From b90efd2258749e04e1b3f71ef0d716f2ac2337e0 Mon Sep 17 00:00:00 2001 From: Willem de Bruijn Date: Thu, 7 Feb 2019 14:54:16 -0500 Subject: bpf: only adjust gso_size on bytestream protocols bpf_skb_change_proto and bpf_skb_adjust_room change skb header length. For GSO packets they adjust gso_size to maintain the same MTU. The gso size can only be safely adjusted on bytestream protocols. Commit d02f51cbcf12 ("bpf: fix bpf_skb_adjust_net/bpf_skb_proto_xlat to deal with gso sctp skbs") excluded SKB_GSO_SCTP. Since then type SKB_GSO_UDP_L4 has been added, whose contents are one gso_size unit per datagram. Also exclude these. Move from a blacklist to a whitelist check to future proof against additional such new GSO types, e.g., for fraglist based GRO. Fixes: bec1f6f69736 ("udp: generate gso with UDP_SEGMENT") Signed-off-by: Willem de Bruijn Signed-off-by: Alexei Starovoitov --- net/core/filter.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) (limited to 'net/core') diff --git a/net/core/filter.c b/net/core/filter.c index 7a54dc11ac2d..f7d0004fc160 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -2789,8 +2789,7 @@ static int bpf_skb_proto_4_to_6(struct sk_buff *skb) u32 off = skb_mac_header_len(skb); int ret; - /* SCTP uses GSO_BY_FRAGS, thus cannot adjust it. */ - if (skb_is_gso(skb) && unlikely(skb_is_gso_sctp(skb))) + if (!skb_is_gso_tcp(skb)) return -ENOTSUPP; ret = skb_cow(skb, len_diff); @@ -2831,8 +2830,7 @@ static int bpf_skb_proto_6_to_4(struct sk_buff *skb) u32 off = skb_mac_header_len(skb); int ret; - /* SCTP uses GSO_BY_FRAGS, thus cannot adjust it. */ - if (skb_is_gso(skb) && unlikely(skb_is_gso_sctp(skb))) + if (!skb_is_gso_tcp(skb)) return -ENOTSUPP; ret = skb_unclone(skb, GFP_ATOMIC); @@ -2957,8 +2955,7 @@ static int bpf_skb_net_grow(struct sk_buff *skb, u32 len_diff) u32 off = skb_mac_header_len(skb) + bpf_skb_net_base_len(skb); int ret; - /* SCTP uses GSO_BY_FRAGS, thus cannot adjust it. */ - if (skb_is_gso(skb) && unlikely(skb_is_gso_sctp(skb))) + if (!skb_is_gso_tcp(skb)) return -ENOTSUPP; ret = skb_cow(skb, len_diff); @@ -2987,8 +2984,7 @@ static int bpf_skb_net_shrink(struct sk_buff *skb, u32 len_diff) u32 off = skb_mac_header_len(skb) + bpf_skb_net_base_len(skb); int ret; - /* SCTP uses GSO_BY_FRAGS, thus cannot adjust it. */ - if (skb_is_gso(skb) && unlikely(skb_is_gso_sctp(skb))) + if (!skb_is_gso_tcp(skb)) return -ENOTSUPP; ret = skb_unclone(skb, GFP_ATOMIC); -- cgit v1.2.3 From 3b89ea9c5902acccdbbdec307c85edd1bf52515e Mon Sep 17 00:00:00 2001 From: Hauke Mehrtens Date: Fri, 15 Feb 2019 17:58:54 +0100 Subject: net: Fix for_each_netdev_feature on Big endian The features attribute is of type u64 and stored in the native endianes on the system. The for_each_set_bit() macro takes a pointer to a 32 bit array and goes over the bits in this area. On little Endian systems this also works with an u64 as the most significant bit is on the highest address, but on big endian the words are swapped. When we expect bit 15 here we get bit 47 (15 + 32). This patch converts it more or less to its own for_each_set_bit() implementation which works on 64 bit integers directly. This is then completely in host endianness and should work like expected. Fixes: fd867d51f ("net/core: generic support for disabling netdev features down stack") Signed-off-by: Hauke Mehrtens Signed-off-by: David S. Miller --- include/linux/netdev_features.h | 23 +++++++++++++++++++++-- net/core/dev.c | 4 ++-- 2 files changed, 23 insertions(+), 4 deletions(-) (limited to 'net/core') diff --git a/include/linux/netdev_features.h b/include/linux/netdev_features.h index 2b2a6dce1630..fce28562bed2 100644 --- a/include/linux/netdev_features.h +++ b/include/linux/netdev_features.h @@ -11,6 +11,7 @@ #define _LINUX_NETDEV_FEATURES_H #include +#include typedef u64 netdev_features_t; @@ -154,8 +155,26 @@ enum { #define NETIF_F_HW_TLS_TX __NETIF_F(HW_TLS_TX) #define NETIF_F_HW_TLS_RX __NETIF_F(HW_TLS_RX) -#define for_each_netdev_feature(mask_addr, bit) \ - for_each_set_bit(bit, (unsigned long *)mask_addr, NETDEV_FEATURE_COUNT) +/* Finds the next feature with the highest number of the range of start till 0. + */ +static inline int find_next_netdev_feature(u64 feature, unsigned long start) +{ + /* like BITMAP_LAST_WORD_MASK() for u64 + * this sets the most significant 64 - start to 0. + */ + feature &= ~0ULL >> (-start & ((sizeof(feature) * 8) - 1)); + + return fls64(feature) - 1; +} + +/* This goes for the MSB to the LSB through the set feature bits, + * mask_addr should be a u64 and bit an int + */ +#define for_each_netdev_feature(mask_addr, bit) \ + for ((bit) = find_next_netdev_feature((mask_addr), \ + NETDEV_FEATURE_COUNT); \ + (bit) >= 0; \ + (bit) = find_next_netdev_feature((mask_addr), (bit) - 1)) /* Features valid for ethtool to change */ /* = all defined minus driver/device-class-related */ diff --git a/net/core/dev.c b/net/core/dev.c index 8e276e0192a1..5d03889502eb 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -8152,7 +8152,7 @@ static netdev_features_t netdev_sync_upper_features(struct net_device *lower, netdev_features_t feature; int feature_bit; - for_each_netdev_feature(&upper_disables, feature_bit) { + for_each_netdev_feature(upper_disables, feature_bit) { feature = __NETIF_F_BIT(feature_bit); if (!(upper->wanted_features & feature) && (features & feature)) { @@ -8172,7 +8172,7 @@ static void netdev_sync_lower_features(struct net_device *upper, netdev_features_t feature; int feature_bit; - for_each_netdev_feature(&upper_disables, feature_bit) { + for_each_netdev_feature(upper_disables, feature_bit) { feature = __NETIF_F_BIT(feature_bit); if (!(features & feature) && (lower->features & feature)) { netdev_dbg(upper, "Disabling feature %pNF on lower dev %s.\n", -- cgit v1.2.3 From 3bed3cc4156eedf652b4df72bdb35d4f1a2a739d Mon Sep 17 00:00:00 2001 From: Alexander Duyck Date: Fri, 15 Feb 2019 14:44:18 -0800 Subject: net: Do not allocate page fragments that are not skb aligned This patch addresses the fact that there are drivers, specifically tun, that will call into the network page fragment allocators with buffer sizes that are not cache aligned. Doing this could result in data alignment and DMA performance issues as these fragment pools are also shared with the skb allocator and any other devices that will use napi_alloc_frags or netdev_alloc_frags. Fixes: ffde7328a36d ("net: Split netdev_alloc_frag into __alloc_page_frag and add __napi_alloc_frag") Reported-by: Jann Horn Signed-off-by: Alexander Duyck Signed-off-by: David S. Miller --- net/core/skbuff.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'net/core') diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 26d848484912..2415d9cb9b89 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -356,6 +356,8 @@ static void *__netdev_alloc_frag(unsigned int fragsz, gfp_t gfp_mask) */ void *netdev_alloc_frag(unsigned int fragsz) { + fragsz = SKB_DATA_ALIGN(fragsz); + return __netdev_alloc_frag(fragsz, GFP_ATOMIC); } EXPORT_SYMBOL(netdev_alloc_frag); @@ -369,6 +371,8 @@ static void *__napi_alloc_frag(unsigned int fragsz, gfp_t gfp_mask) void *napi_alloc_frag(unsigned int fragsz) { + fragsz = SKB_DATA_ALIGN(fragsz); + return __napi_alloc_frag(fragsz, GFP_ATOMIC); } EXPORT_SYMBOL(napi_alloc_frag); -- cgit v1.2.3