1. 14 Jun, 2008 1 commit
  2. 13 Jun, 2008 1 commit
  3. 12 Jun, 2008 1 commit
    • David S. Miller's avatar
      tcp: Revert 'process defer accept as established' changes. · ec0a1966
      David S. Miller authored
      This reverts two changesets, ec3c0982
      ("[TCP]: TCP_DEFER_ACCEPT updates - process as established") and
      the follow-on bug fix 9ae27e0a
      
      
      ("tcp: Fix slab corruption with ipv6 and tcp6fuzz").
      
      This change causes several problems, first reported by Ingo Molnar
      as a distcc-over-loopback regression where connections were getting
      stuck.
      
      Ilpo Järvinen first spotted the locking problems.  The new function
      added by this code, tcp_defer_accept_check(), only has the
      child socket locked, yet it is modifying state of the parent
      listening socket.
      
      Fixing that is non-trivial at best, because we can't simply just grab
      the parent listening socket lock at this point, because it would
      create an ABBA deadlock.  The normal ordering is parent listening
      socket --> child socket, but this code path would require the
      reverse lock ordering.
      
      Next is a problem noticed by Vitaliy Gusev, he noted:
      
      ----------------------------------------
      >--- a/net/ipv4/tcp_timer.c
      >+++ b/net/ipv4/tcp_timer.c
      >@@ -481,6 +481,11 @@ static void tcp_keepalive_timer (unsigned long data)
      > 		goto death;
      > 	}
      >
      >+	if (tp->defer_tcp_accept.request && sk->sk_state == TCP_ESTABLISHED) {
      >+		tcp_send_active_reset(sk, GFP_ATOMIC);
      >+		goto death;
      
      Here socket sk is not attached to listening socket's request queue. tcp_done()
      will not call inet_csk_destroy_sock() (and tcp_v4_destroy_sock() which should
      release this sk) as socket is not DEAD. Therefore socket sk will be lost for
      freeing.
      ----------------------------------------
      
      Finally, Alexey Kuznetsov argues that there might not even be any
      real value or advantage to these new semantics even if we fix all
      of the bugs:
      
      ----------------------------------------
      Hiding from accept() sockets with only out-of-order data only
      is the only thing which is impossible with old approach. Is this really
      so valuable? My opinion: no, this is nothing but a new loophole
      to consume memory without control.
      ----------------------------------------
      
      So revert this thing for now.
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      ec0a1966
  4. 11 Jun, 2008 7 commits
  5. 10 Jun, 2008 2 commits
  6. 09 Jun, 2008 5 commits
  7. 04 Jun, 2008 7 commits
  8. 03 Jun, 2008 2 commits
  9. 22 May, 2008 1 commit
  10. 21 May, 2008 7 commits
  11. 19 May, 2008 2 commits
  12. 16 May, 2008 1 commit
    • Ivo van Doorn's avatar
      mac80211: Add RTNL version of ieee80211_iterate_active_interfaces · 2f561feb
      Ivo van Doorn authored
      Since commit e38bad47
      
      
      	mac80211: make ieee80211_iterate_active_interfaces not need rtnl
      rt2500usb and rt73usb broke down due to attempting register access
      in atomic context (which is not possible for USB hardware).
      
      This patch restores ieee80211_iterate_active_interfaces() to use RTNL lock,
      and provides the non-RTNL version under a new name:
      	ieee80211_iterate_active_interfaces_atomic()
      
      So far only rt2x00 uses ieee80211_iterate_active_interfaces(), and those
      drivers require the RTNL version of ieee80211_iterate_active_interfaces().
      Since they already call that function directly, this patch will automatically
      fix the USB rt2x00 drivers.
      
      v2: Rename ieee80211_iterate_active_interfaces_rtnl
      Signed-off-by: default avatarIvo van Doorn <IvDoorn@gmail.com>
      Acked-by: default avatarJohannes Berg <johannes@sipsolutions.net>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      2f561feb
  13. 14 May, 2008 3 commits
    • Eric Van Hensbergen's avatar
      9p: fix error path during early mount · 887b3ece
      Eric Van Hensbergen authored
      
      
      There was some cleanup issues during early mount which would trigger
      a kernel bug for certain types of failure.  This patch reorganizes the
      cleanup to get rid of the bad behavior.
      
      This also merges the 9pnet and 9pnet_fd modules for the purpose of
      configuration and initialization.  Keeping the fd transport separate
      from the core 9pnet code seemed like a good idea at the time, but in
      practice has caused more harm and confusion than good.
      Signed-off-by: default avatarEric Van Hensbergen <ericvh@gmail.com>
      887b3ece
    • Eric Van Hensbergen's avatar
      9p: Documentation updates · ee443996
      Eric Van Hensbergen authored
      
      
      The kernel-doc comments of much of the 9p system have been in disarray since
      reorganization.  This patch fixes those problems, adds additional documentation
      and a template book which collects the 9p information.
      Signed-off-by: default avatarEric Van Hensbergen <ericvh@gmail.com>
      ee443996
    • Bruno Randolf's avatar
      mac80211: use hardware flags for signal/noise units · 566bfe5a
      Bruno Randolf authored
      
      
      trying to clean up the signal/noise code. the previous code in mac80211 had
      confusing names for the related variables, did not have much definition of
      what units of signal and noise were provided and used implicit mechanisms from
      the wireless extensions.
      
      this patch introduces hardware capability flags to let the hardware specify
      clearly if it can provide signal and noise level values and which units it can
      provide. this also anticipates possible new units like RCPI in the future.
      
      for signal:
      
        IEEE80211_HW_SIGNAL_UNSPEC - unspecified, unknown, hw specific
        IEEE80211_HW_SIGNAL_DB     - dB difference to unspecified reference point
        IEEE80211_HW_SIGNAL_DBM    - dBm, difference to 1mW
      
      for noise we currently only have dBm:
      
        IEEE80211_HW_NOISE_DBM     - dBm, difference to 1mW
      
      if IEEE80211_HW_SIGNAL_UNSPEC or IEEE80211_HW_SIGNAL_DB is used the driver has
      to provide the maximum value (max_signal) it reports in order for applications
      to make sense of the signal values.
      
      i tried my best to find out for each driver what it can provide and update it
      but i'm not sure (?) for some of them and used the more conservative guess in
      doubt. this can be fixed easily after this patch has been merged by changing
      the hardware flags of the driver.
      
      DRIVER          SIGNAL    MAX	NOISE   QUAL
      -----------------------------------------------------------------
      adm8211         unspec(?) 100   n/a     missing
      at76_usb        unspec(?) (?)   unused  missing
      ath5k           dBm             dBm     percent rssi
      b43legacy       dBm             dBm     percent jssi(?)
      b43             dBm             dBm     percent jssi(?)
      iwl-3945        dBm             dBm     percent snr+more
      iwl-4965        dBm             dBm     percent snr+more
      p54             unspec    127   n/a     missing
      rt2x00          dBm	        n/a     percent rssi+tx/rx frame success
        rt2400        dBm             n/a
        rt2500pci     dBm             n/a
        rt2500usb     dBm             n/a
        rt61pci       dBm             n/a
        rt73usb       dBm             n/a
      rtl8180         unspec(?) 65    n/a     (?)
      rtl8187         unspec(?) 65    (?)     noise(?)
      zd1211          dB(?)     100   n/a     percent
      
      drivers/net/wireless/ath5k/base.c:      Changes-licensed-under: 3-Clause-BSD
      Signed-off-by: default avatarBruno Randolf <br1@einfach.org>
      Signed-off-by: default avatarJohn W. Linville <linville@tuxdriver.com>
      566bfe5a