From a22bb0b9b9b09b4cc711f6d577679773e074dde9 Mon Sep 17 00:00:00 2001 From: Francesco Ruggeri Date: Wed, 22 Oct 2014 15:29:24 +0000 Subject: [PATCH 1/4] e1000: unset IFF_UNICAST_FLT on WMware 82545EM VMWare's e1000 implementation does not seem to support unicast filtering. This can be observed by configuring a macvlan interface on eth0 in a VM in VMWare Fusion 5.0.5, and trying to use that interface instead of eth0. Tested on 3.16. Signed-off-by: Francesco Ruggeri Tested-by: Aaron Brown Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/e1000/e1000_main.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c index 5f6aded512f5..24f3986cfae2 100644 --- a/drivers/net/ethernet/intel/e1000/e1000_main.c +++ b/drivers/net/ethernet/intel/e1000/e1000_main.c @@ -1075,7 +1075,10 @@ static int e1000_probe(struct pci_dev *pdev, const struct pci_device_id *ent) NETIF_F_HW_CSUM | NETIF_F_SG); - netdev->priv_flags |= IFF_UNICAST_FLT; + /* Do not set IFF_UNICAST_FLT for VMWare's 82545EM */ + if (hw->device_id != E1000_DEV_ID_82545EM_COPPER || + hw->subsystem_vendor_id != PCI_VENDOR_ID_VMWARE) + netdev->priv_flags |= IFF_UNICAST_FLT; adapter->en_mng_pt = e1000_enable_mng_pass_thru(hw); From bc16e47f03a7dce9ad68029b21519265c334eb12 Mon Sep 17 00:00:00 2001 From: Roman Gushchin Date: Thu, 23 Oct 2014 03:32:27 +0000 Subject: [PATCH 2/4] igb: don't reuse pages with pfmemalloc flag Incoming packet is dropped silently by sk_filter(), if the skb was allocated from pfmemalloc reserves and the corresponding socket is not marked with the SOCK_MEMALLOC flag. Igb driver allocates pages for DMA with __skb_alloc_page(), which calls alloc_pages_node() with the __GFP_MEMALLOC flag. So, in case of OOM condition, igb can get pages with pfmemalloc flag set. If an incoming packet hits the pfmemalloc page and is large enough (small packets are copying into the memory, allocated with netdev_alloc_skb_ip_align(), so they are not affected), it will be dropped. This behavior is ok under high memory pressure, but the problem is that the igb driver reuses these mapped pages. So, packets are still dropping even if all memory issues are gone and there is a plenty of free memory. In my case, some TCP sessions hang on a small percentage (< 0.1%) of machines days after OOMs. Fix this by avoiding reuse of such pages. Signed-off-by: Roman Gushchin Tested-by: Aaron Brown "aaron.f.brown@intel.com" Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/igb/igb_main.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c index a21b14495ebd..a2d72a87cbde 100644 --- a/drivers/net/ethernet/intel/igb/igb_main.c +++ b/drivers/net/ethernet/intel/igb/igb_main.c @@ -6537,6 +6537,9 @@ static bool igb_can_reuse_rx_page(struct igb_rx_buffer *rx_buffer, if (unlikely(page_to_nid(page) != numa_node_id())) return false; + if (unlikely(page->pfmemalloc)) + return false; + #if (PAGE_SIZE < 8192) /* if we are only owner of page we can reuse it */ if (unlikely(page_count(page) != 1)) @@ -6603,7 +6606,8 @@ static bool igb_add_rx_frag(struct igb_ring *rx_ring, memcpy(__skb_put(skb, size), va, ALIGN(size, sizeof(long))); /* we can reuse buffer as-is, just make sure it is local */ - if (likely(page_to_nid(page) == numa_node_id())) + if (likely((page_to_nid(page) == numa_node_id()) && + !page->pfmemalloc)) return true; /* this page cannot be reused so discard it */ From 4d2fcfbcf8141cdf70245a0c0612b8076f4b7e32 Mon Sep 17 00:00:00 2001 From: Junwei Zhang Date: Wed, 22 Oct 2014 15:29:03 +0000 Subject: [PATCH 3/4] ixgbe: need not repeat init skb with NULL Signed-off-by: Martin Zhang Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c index fec5212d4337..d2df4e3d1032 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c @@ -4321,8 +4321,8 @@ static void ixgbe_clean_rx_ring(struct ixgbe_ring *rx_ring) IXGBE_CB(skb)->page_released = false; } dev_kfree_skb(skb); + rx_buffer->skb = NULL; } - rx_buffer->skb = NULL; if (rx_buffer->dma) dma_unmap_page(dev, rx_buffer->dma, ixgbe_rx_pg_size(rx_ring), From e3215f0ac77ec23b052cb0bf511143038ac2ad7b Mon Sep 17 00:00:00 2001 From: Emil Tantilov Date: Tue, 28 Oct 2014 05:50:03 +0000 Subject: [PATCH 4/4] ixgbe: fix race when setting advertised speed Following commands: modprobe ixgbe ifconfig ethX up ethtool -s ethX advertise 0x020 can lead to "setup link failed with code -14" error due to the setup_link call racing with the SFP detection routine in the watchdog. This patch resolves this issue by protecting the setup_link call with check for __IXGBE_IN_SFP_INIT. Reported-by: Scott Harrison Signed-off-by: Emil Tantilov Tested-by: Phil Schmitt Signed-off-by: Jeff Kirsher --- drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c index 3ce4a258f945..0ae038b9af90 100644 --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c @@ -342,12 +342,16 @@ static int ixgbe_set_settings(struct net_device *netdev, if (old == advertised) return err; /* this sets the link speed and restarts auto-neg */ + while (test_and_set_bit(__IXGBE_IN_SFP_INIT, &adapter->state)) + usleep_range(1000, 2000); + hw->mac.autotry_restart = true; err = hw->mac.ops.setup_link(hw, advertised, true); if (err) { e_info(probe, "setup link failed with code %d\n", err); hw->mac.ops.setup_link(hw, old, true); } + clear_bit(__IXGBE_IN_SFP_INIT, &adapter->state); } else { /* in this case we currently only support 10Gb/FULL */ u32 speed = ethtool_cmd_speed(ecmd);