From 4c4ed0748f82e26d8b884383e6737cf5a861ed6f Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Tue, 14 Apr 2015 16:44:14 +0200 Subject: netfilter: nf_tables: fix wrong length for jump/goto verdicts NFT_JUMP/GOTO erronously sets length to sizeof(void *). We then allocate insufficient memory when such element is added to a vmap. Suggested-by: Patrick McHardy Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_tables_api.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 78af83bc9c8e..ad9d11fb29fd 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -4340,7 +4340,6 @@ static int nft_verdict_init(const struct nft_ctx *ctx, struct nft_data *data, case NFT_CONTINUE: case NFT_BREAK: case NFT_RETURN: - desc->len = sizeof(data->verdict); break; case NFT_JUMP: case NFT_GOTO: @@ -4355,10 +4354,10 @@ static int nft_verdict_init(const struct nft_ctx *ctx, struct nft_data *data, chain->use++; data->verdict.chain = chain; - desc->len = sizeof(data); break; } + desc->len = sizeof(data->verdict); desc->type = NFT_DATA_VERDICT; return 0; } -- cgit v1.2.3 From 547c4b547e07dcc60874b6ef6252dd49ff74aec1 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 20 Apr 2015 12:35:47 +0200 Subject: netfilter: bridge: fix NULL deref in physin/out ifindex helpers Might not have an outdev yet. We'll oops when iface goes down while skbs are still nfqueue'd: RIP: 0010:[] [] dev_cmp+0x4f/0x80 nfqnl_rcv_dev_event+0xe2/0x150 notifier_call_chain+0x53/0xa0 Fixes: c737b7c4510026 ("netfilter: bridge: add helpers for fetching physin/outdev") Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/linux/netfilter_bridge.h | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/include/linux/netfilter_bridge.h b/include/linux/netfilter_bridge.h index ab8f76dba668..f2fdb5a52070 100644 --- a/include/linux/netfilter_bridge.h +++ b/include/linux/netfilter_bridge.h @@ -39,12 +39,24 @@ static inline void br_drop_fake_rtable(struct sk_buff *skb) static inline int nf_bridge_get_physinif(const struct sk_buff *skb) { - return skb->nf_bridge ? skb->nf_bridge->physindev->ifindex : 0; + struct nf_bridge_info *nf_bridge; + + if (skb->nf_bridge == NULL) + return 0; + + nf_bridge = skb->nf_bridge; + return nf_bridge->physindev ? nf_bridge->physindev->ifindex : 0; } static inline int nf_bridge_get_physoutif(const struct sk_buff *skb) { - return skb->nf_bridge ? skb->nf_bridge->physoutdev->ifindex : 0; + struct nf_bridge_info *nf_bridge; + + if (skb->nf_bridge == NULL) + return 0; + + nf_bridge = skb->nf_bridge; + return nf_bridge->physoutdev ? nf_bridge->physoutdev->ifindex : 0; } static inline struct net_device * -- cgit v1.2.3 From 8e046d68ba719a5bb95912c154c6314472edccfa Mon Sep 17 00:00:00 2001 From: "Karicheri, Muralidharan" Date: Mon, 27 Apr 2015 14:12:43 -0400 Subject: net: netcp: remove call to netif_carrier_(on/off) for MAC to Phy interface Currently when interface type is MAC to Phy, netif_carrier_(on/off) is called which is not needed as Phy lib already updates the carrier status to net stack. This is needed only for other interface types Signed-off-by: Murali Karicheri Signed-off-by: David S. Miller --- drivers/net/ethernet/ti/netcp_ethss.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/ti/netcp_ethss.c b/drivers/net/ethernet/ti/netcp_ethss.c index 2bef655279f3..9b7e0a34c98b 100644 --- a/drivers/net/ethernet/ti/netcp_ethss.c +++ b/drivers/net/ethernet/ti/netcp_ethss.c @@ -1765,7 +1765,9 @@ static void netcp_ethss_link_state_action(struct gbe_priv *gbe_dev, ALE_PORT_STATE, ALE_PORT_STATE_FORWARD); - if (ndev && slave->open) + if (ndev && slave->open && + slave->link_interface != SGMII_LINK_MAC_PHY && + slave->link_interface != XGMII_LINK_MAC_PHY) netif_carrier_on(ndev); } else { writel(mac_control, GBE_REG_ADDR(slave, emac_regs, @@ -1773,7 +1775,9 @@ static void netcp_ethss_link_state_action(struct gbe_priv *gbe_dev, cpsw_ale_control_set(gbe_dev->ale, slave->port_num, ALE_PORT_STATE, ALE_PORT_STATE_DISABLE); - if (ndev) + if (ndev && + slave->link_interface != SGMII_LINK_MAC_PHY && + slave->link_interface != XGMII_LINK_MAC_PHY) netif_carrier_off(ndev); } -- cgit v1.2.3 From 876a7ae65b86d8cec8efe7d15d050ac61116874e Mon Sep 17 00:00:00 2001 From: Alexei Starovoitov Date: Mon, 27 Apr 2015 14:40:37 -0700 Subject: bpf: fix 64-bit divide ALU64_DIV instruction should be dividing 64-bit by 64-bit, whereas do_div() does 64-bit by 32-bit divide. x64 and arm64 JITs correctly implement 64 by 64 unsigned divide. llvm BPF backend emits code assuming that ALU64_DIV does 64 by 64. Fixes: 89aa075832b0 ("net: sock: allow eBPF programs to be attached to sockets") Reported-by: Michael Holzheu Acked-by: Daniel Borkmann Signed-off-by: Alexei Starovoitov Signed-off-by: David S. Miller --- kernel/bpf/core.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c index 4139a0f8b558..54f0e7fcd0e2 100644 --- a/kernel/bpf/core.c +++ b/kernel/bpf/core.c @@ -357,8 +357,8 @@ select_insn: ALU64_MOD_X: if (unlikely(SRC == 0)) return 0; - tmp = DST; - DST = do_div(tmp, SRC); + div64_u64_rem(DST, SRC, &tmp); + DST = tmp; CONT; ALU_MOD_X: if (unlikely(SRC == 0)) @@ -367,8 +367,8 @@ select_insn: DST = do_div(tmp, (u32) SRC); CONT; ALU64_MOD_K: - tmp = DST; - DST = do_div(tmp, IMM); + div64_u64_rem(DST, IMM, &tmp); + DST = tmp; CONT; ALU_MOD_K: tmp = (u32) DST; @@ -377,7 +377,7 @@ select_insn: ALU64_DIV_X: if (unlikely(SRC == 0)) return 0; - do_div(DST, SRC); + DST = div64_u64(DST, SRC); CONT; ALU_DIV_X: if (unlikely(SRC == 0)) @@ -387,7 +387,7 @@ select_insn: DST = (u32) tmp; CONT; ALU64_DIV_K: - do_div(DST, IMM); + DST = div64_u64(DST, IMM); CONT; ALU_DIV_K: tmp = (u32) DST; -- cgit v1.2.3 From cb6ccf09d6b94bec4def1ac5cf4678d12b216474 Mon Sep 17 00:00:00 2001 From: Herbert Xu Date: Tue, 28 Apr 2015 11:43:15 +0800 Subject: route: Use ipv4_mtu instead of raw rt_pmtu The commit 3cdaa5be9e81a914e633a6be7b7d2ef75b528562 ("ipv4: Don't increase PMTU with Datagram Too Big message") broke PMTU in cases where the rt_pmtu value has expired but is smaller than the new PMTU value. This obsolete rt_pmtu then prevents the new PMTU value from being installed. Fixes: 3cdaa5be9e81 ("ipv4: Don't increase PMTU with Datagram Too Big message") Reported-by: Gerd v. Egidy Signed-off-by: Herbert Xu Signed-off-by: David S. Miller --- net/ipv4/route.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/net/ipv4/route.c b/net/ipv4/route.c index a78540f28276..bff62fc87b8e 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -962,10 +962,7 @@ static void __ip_rt_update_pmtu(struct rtable *rt, struct flowi4 *fl4, u32 mtu) if (dst_metric_locked(dst, RTAX_MTU)) return; - if (dst->dev->mtu < mtu) - return; - - if (rt->rt_pmtu && rt->rt_pmtu < mtu) + if (ipv4_mtu(dst) < mtu) return; if (mtu < ip_rt_min_pmtu) -- cgit v1.2.3 From 3f300ff41d89fe9674b8dbab950ba2572639ee8d Mon Sep 17 00:00:00 2001 From: Simon Xiao Date: Tue, 28 Apr 2015 01:05:17 -0700 Subject: hv_netvsc: introduce netif-msg into netvsc module 1. Introduce netif-msg to netvsc to control debug logging output and keep msg_enable in netvsc_device_context so that it is kept persistently. 2. Only call dump_rndis_message() when NETIF_MSG_RX_ERR or above is specified in netvsc module debug param. In non-debug mode, in current code, dump_rndis_message() will not dump anything but it still initialize some local variables and process the switch logic which is unnecessary, especially in high network throughput situation. Signed-off-by: Simon Xiao Reviewed-by: K. Y. Srinivasan Reviewed-by: Haiyang Zhang Signed-off-by: David S. Miller --- drivers/net/hyperv/hyperv_net.h | 12 ++++++++++++ drivers/net/hyperv/netvsc.c | 3 +++ drivers/net/hyperv/netvsc_drv.c | 20 ++++++++++++++------ drivers/net/hyperv/rndis_filter.c | 3 ++- 4 files changed, 31 insertions(+), 7 deletions(-) diff --git a/drivers/net/hyperv/hyperv_net.h b/drivers/net/hyperv/hyperv_net.h index a10b31664709..e55c8f4f55ef 100644 --- a/drivers/net/hyperv/hyperv_net.h +++ b/drivers/net/hyperv/hyperv_net.h @@ -612,6 +612,15 @@ struct multi_send_data { u32 count; /* counter of batched packets */ }; +/* The context of the netvsc device */ +struct net_device_context { + /* point back to our device context */ + struct hv_device *device_ctx; + struct delayed_work dwork; + struct work_struct work; + u32 msg_enable; /* debug level */ +}; + /* Per netvsc device */ struct netvsc_device { struct hv_device *dev; @@ -667,6 +676,9 @@ struct netvsc_device { struct multi_send_data msd[NR_CPUS]; u32 max_pkt; /* max number of pkt in one send, e.g. 8 */ u32 pkt_align; /* alignment bytes, e.g. 8 */ + + /* The net device context */ + struct net_device_context *nd_ctx; }; /* NdisInitialize message */ diff --git a/drivers/net/hyperv/netvsc.c b/drivers/net/hyperv/netvsc.c index 2e8ad0636b46..c651d4d8b747 100644 --- a/drivers/net/hyperv/netvsc.c +++ b/drivers/net/hyperv/netvsc.c @@ -1197,6 +1197,9 @@ int netvsc_device_add(struct hv_device *device, void *additional_info) */ ndev = net_device->ndev; + /* Add netvsc_device context to netvsc_device */ + net_device->nd_ctx = netdev_priv(ndev); + /* Initialize the NetVSC channel extension */ init_completion(&net_device->channel_init_wait); diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c index a3a9d3898a6e..66c4b0c8d108 100644 --- a/drivers/net/hyperv/netvsc_drv.c +++ b/drivers/net/hyperv/netvsc_drv.c @@ -40,18 +40,21 @@ #include "hyperv_net.h" -struct net_device_context { - /* point back to our device context */ - struct hv_device *device_ctx; - struct delayed_work dwork; - struct work_struct work; -}; #define RING_SIZE_MIN 64 static int ring_size = 128; module_param(ring_size, int, S_IRUGO); MODULE_PARM_DESC(ring_size, "Ring buffer size (# of pages)"); +static const u32 default_msg = NETIF_MSG_DRV | NETIF_MSG_PROBE | + NETIF_MSG_LINK | NETIF_MSG_IFUP | + NETIF_MSG_IFDOWN | NETIF_MSG_RX_ERR | + NETIF_MSG_TX_ERR; + +static int debug = -1; +module_param(debug, int, S_IRUGO); +MODULE_PARM_DESC(debug, "Debug level (0=none,...,16=all)"); + static void do_set_multicast(struct work_struct *w) { struct net_device_context *ndevctx = @@ -888,6 +891,11 @@ static int netvsc_probe(struct hv_device *dev, net_device_ctx = netdev_priv(net); net_device_ctx->device_ctx = dev; + net_device_ctx->msg_enable = netif_msg_init(debug, default_msg); + if (netif_msg_probe(net_device_ctx)) + netdev_dbg(net, "netvsc msg_enable: %d\n", + net_device_ctx->msg_enable); + hv_set_drvdata(dev, net); INIT_DELAYED_WORK(&net_device_ctx->dwork, netvsc_link_change); INIT_WORK(&net_device_ctx->work, do_set_multicast); diff --git a/drivers/net/hyperv/rndis_filter.c b/drivers/net/hyperv/rndis_filter.c index 0d92efefd796..9118cea91882 100644 --- a/drivers/net/hyperv/rndis_filter.c +++ b/drivers/net/hyperv/rndis_filter.c @@ -429,7 +429,8 @@ int rndis_filter_receive(struct hv_device *dev, rndis_msg = pkt->data; - dump_rndis_message(dev, rndis_msg); + if (netif_msg_rx_err(net_dev->nd_ctx)) + dump_rndis_message(dev, rndis_msg); switch (rndis_msg->ndis_msg_type) { case RNDIS_MSG_PACKET: -- cgit v1.2.3 From d9b9e860cef1d904832bc6e92e444adfeabe49e9 Mon Sep 17 00:00:00 2001 From: Michal Schmidt Date: Tue, 28 Apr 2015 11:34:21 +0200 Subject: bnx2x: mark LRO as a fixed disabled feature if disable_tpa is set If disable_tpa is set, remove NETIF_F_LRO from hw_features, so ethtool sees it as "off [fixed]". Note that setting the NETIF_F_LRO bit in dev->features in the 'else' branch is not needed, because the bit was already set by bnx2x_init_dev(). Then the check for disable_tpa in in bnx2x_fix_features() becomes unnecessary. Signed-off-by: Michal Schmidt Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 4 ---- drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 2 +- 2 files changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c index 3558a36b1c2d..dcdbd00d0259 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c @@ -4834,10 +4834,6 @@ netdev_features_t bnx2x_fix_features(struct net_device *dev, features &= ~NETIF_F_GRO; } - /* Note: do not disable SW GRO in kernel when HW GRO is off */ - if (bp->disable_tpa) - features &= ~NETIF_F_LRO; - return features; } diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c index b9f85fccb419..2d067c937bd1 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c @@ -12108,10 +12108,10 @@ static int bnx2x_init_bp(struct bnx2x *bp) /* Set TPA flags */ if (bp->disable_tpa) { bp->flags &= ~(TPA_ENABLE_FLAG | GRO_ENABLE_FLAG); + bp->dev->hw_features &= ~NETIF_F_LRO; bp->dev->features &= ~NETIF_F_LRO; } else { bp->flags |= (TPA_ENABLE_FLAG | GRO_ENABLE_FLAG); - bp->dev->features |= NETIF_F_LRO; } if (CHIP_IS_E1(bp)) -- cgit v1.2.3 From 7e6b4d440b0ae9062b84dfb417ea6d51a45dab76 Mon Sep 17 00:00:00 2001 From: Michal Schmidt Date: Tue, 28 Apr 2015 11:34:22 +0200 Subject: bnx2x: merge fp->disable_tpa with fp->mode It is simpler to have the TPA mode as one three-state variable. Signed-off-by: Michal Schmidt Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 2 +- drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 26 ++++++++++++------------ drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h | 2 +- drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 4 ++-- drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c | 2 +- 5 files changed, 18 insertions(+), 18 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h index 355d5fea5be9..2a41604cdd0e 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h @@ -521,6 +521,7 @@ struct bnx2x_fp_txdata { }; enum bnx2x_tpa_mode_t { + TPA_MODE_DISABLED, TPA_MODE_LRO, TPA_MODE_GRO }; @@ -589,7 +590,6 @@ struct bnx2x_fastpath { /* TPA related */ struct bnx2x_agg_info *tpa_info; - u8 disable_tpa; #ifdef BNX2X_STOP_ON_ERROR u64 tpa_queue_used; #endif diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c index dcdbd00d0259..84612b082b86 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c @@ -947,10 +947,10 @@ static int bnx2x_rx_int(struct bnx2x_fastpath *fp, int budget) u16 frag_size, pages; #ifdef BNX2X_STOP_ON_ERROR /* sanity check */ - if (fp->disable_tpa && + if (fp->mode == TPA_MODE_DISABLED && (CQE_TYPE_START(cqe_fp_type) || CQE_TYPE_STOP(cqe_fp_type))) - BNX2X_ERR("START/STOP packet while disable_tpa type %x\n", + BNX2X_ERR("START/STOP packet while TPA disabled, type %x\n", CQE_TYPE(cqe_fp_type)); #endif @@ -1396,7 +1396,7 @@ void bnx2x_init_rx_rings(struct bnx2x *bp) DP(NETIF_MSG_IFUP, "mtu %d rx_buf_size %d\n", bp->dev->mtu, fp->rx_buf_size); - if (!fp->disable_tpa) { + if (fp->mode != TPA_MODE_DISABLED) { /* Fill the per-aggregation pool */ for (i = 0; i < MAX_AGG_QS(bp); i++) { struct bnx2x_agg_info *tpa_info = @@ -1410,7 +1410,7 @@ void bnx2x_init_rx_rings(struct bnx2x *bp) BNX2X_ERR("Failed to allocate TPA skb pool for queue[%d] - disabling TPA on this queue!\n", j); bnx2x_free_tpa_pool(bp, fp, i); - fp->disable_tpa = 1; + fp->mode = TPA_MODE_DISABLED; break; } dma_unmap_addr_set(first_buf, mapping, 0); @@ -1438,7 +1438,7 @@ void bnx2x_init_rx_rings(struct bnx2x *bp) ring_prod); bnx2x_free_tpa_pool(bp, fp, MAX_AGG_QS(bp)); - fp->disable_tpa = 1; + fp->mode = TPA_MODE_DISABLED; ring_prod = 0; break; } @@ -1560,7 +1560,7 @@ static void bnx2x_free_rx_skbs(struct bnx2x *bp) bnx2x_free_rx_bds(fp); - if (!fp->disable_tpa) + if (fp->mode != TPA_MODE_DISABLED) bnx2x_free_tpa_pool(bp, fp, MAX_AGG_QS(bp)); } } @@ -2477,19 +2477,19 @@ static void bnx2x_bz_fp(struct bnx2x *bp, int index) /* set the tpa flag for each queue. The tpa flag determines the queue * minimal size so it must be set prior to queue memory allocation */ - fp->disable_tpa = !(bp->flags & TPA_ENABLE_FLAG || - (bp->flags & GRO_ENABLE_FLAG && - bnx2x_mtu_allows_gro(bp->dev->mtu))); if (bp->flags & TPA_ENABLE_FLAG) fp->mode = TPA_MODE_LRO; - else if (bp->flags & GRO_ENABLE_FLAG) + else if (bp->flags & GRO_ENABLE_FLAG && + bnx2x_mtu_allows_gro(bp->dev->mtu)) fp->mode = TPA_MODE_GRO; + else + fp->mode = TPA_MODE_DISABLED; /* We don't want TPA if it's disabled in bp * or if this is an FCoE L2 ring. */ if (bp->disable_tpa || IS_FCOE_FP(fp)) - fp->disable_tpa = 1; + fp->mode = TPA_MODE_DISABLED; } int bnx2x_load_cnic(struct bnx2x *bp) @@ -2610,7 +2610,7 @@ int bnx2x_nic_load(struct bnx2x *bp, int load_mode) /* * Zero fastpath structures preserving invariants like napi, which are * allocated only once, fp index, max_cos, bp pointer. - * Also set fp->disable_tpa and txdata_ptr. + * Also set fp->mode and txdata_ptr. */ DP(NETIF_MSG_IFUP, "num queues: %d", bp->num_queues); for_each_queue(bp, i) @@ -4545,7 +4545,7 @@ alloc_mem_err: * In these cases we disable the queue * Min size is different for OOO, TPA and non-TPA queues */ - if (ring_size < (fp->disable_tpa ? + if (ring_size < (fp->mode == TPA_MODE_DISABLED ? MIN_RX_SIZE_NONTPA : MIN_RX_SIZE_TPA)) { /* release memory allocated for this queue */ bnx2x_free_fp_mem_at(bp, index); diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h index adcacda7af7b..d7a71758e876 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h @@ -969,7 +969,7 @@ static inline void bnx2x_free_rx_sge_range(struct bnx2x *bp, { int i; - if (fp->disable_tpa) + if (fp->mode == TPA_MODE_DISABLED) return; for (i = 0; i < last; i++) diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c index 2d067c937bd1..64ea5adf72eb 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c @@ -3128,7 +3128,7 @@ static unsigned long bnx2x_get_q_flags(struct bnx2x *bp, __set_bit(BNX2X_Q_FLG_FORCE_DEFAULT_PRI, &flags); } - if (!fp->disable_tpa) { + if (fp->mode != TPA_MODE_DISABLED) { __set_bit(BNX2X_Q_FLG_TPA, &flags); __set_bit(BNX2X_Q_FLG_TPA_IPV6, &flags); if (fp->mode == TPA_MODE_GRO) @@ -3176,7 +3176,7 @@ static void bnx2x_pf_rx_q_prep(struct bnx2x *bp, u16 sge_sz = 0; u16 tpa_agg_size = 0; - if (!fp->disable_tpa) { + if (fp->mode != TPA_MODE_DISABLED) { pause->sge_th_lo = SGE_TH_LO(bp); pause->sge_th_hi = SGE_TH_HI(bp); diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c index 15b2d1647560..06b8c0d8fd3b 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.c @@ -594,7 +594,7 @@ int bnx2x_vfpf_setup_q(struct bnx2x *bp, struct bnx2x_fastpath *fp, bnx2x_vfpf_prep(bp, &req->first_tlv, CHANNEL_TLV_SETUP_Q, sizeof(*req)); /* select tpa mode to request */ - if (!fp->disable_tpa) { + if (fp->mode != TPA_MODE_DISABLED) { flags |= VFPF_QUEUE_FLG_TPA; flags |= VFPF_QUEUE_FLG_TPA_IPV6; if (fp->mode == TPA_MODE_GRO) -- cgit v1.2.3 From f8dcb5e3365a23b620f5744f23f1918b9c38d883 Mon Sep 17 00:00:00 2001 From: Michal Schmidt Date: Tue, 28 Apr 2015 11:34:23 +0200 Subject: bnx2x: remove {TPA,GRO}_ENABLE_FLAG These flags are redundant with dev->features. Remove them. Just make sure to set dev->features ourselves in bnx2x_set_features() before performing the reload of the card. Signed-off-by: Michal Schmidt Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnx2x/bnx2x.h | 2 -- drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c | 39 +++++++++--------------- drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 5 +-- 3 files changed, 15 insertions(+), 31 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h index 2a41604cdd0e..a3b0f7a0c61e 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h @@ -1545,9 +1545,7 @@ struct bnx2x { #define USING_MSIX_FLAG (1 << 5) #define USING_MSI_FLAG (1 << 6) #define DISABLE_MSI_FLAG (1 << 7) -#define TPA_ENABLE_FLAG (1 << 8) #define NO_MCP_FLAG (1 << 9) -#define GRO_ENABLE_FLAG (1 << 10) #define MF_FUNC_DIS (1 << 11) #define OWN_CNIC_IRQ (1 << 12) #define NO_ISCSI_OOO_FLAG (1 << 13) diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c index 84612b082b86..a8bb8f664d3d 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c @@ -2477,9 +2477,9 @@ static void bnx2x_bz_fp(struct bnx2x *bp, int index) /* set the tpa flag for each queue. The tpa flag determines the queue * minimal size so it must be set prior to queue memory allocation */ - if (bp->flags & TPA_ENABLE_FLAG) + if (bp->dev->features & NETIF_F_LRO) fp->mode = TPA_MODE_LRO; - else if (bp->flags & GRO_ENABLE_FLAG && + else if (bp->dev->features & NETIF_F_GRO && bnx2x_mtu_allows_gro(bp->dev->mtu)) fp->mode = TPA_MODE_GRO; else @@ -3249,7 +3249,7 @@ int bnx2x_low_latency_recv(struct napi_struct *napi) if ((bp->state == BNX2X_STATE_CLOSED) || (bp->state == BNX2X_STATE_ERROR) || - (bp->flags & (TPA_ENABLE_FLAG | GRO_ENABLE_FLAG))) + (bp->dev->features & (NETIF_F_LRO | NETIF_F_GRO))) return LL_FLUSH_FAILED; if (!bnx2x_fp_lock_poll(fp)) @@ -4840,19 +4840,9 @@ netdev_features_t bnx2x_fix_features(struct net_device *dev, int bnx2x_set_features(struct net_device *dev, netdev_features_t features) { struct bnx2x *bp = netdev_priv(dev); - u32 flags = bp->flags; - u32 changes; + netdev_features_t changes = features ^ dev->features; bool bnx2x_reload = false; - - if (features & NETIF_F_LRO) - flags |= TPA_ENABLE_FLAG; - else - flags &= ~TPA_ENABLE_FLAG; - - if (features & NETIF_F_GRO) - flags |= GRO_ENABLE_FLAG; - else - flags &= ~GRO_ENABLE_FLAG; + int rc; /* VFs or non SRIOV PFs should be able to change loopback feature */ if (!pci_num_vf(bp->pdev)) { @@ -4869,24 +4859,23 @@ int bnx2x_set_features(struct net_device *dev, netdev_features_t features) } } - changes = flags ^ bp->flags; - /* if GRO is changed while LRO is enabled, don't force a reload */ - if ((changes & GRO_ENABLE_FLAG) && (flags & TPA_ENABLE_FLAG)) - changes &= ~GRO_ENABLE_FLAG; + if ((changes & NETIF_F_GRO) && (features & NETIF_F_LRO)) + changes &= ~NETIF_F_GRO; /* if GRO is changed while HW TPA is off, don't force a reload */ - if ((changes & GRO_ENABLE_FLAG) && bp->disable_tpa) - changes &= ~GRO_ENABLE_FLAG; + if ((changes & NETIF_F_GRO) && bp->disable_tpa) + changes &= ~NETIF_F_GRO; if (changes) bnx2x_reload = true; - bp->flags = flags; - if (bnx2x_reload) { - if (bp->recovery_state == BNX2X_RECOVERY_DONE) - return bnx2x_reload_if_running(dev); + if (bp->recovery_state == BNX2X_RECOVERY_DONE) { + dev->features = features; + rc = bnx2x_reload_if_running(dev); + return rc ? rc : 1; + } /* else: bnx2x_nic_load() will be called at end of recovery */ } diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c index 64ea5adf72eb..60d2b8ebba16 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c @@ -3304,7 +3304,7 @@ static void bnx2x_pf_init(struct bnx2x *bp) /* This flag is relevant for E1x only. * E2 doesn't have a TPA configuration in a function level. */ - flags |= (bp->flags & TPA_ENABLE_FLAG) ? FUNC_FLG_TPA : 0; + flags |= (bp->dev->features & NETIF_F_LRO) ? FUNC_FLG_TPA : 0; func_init.func_flgs = flags; func_init.pf_id = BP_FUNC(bp); @@ -12107,11 +12107,8 @@ static int bnx2x_init_bp(struct bnx2x *bp) /* Set TPA flags */ if (bp->disable_tpa) { - bp->flags &= ~(TPA_ENABLE_FLAG | GRO_ENABLE_FLAG); bp->dev->hw_features &= ~NETIF_F_LRO; bp->dev->features &= ~NETIF_F_LRO; - } else { - bp->flags |= (TPA_ENABLE_FLAG | GRO_ENABLE_FLAG); } if (CHIP_IS_E1(bp)) -- cgit v1.2.3 From 18fe369a5a4f175ee652ba8b7d8d7e49f140e281 Mon Sep 17 00:00:00 2001 From: Antonio Ospite Date: Tue, 28 Apr 2015 13:11:27 +0200 Subject: trivial: net: atl1e: atl1e_hw.h: fix 0x0x prefix Fix the 0x0x prefix in an integer constant. Signed-off-by: Antonio Ospite Cc: Jay Cliburn Cc: Chris Snook Cc: netdev@vger.kernel.org Signed-off-by: David S. Miller --- drivers/net/ethernet/atheros/atl1e/atl1e_hw.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/atheros/atl1e/atl1e_hw.h b/drivers/net/ethernet/atheros/atl1e/atl1e_hw.h index 74df16aef793..88a6271de5bc 100644 --- a/drivers/net/ethernet/atheros/atl1e/atl1e_hw.h +++ b/drivers/net/ethernet/atheros/atl1e/atl1e_hw.h @@ -129,7 +129,7 @@ s32 atl1e_restart_autoneg(struct atl1e_hw *hw); #define TWSI_CTRL_LD_SLV_ADDR_SHIFT 8 #define TWSI_CTRL_SW_LDSTART 0x800 #define TWSI_CTRL_HW_LDSTART 0x1000 -#define TWSI_CTRL_SMB_SLV_ADDR_MASK 0x0x7F +#define TWSI_CTRL_SMB_SLV_ADDR_MASK 0x7F #define TWSI_CTRL_SMB_SLV_ADDR_SHIFT 15 #define TWSI_CTRL_LD_EXIST 0x400000 #define TWSI_CTRL_READ_FREQ_SEL_MASK 0x3 -- cgit v1.2.3 From 165996bd351f0e711feff92479c0796098afaa38 Mon Sep 17 00:00:00 2001 From: Antonio Ospite Date: Tue, 28 Apr 2015 13:11:29 +0200 Subject: trivial: net: systemport: bcmsysport.h: fix 0x0x prefix Fix the 0x0x prefix in an integer constant. In this case, while at it, also fix a typo (s/unitcast/unicast/). Signed-off-by: Antonio Ospite Cc: Florian Fainelli Cc: netdev@vger.kernel.org Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bcmsysport.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/broadcom/bcmsysport.h b/drivers/net/ethernet/broadcom/bcmsysport.h index 7e3d87a88c76..e2c043eabbf3 100644 --- a/drivers/net/ethernet/broadcom/bcmsysport.h +++ b/drivers/net/ethernet/broadcom/bcmsysport.h @@ -543,7 +543,7 @@ struct bcm_sysport_tx_counters { u32 jbr; /* RO # of xmited jabber count*/ u32 bytes; /* RO # of xmited byte count */ u32 pok; /* RO # of xmited good pkt */ - u32 uc; /* RO (0x0x4f0)# of xmited unitcast pkt */ + u32 uc; /* RO (0x4f0) # of xmited unicast pkt */ }; struct bcm_sysport_mib { -- cgit v1.2.3 From 2b70fe5aba0dd00d81173243c7ab04c66aeb67d8 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Tue, 28 Apr 2015 13:33:21 +0200 Subject: net: sched: act_connmark: don't zap skb->nfct This action is meant to be passive, i.e. we should not alter skb->nfct: If nfct is present just leave it alone. Compile tested only. Cc: Jamal Hadi Salim Signed-off-by: Florian Westphal Acked-by: Pablo Neira Ayuso Signed-off-by: David S. Miller --- net/sched/act_connmark.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/net/sched/act_connmark.c b/net/sched/act_connmark.c index 8e472518f9f6..295d14bd6c67 100644 --- a/net/sched/act_connmark.c +++ b/net/sched/act_connmark.c @@ -63,7 +63,6 @@ static int tcf_connmark(struct sk_buff *skb, const struct tc_action *a, skb->mark = c->mark; /* using overlimits stats to count how many packets marked */ ca->tcf_qstats.overlimits++; - nf_ct_put(c); goto out; } @@ -82,7 +81,6 @@ static int tcf_connmark(struct sk_buff *skb, const struct tc_action *a, nf_ct_put(c); out: - skb->nfct = NULL; spin_unlock(&ca->tcf_lock); return ca->tcf_action; } -- cgit v1.2.3 From 821996795973fd52703c35811a03db9fec1ac141 Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Tue, 28 Apr 2015 18:33:48 +0200 Subject: bridge/mdb: remove wrong use of NLM_F_MULTI NLM_F_MULTI must be used only when a NLMSG_DONE message is sent. In fact, it is sent only at the end of a dump. Libraries like libnl will wait forever for NLMSG_DONE. Fixes: 37a393bc4932 ("bridge: notify mdb changes via netlink") CC: Cong Wang CC: Stephen Hemminger CC: bridge@lists.linux-foundation.org Signed-off-by: Nicolas Dichtel Signed-off-by: David S. Miller --- net/bridge/br_mdb.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/bridge/br_mdb.c b/net/bridge/br_mdb.c index 409608960899..e29ad70b3000 100644 --- a/net/bridge/br_mdb.c +++ b/net/bridge/br_mdb.c @@ -170,7 +170,7 @@ static int nlmsg_populate_mdb_fill(struct sk_buff *skb, struct br_port_msg *bpm; struct nlattr *nest, *nest2; - nlh = nlmsg_put(skb, pid, seq, type, sizeof(*bpm), NLM_F_MULTI); + nlh = nlmsg_put(skb, pid, seq, type, sizeof(*bpm), 0); if (!nlh) return -EMSGSIZE; -- cgit v1.2.3 From 46c264daaaa569e24f8aba877d0fd8167c42a9a4 Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Tue, 28 Apr 2015 18:33:49 +0200 Subject: bridge/nl: remove wrong use of NLM_F_MULTI NLM_F_MULTI must be used only when a NLMSG_DONE message is sent. In fact, it is sent only at the end of a dump. Libraries like libnl will wait forever for NLMSG_DONE. Fixes: e5a55a898720 ("net: create generic bridge ops") Fixes: 815cccbf10b2 ("ixgbe: add setlink, getlink support to ixgbe and ixgbevf") CC: John Fastabend CC: Sathya Perla CC: Subbu Seetharaman CC: Ajit Khaparde CC: Jeff Kirsher CC: intel-wired-lan@lists.osuosl.org CC: Jiri Pirko CC: Scott Feldman CC: Stephen Hemminger CC: bridge@lists.linux-foundation.org Signed-off-by: Nicolas Dichtel Signed-off-by: David S. Miller --- drivers/net/ethernet/emulex/benet/be_main.c | 5 +++-- drivers/net/ethernet/intel/i40e/i40e_main.c | 7 ++++--- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 4 ++-- drivers/net/ethernet/rocker/rocker.c | 5 +++-- include/linux/netdevice.h | 6 ++++-- include/linux/rtnetlink.h | 2 +- net/bridge/br_netlink.c | 4 ++-- net/bridge/br_private.h | 2 +- net/core/rtnetlink.c | 12 +++++++----- 9 files changed, 27 insertions(+), 20 deletions(-) diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c index fb0bc3c3620e..a6dcbf850c1f 100644 --- a/drivers/net/ethernet/emulex/benet/be_main.c +++ b/drivers/net/ethernet/emulex/benet/be_main.c @@ -4846,7 +4846,8 @@ err: } static int be_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, - struct net_device *dev, u32 filter_mask) + struct net_device *dev, u32 filter_mask, + int nlflags) { struct be_adapter *adapter = netdev_priv(dev); int status = 0; @@ -4868,7 +4869,7 @@ static int be_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, return ndo_dflt_bridge_getlink(skb, pid, seq, dev, hsw_mode == PORT_FWD_TYPE_VEPA ? BRIDGE_MODE_VEPA : BRIDGE_MODE_VEB, - 0, 0); + 0, 0, nlflags); } #ifdef CONFIG_BE2NET_VXLAN diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index 24481cd7e59a..a54c14491e3b 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -8053,10 +8053,10 @@ static int i40e_ndo_bridge_setlink(struct net_device *dev, #ifdef HAVE_BRIDGE_FILTER static int i40e_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, - u32 __always_unused filter_mask) + u32 __always_unused filter_mask, int nlflags) #else static int i40e_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, - struct net_device *dev) + struct net_device *dev, int nlflags) #endif /* HAVE_BRIDGE_FILTER */ { struct i40e_netdev_priv *np = netdev_priv(dev); @@ -8078,7 +8078,8 @@ static int i40e_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, if (!veb) return 0; - return ndo_dflt_bridge_getlink(skb, pid, seq, dev, veb->bridge_mode); + return ndo_dflt_bridge_getlink(skb, pid, seq, dev, veb->bridge_mode, + nlflags); } #endif /* HAVE_BRIDGE_ATTRIBS */ diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index d3f4b0ceb3f7..5be12a00e1f4 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -8044,7 +8044,7 @@ static int ixgbe_ndo_bridge_setlink(struct net_device *dev, static int ixgbe_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, - u32 filter_mask) + u32 filter_mask, int nlflags) { struct ixgbe_adapter *adapter = netdev_priv(dev); @@ -8052,7 +8052,7 @@ static int ixgbe_ndo_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, return 0; return ndo_dflt_bridge_getlink(skb, pid, seq, dev, - adapter->bridge_mode, 0, 0); + adapter->bridge_mode, 0, 0, nlflags); } static void *ixgbe_fwd_add(struct net_device *pdev, struct net_device *vdev) diff --git a/drivers/net/ethernet/rocker/rocker.c b/drivers/net/ethernet/rocker/rocker.c index a570a60533be..ec251531bd9f 100644 --- a/drivers/net/ethernet/rocker/rocker.c +++ b/drivers/net/ethernet/rocker/rocker.c @@ -4176,14 +4176,15 @@ static int rocker_port_bridge_setlink(struct net_device *dev, static int rocker_port_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, - u32 filter_mask) + u32 filter_mask, int nlflags) { struct rocker_port *rocker_port = netdev_priv(dev); u16 mode = BRIDGE_MODE_UNDEF; u32 mask = BR_LEARNING | BR_LEARNING_SYNC; return ndo_dflt_bridge_getlink(skb, pid, seq, dev, mode, - rocker_port->brport_flags, mask); + rocker_port->brport_flags, mask, + nlflags); } static int rocker_port_get_phys_port_name(struct net_device *dev, diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index dbad4d728b4b..1899c74a7127 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -977,7 +977,8 @@ typedef u16 (*select_queue_fallback_t)(struct net_device *dev, * int (*ndo_bridge_setlink)(struct net_device *dev, struct nlmsghdr *nlh, * u16 flags) * int (*ndo_bridge_getlink)(struct sk_buff *skb, u32 pid, u32 seq, - * struct net_device *dev, u32 filter_mask) + * struct net_device *dev, u32 filter_mask, + * int nlflags) * int (*ndo_bridge_dellink)(struct net_device *dev, struct nlmsghdr *nlh, * u16 flags); * @@ -1173,7 +1174,8 @@ struct net_device_ops { int (*ndo_bridge_getlink)(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, - u32 filter_mask); + u32 filter_mask, + int nlflags); int (*ndo_bridge_dellink)(struct net_device *dev, struct nlmsghdr *nlh, u16 flags); diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h index 2da5d1081ad9..7b8e260c4a27 100644 --- a/include/linux/rtnetlink.h +++ b/include/linux/rtnetlink.h @@ -122,5 +122,5 @@ extern int ndo_dflt_fdb_del(struct ndmsg *ndm, extern int ndo_dflt_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, u16 mode, - u32 flags, u32 mask); + u32 flags, u32 mask, int nlflags); #endif /* __LINUX_RTNETLINK_H */ diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c index 0e4ddb81610d..4b5c236998ff 100644 --- a/net/bridge/br_netlink.c +++ b/net/bridge/br_netlink.c @@ -394,7 +394,7 @@ errout: * Dump information about all ports, in response to GETLINK */ int br_getlink(struct sk_buff *skb, u32 pid, u32 seq, - struct net_device *dev, u32 filter_mask) + struct net_device *dev, u32 filter_mask, int nlflags) { struct net_bridge_port *port = br_port_get_rtnl(dev); @@ -402,7 +402,7 @@ int br_getlink(struct sk_buff *skb, u32 pid, u32 seq, !(filter_mask & RTEXT_FILTER_BRVLAN_COMPRESSED)) return 0; - return br_fill_ifinfo(skb, port, pid, seq, RTM_NEWLINK, NLM_F_MULTI, + return br_fill_ifinfo(skb, port, pid, seq, RTM_NEWLINK, nlflags, filter_mask, dev); } diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 6ca0251cb478..3362c29400f1 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -828,7 +828,7 @@ void br_ifinfo_notify(int event, struct net_bridge_port *port); int br_setlink(struct net_device *dev, struct nlmsghdr *nlmsg, u16 flags); int br_dellink(struct net_device *dev, struct nlmsghdr *nlmsg, u16 flags); int br_getlink(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, - u32 filter_mask); + u32 filter_mask, int nlflags); #ifdef CONFIG_SYSFS /* br_sysfs_if.c */ diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 358d52a38533..666e0928ba40 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -2854,7 +2854,7 @@ static int brport_nla_put_flag(struct sk_buff *skb, u32 flags, u32 mask, int ndo_dflt_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, struct net_device *dev, u16 mode, - u32 flags, u32 mask) + u32 flags, u32 mask, int nlflags) { struct nlmsghdr *nlh; struct ifinfomsg *ifm; @@ -2863,7 +2863,7 @@ int ndo_dflt_bridge_getlink(struct sk_buff *skb, u32 pid, u32 seq, u8 operstate = netif_running(dev) ? dev->operstate : IF_OPER_DOWN; struct net_device *br_dev = netdev_master_upper_dev_get(dev); - nlh = nlmsg_put(skb, pid, seq, RTM_NEWLINK, sizeof(*ifm), NLM_F_MULTI); + nlh = nlmsg_put(skb, pid, seq, RTM_NEWLINK, sizeof(*ifm), nlflags); if (nlh == NULL) return -EMSGSIZE; @@ -2969,7 +2969,8 @@ static int rtnl_bridge_getlink(struct sk_buff *skb, struct netlink_callback *cb) if (br_dev && br_dev->netdev_ops->ndo_bridge_getlink) { if (idx >= cb->args[0] && br_dev->netdev_ops->ndo_bridge_getlink( - skb, portid, seq, dev, filter_mask) < 0) + skb, portid, seq, dev, filter_mask, + NLM_F_MULTI) < 0) break; idx++; } @@ -2977,7 +2978,8 @@ static int rtnl_bridge_getlink(struct sk_buff *skb, struct netlink_callback *cb) if (ops->ndo_bridge_getlink) { if (idx >= cb->args[0] && ops->ndo_bridge_getlink(skb, portid, seq, dev, - filter_mask) < 0) + filter_mask, + NLM_F_MULTI) < 0) break; idx++; } @@ -3018,7 +3020,7 @@ static int rtnl_bridge_notify(struct net_device *dev) goto errout; } - err = dev->netdev_ops->ndo_bridge_getlink(skb, 0, 0, dev, 0); + err = dev->netdev_ops->ndo_bridge_getlink(skb, 0, 0, dev, 0, 0); if (err < 0) goto errout; -- cgit v1.2.3 From f2f67390a4b961dae83733732e96e1a394a53c4e Mon Sep 17 00:00:00 2001 From: Nicolas Dichtel Date: Tue, 28 Apr 2015 18:33:50 +0200 Subject: tipc: remove wrong use of NLM_F_MULTI NLM_F_MULTI must be used only when a NLMSG_DONE message is sent. In fact, it is sent only at the end of a dump. Libraries like libnl will wait forever for NLMSG_DONE. Fixes: 35b9dd7607f0 ("tipc: add bearer get/dump to new netlink api") Fixes: 7be57fc69184 ("tipc: add link get/dump to new netlink api") Fixes: 46f15c6794fb ("tipc: add media get/dump to new netlink api") CC: Richard Alpe CC: Jon Maloy CC: Ying Xue CC: tipc-discussion@lists.sourceforge.net Signed-off-by: Nicolas Dichtel Signed-off-by: David S. Miller --- net/tipc/bearer.c | 17 +++++++++-------- net/tipc/link.c | 8 ++++---- 2 files changed, 13 insertions(+), 12 deletions(-) diff --git a/net/tipc/bearer.c b/net/tipc/bearer.c index 3613e72e858e..70e3dacbf84a 100644 --- a/net/tipc/bearer.c +++ b/net/tipc/bearer.c @@ -591,14 +591,14 @@ void tipc_bearer_stop(struct net *net) /* Caller should hold rtnl_lock to protect the bearer */ static int __tipc_nl_add_bearer(struct tipc_nl_msg *msg, - struct tipc_bearer *bearer) + struct tipc_bearer *bearer, int nlflags) { void *hdr; struct nlattr *attrs; struct nlattr *prop; hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family, - NLM_F_MULTI, TIPC_NL_BEARER_GET); + nlflags, TIPC_NL_BEARER_GET); if (!hdr) return -EMSGSIZE; @@ -657,7 +657,7 @@ int tipc_nl_bearer_dump(struct sk_buff *skb, struct netlink_callback *cb) if (!bearer) continue; - err = __tipc_nl_add_bearer(&msg, bearer); + err = __tipc_nl_add_bearer(&msg, bearer, NLM_F_MULTI); if (err) break; } @@ -705,7 +705,7 @@ int tipc_nl_bearer_get(struct sk_buff *skb, struct genl_info *info) goto err_out; } - err = __tipc_nl_add_bearer(&msg, bearer); + err = __tipc_nl_add_bearer(&msg, bearer, 0); if (err) goto err_out; rtnl_unlock(); @@ -857,14 +857,14 @@ int tipc_nl_bearer_set(struct sk_buff *skb, struct genl_info *info) } static int __tipc_nl_add_media(struct tipc_nl_msg *msg, - struct tipc_media *media) + struct tipc_media *media, int nlflags) { void *hdr; struct nlattr *attrs; struct nlattr *prop; hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family, - NLM_F_MULTI, TIPC_NL_MEDIA_GET); + nlflags, TIPC_NL_MEDIA_GET); if (!hdr) return -EMSGSIZE; @@ -916,7 +916,8 @@ int tipc_nl_media_dump(struct sk_buff *skb, struct netlink_callback *cb) rtnl_lock(); for (; media_info_array[i] != NULL; i++) { - err = __tipc_nl_add_media(&msg, media_info_array[i]); + err = __tipc_nl_add_media(&msg, media_info_array[i], + NLM_F_MULTI); if (err) break; } @@ -963,7 +964,7 @@ int tipc_nl_media_get(struct sk_buff *skb, struct genl_info *info) goto err_out; } - err = __tipc_nl_add_media(&msg, media); + err = __tipc_nl_add_media(&msg, media, 0); if (err) goto err_out; rtnl_unlock(); diff --git a/net/tipc/link.c b/net/tipc/link.c index 57be6e6aff99..a11cc5e5e0ae 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -2013,7 +2013,7 @@ msg_full: /* Caller should hold appropriate locks to protect the link */ static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg, - struct tipc_link *link) + struct tipc_link *link, int nlflags) { int err; void *hdr; @@ -2022,7 +2022,7 @@ static int __tipc_nl_add_link(struct net *net, struct tipc_nl_msg *msg, struct tipc_net *tn = net_generic(net, tipc_net_id); hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family, - NLM_F_MULTI, TIPC_NL_LINK_GET); + nlflags, TIPC_NL_LINK_GET); if (!hdr) return -EMSGSIZE; @@ -2095,7 +2095,7 @@ static int __tipc_nl_add_node_links(struct net *net, struct tipc_nl_msg *msg, if (!node->links[i]) continue; - err = __tipc_nl_add_link(net, msg, node->links[i]); + err = __tipc_nl_add_link(net, msg, node->links[i], NLM_F_MULTI); if (err) return err; } @@ -2209,7 +2209,7 @@ int tipc_nl_link_get(struct sk_buff *skb, struct genl_info *info) goto err_out; } - err = __tipc_nl_add_link(net, &msg, link); + err = __tipc_nl_add_link(net, &msg, link, 0); if (err) goto err_out; -- cgit v1.2.3 From 0d699f28ee5d0641470a603ab5904e463cb1532a Mon Sep 17 00:00:00 2001 From: Jon Paul Maloy Date: Tue, 28 Apr 2015 16:59:04 -0400 Subject: tipc: fix problem with parallel link synchronization mechanism Currently, we try to accumulate arrived packets in the links's 'deferred' queue during the parallel link syncronization phase. This entails two problems: - With an unlucky combination of arriving packets the algorithm may go into a lockstep with the out-of-sequence handling function, where the synch mechanism is adding a packet to the deferred queue, while the out-of-sequence handling is retrieving it again, thus ending up in a loop inside the node_lock scope. - Even if this is avoided, the link will very often send out unnecessary protocol messages, in the worst case leading to redundant retransmissions. We fix this by just dropping arriving packets on the upcoming link during the synchronization phase, thus relying on the retransmission protocol to resolve the situation once the two links have arrived to a synchronized state. Reviewed-by: Erik Hugne Reviewed-by: Ying Xue Signed-off-by: Jon Maloy Signed-off-by: David S. Miller --- net/tipc/link.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/net/tipc/link.c b/net/tipc/link.c index a11cc5e5e0ae..43a515dc97b0 100644 --- a/net/tipc/link.c +++ b/net/tipc/link.c @@ -1145,11 +1145,8 @@ void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b_ptr) } /* Synchronize with parallel link if applicable */ if (unlikely((l_ptr->flags & LINK_SYNCHING) && !msg_dup(msg))) { - link_handle_out_of_seq_msg(l_ptr, skb); - if (link_synch(l_ptr)) - link_retrieve_defq(l_ptr, &head); - skb = NULL; - goto unlock; + if (!link_synch(l_ptr)) + goto unlock; } l_ptr->next_in_no++; if (unlikely(!skb_queue_empty(&l_ptr->deferdq))) -- cgit v1.2.3 From 42eab005a5dd5d7ea2b0328aecc4d6cc0c23c9c2 Mon Sep 17 00:00:00 2001 From: Benjamin Poirier Date: Tue, 28 Apr 2015 14:49:29 -0700 Subject: mlx4: Fix tx ring affinity_mask creation By default, the number of tx queues is limited by the number of online cpus in mlx4_en_get_profile(). However, this limit no longer holds after the ethtool .set_channels method has been called. In that situation, the driver may access invalid bits of certain cpumask variables when queue_index >= nr_cpu_ids. Signed-off-by: Benjamin Poirier Acked-by: Ido Shamay Fixes: d03a68f ("net/mlx4_en: Configure the XPS queue mapping on driver load") Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_tx.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/en_tx.c b/drivers/net/ethernet/mellanox/mlx4/en_tx.c index 1783705273d8..f7bf312fb443 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_tx.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_tx.c @@ -143,8 +143,10 @@ int mlx4_en_create_tx_ring(struct mlx4_en_priv *priv, ring->hwtstamp_tx_type = priv->hwtstamp_config.tx_type; ring->queue_index = queue_index; - if (queue_index < priv->num_tx_rings_p_up && cpu_online(queue_index)) - cpumask_set_cpu(queue_index, &ring->affinity_mask); + if (queue_index < priv->num_tx_rings_p_up) + cpumask_set_cpu_local_first(queue_index, + priv->mdev->dev->numa_node, + &ring->affinity_mask); *pring = ring; return 0; @@ -213,7 +215,7 @@ int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv, err = mlx4_qp_to_ready(mdev->dev, &ring->wqres.mtt, &ring->context, &ring->qp, &ring->qp_state); - if (!user_prio && cpu_online(ring->queue_index)) + if (!cpumask_empty(&ring->affinity_mask)) netif_set_xps_queue(priv->dev, &ring->affinity_mask, ring->queue_index); -- cgit v1.2.3 From 48734994ac268eb391a66dca4cde7d08a08aba08 Mon Sep 17 00:00:00 2001 From: Vlastimil Setka Date: Wed, 29 Apr 2015 00:17:11 +0200 Subject: altera_tse: Correct rx packet length Altera TSE MAC rx DMA transfer starts with the 2 additional bytes for IP payload alignment. This patch fixes tse_rx() function loop which reads DMA rx status and extracts packet length from it. Status signalises a whole DMA transfer length, which is 2 bytes longer than the packet itself. Signed-off-by: Vlastimil Setka Signed-off-by: David S. Miller --- drivers/net/ethernet/altera/altera_tse_main.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/net/ethernet/altera/altera_tse_main.c b/drivers/net/ethernet/altera/altera_tse_main.c index 0533c051a3e5..da48e66377b5 100644 --- a/drivers/net/ethernet/altera/altera_tse_main.c +++ b/drivers/net/ethernet/altera/altera_tse_main.c @@ -391,6 +391,12 @@ static int tse_rx(struct altera_tse_private *priv, int limit) "RCV pktstatus %08X pktlength %08X\n", pktstatus, pktlength); + /* DMA trasfer from TSE starts with 2 aditional bytes for + * IP payload alignment. Status returned by get_rx_status() + * contains DMA transfer length. Packet is 2 bytes shorter. + */ + pktlength -= 2; + count++; next_entry = (++priv->rx_cons) % priv->rx_ring_size; -- cgit v1.2.3 From b56fc3c536541c8081cd5f1f1d101a16c002a365 Mon Sep 17 00:00:00 2001 From: KY Srinivasan Date: Tue, 28 Apr 2015 17:59:48 -0700 Subject: hv_netvsc: Fix a bug in netvsc_start_xmit() Commit b08cc79155fc26d0d112b1470d1ece5034651a4b eliminated memory allocation in the packet send path: "hv_netvsc: Eliminate memory allocation in the packet send path The network protocol used to communicate with the host is the remote ndis (rndis) protocol. We need to decorate each outgoing packet with a rndis header and additional rndis state (rndis per-packet state). To manage this state, we currently allocate memory in the transmit path. Eliminate this allocation by requesting additional head room in the skb." This commit introduced a bug since it did not account for the case if the skb was cloned. Fix this bug. Signed-off-by: K. Y. Srinivasan Tested-by: Dexuan Cui Signed-off-by: David S. Miller --- drivers/net/hyperv/hyperv_net.h | 1 - drivers/net/hyperv/netvsc.c | 5 ----- drivers/net/hyperv/netvsc_drv.c | 27 +++++++-------------------- 3 files changed, 7 insertions(+), 26 deletions(-) diff --git a/drivers/net/hyperv/hyperv_net.h b/drivers/net/hyperv/hyperv_net.h index e55c8f4f55ef..41071d32bc8e 100644 --- a/drivers/net/hyperv/hyperv_net.h +++ b/drivers/net/hyperv/hyperv_net.h @@ -128,7 +128,6 @@ struct ndis_tcp_ip_checksum_info; struct hv_netvsc_packet { /* Bookkeeping stuff */ u32 status; - bool part_of_skb; bool is_data_pkt; bool xmit_more; /* from skb */ diff --git a/drivers/net/hyperv/netvsc.c b/drivers/net/hyperv/netvsc.c index c651d4d8b747..2d9ef533cc48 100644 --- a/drivers/net/hyperv/netvsc.c +++ b/drivers/net/hyperv/netvsc.c @@ -889,11 +889,6 @@ int netvsc_send(struct hv_device *device, } else { packet->page_buf_cnt = 0; packet->total_data_buflen += msd_len; - if (!packet->part_of_skb) { - skb = (struct sk_buff *)(unsigned long)packet-> - send_completion_tid; - packet->send_completion_tid = 0; - } } if (msdp->pkt) diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c index 66c4b0c8d108..5993c7e2d723 100644 --- a/drivers/net/hyperv/netvsc_drv.c +++ b/drivers/net/hyperv/netvsc_drv.c @@ -238,9 +238,6 @@ void netvsc_xmit_completion(void *context) struct sk_buff *skb = (struct sk_buff *) (unsigned long)packet->send_completion_tid; - if (!packet->part_of_skb) - kfree(packet); - if (skb) dev_kfree_skb_any(skb); } @@ -392,7 +389,6 @@ static int netvsc_start_xmit(struct sk_buff *skb, struct net_device *net) u32 net_trans_info; u32 hash; u32 skb_length; - u32 head_room; u32 pkt_sz; struct hv_page_buffer page_buf[MAX_PAGE_BUFFER_COUNT]; @@ -405,7 +401,6 @@ static int netvsc_start_xmit(struct sk_buff *skb, struct net_device *net) check_size: skb_length = skb->len; - head_room = skb_headroom(skb); num_data_pgs = netvsc_get_slots(skb) + 2; if (num_data_pgs > MAX_PAGE_BUFFER_COUNT && linear) { net_alert_ratelimited("packet too big: %u pages (%u bytes)\n", @@ -424,20 +419,14 @@ check_size: pkt_sz = sizeof(struct hv_netvsc_packet) + RNDIS_AND_PPI_SIZE; - if (head_room < pkt_sz) { - packet = kmalloc(pkt_sz, GFP_ATOMIC); - if (!packet) { - /* out of memory, drop packet */ - netdev_err(net, "unable to alloc hv_netvsc_packet\n"); - ret = -ENOMEM; - goto drop; - } - packet->part_of_skb = false; - } else { - /* Use the headroom for building up the packet */ - packet = (struct hv_netvsc_packet *)skb->head; - packet->part_of_skb = true; + ret = skb_cow_head(skb, pkt_sz); + if (ret) { + netdev_err(net, "unable to alloc hv_netvsc_packet\n"); + ret = -ENOMEM; + goto drop; } + /* Use the headroom for building up the packet */ + packet = (struct hv_netvsc_packet *)skb->head; packet->status = 0; packet->xmit_more = skb->xmit_more; @@ -594,8 +583,6 @@ drop: net->stats.tx_bytes += skb_length; net->stats.tx_packets++; } else { - if (packet && !packet->part_of_skb) - kfree(packet); if (ret != -EAGAIN) { dev_kfree_skb_any(skb); net->stats.tx_dropped++; -- cgit v1.2.3 From d4c216c54197d741ed8b7ca54f13645dfb3eacde Mon Sep 17 00:00:00 2001 From: Punnaiah Choudary Kalluri Date: Wed, 29 Apr 2015 08:34:46 +0530 Subject: net: macb: Fix race condition in driver when Rx frame is dropped Under heavy Rx load, observed that the Hw is updating the USED bit and it is not updating the received frame status to the BD control field. This could be lack of resources for processing the BDs at high data rates. Driver drops the frame associated with this BD but not clearing the USED bit. So, this is causing hang condition as Hw expects USED bit to be cleared for this BD. Signed-off-by: Punnaiah Choudary Kalluri Signed-off-by: David S. Miller --- drivers/net/ethernet/cadence/macb.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/net/ethernet/cadence/macb.c b/drivers/net/ethernet/cadence/macb.c index 665c29098e3c..4104d49f005d 100644 --- a/drivers/net/ethernet/cadence/macb.c +++ b/drivers/net/ethernet/cadence/macb.c @@ -707,6 +707,9 @@ static void gem_rx_refill(struct macb *bp) /* properly align Ethernet header */ skb_reserve(skb, NET_IP_ALIGN); + } else { + bp->rx_ring[entry].addr &= ~MACB_BIT(RX_USED); + bp->rx_ring[entry].ctrl = 0; } } -- cgit v1.2.3 From 50d4964f1d0411d82cca593f2664bfab7f82dbbf Mon Sep 17 00:00:00 2001 From: Guenter Roeck Date: Wed, 29 Apr 2015 10:56:15 -0700 Subject: net: dsa: Fix scope of eeprom-length property eeprom-length is a switch property, not a dsa property, and thus needs to be attached to the switch node, not to the dsa node. Reported-by: Andrew Lunn Fixes: 6793abb4e849 ("net: dsa: Add support for switch EEPROM access") Signed-off-by: Guenter Roeck Acked-by: Andrew Lunn Signed-off-by: David S. Miller --- net/dsa/dsa.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/dsa/dsa.c b/net/dsa/dsa.c index 079a224471e7..e6f6cc3a1bcf 100644 --- a/net/dsa/dsa.c +++ b/net/dsa/dsa.c @@ -633,7 +633,7 @@ static int dsa_of_probe(struct device *dev) if (cd->sw_addr > PHY_MAX_ADDR) continue; - if (!of_property_read_u32(np, "eeprom-length", &eeprom_len)) + if (!of_property_read_u32(child, "eeprom-length", &eeprom_len)) cd->eeprom_len = eeprom_len; for_each_available_child_of_node(child, port) { -- cgit v1.2.3 From e913fb279c564f2af05658b3f01041757d2e9303 Mon Sep 17 00:00:00 2001 From: Pai Date: Wed, 29 Apr 2015 14:24:23 -0400 Subject: net: Fix Kernel Panic in bonding driver debugfs file: rlb_hash_table This patch fixes a Kernel Panic in bonding driver debugfs file: rlb_hash_table. $> modprobe bonding mode=6 $> cat /sys/kernel/debug/bonding/bond0/rlb_hash_table This will crash the kernel. The struct alb_bond_info is initialized only when the bonding interface is initialized (ip link set bond0 up) and not at the time it is allocated. If we try to read the table before that, it'll result in a kernel panic. The patch applies against both net and net-next Signed-off-by: Vishwanath Pai Signed-off-by: Andy Gospodarek Signed-off-by: David S. Miller --- drivers/net/bonding/bond_main.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 3a10551d64cf..d5fe5d5f490f 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -4544,6 +4544,8 @@ unsigned int bond_get_num_tx_queues(void) int bond_create(struct net *net, const char *name) { struct net_device *bond_dev; + struct bonding *bond; + struct alb_bond_info *bond_info; int res; rtnl_lock(); @@ -4557,6 +4559,14 @@ int bond_create(struct net *net, const char *name) return -ENOMEM; } + /* + * Initialize rx_hashtbl_used_head to RLB_NULL_INDEX. + * It is set to 0 by default which is wrong. + */ + bond = netdev_priv(bond_dev); + bond_info = &(BOND_ALB_INFO(bond)); + bond_info->rx_hashtbl_used_head = RLB_NULL_INDEX; + dev_net_set(bond_dev, net); bond_dev->rtnl_link_ops = &bond_link_ops; -- cgit v1.2.3 From 12a8541d5c82d17d42424d46ec36929cfef06a10 Mon Sep 17 00:00:00 2001 From: Yuval Mintz Date: Wed, 29 Apr 2015 08:09:49 +0300 Subject: bnx2x: Delay during kdump load In a kdump environment interfaces might be re-loaded without a proper unload sequence in the previous running kernel. bnx2x management FW and driver maintains a `pulse' that notifies the FW that the driver is still up and running. Driver load on the kdump kernel should be performed only after the pulse has been out-of-sync long enough for the management FW to identify that the driver has crashed, on which point it will perform some necessary cleanup of the HW. In today's distros kdump loading is quite fast, sometimes too fast for our FW to get out-of-sync. This patch delays the bnx2x's probe during kdump to allow a proper re-load on the kdump kernel. Signed-off-by: Yuval Mintz Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c index 60d2b8ebba16..556dcc162a62 100644 --- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c +++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c @@ -13368,6 +13368,12 @@ static int bnx2x_init_one(struct pci_dev *pdev, bool is_vf; int cnic_cnt; + /* Management FW 'remembers' living interfaces. Allow it some time + * to forget previously living interfaces, allowing a proper re-load. + */ + if (is_kdump_kernel()) + msleep(5000); + /* An estimated maximum supported CoS number according to the chip * version. * We will try to roughly estimate the maximum number of CoSes this chip -- cgit v1.2.3 From 7f0b8a56c978b0a3315ac84c6cbb065413afb8e9 Mon Sep 17 00:00:00 2001 From: Hariprasad Shenai Date: Wed, 29 Apr 2015 17:19:05 +0530 Subject: cxgb4: Fix MC1 memory offset calculation Commit 6559a7e8296002b4 ("cxgb4: Cleanup macros so they follow the same style and look consistent") introduced a regression where reading MC1 memory in adapters where MC0 isn't present or MC0 size is not equal to MC1 size caused the adapter to crash due to incorrect computation of memoffset. Fix is to read the size of MC0 instead of MC1 for offset calculation Signed-off-by: Steve Wise Signed-off-by: Hariprasad Shenai Signed-off-by: David S. Miller --- drivers/net/ethernet/chelsio/cxgb4/t4_hw.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c b/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c index 5959e3ae72da..e8578a742f2a 100644 --- a/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c +++ b/drivers/net/ethernet/chelsio/cxgb4/t4_hw.c @@ -492,7 +492,7 @@ int t4_memory_rw(struct adapter *adap, int win, int mtype, u32 addr, memoffset = (mtype * (edc_size * 1024 * 1024)); else { mc_size = EXT_MEM0_SIZE_G(t4_read_reg(adap, - MA_EXT_MEMORY1_BAR_A)); + MA_EXT_MEMORY0_BAR_A)); memoffset = (MEM_MC0 * edc_size + mc_size) * 1024 * 1024; } -- cgit v1.2.3 From f94813f3c1d02090cc02dcfcbed339897830acb8 Mon Sep 17 00:00:00 2001 From: Benjamin Poirier Date: Wed, 29 Apr 2015 15:59:35 -0700 Subject: mlx4_en: Use correct loop cursor in error path. Signed-off-by: Benjamin Poirier Fixes: 9e311e7 ("net/mlx4_en: Use affinity hint") Acked-by: Amir Vadai Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index 0f1afc085d58..bf173d7b873b 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -1721,7 +1721,7 @@ mac_err: cq_err: while (rx_index--) { mlx4_en_deactivate_cq(priv, priv->rx_cq[rx_index]); - mlx4_en_free_affinity_hint(priv, i); + mlx4_en_free_affinity_hint(priv, rx_index); } for (i = 0; i < priv->rx_ring_num; i++) mlx4_en_deactivate_rx_ring(priv, priv->rx_ring[i]); -- cgit v1.2.3 From 17d5ceb6e43ea545d6d92db2f3ddb035233ba335 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Wed, 29 Apr 2015 16:52:51 -0400 Subject: net/mlx4_core: Fix unaligned accesses Addresses the following kernel logs seen during boot: Kernel unaligned access at TPC[100ee150] mlx4_QUERY_HCA+0x80/0x248 [mlx4_core] Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core] Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core] Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core] Kernel unaligned access at TPC[100f071c] mlx4_QUERY_ADAPTER+0x100/0x12c [mlx4_core] Signed-off-by: David Ahern Acked-by: Or Gerlitz Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/fw.c | 18 ++++++++++++++---- 1 file changed, 14 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c index a4079811b176..e30bf57ad7a1 100644 --- a/drivers/net/ethernet/mellanox/mlx4/fw.c +++ b/drivers/net/ethernet/mellanox/mlx4/fw.c @@ -56,11 +56,13 @@ MODULE_PARM_DESC(enable_qos, "Enable Enhanced QoS support (default: on)"); #define MLX4_GET(dest, source, offset) \ do { \ void *__p = (char *) (source) + (offset); \ + u64 val; \ switch (sizeof (dest)) { \ case 1: (dest) = *(u8 *) __p; break; \ case 2: (dest) = be16_to_cpup(__p); break; \ case 4: (dest) = be32_to_cpup(__p); break; \ - case 8: (dest) = be64_to_cpup(__p); break; \ + case 8: val = get_unaligned((u64 *)__p); \ + (dest) = be64_to_cpu(val); break; \ default: __buggy_use_of_MLX4_GET(); \ } \ } while (0) @@ -1605,9 +1607,17 @@ static void get_board_id(void *vsd, char *board_id) * swaps each 4-byte word before passing it back to * us. Therefore we need to swab it before printing. */ - for (i = 0; i < 4; ++i) - ((u32 *) board_id)[i] = - swab32(*(u32 *) (vsd + VSD_OFFSET_MLX_BOARD_ID + i * 4)); + u32 *bid_u32 = (u32 *)board_id; + + for (i = 0; i < 4; ++i) { + u32 *addr; + u32 val; + + addr = (u32 *) (vsd + VSD_OFFSET_MLX_BOARD_ID + i * 4); + val = get_unaligned(addr); + val = swab32(val); + put_unaligned(val, &bid_u32[i]); + } } } -- cgit v1.2.3 From c232d8a8bb1416f7ec21bb1aabc7c4ec8a5a899e Mon Sep 17 00:00:00 2001 From: Tony Camuso Date: Thu, 30 Apr 2015 07:51:27 -0400 Subject: netxen_nic: use spin_[un]lock_bh around tx_clean_lock While testing this driver with DEBUG_LOCKDEP and DEBUG_SPINLOCK enabled did not produce any traces, it would be more prudent in the case of tx_clean_lock to use spin_[un]lock_bh, since this lock is manipulated in both the process and softirq contexts. This patch was tested for functionality and regressions with netperf and DEBUG_LOCKDEP and DEBUG_SPINLOCK enabled. Signed-off-by: Tony Camuso Acked-by: Neil Horman Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/netxen/netxen_nic_init.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/qlogic/netxen/netxen_nic_init.c b/drivers/net/ethernet/qlogic/netxen/netxen_nic_init.c index 5c4068353f66..8da7c3faf817 100644 --- a/drivers/net/ethernet/qlogic/netxen/netxen_nic_init.c +++ b/drivers/net/ethernet/qlogic/netxen/netxen_nic_init.c @@ -135,7 +135,7 @@ void netxen_release_tx_buffers(struct netxen_adapter *adapter) int i, j; struct nx_host_tx_ring *tx_ring = adapter->tx_ring; - spin_lock(&adapter->tx_clean_lock); + spin_lock_bh(&adapter->tx_clean_lock); cmd_buf = tx_ring->cmd_buf_arr; for (i = 0; i < tx_ring->num_desc; i++) { buffrag = cmd_buf->frag_array; @@ -159,7 +159,7 @@ void netxen_release_tx_buffers(struct netxen_adapter *adapter) } cmd_buf++; } - spin_unlock(&adapter->tx_clean_lock); + spin_unlock_bh(&adapter->tx_clean_lock); } void netxen_free_sw_resources(struct netxen_adapter *adapter) -- cgit v1.2.3 From 07841f9d94c11afe00c0498cf242edf4075729f4 Mon Sep 17 00:00:00 2001 From: Ido Shamay Date: Thu, 30 Apr 2015 17:32:46 +0300 Subject: net/mlx4_en: Schedule napi when RX buffers allocation fails When system is out of memory, refilling of RX buffers fails while the driver continue to pass the received packets to the kernel stack. At some point, when all RX buffers deplete, driver may fall into a sleep, and not recover when memory for new RX buffers is once again availible. This is because hardware does not have valid descriptors, so no interrupt will be generated for the driver to return to work in napi context. Fix it by schedule the napi poll function from stats_task delayed workqueue, as long as the allocations fail. Signed-off-by: Ido Shamay Signed-off-by: Amir Vadai Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 1 + drivers/net/ethernet/mellanox/mlx4/en_rx.c | 26 ++++++++++++++++++++++++-- drivers/net/ethernet/mellanox/mlx4/mlx4_en.h | 1 + 3 files changed, 26 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index bf173d7b873b..32f5ec737472 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -1467,6 +1467,7 @@ static void mlx4_en_service_task(struct work_struct *work) if (mdev->dev->caps.flags2 & MLX4_DEV_CAP_FLAG2_TS) mlx4_en_ptp_overflow_check(mdev); + mlx4_en_recover_from_oom(priv); queue_delayed_work(mdev->workqueue, &priv->service_task, SERVICE_TASK_DELAY); } diff --git a/drivers/net/ethernet/mellanox/mlx4/en_rx.c b/drivers/net/ethernet/mellanox/mlx4/en_rx.c index 4fdd3c37e47b..2a77a6b19121 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_rx.c @@ -244,6 +244,12 @@ static int mlx4_en_prepare_rx_desc(struct mlx4_en_priv *priv, return mlx4_en_alloc_frags(priv, rx_desc, frags, ring->page_alloc, gfp); } +static inline bool mlx4_en_is_ring_empty(struct mlx4_en_rx_ring *ring) +{ + BUG_ON((u32)(ring->prod - ring->cons) > ring->actual_size); + return ring->prod == ring->cons; +} + static inline void mlx4_en_update_rx_prod_db(struct mlx4_en_rx_ring *ring) { *ring->wqres.db.db = cpu_to_be32(ring->prod & 0xffff); @@ -315,8 +321,7 @@ static void mlx4_en_free_rx_buf(struct mlx4_en_priv *priv, ring->cons, ring->prod); /* Unmap and free Rx buffers */ - BUG_ON((u32) (ring->prod - ring->cons) > ring->actual_size); - while (ring->cons != ring->prod) { + while (!mlx4_en_is_ring_empty(ring)) { index = ring->cons & ring->size_mask; en_dbg(DRV, priv, "Processing descriptor:%d\n", index); mlx4_en_free_rx_desc(priv, ring, index); @@ -491,6 +496,23 @@ err_allocator: return err; } +/* We recover from out of memory by scheduling our napi poll + * function (mlx4_en_process_cq), which tries to allocate + * all missing RX buffers (call to mlx4_en_refill_rx_buffers). + */ +void mlx4_en_recover_from_oom(struct mlx4_en_priv *priv) +{ + int ring; + + if (!priv->port_up) + return; + + for (ring = 0; ring < priv->rx_ring_num; ring++) { + if (mlx4_en_is_ring_empty(priv->rx_ring[ring])) + napi_reschedule(&priv->rx_cq[ring]->napi); + } +} + void mlx4_en_destroy_rx_ring(struct mlx4_en_priv *priv, struct mlx4_en_rx_ring **pring, u32 size, u16 stride) diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h index 9de30216b146..d021f079f181 100644 --- a/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h +++ b/drivers/net/ethernet/mellanox/mlx4/mlx4_en.h @@ -774,6 +774,7 @@ int mlx4_en_activate_tx_ring(struct mlx4_en_priv *priv, void mlx4_en_deactivate_tx_ring(struct mlx4_en_priv *priv, struct mlx4_en_tx_ring *ring); void mlx4_en_set_num_rx_rings(struct mlx4_en_dev *mdev); +void mlx4_en_recover_from_oom(struct mlx4_en_priv *priv); int mlx4_en_create_rx_ring(struct mlx4_en_priv *priv, struct mlx4_en_rx_ring **pring, u32 size, u16 stride, int node); -- cgit v1.2.3 From e813bb2b955d9f72c94be7d592746b49929a499b Mon Sep 17 00:00:00 2001 From: Markus Pargmann Date: Thu, 30 Apr 2015 17:07:50 +0200 Subject: net: fec: Fix RGMII-ID mode RGMII-ID uses an internal delay within the transmitter or receiver. This feature is phy specific. The rest of the communication is normal RGMII. So the fec driver has to check for all RGMII modes, not only 'PHY_INTERFACE_MODE_RGMII'. Signed-off-by: Markus Pargmann Signed-off-by: David S. Miller --- drivers/net/ethernet/freescale/fec_main.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c index f6a3a7abd468..66d47e448e4d 100644 --- a/drivers/net/ethernet/freescale/fec_main.c +++ b/drivers/net/ethernet/freescale/fec_main.c @@ -988,7 +988,10 @@ fec_restart(struct net_device *ndev) rcntl |= 0x40000000 | 0x00000020; /* RGMII, RMII or MII */ - if (fep->phy_interface == PHY_INTERFACE_MODE_RGMII) + if (fep->phy_interface == PHY_INTERFACE_MODE_RGMII || + fep->phy_interface == PHY_INTERFACE_MODE_RGMII_ID || + fep->phy_interface == PHY_INTERFACE_MODE_RGMII_RXID || + fep->phy_interface == PHY_INTERFACE_MODE_RGMII_TXID) rcntl |= (1 << 6); else if (fep->phy_interface == PHY_INTERFACE_MODE_RMII) rcntl |= (1 << 8); -- cgit v1.2.3 From a134f083e79fb4c3d0a925691e732c56911b4326 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Fri, 1 May 2015 22:02:47 -0400 Subject: ipv4: Missing sk_nulls_node_init() in ping_unhash(). If we don't do that, then the poison value is left in the ->pprev backlink. This can cause crashes if we do a disconnect, followed by a connect(). Tested-by: Linus Torvalds Reported-by: Wen Xu Signed-off-by: David S. Miller --- net/ipv4/ping.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/ipv4/ping.c b/net/ipv4/ping.c index a93f260cf24c..05ff44b758df 100644 --- a/net/ipv4/ping.c +++ b/net/ipv4/ping.c @@ -158,6 +158,7 @@ void ping_unhash(struct sock *sk) if (sk_hashed(sk)) { write_lock_bh(&ping_table.lock); hlist_nulls_del(&sk->sk_nulls_node); + sk_nulls_node_init(&sk->sk_nulls_node); sock_put(sk); isk->inet_num = 0; isk->inet_sport = 0; -- cgit v1.2.3