1. 20 May, 2016 32 commits
  2. 19 May, 2016 8 commits
    • David S. Miller's avatar
      Merge tag 'batman-adv-fix-for-davem' of git://git.open-mesh.org/linux-merge · e22cb5e7
      David S. Miller authored
      Antonio Quartulli says:
      
      ====================
      During the Wireless Battle Mesh v9 in Porto (PT) at the beginning of
      May, we managed to uncover and fix some important bugs in our
      new B.A.T.M.A.N. V algorithm. These are the fixes we came up with
      together with others that I collected in the past weeks:
      - avoid potential crash due to NULL pointer dereference in
        B.A.T.M.A.N. V routine when a neigh_ifinfo object is not found, by
        Sven Eckelmann
      - avoid use-after-free of skb when counting outgoing bytes, by Florian
        Westphal
      - fix neigh_ifinfo object reference counting imbalance when using
        B.A.T.M.A.N. V, by Sven Eckelmann. Such imbalance may lead to the
        impossibility of releasing the related netdev object on shutdown
      - avoid invalid memory access in case of error while allocating
        bcast_own_sum when a new hard-interface is added, by Sven Eckelmann
      - ensure originator address is updated in OMG/ELP packet content upon
        primary interface address change, by Antonio Quartulli
      - fix integer overflow when computing TQ metric (B.A.T.M.A.N. IV), by
        Sven Eckelmann
      - avoid race condition while adding new neigh_node which would result
        in having two objects mapping to the same physical neighbour, by
        Linus Lüssing
      - ensure originator address is initialized in ELP packet content on
        secondary interfaces, by Marek Lindner
      ====================
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      e22cb5e7
    • Amit Ghadge's avatar
      net: Fix coding style warnings and errors. · 6104503c
      Amit Ghadge authored
      This is a patch to clean checkpatch warnings and errors
      in the Space.c file.
      Clean up the following warnings and errors.
      
      WARNING :
      * Block comments use * on subsequent lines
      * Missing a blank line after declarations
      * networking block comments don't use an empty /* line, use /*
      * please, no space before tabs
      * please, no spaces at the start of a line
      * line over 80 characters
      
      ERROR :
      * code indent should use tabs where possible
      * space prohibited after that open parenthesis '('
      Signed-off-by: default avatarAmit Ghadge <amitg.b14@gmail.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      6104503c
    • David S. Miller's avatar
      Merge branch 'tcp_bh_fixes' · 87553aa5
      David S. Miller authored
      Eric Dumazet says:
      
      ====================
      net: block BH in TCP callbacks
      
      Four layers using TCP stack were assuming sk_callback_lock could
      be locked using read_lock() in their handlers because TCP stack
      was running with BH disabled.
      
      This is no longer the case. Since presumably the rest could
      also depend on BH being disabled, just use read_lock_bh().
      
      Then each layer might consider switching to RCU protection
      and no longer depend on BH.
      ====================
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      87553aa5
    • Eric Dumazet's avatar
      tipc: block BH in TCP callbacks · b91083a4
      Eric Dumazet authored
      TCP stack can now run from process context.
      
      Use read_lock_bh(&sk->sk_callback_lock) variant to restore previous
      assumption.
      
      Fixes: 5413d1ba ("net: do not block BH while processing socket backlog")
      Fixes: d41a69f1 ("tcp: make tcp_sendmsg() aware of socket backlog")
      Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
      Cc: Jon Maloy <jon.maloy@ericsson.com>
      Cc: Ying Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      b91083a4
    • Eric Dumazet's avatar
      rds: tcp: block BH in TCP callbacks · 38036629
      Eric Dumazet authored
      TCP stack can now run from process context.
      
      Use read_lock_bh(&sk->sk_callback_lock) variant to restore previous
      assumption.
      
      Fixes: 5413d1ba ("net: do not block BH while processing socket backlog")
      Fixes: d41a69f1 ("tcp: make tcp_sendmsg() aware of socket backlog")
      Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      38036629
    • Eric Dumazet's avatar
      ocfs2/cluster: block BH in TCP callbacks · e1daca28
      Eric Dumazet authored
      TCP stack can now run from process context.
      
      Use read_lock_bh() variant to restore previous assumption.
      
      Fixes: 5413d1ba ("net: do not block BH while processing socket backlog")
      Fixes: d41a69f1 ("tcp: make tcp_sendmsg() aware of socket backlog")
      Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      e1daca28
    • Eric Dumazet's avatar
      scsi_tcp: block BH in TCP callbacks · 7cb001d4
      Eric Dumazet authored
      iscsi_sw_tcp_data_ready() and iscsi_sw_tcp_state_change() were
      using read_lock(&sk->sk_callback_lock) which is fine if caller
      disabled BH.
      
      TCP stack no longer has this requirement and can run from
      process context.
      
      Use read_lock_bh() variant to restore previous assumption.
      
      Ideally this code could use RCU instead...
      
      Fixes: 5413d1ba ("net: do not block BH while processing socket backlog")
      Fixes: d41a69f1 ("tcp: make tcp_sendmsg() aware of socket backlog")
      Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
      Cc: Mike Christie <michaelc@cs.wisc.edu>
      Cc: Venkatesh Srinivas <venkateshs@google.com>
      Acked-by: default avatarMike Christie <michaelc@cs.wisc.edu>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      7cb001d4
    • xypron.glpk@gmx.de's avatar
      net: pegasus: remove dead coding · e00be9e4
      xypron.glpk@gmx.de authored
      (!count || count < 4) is always true.
      So let's remove the coding which is dead at least since 2005.
      Signed-off-by: default avatarHeinrich Schuchardt <xypron.glpk@gmx.de>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      e00be9e4