

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

Quoted 5 lines
  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

Hi Dan,

Thanks for the bug report!

Quoted 8 lines
Quoted 4 lines

This looks a bit strange to me because none of the other users of this API handle positive error code in any special way (including reference netdevsim implementation). Maybe API itself should be fixed? Jiri, Ido, what do you think?

The other functions that call register_fib_notifier() return an int, but mlx5e_tc_tun_init() returns a pointer. I think that's why it was
flagged: "error: passing non negative 1 to ERR_PTR".

Actually, some of them do. I mentioned netdevsim specifically because nsim_fib_create() returns pointer and also casts return value of register_fib_notifier() with ERR_PTR.

Quoted 27 lines

fib6_walk_continue() cannot return a positive value when called from register_fib_notifier(), but ignoring it means that you will keep getting the same static analysis error. What about:

net/ipv6/ip6_fib.c +3 -2
index f43e27555725..ef9d022e693f 100644
@@ -499,7 +499,7 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb,
 
                hlist_for_each_entry_rcu(tb, head, tb6_hlist) {
                        err = fib6_table_dump(net, tb, w);
-                       if (err < 0)
+                       if (err)
                                goto out;
                }
        }
@@ -507,7 +507,8 @@ int fib6_tables_dump(struct net *net, struct notifier_block *nb,
 out:
        kfree(w);
 
-       return err;
+       /* The tree traversal function should never return a positive value. */
+       return err > 0 ? -EINVAL : err;
 }
 
Quoted 4 lines

net/ipv6/ip6_fib.c int fib6_tables_dump(struct net *net, struct notifier_block *nb, · line 514

Quoted 2 lines
net/ipv6/ip6_fib.c +0 -0
 static int fib6_dump_node(struct fib6_walker *w)

Makes sense. You want me to send the fix or you will do it?

Can you send to linux-internal and then I will apply it to our regression and let you know tomorrow morning?

Quoted 15 lines
Quoted 14 lines
Quoted 14 lines

Regards,
Vlad

Quoted 8 lines

1640 }
1641
1642 return encap;
1643 }

regards,
dan carpenter

#5 lore raw
In reply to Vlad Buslov
Quoted 49 lines · Vlad Buslov
Quoted 47 lines · Ido Schimmel
Quoted 45 lines · Vlad Buslov
Quoted 30 lines · Dan Carpenter

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

Quoted 5 lines
  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

Hi Dan,

Thanks for the bug report!

This looks a bit strange to me because none of the other users of this API handle positive error code in any special way (including reference netdevsim implementation). Maybe API itself should be fixed? Jiri, Ido, what do you think?

Quoted 8 lines
Quoted 3 lines

The other functions that call register_fib_notifier() return an int, but mlx5e_tc_tun_init() returns a pointer. I think that's why it was
flagged: "error: passing non negative 1 to ERR_PTR".

Actually, some of them do. I mentioned netdevsim specifically because nsim_fib_create() returns pointer and also casts return value of register_fib_notifier() with ERR_PTR.

Huh. Smatch prints a warning for nsim_fib_create() but I never forwarded the warning to anyone apparently.

regards,
dan carpenter

#6 lore raw
In reply to Ido Schimmel
Quoted 47 lines · Ido Schimmel
Quoted 37 lines · Vlad Buslov
Quoted 30 lines · Dan Carpenter

Hello Vlad Buslov,

The patch 8914add2c9e5: "net/mlx5e: Handle FIB events to update tunnel endpoint device" from Jan 25, 2021, leads to the following static checker warning:

	drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c:1639 mlx5e_tc_tun_init()
	error: passing non negative 1 to ERR_PTR

drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_encap.c
1622 struct mlx5e_tc_tun_encap *mlx5e_tc_tun_init(struct mlx5e_priv *priv) 1623 {
1624 struct mlx5e_tc_tun_encap *encap;
1625 int err;
1626
1627 encap = kvzalloc(sizeof(*encap), GFP_KERNEL);
1628 if (!encap)
1629 return ERR_PTR(-ENOMEM);
1630

  1631          encap->priv = priv;
  1632          encap->fib_nb.notifier_call = mlx5e_tc_tun_fib_event;
  1633          spin_lock_init(&encap->route_lock);
  1634          hash_init(encap->route_tbl);
  1635          err = register_fib_notifier(dev_net(priv->netdev), &encap->fib_nb,

1636 NULL, NULL);

register_fib_notifier() calls fib_net_dump() which eventually calls fib6_walk_continue() which can return 1 if "walk is incomplete (i.e. suspended)".

Quoted 5 lines
  1637          if (err) {
  1638                  kvfree(encap);
  1639                  return ERR_PTR(err);

If this returns 1 it will eventually lead to an Oops.

Quoted 8 lines

Hi Dan,

Thanks for the bug report!

This looks a bit strange to me because none of the other users of this API handle positive error code in any special way (including reference netdevsim implementation). Maybe API itself should be fixed? Jiri, Ido, what do you think?

Quoted 6 lines

The other functions that call register_fib_notifier() return an int, but mlx5e_tc_tun_init() returns a pointer. I think that's why it was
flagged: "error: passing non negative 1 to ERR_PTR".

fib6_walk_continue() cannot return a positive value when called from register_fib_notifier()

Ideally Smatch would be able to figure out this rule automatically.

I see now that fib6_tables_dump() can't return positive because it returns either zero or the return from the "w->func = fib6_node_dump;" and fib6_node_dump() doesn't return positive.

Handling it properly is something I've thought about but I haven't figured out how. What I will do is hard code this into Smatch by adding a line into the smatch_data/db/fixup_kernel.sh file:

delete from return_states where function = 'fib6_tables_dump' and return = '1';

regards,
dan carpenter