From patchwork Thu Mar 3 23:19:52 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Samudrala, Sridhar" X-Patchwork-Id: 591643 X-Patchwork-Delegate: davem@davemloft.net Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id D0BED141B02 for ; Fri, 4 Mar 2016 10:24:00 +1100 (AEDT) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758613AbcCCXXz (ORCPT ); Thu, 3 Mar 2016 18:23:55 -0500 Received: from mga03.intel.com ([134.134.136.65]:27501 "EHLO mga03.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754573AbcCCXXz (ORCPT ); Thu, 3 Mar 2016 18:23:55 -0500 Received: from orsmga001.jf.intel.com ([10.7.209.18]) by orsmga103.jf.intel.com with ESMTP; 03 Mar 2016 15:23:49 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.22,533,1449561600"; d="scan'208";a="900789163" Received: from sri02.jf.intel.com ([10.166.39.21]) by orsmga001.jf.intel.com with ESMTP; 03 Mar 2016 15:23:48 -0800 From: Sridhar Samudrala To: john.r.fastabend@intel.com, davem@davemloft.net, jeffrey.t.kirsher@intel.com, netdev@vger.kernel.org Subject: [PATCH net-next] net: ixgbe: fix error handling in tc cls_u32 offload routines. Date: Thu, 3 Mar 2016 15:19:52 -0800 Message-Id: <1457047192-17092-1-git-send-email-sridhar.samudrala@intel.com> X-Mailer: git-send-email 2.1.0 Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org Check for handle ids when adding/deleting hash nodes OR adding/deleting filter entries and limit them to max number of links or header nodes supported(IXGBE_MAX_LINK_HANDLE). Start from bit 0 when setting hash table bit-map.(adapter->tables) Signed-off-by: Sridhar Samudrala --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 52 +++++++++++++++++---------- 1 file changed, 34 insertions(+), 18 deletions(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index b893ff8..10ccd96 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -8209,10 +8209,17 @@ int ixgbe_setup_tc(struct net_device *dev, u8 tc) static int ixgbe_delete_clsu32(struct ixgbe_adapter *adapter, struct tc_cls_u32_offload *cls) { + u32 uhtid = TC_U32_USERHTID(cls->knode.handle); + u32 loc; int err; + if ((uhtid != 0x800) && (uhtid >= IXGBE_MAX_LINK_HANDLE)) + return -EINVAL; + + loc = cls->knode.handle & 0xfffff; + spin_lock(&adapter->fdir_perfect_lock); - err = ixgbe_update_ethtool_fdir_entry(adapter, NULL, cls->knode.handle); + err = ixgbe_update_ethtool_fdir_entry(adapter, NULL, loc); spin_unlock(&adapter->fdir_perfect_lock); return err; } @@ -8221,20 +8228,30 @@ static int ixgbe_configure_clsu32_add_hnode(struct ixgbe_adapter *adapter, __be16 protocol, struct tc_cls_u32_offload *cls) { + u32 uhtid = TC_U32_USERHTID(cls->hnode.handle); + + if (uhtid >= IXGBE_MAX_LINK_HANDLE) + return -EINVAL; + /* This ixgbe devices do not support hash tables at the moment * so abort when given hash tables. */ if (cls->hnode.divisor > 0) return -EINVAL; - set_bit(TC_U32_USERHTID(cls->hnode.handle), &adapter->tables); + set_bit(uhtid-1, &adapter->tables); return 0; } static int ixgbe_configure_clsu32_del_hnode(struct ixgbe_adapter *adapter, struct tc_cls_u32_offload *cls) { - clear_bit(TC_U32_USERHTID(cls->hnode.handle), &adapter->tables); + u32 uhtid = TC_U32_USERHTID(cls->hnode.handle); + + if (uhtid >= IXGBE_MAX_LINK_HANDLE) + return -EINVAL; + + clear_bit(uhtid-1, &adapter->tables); return 0; } @@ -8252,27 +8269,29 @@ static int ixgbe_configure_clsu32(struct ixgbe_adapter *adapter, #endif int i, err = 0; u8 queue; - u32 handle; + u32 uhtid, link_uhtid; memset(&mask, 0, sizeof(union ixgbe_atr_input)); - handle = cls->knode.handle; + uhtid = TC_U32_USERHTID(cls->knode.handle); + link_uhtid = TC_U32_USERHTID(cls->knode.link_handle); - /* At the moment cls_u32 jumps to transport layer and skips past + /* At the moment cls_u32 jumps to network layer and skips past * L2 headers. The canonical method to match L2 frames is to use * negative values. However this is error prone at best but really * just broken because there is no way to "know" what sort of hdr - * is in front of the transport layer. Fix cls_u32 to support L2 + * is in front of the network layer. Fix cls_u32 to support L2 * headers when needed. */ if (protocol != htons(ETH_P_IP)) return -EINVAL; - if (cls->knode.link_handle || - cls->knode.link_handle >= IXGBE_MAX_LINK_HANDLE) { + if (link_uhtid) { struct ixgbe_nexthdr *nexthdr = ixgbe_ipv4_jumps; - u32 uhtid = TC_U32_USERHTID(cls->knode.link_handle); - if (!test_bit(uhtid, &adapter->tables)) + if (link_uhtid >= IXGBE_MAX_LINK_HANDLE) + return -EINVAL; + + if (!test_bit(link_uhtid-1, &adapter->tables)) return -EINVAL; for (i = 0; nexthdr[i].jump; i++) { @@ -8288,10 +8307,7 @@ static int ixgbe_configure_clsu32(struct ixgbe_adapter *adapter, nexthdr->mask != cls->knode.sel->keys[0].mask) return -EINVAL; - if (uhtid >= IXGBE_MAX_LINK_HANDLE) - return -EINVAL; - - adapter->jump_tables[uhtid] = nexthdr->jump; + adapter->jump_tables[link_uhtid] = nexthdr->jump; } return 0; } @@ -8308,13 +8324,13 @@ static int ixgbe_configure_clsu32(struct ixgbe_adapter *adapter, * To add support for new nodes update ixgbe_model.h parse structures * this function _should_ be generic try not to hardcode values here. */ - if (TC_U32_USERHTID(handle) == 0x800) { + if (uhtid == 0x800) { field_ptr = adapter->jump_tables[0]; } else { - if (TC_U32_USERHTID(handle) >= ARRAY_SIZE(adapter->jump_tables)) + if (uhtid >= IXGBE_MAX_LINK_HANDLE) return -EINVAL; - field_ptr = adapter->jump_tables[TC_U32_USERHTID(handle)]; + field_ptr = adapter->jump_tables[uhtid]; } if (!field_ptr)