| Message ID | 20240927123801.14853-1-wojciech.drewek@intel.com |
|---|---|
| State | Accepted |
| Delegated to: | Anthony Nguyen |
| Headers | show |
| Series | [iwl-net,v2] ice: Flush FDB entries before reset | expand |
> -----Original Message----- > From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of > Wojciech Drewek > Sent: Friday, September 27, 2024 6:08 PM > To: netdev@vger.kernel.org > Cc: Nguyen, Anthony L <anthony.l.nguyen@intel.com>; > marcin.szycik@linux.intel.com; intel-wired-lan@lists.osuosl.org; > horms@kernel.org; michal.swiatkowski@linux.intel.com > Subject: [Intel-wired-lan] [PATCH iwl-net v2] ice: Flush FDB entries before > reset > > Triggering the reset while in switchdev mode causes errors[1]. Rules are > already removed by this time because switch content is flushed in case of the > reset. > This means that rules were deleted from HW but SW still thinks they exist so > when we get SWITCHDEV_FDB_DEL_TO_DEVICE notification we try to delete > not existing rule. > > We can avoid these errors by clearing the rules early in the reset flow before > they are removed from HW. > Switchdev API will get notified that the rule was removed so we won't get > SWITCHDEV_FDB_DEL_TO_DEVICE notification. > Remove unnecessary ice_clear_sw_switch_recipes. > > [1] > ice 0000:01:00.0: Failed to delete FDB forward rule, err: -2 ice 0000:01:00.0: > Failed to delete FDB guard rule, err: -2 > > Fixes: 7c945a1a8e5f ("ice: Switchdev FDB events support") > Reviewed-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com> > Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com> > --- > v2: extend commit msg, add NULL pointer check > --- > .../net/ethernet/intel/ice/ice_eswitch_br.c | 5 +++- > .../net/ethernet/intel/ice/ice_eswitch_br.h | 1 + > drivers/net/ethernet/intel/ice/ice_main.c | 24 +++---------------- > 3 files changed, 8 insertions(+), 22 deletions(-) > Tested-by: Sujai Buvaneswaran <sujai.buvaneswaran@intel.com>
diff --git a/drivers/net/ethernet/intel/ice/ice_eswitch_br.c b/drivers/net/ethernet/intel/ice/ice_eswitch_br.c index f5aceb32bf4d..cccb7ddf61c9 100644 --- a/drivers/net/ethernet/intel/ice/ice_eswitch_br.c +++ b/drivers/net/ethernet/intel/ice/ice_eswitch_br.c @@ -582,10 +582,13 @@ ice_eswitch_br_switchdev_event(struct notifier_block *nb, return NOTIFY_DONE; } -static void ice_eswitch_br_fdb_flush(struct ice_esw_br *bridge) +void ice_eswitch_br_fdb_flush(struct ice_esw_br *bridge) { struct ice_esw_br_fdb_entry *entry, *tmp; + if (!bridge) + return; + list_for_each_entry_safe(entry, tmp, &bridge->fdb_list, list) ice_eswitch_br_fdb_entry_notify_and_cleanup(bridge, entry); } diff --git a/drivers/net/ethernet/intel/ice/ice_eswitch_br.h b/drivers/net/ethernet/intel/ice/ice_eswitch_br.h index c15c7344d7f8..66a2c804338f 100644 --- a/drivers/net/ethernet/intel/ice/ice_eswitch_br.h +++ b/drivers/net/ethernet/intel/ice/ice_eswitch_br.h @@ -117,5 +117,6 @@ void ice_eswitch_br_offloads_deinit(struct ice_pf *pf); int ice_eswitch_br_offloads_init(struct ice_pf *pf); +void ice_eswitch_br_fdb_flush(struct ice_esw_br *bridge); #endif /* _ICE_ESWITCH_BR_H_ */ diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index c7db88b517da..dcd4c8204753 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -519,25 +519,6 @@ static void ice_pf_dis_all_vsi(struct ice_pf *pf, bool locked) pf->vf_agg_node[node].num_vsis = 0; } -/** - * ice_clear_sw_switch_recipes - clear switch recipes - * @pf: board private structure - * - * Mark switch recipes as not created in sw structures. There are cases where - * rules (especially advanced rules) need to be restored, either re-read from - * hardware or added again. For example after the reset. 'recp_created' flag - * prevents from doing that and need to be cleared upfront. - */ -static void ice_clear_sw_switch_recipes(struct ice_pf *pf) -{ - struct ice_sw_recipe *recp; - u8 i; - - recp = pf->hw.switch_info->recp_list; - for (i = 0; i < ICE_MAX_NUM_RECIPES; i++) - recp[i].recp_created = false; -} - /** * ice_prepare_for_reset - prep for reset * @pf: board private structure @@ -574,8 +555,9 @@ ice_prepare_for_reset(struct ice_pf *pf, enum ice_reset_req reset_type) mutex_unlock(&pf->vfs.table_lock); if (ice_is_eswitch_mode_switchdev(pf)) { - if (reset_type != ICE_RESET_PFR) - ice_clear_sw_switch_recipes(pf); + rtnl_lock(); + ice_eswitch_br_fdb_flush(pf->eswitch.br_offloads->bridge); + rtnl_unlock(); } /* release ADQ specific HW and SW resources */
Triggering the reset while in switchdev mode causes errors[1]. Rules are already removed by this time because switch content is flushed in case of the reset. This means that rules were deleted from HW but SW still thinks they exist so when we get SWITCHDEV_FDB_DEL_TO_DEVICE notification we try to delete not existing rule. We can avoid these errors by clearing the rules early in the reset flow before they are removed from HW. Switchdev API will get notified that the rule was removed so we won't get SWITCHDEV_FDB_DEL_TO_DEVICE notification. Remove unnecessary ice_clear_sw_switch_recipes. [1] ice 0000:01:00.0: Failed to delete FDB forward rule, err: -2 ice 0000:01:00.0: Failed to delete FDB guard rule, err: -2 Fixes: 7c945a1a8e5f ("ice: Switchdev FDB events support") Reviewed-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com> Signed-off-by: Wojciech Drewek <wojciech.drewek@intel.com> --- v2: extend commit msg, add NULL pointer check --- .../net/ethernet/intel/ice/ice_eswitch_br.c | 5 +++- .../net/ethernet/intel/ice/ice_eswitch_br.h | 1 + drivers/net/ethernet/intel/ice/ice_main.c | 24 +++---------------- 3 files changed, 8 insertions(+), 22 deletions(-)