Re: [PATCH net-next 6/6] net: bridge: avoid uselessly making offloaded ports promiscuous

Re: [PATCH net-next 6/6] net: bridge: avoid uselessly making offloaded ports promiscuous

[kernel test robot]

Hi Vladimir,

I love your patch! Yet something to improve:

[auto build test ERROR on net-next/master]

url:    https://github.com/intel-lab-lkp/linux/commits/Vladimir-Oltean/Disable-host-flooding-for-DSA-ports-under-a-bridge/20220409-041556
base:   https://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git e8bd70250a821edb541c3abe1eacdad9f8dc7adf
config: x86_64-randconfig-a003 (https://download.01.org/0day-ci/archive/20220409/202204091856.0PBgeBSa-lkp@intel.com/config)
compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project 893e1c18b98d8bbc7b8d7d22cc2c348f65c72ad9)
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/intel-lab-lkp/linux/commit/2b24e24c704fa129c753dbc8fb764c95f4a3562c
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Vladimir-Oltean/Disable-host-flooding-for-DSA-ports-under-a-bridge/20220409-041556
        git checkout 2b24e24c704fa129c753dbc8fb764c95f4a3562c
        # save the config file to linux build tree
        mkdir build_dir
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash net/bridge/

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

All errors (new ones prefixed by >>):

>> net/bridge/br_if.c:145:10: error: no member named 'offload_count' in 'struct net_bridge_port'
                   if (p->offload_count) {
                       ~  ^
   1 error generated.


vim +145 net/bridge/br_if.c

   129	
   130	/* When a port is added or removed or when certain port flags
   131	 * change, this function is called to automatically manage
   132	 * promiscuity setting of all the bridge ports.  We are always called
   133	 * under RTNL so can skip using rcu primitives.
   134	 */
   135	void br_manage_promisc(struct net_bridge *br)
   136	{
   137		struct net_bridge_port *p;
   138	
   139		list_for_each_entry(p, &br->port_list, list) {
   140			/* Offloaded ports have a separate address database for
   141			 * forwarding, which is managed through switchdev and not
   142			 * through dev_uc_add(), so the promiscuous concept makes no
   143			 * sense for them. Avoid updating promiscuity in that case.
   144			 */
 > 145			if (p->offload_count) {
   146				br_port_clear_promisc(p);
   147				continue;
   148			}
   149	
   150			/* If bridge is promiscuous, unconditionally place all ports
   151			 * in promiscuous mode too. This allows the bridge device to
   152			 * locally receive all unknown traffic.
   153			 */
   154			if (br->dev->flags & IFF_PROMISC) {
   155				br_port_set_promisc(p);
   156				continue;
   157			}
   158	
   159			/* If vlan filtering is disabled, place all ports in
   160			 * promiscuous mode.
   161			 */
   162			if (!br_vlan_enabled(br->dev)) {
   163				br_port_set_promisc(p);
   164				continue;
   165			}
   166	
   167			/* If the number of auto-ports is <= 1, then all other ports
   168			 * will have their output configuration statically specified
   169			 * through fdbs. Since ingress on the auto-port becomes
   170			 * forwarding/egress to other ports and egress configuration is
   171			 * statically known, we can say that ingress configuration of
   172			 * the auto-port is also statically known.
   173			 * This lets us disable promiscuous mode and write this config
   174			 * to hw.
   175			 */
   176			if (br->auto_cnt == 0 ||
   177			    (br->auto_cnt == 1 && br_auto_port(p)))
   178				br_port_clear_promisc(p);
   179			else
   180				br_port_set_promisc(p);
   181		}
   182	}
   183	

-- 
0-DAY CI Kernel Test Service
https://01.org/lkp

Re: [PATCH net-next 6/6] net: bridge: avoid uselessly making offloaded ports promiscuous

[Vladimir Oltean]

On Sat, Apr 09, 2022 at 06:17:15PM +0800, kernel test robot wrote:
> >> net/bridge/br_if.c:145:10: error: no member named 'offload_count' in 'struct net_bridge_port'
>                    if (p->offload_count) {
>                        ~  ^
>    130	/* When a port is added or removed or when certain port flags
>    131	 * change, this function is called to automatically manage
>    132	 * promiscuity setting of all the bridge ports.  We are always called
>    133	 * under RTNL so can skip using rcu primitives.
>    134	 */
>    135	void br_manage_promisc(struct net_bridge *br)
>    136	{
>    137		struct net_bridge_port *p;
>    138	
>    139		list_for_each_entry(p, &br->port_list, list) {
>    140			/* Offloaded ports have a separate address database for
>    141			 * forwarding, which is managed through switchdev and not
>    142			 * through dev_uc_add(), so the promiscuous concept makes no
>    143			 * sense for them. Avoid updating promiscuity in that case.
>    144			 */
>  > 145			if (p->offload_count) {

Good point. Please imagine there's a static inline nbp_get_offload_count()
that returns 0 when CONFIG_NET_SWITCHDEV=n. I'll address this for v2.

>    146				br_port_clear_promisc(p);
>    147				continue;
>    148			}
>    149	
>    150			/* If bridge is promiscuous, unconditionally place all ports
>    151			 * in promiscuous mode too. This allows the bridge device to
>    152			 * locally receive all unknown traffic.
>    153			 */
>    154			if (br->dev->flags & IFF_PROMISC) {
>    155				br_port_set_promisc(p);
>    156				continue;
>    157			}
>    158	
>    159			/* If vlan filtering is disabled, place all ports in
>    160			 * promiscuous mode.
>    161			 */
>    162			if (!br_vlan_enabled(br->dev)) {
>    163				br_port_set_promisc(p);
>    164				continue;
>    165			}
>    166	
>    167			/* If the number of auto-ports is <= 1, then all other ports
>    168			 * will have their output configuration statically specified
>    169			 * through fdbs. Since ingress on the auto-port becomes
>    170			 * forwarding/egress to other ports and egress configuration is
>    171			 * statically known, we can say that ingress configuration of
>    172			 * the auto-port is also statically known.
>    173			 * This lets us disable promiscuous mode and write this config
>    174			 * to hw.
>    175			 */
>    176			if (br->auto_cnt == 0 ||
>    177			    (br->auto_cnt == 1 && br_auto_port(p)))
>    178				br_port_clear_promisc(p);
>    179			else
>    180				br_port_set_promisc(p);
>    181		}
>    182	}
>    183