Cumulative patch for avoiding unaligned access traps on ar71xx hardware using the AG71xx network driver and the Linux 3.7.1 kernel. This patch contains OpenWRT's existing 902-unaligned_access_hacks.patch plus additional modifications from CeroWRT. The result
is a significant increase in performance with both IPv4 and IPv6 traffic.

This is an update of the previous patch at http://patchwork.openwrt.org/patch/2147/.

(Note: I use __get_unaligned_cpu32() in this code since it mirrors the existing pointer
dereference better than using __get_unaligned_be32() would.)

Signed-off-by: Robert Bradley <robert.bradl...@gmail.com>
---
.../patches-3.7/902-unaligned_access_hacks.patch | 746 +++++++++++++++++++-
 1 file changed, 739 insertions(+), 7 deletions(-)

diff --git a/target/linux/ar71xx/patches-3.7/902-unaligned_access_hacks.patch b/target/linux/ar71xx/patches-3.7/902-unaligned_access_hacks.patch
index 23ee8a8..3b367b9 100644
--- a/target/linux/ar71xx/patches-3.7/902-unaligned_access_hacks.patch
+++ b/target/linux/ar71xx/patches-3.7/902-unaligned_access_hacks.patch
@@ -1,3 +1,5 @@
+diff --git a/arch/mips/include/asm/checksum.h b/arch/mips/include/asm/checksum.h
+index f2f7c6c..9be4201 100644
 --- a/arch/mips/include/asm/checksum.h
 +++ b/arch/mips/include/asm/checksum.h
 @@ -12,6 +12,7 @@
@@ -8,7 +10,7 @@

  #include <asm/uaccess.h>

-@@ -104,26 +105,30 @@ static inline __sum16 ip_fast_csum(const
+@@ -104,26 +105,30 @@ static inline __sum16 ip_fast_csum(const void *iph, unsigned int ihl)
      const unsigned int *stop = word + ihl;
      unsigned int csum;
      int carry;
@@ -50,6 +52,243 @@
      } while (word != stop);

      return csum_fold(csum);
+diff --git a/include/linux/if_vlan.h b/include/linux/if_vlan.h
+index d06cc5c..dabd8eb 100644
+--- a/include/linux/if_vlan.h
++++ b/include/linux/if_vlan.h
+@@ -38,7 +38,7 @@
+ struct vlan_hdr {
+     __be16    h_vlan_TCI;
+     __be16    h_vlan_encapsulated_proto;
+-};
++} __packed;
+
+ /**
+  *    struct vlan_ethhdr - vlan ethernet header (ethhdr + vlan_hdr)
+diff --git a/include/net/flow_keys.h b/include/net/flow_keys.h
+index 80461c1..ffd6fbd 100644
+--- a/include/net/flow_keys.h
++++ b/include/net/flow_keys.h
+@@ -10,7 +10,7 @@ struct flow_keys {
+         __be16 port16[2];
+     };
+     u8 ip_proto;
+-};
++} __packed;
+
+ extern bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow);
+ #endif
+diff --git a/include/net/inet_ecn.h b/include/net/inet_ecn.h
+index aab7375..e5e9f46 100644
+--- a/include/net/inet_ecn.h
++++ b/include/net/inet_ecn.h
+@@ -113,15 +113,21 @@ struct ipv6hdr;
+
+ static inline int IP6_ECN_set_ce(struct ipv6hdr *iph)
+ {
++    __be32 dsfield;
+     if (INET_ECN_is_not_ect(ipv6_get_dsfield(iph)))
+         return 0;
+-    *(__be32*)iph |= htonl(INET_ECN_CE << 20);
++    dsfield = __get_unaligned_cpu32((__be32*)iph);
++    dsfield |= htonl(INET_ECN_CE << 20);
++    __put_unaligned_cpu32(dsfield, (__be32*)iph);
+     return 1;
+ }
+
+ static inline void IP6_ECN_clear(struct ipv6hdr *iph)
+ {
+-    *(__be32*)iph &= ~htonl(INET_ECN_MASK << 20);
++    __be32 dsfield;
++    dsfield = __get_unaligned_cpu32((__be32*)iph);
++    dsfield &= ~htonl(INET_ECN_MASK << 20);
++    __put_unaligned_cpu32(dsfield, (__be32*)iph);
+ }
+
+ static inline void ipv6_copy_dscp(unsigned int dscp, struct ipv6hdr *inner)
+diff --git a/include/net/ipv6.h b/include/net/ipv6.h
+index 979bf6c..0259560 100644
+--- a/include/net/ipv6.h
++++ b/include/net/ipv6.h
+@@ -106,7 +106,7 @@ struct frag_hdr {
+     __u8    reserved;
+     __be16    frag_off;
+     __be32    identification;
+-};
++} __packed __attribute__((aligned(2)));
+
+ #define    IP6_MF    0x0001
+
+@@ -384,7 +384,7 @@ static inline bool __ipv6_prefix_equal(const __be32 *a1, const __be32 *a2,
+
+     /* check incomplete u32 in prefix */
+     pbi = prefixlen & 0x1f;
+-    if (pbi && ((a1[pdw] ^ a2[pdw]) & htonl((0xffffffff) << (32 - pbi))))
++ if (pbi && ((__get_unaligned_cpu32(&a1[pdw]) ^ __get_unaligned_cpu32(&a2[pdw])) & htonl((0xffffffff) << (32 - pbi))))
+         return false;
+
+     return true;
+@@ -506,7 +506,7 @@ static inline int __ipv6_addr_diff(const void *token1, const void *token2, int a
+     addrlen >>= 2;
+
+     for (i = 0; i < addrlen; i++) {
+-        __be32 xb = a1[i] ^ a2[i];
++ __be32 xb = __get_unaligned_cpu32(&a1[i]) ^ __get_unaligned_cpu32(&a2[i]);
+         if (xb)
+             return i * 32 + 31 - __fls(ntohl(xb));
+     }
+diff --git a/include/net/ndisc.h b/include/net/ndisc.h
+index 980d263..f06c42b 100644
+--- a/include/net/ndisc.h
++++ b/include/net/ndisc.h
+@@ -65,23 +65,23 @@ struct nd_msg {
+         struct icmp6hdr    icmph;
+         struct in6_addr    target;
+     __u8        opt[0];
+-};
++} __packed __attribute__((aligned(2)));
+
+ struct rs_msg {
+     struct icmp6hdr    icmph;
+     __u8        opt[0];
+-};
++} __packed __attribute__((aligned(2)));
+
+ struct ra_msg {
+         struct icmp6hdr        icmph;
+     __be32            reachable_time;
+     __be32            retrans_timer;
+-};
++} __packed __attribute__((aligned(2)));
+
+ struct nd_opt_hdr {
+     __u8        nd_opt_type;
+     __u8        nd_opt_len;
+-} __packed;
++} __packed __attribute__((aligned(2)));
+
+ /* ND options */
+ struct ndisc_options {
+@@ -135,10 +135,10 @@ static inline u32 ndisc_hashfn(const void *pkey, const struct net_device *dev, _
+ {
+     const u32 *p32 = pkey;
+
+-    return (((p32[0] ^ hash32_ptr(dev)) * hash_rnd[0]) +
+-        (p32[1] * hash_rnd[1]) +
+-        (p32[2] * hash_rnd[2]) +
+-        (p32[3] * hash_rnd[3]));
++ return (((__get_unaligned_cpu32(&p32[0]) ^ hash32_ptr(dev)) * hash_rnd[0]) +
++        (__get_unaligned_cpu32(&p32[1]) * hash_rnd[1]) +
++        (__get_unaligned_cpu32(&p32[2]) * hash_rnd[2]) +
++        (__get_unaligned_cpu32(&p32[3]) * hash_rnd[3]));
+ }
+
+ static inline struct neighbour *__ipv6_neigh_lookup(struct neigh_table *tbl, struct net_device *dev, const void *pkey) +@@ -156,8 +156,8 @@ static inline struct neighbour *__ipv6_neigh_lookup(struct neigh_table *tbl, str
+          n = rcu_dereference_bh(n->next)) {
+         u32 *n32 = (u32 *) n->primary_key;
+         if (n->dev == dev &&
+-            ((n32[0] ^ p32[0]) | (n32[1] ^ p32[1]) |
+-             (n32[2] ^ p32[2]) | (n32[3] ^ p32[3])) == 0) {
++ ((n32[0] ^ __get_unaligned_cpu32(&p32[0])) | (n32[1] ^ __get_unaligned_cpu32(&p32[1])) | ++ (n32[2] ^ __get_unaligned_cpu32(&p32[2])) | (n32[3] ^ __get_unaligned_cpu32(&p32[3]))) == 0) {
+             if (!atomic_inc_not_zero(&n->refcnt))
+                 n = NULL;
+             break;
+diff --git a/include/uapi/linux/icmp.h b/include/uapi/linux/icmp.h
+index 16fff05..75c622e 100644
+--- a/include/uapi/linux/icmp.h
++++ b/include/uapi/linux/icmp.h
+@@ -80,7 +80,7 @@ struct icmphdr {
+         __be16    mtu;
+     } frag;
+   } un;
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+
+ /*
+diff --git a/include/uapi/linux/icmpv6.h b/include/uapi/linux/icmpv6.h
+index e0133c7..fd15bc2e 100644
+--- a/include/uapi/linux/icmpv6.h
++++ b/include/uapi/linux/icmpv6.h
+@@ -76,7 +76,7 @@ struct icmp6hdr {
+ #define icmp6_addrconf_other    icmp6_dataun.u_nd_ra.other
+ #define icmp6_rt_lifetime    icmp6_dataun.u_nd_ra.rt_lifetime
+ #define icmp6_router_pref    icmp6_dataun.u_nd_ra.router_pref
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+
+ #define ICMPV6_ROUTER_PREF_LOW        0x3
+diff --git a/include/uapi/linux/igmp.h b/include/uapi/linux/igmp.h
+index ccbb32a..564666a 100644
+--- a/include/uapi/linux/igmp.h
++++ b/include/uapi/linux/igmp.h
+@@ -32,7 +32,7 @@ struct igmphdr {
+     __u8 code;        /* For newer IGMP */
+     __sum16 csum;
+     __be32 group;
+-};
++} __packed __attribute__((aligned(2)));
+
+ /* V3 group record types [grec_type] */
+ #define IGMPV3_MODE_IS_INCLUDE        1
+@@ -48,7 +48,7 @@ struct igmpv3_grec {
+     __be16    grec_nsrcs;
+     __be32    grec_mca;
+     __be32    grec_src[0];
+-};
++} __packed __attribute__((aligned(2)));
+
+ struct igmpv3_report {
+     __u8 type;
+@@ -57,7 +57,7 @@ struct igmpv3_report {
+     __be16 resv2;
+     __be16 ngrec;
+     struct igmpv3_grec grec[0];
+-};
++} __packed __attribute__((aligned(2)));
+
+ struct igmpv3_query {
+     __u8 type;
+@@ -78,7 +78,7 @@ struct igmpv3_query {
+     __u8 qqic;
+     __be16 nsrcs;
+     __be32 srcs[0];
+-};
++} __packed __attribute__((aligned(2)));
+
+ #define IGMP_HOST_MEMBERSHIP_QUERY    0x11    /* From RFC1112 */
+ #define IGMP_HOST_MEMBERSHIP_REPORT    0x12    /* Ditto */
+diff --git a/include/uapi/linux/in.h b/include/uapi/linux/in.h
+index 9edb441..dd95bfa 100644
+--- a/include/uapi/linux/in.h
++++ b/include/uapi/linux/in.h
+@@ -55,7 +55,7 @@ enum {
+ /* Internet address. */
+ struct in_addr {
+     __be32    s_addr;
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ #define IP_TOS        1
+ #define IP_TTL        2
+diff --git a/include/uapi/linux/in6.h b/include/uapi/linux/in6.h
+index 1e31599..fc75ab1 100644
+--- a/include/uapi/linux/in6.h
++++ b/include/uapi/linux/in6.h
+@@ -36,7 +36,7 @@ struct in6_addr {
+ #define s6_addr            in6_u.u6_addr8
+ #define s6_addr16        in6_u.u6_addr16
+ #define s6_addr32        in6_u.u6_addr32
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ /* IPv6 Wildcard Address (::) and Loopback Address (::1) defined in RFC2553 + * NOTE: Be aware the IN6ADDR_* constants and in6addr_* externals are defined
+diff --git a/include/uapi/linux/ip.h b/include/uapi/linux/ip.h
+index 6cf06bf..af4401e 100644
 --- a/include/uapi/linux/ip.h
 +++ b/include/uapi/linux/ip.h
 @@ -102,7 +102,7 @@ struct iphdr {
@@ -57,21 +296,101 @@
      __be32    daddr;
      /*The options start here. */
 -};
-+} __packed __attribute__((aligned(2)));
++} __attribute__((packed)) __attribute__((aligned(2)));


  struct ip_auth_hdr {
+@@ -112,25 +112,25 @@ struct ip_auth_hdr {
+     __be32 spi;
+     __be32 seq_no;        /* Sequence number */
+ __u8 auth_data[0]; /* Variable len but >=4. Mind the 64 bit alignment! */
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ struct ip_esp_hdr {
+     __be32 spi;
+     __be32 seq_no;        /* Sequence number */
+ __u8 enc_data[0]; /* Variable len but >=8. Mind the 64 bit alignment! */
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ struct ip_comp_hdr {
+     __u8 nexthdr;
+     __u8 flags;
+     __be16 cpi;
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ struct ip_beet_phdr {
+     __u8 nexthdr;
+     __u8 hdrlen;
+     __u8 padlen;
+     __u8 reserved;
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ #endif /* _UAPI_LINUX_IP_H */
+diff --git a/include/uapi/linux/ipv6.h b/include/uapi/linux/ipv6.h
+index a6d7d1c..6247ba2 100644
 --- a/include/uapi/linux/ipv6.h
 +++ b/include/uapi/linux/ipv6.h
-@@ -123,7 +123,7 @@ struct ipv6hdr {
+@@ -49,8 +49,7 @@ struct ipv6_rt_hdr {
+      *    type specific data
+      *    variable length field
+      */
+-};
+-
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ struct ipv6_opt_hdr {
+     __u8         nexthdr;
+@@ -58,7 +57,7 @@ struct ipv6_opt_hdr {
+     /*
+      * TLV encoded option data follows.
+      */
+-} __attribute__((packed));    /* required for some archs */
++} __attribute__((packed)) __attribute__((aligned(2))); /* required for some archs */
+
+ #define ipv6_destopt_hdr ipv6_opt_hdr
+ #define ipv6_hopopt_hdr  ipv6_opt_hdr
+@@ -74,7 +73,7 @@ struct rt0_hdr {
+     struct in6_addr        addr[0];
+
+ #define rt0_type        rt_hdr.type
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ /*
+  *    routing header type 2
+@@ -86,7 +85,7 @@ struct rt2_hdr {
+     struct in6_addr        addr;
+
+ #define rt2_type        rt_hdr.type
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ /*
+  *    home address option in destination options header
+@@ -96,7 +95,7 @@ struct ipv6_destopt_hao {
+     __u8            type;
+     __u8            length;
+     struct in6_addr        addr;
+-} __attribute__((packed));
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+ /*
+  *    IPv6 fixed header
+@@ -123,7 +122,7 @@ struct ipv6hdr {

      struct    in6_addr    saddr;
      struct    in6_addr    daddr;
 -};
-+} __packed __attribute__((aligned(2)));
++} __attribute__((packed)) __attribute__((aligned(2)));


  /* index values for the variables in ipv6_devconf */
+diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h
+index e962faa..fa6dc5c 100644
 --- a/include/uapi/linux/tcp.h
 +++ b/include/uapi/linux/tcp.h
 @@ -54,7 +54,7 @@ struct tcphdr {
@@ -79,10 +398,24 @@
      __sum16    check;
      __be16    urg_ptr;
 -};
-+} __packed __attribute__((aligned(2)));
++} __attribute__((packed)) __attribute__((aligned(2)));

  /*
   *    The union cast uses a gcc extension to avoid aliasing problems
+@@ -64,9 +64,9 @@ struct tcphdr {
+ union tcp_word_hdr {
+     struct tcphdr hdr;
+     __be32           words[5];
+-};
++} __attribute__((packed)) __attribute__((aligned(2)));
+
+-#define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3])
++#define tcp_flag_word(tp) ( __get_unaligned_cpu32(&(((union tcp_word_hdr *)(tp))->words [3])))
+
+ enum {
+     TCP_FLAG_CWR = __constant_cpu_to_be32(0x00800000),
+diff --git a/include/uapi/linux/udp.h b/include/uapi/linux/udp.h
+index e2bcfd7..c7eb65b 100644
 --- a/include/uapi/linux/udp.h
 +++ b/include/uapi/linux/udp.h
 @@ -24,7 +24,7 @@ struct udphdr {
@@ -90,10 +423,58 @@
      __be16    len;
      __sum16    check;
 -};
-+} __packed __attribute__((aligned(2)));
++} __attribute__((packed)) __attribute__((aligned(2)));

  /* UDP socket options */
  #define UDP_CORK    1    /* Never send partially complete segments */
+diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
+index 466820b..f978bf9 100644
+--- a/net/core/flow_dissector.c
++++ b/net/core/flow_dissector.c
+@@ -18,7 +18,10 @@ static void iph_to_flow_copy_addrs(struct flow_keys *flow, const struct iphdr *i
+ {
+     BUILD_BUG_ON(offsetof(typeof(*flow), dst) !=
+              offsetof(typeof(*flow), src) + sizeof(flow->src));
+- memcpy(&flow->src, &iph->saddr, sizeof(flow->src) + sizeof(flow->dst)); ++ /* memcpy(&flow->src, &iph->saddr, sizeof(flow->src) + sizeof(flow->dst)); */
++    flow->src = iph->saddr;
++    flow->dst = iph->daddr;
++
+ }
+
+ bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow)
+@@ -137,7 +140,7 @@ ipv6:
+         nhoff += poff;
+         ports = skb_header_pointer(skb, nhoff, sizeof(_ports), &_ports);
+         if (ports)
+-            flow->ports = *ports;
++          flow->ports = __get_unaligned_cpu32(ports);
+     }
+
+     return true;
+diff --git a/net/core/secure_seq.c b/net/core/secure_seq.c
+index e61a8bb..f7beb31 100644
+--- a/net/core/secure_seq.c
++++ b/net/core/secure_seq.c
+@@ -7,6 +7,7 @@
+ #include <linux/hrtimer.h>
+ #include <linux/ktime.h>
+ #include <linux/string.h>
++#include <linux/unaligned/packed_struct.h>
+
+ #include <net/secure_seq.h>
+
+@@ -46,7 +47,7 @@ __u32 secure_tcpv6_sequence_number(const __be32 *saddr, const __be32 *daddr,
+
+     memcpy(hash, saddr, 16);
+     for (i = 0; i < 4; i++)
+-        secret[i] = net_secret[i] + (__force u32)daddr[i];
++ secret[i] = net_secret[i] + (__force u32) __get_unaligned_cpu32(&daddr[i]);
+     secret[4] = net_secret[4] +
+         (((__force u16)sport << 16) + (__force u16)dport);
+     for (i = 5; i < MD5_MESSAGE_BYTES / 4; i++)
+diff --git a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
+index fcdd0c2..e8d93c8 100644
 --- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
 +++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
 @@ -14,6 +14,7 @@
@@ -104,7 +485,7 @@
  #include <net/route.h>
  #include <net/ip.h>

-@@ -39,8 +40,8 @@ static bool ipv4_pkt_to_tuple(const stru
+@@ -39,8 +40,8 @@ static bool ipv4_pkt_to_tuple(const struct sk_buff *skb, unsigned int nhoff,
      if (ap == NULL)
          return false;

@@ -115,3 +496,354 @@

      return true;
  }
+diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
+index e457c7a..55ec75c 100644
+--- a/net/ipv4/tcp.c
++++ b/net/ipv4/tcp.c
+@@ -3149,7 +3149,7 @@ found:
+
+     p = *head;
+     th2 = tcp_hdr(p);
+-    tcp_flag_word(th2) |= flags & (TCP_FLAG_FIN | TCP_FLAG_PSH);
++ tcp_flag_word2(th2) = tcp_flag_word(th2) | flags & (TCP_FLAG_FIN | TCP_FLAG_PSH);
+
+ out_check_final:
+     flush = len < mss;
+diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
+index 181fc82..490d83e 100644
+--- a/net/ipv4/tcp_input.c
++++ b/net/ipv4/tcp_input.c
+@@ -3842,13 +3842,13 @@ static bool tcp_parse_aligned_timestamp(struct tcp_sock *tp, const struct tcphdr
+ {
+     const __be32 *ptr = (const __be32 *)(th + 1);
+
+-    if (*ptr == htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16)
++ if (__get_unaligned_cpu32(ptr) == htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16)
+               | (TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP)) {
+         tp->rx_opt.saw_tstamp = 1;
+         ++ptr;
+-        tp->rx_opt.rcv_tsval = ntohl(*ptr);
++        tp->rx_opt.rcv_tsval = ntohl(__get_unaligned_cpu32(ptr));
+         ++ptr;
+-        tp->rx_opt.rcv_tsecr = ntohl(*ptr);
++        tp->rx_opt.rcv_tsecr = ntohl(__get_unaligned_cpu32(ptr));
+         return true;
+     }
+     return false;
+diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
+index a974247..5e7721f 100644
+--- a/net/ipv6/af_inet6.c
++++ b/net/ipv6/af_inet6.c
+@@ -687,7 +687,7 @@ bool ipv6_opt_accepted(const struct sock *sk, const struct sk_buff *skb)
+         if ((opt->hop && (np->rxopt.bits.hopopts ||
+                   np->rxopt.bits.ohopopts)) ||
+             ((IPV6_FLOWINFO_MASK &
+-              *(__be32 *)skb_network_header(skb)) &&
++              __get_unaligned_cpu32((__be32 *)skb_network_header(skb))) &&
+              np->rxopt.bits.rxflow) ||
+             (opt->srcrt && (np->rxopt.bits.srcrt ||
+              np->rxopt.bits.osrcrt)) ||
+diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c
+index be2b67d6..6cbd2b7 100644
+--- a/net/ipv6/datagram.c
++++ b/net/ipv6/datagram.c
+@@ -360,12 +360,12 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len)
+                 *(struct in6_addr *)(nh + serr->addr_offset);
+             if (np->sndflow)
+                 sin->sin6_flowinfo =
+-                    (*(__be32 *)(nh + serr->addr_offset - 24) &
++ (__get_unaligned_cpu32((__be32 *)(nh + serr->addr_offset - 24)) &
+                      IPV6_FLOWINFO_MASK);
+             if (ipv6_addr_type(&sin->sin6_addr) & IPV6_ADDR_LINKLOCAL)
+                 sin->sin6_scope_id = IP6CB(skb)->iif;
+         } else {
+-            ipv6_addr_set_v4mapped(*(__be32 *)(nh + serr->addr_offset),
++ ipv6_addr_set_v4mapped(__get_unaligned_cpu32((__be32 *)(nh + serr->addr_offset)),
+                            &sin->sin6_addr);
+         }
+     }
+@@ -492,8 +492,8 @@ int datagram_recv_ctl(struct sock *sk, struct msghdr *msg, struct sk_buff *skb)
+         put_cmsg(msg, SOL_IPV6, IPV6_TCLASS, sizeof(tclass), &tclass);
+     }
+
+-    if (np->rxopt.bits.rxflow && (*(__be32 *)nh & IPV6_FLOWINFO_MASK)) {
+-        __be32 flowinfo = *(__be32 *)nh & IPV6_FLOWINFO_MASK;
++ if (np->rxopt.bits.rxflow && (__get_unaligned_cpu32((__be32 *)nh) & IPV6_FLOWINFO_MASK)) { ++ __be32 flowinfo = __get_unaligned_cpu32((__be32 *)nh) & IPV6_FLOWINFO_MASK; + put_cmsg(msg, SOL_IPV6, IPV6_FLOWINFO, sizeof(flowinfo), &flowinfo);
+     }
+
+@@ -588,7 +588,7 @@ int datagram_recv_ctl(struct sock *sk, struct msghdr *msg, struct sk_buff *skb)
+
+             sin6.sin6_family = AF_INET6;
+             sin6.sin6_addr = ipv6_hdr(skb)->daddr;
+-            sin6.sin6_port = ports[1];
++            sin6.sin6_port = __get_unaligned_cpu16(&(ports[1]));
+             sin6.sin6_flowinfo = 0;
+             sin6.sin6_scope_id = 0;
+
+@@ -680,12 +680,12 @@ int datagram_send_ctl(struct net *net, struct sock *sk,
+             }
+
+             if (fl6->flowlabel&IPV6_FLOWINFO_MASK) {
+- if ((fl6->flowlabel^*(__be32 *)CMSG_DATA(cmsg))&~IPV6_FLOWINFO_MASK) { ++ if ((fl6->flowlabel^__get_unaligned_cpu32((__be32 *)CMSG_DATA(cmsg)))&~IPV6_FLOWINFO_MASK) {
+                     err = -EINVAL;
+                     goto exit_f;
+                 }
+             }
+- fl6->flowlabel = IPV6_FLOWINFO_MASK & *(__be32 *)CMSG_DATA(cmsg); ++ fl6->flowlabel = IPV6_FLOWINFO_MASK & __get_unaligned_cpu32((__be32 *)CMSG_DATA(cmsg));
+             break;
+
+         case IPV6_2292HOPOPTS:
+diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
+index fa3d9c3..fd0f487 100644
+--- a/net/ipv6/exthdrs.c
++++ b/net/ipv6/exthdrs.c
+@@ -622,7 +622,7 @@ static bool ipv6_hop_jumbo(struct sk_buff *skb, int optoff)
+         goto drop;
+     }
+
+-    pkt_len = ntohl(*(__be32 *)(nh + optoff + 2));
++    pkt_len = ntohl(__get_unaligned_cpu32((__be32 *)(nh + optoff + 2)));
+     if (pkt_len <= IPV6_MAXPLEN) {
+         IP6_INC_STATS_BH(net, ipv6_skb_idev(skb),
+                  IPSTATS_MIB_INHDRERRORS);
+diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
+index 24995a9..ce8bd33 100644
+--- a/net/ipv6/ip6_fib.c
++++ b/net/ipv6/ip6_fib.c
+@@ -144,7 +144,7 @@ static __inline__ __be32 addr_bit_set(const void *token, int fn_bit)
+      * See include/asm-generic/bitops/le.h.
+      */
+ return (__force __be32)(1 << ((~fn_bit ^ BITOP_BE32_SWIZZLE) & 0x1f)) &
+-           addr[fn_bit >> 5];
++           __get_unaligned_cpu32(&addr[fn_bit >> 5]);
+ }
+
+ static __inline__ struct fib6_node * node_alloc(void)
+diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
+index aece3e7..95fb511 100644
+--- a/net/ipv6/ip6_output.c
++++ b/net/ipv6/ip6_output.c
+@@ -216,7 +216,9 @@ int ip6_xmit(struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6,
+     if (hlimit < 0)
+         hlimit = ip6_dst_hoplimit(dst);
+
+-    *(__be32 *)hdr = htonl(0x60000000 | (tclass << 20)) | fl6->flowlabel;
++    __put_unaligned_cpu32(htonl(0x60000000 | (tclass << 20))
++                  | fl6->flowlabel,
++                      (__be32*)hdr);
+
+     hdr->payload_len = htons(seg_len);
+     hdr->nexthdr = proto;
+@@ -267,7 +269,7 @@ int ip6_nd_hdr(struct sock *sk, struct sk_buff *skb, struct net_device *dev,
+     skb_put(skb, sizeof(struct ipv6hdr));
+     hdr = ipv6_hdr(skb);
+
+-    *(__be32*)hdr = htonl(0x60000000);
++    __put_unaligned_cpu32(htonl(0x60000000), (__be32*)hdr);
+
+     hdr->payload_len = htons(len);
+     hdr->nexthdr = proto;
+@@ -1614,8 +1616,9 @@ int ip6_push_pending_frames(struct sock *sk)
+     skb_reset_network_header(skb);
+     hdr = ipv6_hdr(skb);
+
+-    *(__be32*)hdr = fl6->flowlabel |
+-             htonl(0x60000000 | ((int)np->cork.tclass << 20));
++    __put_unaligned_cpu32(fl6->flowlabel |
++                  htonl(0x60000000 | ((int)np->cork.tclass << 20)),
++                  (__be32*)hdr);
+
+     hdr->hop_limit = np->cork.hop_limit;
+     hdr->nexthdr = proto;
+diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
+index cb7e2de..856c241 100644
+--- a/net/ipv6/ip6_tunnel.c
++++ b/net/ipv6/ip6_tunnel.c
+@@ -1004,7 +1004,7 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
+     skb_push(skb, sizeof(struct ipv6hdr));
+     skb_reset_network_header(skb);
+     ipv6h = ipv6_hdr(skb);
+-    *(__be32*)ipv6h = fl6->flowlabel | htonl(0x60000000);
++ __put_unaligned_cpu32(fl6->flowlabel | htonl(0x60000000), (__be32*)ipv6h);
+     dsfield = INET_ECN_encapsulate(0, dsfield);
+     ipv6_change_dsfield(ipv6h, ~INET_ECN_MASK, dsfield);
+     ipv6h->hop_limit = t->parms.hop_limit;
+@@ -1110,9 +1110,9 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
+
+     dsfield = ipv6_get_dsfield(ipv6h);
+     if (t->parms.flags & IP6_TNL_F_USE_ORIG_TCLASS)
+-        fl6.flowlabel |= (*(__be32 *) ipv6h & IPV6_TCLASS_MASK);
++ fl6.flowlabel |= (__get_unaligned_cpu32((__be32 *) ipv6h) & IPV6_TCLASS_MASK);
+     if (t->parms.flags & IP6_TNL_F_USE_ORIG_FLOWLABEL)
+-        fl6.flowlabel |= (*(__be32 *) ipv6h & IPV6_FLOWLABEL_MASK);
++ fl6.flowlabel |= (__get_unaligned_cpu32((__be32 *) ipv6h) & IPV6_FLOWLABEL_MASK);
+     if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
+         fl6.flowi6_mark = skb->mark;
+
+diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c
+index da8a4e3..8947ed1 100644
+--- a/net/ipv6/reassembly.c
++++ b/net/ipv6/reassembly.c
+@@ -124,10 +124,10 @@ void ip6_frag_init(struct inet_frag_queue *q, void *a)
+     struct frag_queue *fq = container_of(q, struct frag_queue, q);
+     struct ip6_create_arg *arg = a;
+
+-    fq->id = arg->id;
+-    fq->user = arg->user;
+-    fq->saddr = *arg->src;
+-    fq->daddr = *arg->dst;
++    fq->id = __get_unaligned_cpu32(&arg->id);
++    fq->user = __get_unaligned_cpu32(&arg->user);
++    memcpy(&fq->saddr, arg->src, sizeof(struct in6_addr));
++    memcpy(&fq->daddr, arg->dst, sizeof(struct in6_addr));
+ }
+ EXPORT_SYMBOL(ip6_frag_init);
+
+diff --git a/net/ipv6/route.c b/net/ipv6/route.c
+index b1e6cf0..bd010a8 100644
+--- a/net/ipv6/route.c
++++ b/net/ipv6/route.c
+@@ -938,7 +938,7 @@ void ip6_route_input(struct sk_buff *skb)
+         .flowi6_iif = skb->dev->ifindex,
+         .daddr = iph->daddr,
+         .saddr = iph->saddr,
+-        .flowlabel = (* (__be32 *) iph) & IPV6_FLOWINFO_MASK,
++        .flowlabel = (__get_unaligned_cpu32(iph)) & IPV6_FLOWINFO_MASK,
+         .flowi6_mark = skb->mark,
+         .flowi6_proto = iph->nexthdr,
+     };
+@@ -1108,7 +1108,7 @@ void ip6_update_pmtu(struct sk_buff *skb, struct net *net, __be32 mtu,
+     fl6.flowi6_flags = 0;
+     fl6.daddr = iph->daddr;
+     fl6.saddr = iph->saddr;
+-    fl6.flowlabel = (*(__be32 *) iph) & IPV6_FLOWINFO_MASK;
++ fl6.flowlabel = __get_unaligned_cpu32((__be32 *) iph) & IPV6_FLOWINFO_MASK;
+
+     dst = ip6_route_output(net, NULL, &fl6);
+     if (!dst->error)
+@@ -1136,7 +1136,7 @@ void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark)
+     fl6.flowi6_flags = 0;
+     fl6.daddr = iph->daddr;
+     fl6.saddr = iph->saddr;
+-    fl6.flowlabel = (*(__be32 *) iph) & IPV6_FLOWINFO_MASK;
++ fl6.flowlabel = __get_unaligned_cpu32((__be32 *) iph) & IPV6_FLOWINFO_MASK;
+
+     dst = ip6_route_output(net, NULL, &fl6);
+     if (!dst->error)
+diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
+index f8c4c08..8ce1aa6 100644
+--- a/net/ipv6/xfrm6_policy.c
++++ b/net/ipv6/xfrm6_policy.c
+@@ -166,8 +166,8 @@ _decode_session6(struct sk_buff *skb, struct flowi *fl, int reverse)
+                  pskb_may_pull(skb, nh + offset + 4 - skb->data))) {
+                 __be16 *ports = (__be16 *)exthdr;
+
+-                fl6->fl6_sport = ports[!!reverse];
+-                fl6->fl6_dport = ports[!reverse];
++                fl6->fl6_sport = __get_unaligned_cpu16(&ports[!!reverse]);
++                fl6->fl6_dport = __get_unaligned_cpu16(&ports[!reverse]);
+             }
+             fl6->flowi6_proto = nexthdr;
+             return;
+diff --git a/net/netfilter/nf_conntrack_proto_tcp.c b/net/netfilter/nf_conntrack_proto_tcp.c
+index 61f9285..13b4b80 100644
+--- a/net/netfilter/nf_conntrack_proto_tcp.c
++++ b/net/netfilter/nf_conntrack_proto_tcp.c
+@@ -449,7 +449,7 @@ static void tcp_sack(const struct sk_buff *skb, unsigned int dataoff,
+
+     /* Fast path for timestamp-only option */
+     if (length == TCPOLEN_TSTAMP_ALIGNED
+-        && *(__be32 *)ptr == htonl((TCPOPT_NOP << 24)
++ && __get_unaligned_cpu32((__be32 *)ptr) == htonl((TCPOPT_NOP << 24)
+                        | (TCPOPT_NOP << 16)
+                        | (TCPOPT_TIMESTAMP << 8)
+                        | TCPOLEN_TIMESTAMP))
+diff --git a/net/netfilter/xt_LOG.c b/net/netfilter/xt_LOG.c
+index fa40096..44946eb 100644
+--- a/net/netfilter/xt_LOG.c
++++ b/net/netfilter/xt_LOG.c
+@@ -521,9 +521,9 @@ static void dump_ipv6_packet(struct sbuff *m,
+     /* Max length: 44 "LEN=65535 TC=255 HOPLIMIT=255 FLOWLBL=FFFFF " */
+     sb_add(m, "LEN=%Zu TC=%u HOPLIMIT=%u FLOWLBL=%u ",
+            ntohs(ih->payload_len) + sizeof(struct ipv6hdr),
+-           (ntohl(*(__be32 *)ih) & 0x0ff00000) >> 20,
++ (ntohl(__get_unaligned_cpu32((__be32 *)ih)) & 0x0ff00000) >> 20,
+            ih->hop_limit,
+-           (ntohl(*(__be32 *)ih) & 0x000fffff));
++           (ntohl(__get_unaligned_cpu32((__be32 *)ih)) & 0x000fffff));
+
+     fragment = 0;
+     ptr = ip6hoff + sizeof(struct ipv6hdr);
+diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c
+index c7c27bc..9d4d5c9 100644
+--- a/net/sched/cls_u32.c
++++ b/net/sched/cls_u32.c
+@@ -142,7 +142,7 @@ next_knode:
+             data = skb_header_pointer(skb, toff, 4, &hdata);
+             if (!data)
+                 goto out;
+-            if ((*data ^ key->val) & key->mask) {
++            if ((__get_unaligned_cpu32(data) ^ key->val) & key->mask) {
+                 n = n->next;
+                 goto next_knode;
+             }
+@@ -193,7 +193,7 @@ check_terminal:
+                           &hdata);
+             if (!data)
+                 goto out;
+-            sel = ht->divisor & u32_hash_fold(*data, &n->sel,
++ sel = ht->divisor & u32_hash_fold(__get_unaligned_cpu32(data), &n->sel,
+                               n->fshift);
+         }
+ if (!(n->sel.flags & (TC_U32_VAROFFSET | TC_U32_OFFSET | TC_U32_EAT)))
+@@ -209,7 +209,7 @@ check_terminal:
+                               2, &hdata);
+                 if (!data)
+                     goto out;
+-                off2 += ntohs(n->sel.offmask & *data) >>
++ off2 += ntohs(n->sel.offmask & __get_unaligned_cpu16(data)) >>
+                     n->sel.offshift;
+             }
+             off2 &= ~3;
+diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
+index a18d975..7ccb37f 100644
+--- a/net/sched/sch_api.c
++++ b/net/sched/sch_api.c
+@@ -1658,7 +1658,7 @@ done:
+ int tc_classify_compat(struct sk_buff *skb, const struct tcf_proto *tp,
+                struct tcf_result *res)
+ {
+-    __be16 protocol = skb->protocol;
++    __be16 protocol = __get_unaligned_cpu16(&skb->protocol);
+     int err;
+
+     for (; tp; tp = tp->next) {
+diff --git a/net/xfrm/xfrm_input.c b/net/xfrm/xfrm_input.c
+index ab2bb42..e32ad98 100644
+--- a/net/xfrm/xfrm_input.c
++++ b/net/xfrm/xfrm_input.c
+@@ -52,6 +52,7 @@ int xfrm_parse_spi(struct sk_buff *skb, u8 nexthdr, __be32 *spi, __be32 *seq)
+ {
+     int offset, offset_seq;
+     int hlen;
++    __be32 *pspi, *pseq;
+
+     switch (nexthdr) {
+     case IPPROTO_AH:
+@@ -77,8 +78,10 @@ int xfrm_parse_spi(struct sk_buff *skb, u8 nexthdr, __be32 *spi, __be32 *seq)
+     if (!pskb_may_pull(skb, hlen))
+         return -EINVAL;
+
+-    *spi = *(__be32*)(skb_transport_header(skb) + offset);
+-    *seq = *(__be32*)(skb_transport_header(skb) + offset_seq);
++    pspi = (__be32*)(skb_transport_header(skb) + offset);
++    pseq = (__be32*)(skb_transport_header(skb) + offset_seq);
++    *spi = __get_unaligned_cpu32(pspi);
++    *seq = __get_unaligned_cpu32(pseq);
+     return 0;
+ }
+

--
Robert Bradley

_______________________________________________
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel

Reply via email to