Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking changes from David S Miller:

 1) Remove the ipv4 routing cache.  Now lookups go directly into the FIB
    trie and use prebuilt routes cached there.

    No more garbage collection, no more rDOS attacks on the routing
    cache.  Instead we now get predictable and consistent performance,
    no matter what the pattern of traffic we service.

    This has been almost 2 years in the making.  Special thanks to
    Julian Anastasov, Eric Dumazet, Steffen Klassert, and others who
    have helped along the way.

    I'm sure that with a change of this magnitude there will be some
    kind of fallout, but such things ought the be simple to fix at this
    point.  Luckily I'm not European so I'll be around all of August to
    fix things :-)

    The major stages of this work here are each fronted by a forced
    merge commit whose commit message contains a top-level description
    of the motivations and implementation issues.

 2) Pre-demux of established ipv4 TCP sockets, saves a route demux on
    input.

 3) TCP SYN/ACK performance tweaks from Eric Dumazet.

 4) Add namespace support for netfilter L4 conntrack helpers, from Gao
    Feng.

 5) Add config mechanism for Energy Efficient Ethernet to ethtool, from
    Yuval Mintz.

 6) Remove quadratic behavior from /proc/net/unix, from Eric Dumazet.

 7) Support for connection tracker helpers in userspace, from Pablo
    Neira Ayuso.

 8) Allow userspace driven TX load balancing functions in TEAM driver,
    from Jiri Pirko.

 9) Kill off NLMSG_PUT and RTA_PUT macros, more gross stuff with
    embedded gotos.

10) TCP Small Queues, essentially minimize the amount of TCP data queued
    up in the packet scheduler layer.  Whereas the existing BQL (Byte
    Queue Limits) limits the pkt_sched --> netdevice queuing levels,
    this controls the TCP --> pkt_sched queueing levels.

    From Eric Dumazet.

11) Reduce the number of get_page/put_page ops done on SKB fragments,
    from Alexander Duyck.

12) Implement protection against blind resets in TCP (RFC 5961), from
    Eric Dumazet.

13) Support the client side of TCP Fast Open, basically the ability to
    send data in the SYN exchange, from Yuchung Cheng.

    Basically, the sender queues up data with a sendmsg() call using
    MSG_FASTOPEN, then they do the connect() which emits the queued up
    fastopen data.

14) Avoid all the problems we get into in TCP when timers or PMTU events
    hit a locked socket.  The TCP Small Queues changes added a
    tcp_release_cb() that allows us to queue work up to the
    release_sock() caller, and that's what we use here too.  From Eric
    Dumazet.

15) Zero copy on TX support for TUN driver, from Michael S. Tsirkin.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1870 commits)
  genetlink: define lockdep_genl_is_held() when CONFIG_LOCKDEP
  r8169: revert "add byte queue limit support".
  ipv4: Change rt->rt_iif encoding.
  net: Make skb->skb_iif always track skb->dev
  ipv4: Prepare for change of rt->rt_iif encoding.
  ipv4: Remove all RTCF_DIRECTSRC handliing.
  ipv4: Really ignore ICMP address requests/replies.
  decnet: Don't set RTCF_DIRECTSRC.
  net/ipv4/ip_vti.c: Fix __rcu warnings detected by sparse.
  ipv4: Remove redundant assignment
  rds: set correct msg_namelen
  openvswitch: potential NULL deref in sample()
  tcp: dont drop MTU reduction indications
  bnx2x: Add new 57840 device IDs
  tcp: avoid oops in tcp_metrics and reset tcpm_stamp
  niu: Change niu_rbr_fill() to use unlikely() to check niu_rbr_add_page() return value
  niu: Fix to check for dma mapping errors.
  net: Fix references to out-of-scope variables in put_cmsg_compat()
  net: ethernet: davinci_emac: add pm_runtime support
  net: ethernet: davinci_emac: Remove unnecessary #include
  ...

1 files changed, 162 insertions, 24 deletions

diff --git a/Documentation/networking/can.txt b/Documentation/networking/can.txt
index ac295399f0d4..820f55344edc 100644
--- a/Documentation/networking/can.txt
+++ b/Documentation/networking/can.txt
@@ -22,7 +22,8 @@ This file contains
       4.1.2 RAW socket option CAN_RAW_ERR_FILTER
       4.1.3 RAW socket option CAN_RAW_LOOPBACK
       4.1.4 RAW socket option CAN_RAW_RECV_OWN_MSGS
-      4.1.5 RAW socket returned message flags
+      4.1.5 RAW socket option CAN_RAW_FD_FRAMES
+      4.1.6 RAW socket returned message flags
     4.2 Broadcast Manager protocol sockets (SOCK_DGRAM)
     4.3 connected transport protocols (SOCK_SEQPACKET)
     4.4 unconnected transport protocols (SOCK_DGRAM)
@@ -41,7 +42,8 @@ This file contains
       6.5.1 Netlink interface to set/get devices properties
       6.5.2 Setting the CAN bit-timing
       6.5.3 Starting and stopping the CAN network device
-    6.6 supported CAN hardware
+    6.6 CAN FD (flexible data rate) driver support
+    6.7 supported CAN hardware
 
   7 Socket CAN resources
 
@@ -232,16 +234,16 @@ solution for a couple of reasons:
   arbitration problems and error frames caused by the different
   ECUs. The occurrence of detected errors are important for diagnosis
   and have to be logged together with the exact timestamp. For this
-  reason the CAN interface driver can generate so called Error Frames
-  that can optionally be passed to the user application in the same
-  way as other CAN frames. Whenever an error on the physical layer
+  reason the CAN interface driver can generate so called Error Message
+  Frames that can optionally be passed to the user application in the
+  same way as other CAN frames. Whenever an error on the physical layer
   or the MAC layer is detected (e.g. by the CAN controller) the driver
-  creates an appropriate error frame. Error frames can be requested by
-  the user application using the common CAN filter mechanisms. Inside
-  this filter definition the (interested) type of errors may be
-  selected. The reception of error frames is disabled by default.
-  The format of the CAN error frame is briefly described in the Linux
-  header file "include/linux/can/error.h".
+  creates an appropriate error message frame. Error messages frames can
+  be requested by the user application using the common CAN filter
+  mechanisms. Inside this filter definition the (interested) type of
+  errors may be selected. The reception of error messages is disabled
+  by default. The format of the CAN error message frame is briefly
+  described in the Linux header file "include/linux/can/error.h".
 
 4. How to use Socket CAN
 ------------------------
@@ -273,7 +275,7 @@ solution for a couple of reasons:
 
     struct can_frame {
             canid_t can_id;  /* 32 bit CAN_ID + EFF/RTR/ERR flags */
-            __u8    can_dlc; /* data length code: 0 .. 8 */
+            __u8    can_dlc; /* frame payload length in byte (0 .. 8) */
             __u8    data[8] __attribute__((aligned(8)));
     };
 
@@ -375,6 +377,51 @@ solution for a couple of reasons:
     nbytes = sendto(s, &frame, sizeof(struct can_frame),
                     0, (struct sockaddr*)&addr, sizeof(addr));
 
+  Remark about CAN FD (flexible data rate) support:
+
+  Generally the handling of CAN FD is very similar to the formerly described
+  examples. The new CAN FD capable CAN controllers support two different
+  bitrates for the arbitration phase and the payload phase of the CAN FD frame
+  and up to 64 bytes of payload. This extended payload length breaks all the
+  kernel interfaces (ABI) which heavily rely on the CAN frame with fixed eight
+  bytes of payload (struct can_frame) like the CAN_RAW socket. Therefore e.g.
+  the CAN_RAW socket supports a new socket option CAN_RAW_FD_FRAMES that
+  switches the socket into a mode that allows the handling of CAN FD frames
+  and (legacy) CAN frames simultaneously (see section 4.1.5).
+
+  The struct canfd_frame is defined in include/linux/can.h:
+
+    struct canfd_frame {
+            canid_t can_id;  /* 32 bit CAN_ID + EFF/RTR/ERR flags */
+            __u8    len;     /* frame payload length in byte (0 .. 64) */
+            __u8    flags;   /* additional flags for CAN FD */
+            __u8    __res0;  /* reserved / padding */
+            __u8    __res1;  /* reserved / padding */
+            __u8    data[64] __attribute__((aligned(8)));
+    };
+
+  The struct canfd_frame and the existing struct can_frame have the can_id,
+  the payload length and the payload data at the same offset inside their
+  structures. This allows to handle the different structures very similar.
+  When the content of a struct can_frame is copied into a struct canfd_frame
+  all structure elements can be used as-is - only the data[] becomes extended.
+
+  When introducing the struct canfd_frame it turned out that the data length
+  code (DLC) of the struct can_frame was used as a length information as the
+  length and the DLC has a 1:1 mapping in the range of 0 .. 8. To preserve
+  the easy handling of the length information the canfd_frame.len element
+  contains a plain length value from 0 .. 64. So both canfd_frame.len and
+  can_frame.can_dlc are equal and contain a length information and no DLC.
+  For details about the distinction of CAN and CAN FD capable devices and
+  the mapping to the bus-relevant data length code (DLC), see chapter 6.6.
+
+  The length of the two CAN(FD) frame structures define the maximum transfer
+  unit (MTU) of the CAN(FD) network interface and skbuff data length. Two
+  definitions are specified for CAN specific MTUs in include/linux/can.h :
+
+  #define CAN_MTU   (sizeof(struct can_frame))   == 16  => 'legacy' CAN frame
+  #define CANFD_MTU (sizeof(struct canfd_frame)) == 72  => CAN FD frame
+
   4.1 RAW protocol sockets with can_filters (SOCK_RAW)
 
   Using CAN_RAW sockets is extensively comparable to the commonly
@@ -383,7 +430,7 @@ solution for a couple of reasons:
   defaults are set at RAW socket binding time:
 
   - The filters are set to exactly one filter receiving everything
-  - The socket only receives valid data frames (=> no error frames)
+  - The socket only receives valid data frames (=> no error message frames)
   - The loopback of sent CAN frames is enabled (see chapter 3.2)
   - The socket does not receive its own sent frames (in loopback mode)
 
@@ -434,7 +481,7 @@ solution for a couple of reasons:
   4.1.2 RAW socket option CAN_RAW_ERR_FILTER
 
   As described in chapter 3.4 the CAN interface driver can generate so
-  called Error Frames that can optionally be passed to the user
+  called Error Message Frames that can optionally be passed to the user
   application in the same way as other CAN frames. The possible
   errors are divided into different error classes that may be filtered
   using the appropriate error mask. To register for every possible
@@ -472,7 +519,69 @@ solution for a couple of reasons:
     setsockopt(s, SOL_CAN_RAW, CAN_RAW_RECV_OWN_MSGS,
                &recv_own_msgs, sizeof(recv_own_msgs));
 
-  4.1.5 RAW socket returned message flags
+  4.1.5 RAW socket option CAN_RAW_FD_FRAMES
+
+  CAN FD support in CAN_RAW sockets can be enabled with a new socket option
+  CAN_RAW_FD_FRAMES which is off by default. When the new socket option is
+  not supported by the CAN_RAW socket (e.g. on older kernels), switching the
+  CAN_RAW_FD_FRAMES option returns the error -ENOPROTOOPT.
+
+  Once CAN_RAW_FD_FRAMES is enabled the application can send both CAN frames
+  and CAN FD frames. OTOH the application has to handle CAN and CAN FD frames
+  when reading from the socket.
+
+    CAN_RAW_FD_FRAMES enabled:  CAN_MTU and CANFD_MTU are allowed
+    CAN_RAW_FD_FRAMES disabled: only CAN_MTU is allowed (default)
+
+  Example:
+    [ remember: CANFD_MTU == sizeof(struct canfd_frame) ]
+
+    struct canfd_frame cfd;
+
+    nbytes = read(s, &cfd, CANFD_MTU);
+
+    if (nbytes == CANFD_MTU) {
+            printf("got CAN FD frame with length %d\n", cfd.len);
+	    /* cfd.flags contains valid data */
+    } else if (nbytes == CAN_MTU) {
+            printf("got legacy CAN frame with length %d\n", cfd.len);
+	    /* cfd.flags is undefined */
+    } else {
+            fprintf(stderr, "read: invalid CAN(FD) frame\n");
+            return 1;
+    }
+
+    /* the content can be handled independently from the received MTU size */
+
+    printf("can_id: %X data length: %d data: ", cfd.can_id, cfd.len);
+    for (i = 0; i < cfd.len; i++)
+            printf("%02X ", cfd.data[i]);
+
+  When reading with size CANFD_MTU only returns CAN_MTU bytes that have
+  been received from the socket a legacy CAN frame has been read into the
+  provided CAN FD structure. Note that the canfd_frame.flags data field is
+  not specified in the struct can_frame and therefore it is only valid in
+  CANFD_MTU sized CAN FD frames.
+
+  As long as the payload length is <=8 the received CAN frames from CAN FD
+  capable CAN devices can be received and read by legacy sockets too. When
+  user-generated CAN FD frames have a payload length <=8 these can be send
+  by legacy CAN network interfaces too. Sending CAN FD frames with payload
+  length > 8 to a legacy CAN network interface returns an -EMSGSIZE error.
+
+  Implementation hint for new CAN applications:
+
+  To build a CAN FD aware application use struct canfd_frame as basic CAN
+  data structure for CAN_RAW based applications. When the application is
+  executed on an older Linux kernel and switching the CAN_RAW_FD_FRAMES
+  socket option returns an error: No problem. You'll get legacy CAN frames
+  or CAN FD frames and can process them the same way.
+
+  When sending to CAN devices make sure that the device is capable to handle
+  CAN FD frames by checking if the device maximum transfer unit is CANFD_MTU.
+  The CAN device MTU can be retrieved e.g. with a SIOCGIFMTU ioctl() syscall.
+
+  4.1.6 RAW socket returned message flags
 
   When using recvmsg() call, the msg->msg_flags may contain following flags:
 
@@ -527,7 +636,7 @@ solution for a couple of reasons:
 
     rcvlist_all - list for unfiltered entries (no filter operations)
     rcvlist_eff - list for single extended frame (EFF) entries
-    rcvlist_err - list for error frames masks
+    rcvlist_err - list for error message frames masks
     rcvlist_fil - list for mask/value filters
     rcvlist_inv - list for mask/value filters (inverse semantic)
     rcvlist_sff - list for single standard frame (SFF) entries
@@ -573,10 +682,13 @@ solution for a couple of reasons:
     dev->type  = ARPHRD_CAN; /* the netdevice hardware type */
     dev->flags = IFF_NOARP;  /* CAN has no arp */
 
-    dev->mtu   = sizeof(struct can_frame);
+    dev->mtu = CAN_MTU; /* sizeof(struct can_frame) -> legacy CAN interface */
 
-  The struct can_frame is the payload of each socket buffer in the
-  protocol family PF_CAN.
+    or alternative, when the controller supports CAN with flexible data rate:
+    dev->mtu = CANFD_MTU; /* sizeof(struct canfd_frame) -> CAN FD interface */
+
+  The struct can_frame or struct canfd_frame is the payload of each socket
+  buffer (skbuff) in the protocol family PF_CAN.
 
   6.2 local loopback of sent frames
 
@@ -784,15 +896,41 @@ solution for a couple of reasons:
     $ ip link set canX type can restart-ms 100
 
   Alternatively, the application may realize the "bus-off" condition
-  by monitoring CAN error frames and do a restart when appropriate with
-  the command:
+  by monitoring CAN error message frames and do a restart when
+  appropriate with the command:
 
     $ ip link set canX type can restart
 
-  Note that a restart will also create a CAN error frame (see also
-  chapter 3.4).
+  Note that a restart will also create a CAN error message frame (see
+  also chapter 3.4).
+
+  6.6 CAN FD (flexible data rate) driver support
+
+  CAN FD capable CAN controllers support two different bitrates for the
+  arbitration phase and the payload phase of the CAN FD frame. Therefore a
+  second bittiming has to be specified in order to enable the CAN FD bitrate.
+
+  Additionally CAN FD capable CAN controllers support up to 64 bytes of
+  payload. The representation of this length in can_frame.can_dlc and
+  canfd_frame.len for userspace applications and inside the Linux network
+  layer is a plain value from 0 .. 64 instead of the CAN 'data length code'.
+  The data length code was a 1:1 mapping to the payload length in the legacy
+  CAN frames anyway. The payload length to the bus-relevant DLC mapping is
+  only performed inside the CAN drivers, preferably with the helper
+  functions can_dlc2len() and can_len2dlc().
+
+  The CAN netdevice driver capabilities can be distinguished by the network
+  devices maximum transfer unit (MTU):
+
+  MTU = 16 (CAN_MTU)   => sizeof(struct can_frame)   => 'legacy' CAN device
+  MTU = 72 (CANFD_MTU) => sizeof(struct canfd_frame) => CAN FD capable device
+
+  The CAN device MTU can be retrieved e.g. with a SIOCGIFMTU ioctl() syscall.
+  N.B. CAN FD capable devices can also handle and send legacy CAN frames.
+
+  FIXME: Add details about the CAN FD controller configuration when available.
 
-  6.6 Supported CAN hardware
+  6.7 Supported CAN hardware
 
   Please check the "Kconfig" file in "drivers/net/can" to get an actual
   list of the support CAN hardware. On the Socket CAN project website